Haematologica, Volume 103, issue 5

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

Ancient Greek

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

Scientific Latin

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

Scientific Latin

haematologicus (adjective) = related to blood

Modern English

haematologica (adjective, plural and neuter, used as a noun) = hematological subjects The oldest hematology journal, publishing the newest research results.

2016 JCR impact factor = 7.702

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

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation Editor-in-Chief Luca Malcovati (Pavia)

Managing Director Antonio Majocchi (Pavia)

Associate Editors Omar I. Abdel-Wahab (New York), Hélène Cavé (Paris), Simon Mendez-Ferrer (Cambridge), Pavan Reddy (Ann Arbor), Andreas Rosenwald (Wuerzburg), Monika Engelhardt (Freiburg), Davide Rossi (Bellinzona), Jacob Rowe (Haifa, Jerusalem), Wyndham Wilson (Bethesda), Paul Kyrle (Vienna), 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 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); 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 2018 are as following: Print edition

Institutional Euro 600

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: Press Up, zona Via Cassia Km 36, 300 Zona Ind.le Settevene - 01036 Nepi (VT)

haematologica calendar of events

Journal of the European Hematology Association Published by the Ferrata Storti Foundation The 4th World Congress on Controversies in Multiple Myeloma Chairs: M Mohty, A Nagler, T Facon May 3-5, 2018 Paris, France EHA Hematology Tutorial on Thalassemia May 10-11, 2018 Shiraz, Iran JACIE Inspector Training Course EBMT / JACIE Chair: E McGrath May 24-25, 2018 Barcelona, Spain

23rd Congress of EHA June 14-17, 2018 Stockholm, Sweden EHA-SAH Hematology Tutorial on lymphoid Malignancies and Plasma Cell Dyscrasias September 14-15, 2018 Buenos Aires, Argentina EHA-SWG Scientific Meeting on Aging and Hematology Chair: D Bron October 12-14, 2018 Warsaw, Poland

35th International Congress of the ISBT The International Society of Blood Transfusion (ISBT) Chairs: K Pavenski, E van der Schoot, E Wood June 2-6, 2018 Toronto, Canada 20th Congress of the European Society for Haemapheresis European Society for Haemapheresis June 13-14, 2018 Valencia, Spain

Calendar of Events updated on April 5, 2018

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

Table of Contents Volume 103, Issue 5: May 2018 Cover Figure

Cluster of osteoblasts with eccentric round to oval nuclei and abundant blue, mottled cytoplasm. Courtesy of Prof. Rosangela Invernizzi.

Editorials 747

(Auto-)immune signature in aplastic anemia Antonio M. Risitano

749

Hematopoietic stem cell mobilization with plerixafor in sickle cell disease Matthew M. Hsieh and John F. Tisdale

751

Age-related clonal hematopoiesis and monoclonal B-cell lymphocytosis / chronic lymphocytic leukemia: a new association?

Adalgisa Condoluci and Davide Rossi

Perspectives 753

Denosumab for bone lesions in multiple myeloma – what is its value?

Daniel A. Goldstein

755

Osteoprotective medication in the era of novel agents: a European perspective on values, risks and future solutions Monika Engelhardt

Articles Bone Marrow Failure

759

Deep sequencing and flow cytometric characterization of expanded effector memory CD8+CD57+ T cells frequently reveals T-cell receptor Vβ oligoclonality and CDR3 homology in acquired aplastic anemia Valentina Giudice et al.

Red Cell Biology & its Disorders

770

Safety and efficacy of plerixafor dose escalation for the mobilization of CD34+ hematopoietic progenitor cells in patients with sickle cell disease: interim results Farid Boulad et al.

778

Plerixafor enables safe, rapid, efficient mobilization of hematopoietic stem cells in sickle cell disease patients after exchange transfusion Chantal Lagresle-Peyrou et al.

787

RON kinase inhibition reduces renal endothelial injury in sickle cell disease mice Alfia Khaibullina et al.

Myeloproliferative Disorders

798

The KIT and PDGFRA switch-control inhibitor DCC-2618 blocks growth and survival of multiple neoplastic cell types in advanced mastocytosis Mathias Schneeweiss et al.

Acute Myeloid Leukemia

810

Distinct protein signatures of acute myeloid leukemia bone marrow-derived stromal cells are prognostic for patient survival Steven M. Kornblau et al.

822

Clinical relevance of IDH1/2 mutant allele burden during follow-up in acute myeloid leukemia. A study by the French ALFA group Yann Ferret et al.

Haematologica 2018; vol. 103 no. 5 - May 2018 http://www.haematologica.org/

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

830

MSH6 haploinsufficiency at relapse contributes to the development of thiopurine resistance in pediatric B-lymphoblastic leukemia Nikki A. Evensen et al.

Hodgkin Lymphoma

840

A phase II study of the oral JAK1/JAK2 inhibitor ruxolitinib in advanced relapsed/refractory Hodgkin lymphoma Eric Van Den Neste et al.

Non-Hodgkin Lymphoma

849

A B-cell receptor-related gene signature predicts survival in mantle cell lymphoma: results from the Fondazione Italiana Linfomi MCL-0208 trial Riccardo Bomben et al.

857

Incidence and risk factors for relapses in HIV-associated non-Hodgkin lymphoma as observed in the German HIV-related lymphoma cohort study Philipp Schommer et al.

Chronic Lymphocytic Leukemia

865

Highly similar genomic landscapes in monoclonal B-cell lymphocytosis and ultra-stable chronic lymphocytic leukemia with low frequency of driver mutations Andreas Agathangelidis et al.

874

Toxicities and outcomes of 616 ibrutinib-treated patients in the United States: a real-world analysis Anthony R. Mato et al.

Plasma Cell Disorders

880

A novel nano-immunoassay method for quantification of proteins from CD138-purified myeloma cells: biological and clinical utility Irena Misiewicz-Krzeminska et al.

890

Impact of extramedullary disease in patients with newly diagnosed multiple myeloma undergoing autologous stem cell transplantation: a study from the Chronic Malignancies Working Party of the EBMT Nico Gagelmann et al.

Hemostasis

898

Immobilized fibrinogen activates human platelets through glycoprotein VI Pierre H Mangin et al.

Coagulation & its Disorders

908

Circulating microRNAs as biomarkers of disease and typification of the atherothrombotic status in antiphospholipid syndrome Carlos Pérez-Sánchez et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/103/5.toc

e184

Iron overload in transfusion-dependent survivors of hemoglobin Bart’s hydrops fetalis.

Ali Amid et al. http://www.haematologica.org/content/103/5/e184

Haematologica 2018; vol. 103 no. 5 - May 2018 http://www.haematologica.org/

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation e188

A defined culture method enabling the establishment of ring sideroblasts from induced pluripotent cells of X-linked sideroblastic anemia Shunsuke Hatta et al. http://www.haematologica.org/content/103/5/e188

e192

The mutational landscape of 18 investigated genes clearly separates four subtypes of myelodysplastic/myeloproliferative neoplasms Manja Meggendorfer et al. http://www.haematologica.org/content/103/5/e192

e196

Clonal evolution in the transition from cutaneous disease to acute leukemia suggested by liquid biopsy in blastic plasmacytoid dendritic cell neoplasm Eleni Ladikou et al. http://www.haematologica.org/content/103/5/e196

e200

Heterozygous carriers of germline c.657_661del5 founder mutation in NBN gene are at risk of central nervous system relapse of B-cell precursor acute lymphoblastic leukemia Bartłomiej Tomasik et al. http://www.haematologica.org/content/103/5/e200

e204

Ibrutinib for chronic lymphocytic leukemia: international experience from a named patient program Peter Hillmen et al. http://www.haematologica.org/content/103/5/e204

e207

IGHV segment utilization in immunoglobulin gene rearrangement differentiates patients with anti-myelin-associated glycoprotein neuropathy from others immunoglobulin M-gammopathies Jean-Sebastien Allain et al. http://www.haematologica.org/content/103/5/e207

e211

Outcomes of patients with relapsed aggressive adult T-cell leukemia-lymphoma: clinical effectiveness of anti-CCR4 antibody and allogeneic hematopoietic stem cell transplantation Shigeo Fuji et al. http://www.haematologica.org/content/103/5/e211

e215

Sequential loss of tumor surface antigens following chimeric antigen receptor T-cell therapies in diffuse large B-cell lymphoma Haneen Shalabi et al. http://www.haematologica.org/content/103/5/e215

Case Reports Case Reports are available online only at www.haematologica.org/content/103/5.toc

e219

Novel hereditary spherocytosis-associated splice site mutation in the ANK1 gene caused by parental gonosomal mosaicism Xiong Wang, et al. http://www.haematologica.org/content/103/5/e219

e223

Severe hemolysis and transfusion reactions after treatment with BGB-3111 and PD-1 antibody for WaldenstrĂśm macroglobulinemia Jad Othman et al. http://www.haematologica.org/content/103/5/e223

e226

Usefulness of initial plasma dabigatran concentration to predict rebound after reversal Nicolas Gendron et al. http://www.haematologica.org/content/103/5/e226

Haematologica 2018; vol. 103 no. 5 - May 2018 http://www.haematologica.org/

EDITORIALS (Auto-)immune signature in aplastic anemia Antonio M. Risitano Hematology, Department of Clinical Medicine and Surgery, Federico II University, Naples, Italy E-mail: amrisita@unina.it doi:10.3324/haematol.2018.190884

A

cquired idiopathic aplastic anemia (IAA) is a rare hematologic disorder characterized by the failure of hematopoiesis secondary to an immune-mediated damage of the bone marrow. IAA is bona fide considered an auto-immune disease with a T-cell-mediated pathophysiology.1 In this issue of the Journal, Giudice et al.2 describe the oligoclonal pattern of effector memory CD8+ CD57+ T cells in IAA patients by using a combined deep sequencing and flow cytometry approach. Indeed, Giudice et al. show that clonally expanded T-cell populations are frequently detectable even within the effector memory compartment, and that they tend to correlate with disease activity. Thus, the characterization of the T-cell receptor (TCR) repertoire by high-resolution techniques may play a role in confirming the diagnosis of immune-mediated IAA and to monitor affected patients during their disease course.

There is widespread clinical and experimental evidence to support the autoimmune pathophysiology of IAA.1 The most striking is that patients with IAA may respond to Tcell-targeted immunosuppressive therapies (IST), with rates of hematologic responses ranging between 50% and 70%.3 Almost four decades of investigations have provided us with a plethora of experimental data corroborating the hypothesis of an immune-mediated pathophysiology. Increased circulating activated T cells were described in IAA patients in the ‘80s.4 These T cells may suppress hematopoiesis through the secretion of different inflammatory cytokines5,6 and/or via cell-mediated direct killing. Among the different inhibitory cytokines, interferon-γ (IFN-γ) plays a major role in suppressing human hematopoietic stem cells (HSC) in vivo, as suggested by in vitro inhibition of cell cycle progression and induction of apoptosis of hematopoietic progenitors.7 More recently, it has been suggested that IFN-γ may also exert its inhibitory effect on HSC impairing the homeostatic survival signal delivered by thrombopoietin through its cognate receptor c-MPL.8 This inhibitory milieu is generated by immune cells, and mostly by T cells that become activated and proliferate in response to an antigen-driven stimulation.

While the search for these putative antigens has remained unsuccessful, the demonstration of clonal expansion of T-cell populations identified by their TCR has been considered robust proof of a T-cell-mediated pathophysiology in IAA.9,10 Our growing understanding of the immune system and the availability of powerful novel techniques has nurtured continuous research in the field of IAA. On the one hand, investigators have tried to further dissect the abnormalities of the immune system in patients suffering from IAA. Indeed, looking at specific functional T-cell subsets, recurrent immune derangements have been found, such as increased T-helper type 17 cells (Th17)11 and reduced regulatory T cells (Treg).12,13 However, irrespective of the deep phenotyping of CD4+ T cells [i.e. by multiparameter mass cytometry, termed cytometry by time-of-flight (CyTOF)],14 haematologica | 2018; 103(5)

only limited data are available about the characterization of specific functional CD8+ T-cell subsets. On the other hand, novel techniques of deep DNA sequencing have become available, and their application in IAA has led to the description of somatic mutations in myeloid cells,15 with the subsequent ongoing debate about their actual meaning,16 but also to high throughput TCR analysis.17 In their study, Giudice et al. specifically investigated the compartment of effector memory T cells (TEM) in IAA patients, looking for possible clonality as assessed by flow cytometry analysis of Vβ usage and sequencing of the hypervariable complementary determining region 3 (CDR3) of the TCR. In agreement with previous reports,1,10 Giudice et al. confirm that AA patients often exhibit a skewed usage of Vβ families, usually within the CD8+ T-cell compartment; these oligoclonal expansions are more frequent in patients with increased percentage of TEM (as defined by co-expression of CD8 and CD57), which are found in approximately 70% of AA patients. These gross abnormalities of the TCR Vβ usage were dissected at the clonal level by deep sequencing of the TCR, which allows the comparison of more than 107 reads corresponding to TCRs harbored by individual T cells.

Indeed, clonal expansions were identified by repetitive use of TCR β variable (TRBV) and joint (TRJV) genes (i.e. redundant TRBV/TRBJ combinations), as well as by CDR3 size and DJ length profiles. Immunodominant clones within different T-cell subsets were invariably detected in all AA patients with increased CD8+ CD57+ TEM; however, the actual magnitude of these clonal expansions was extremely heterogeneous in the different T-cell subsets. Indeed, these clones remain minimally expanded (approx. 3%) within the CD4+ compartment, while they become largely dominant (approx. 18%) in the CD8+ compartment; the expansion appears even larger when the CD8+ CD57+ TEM is analyzed.

This difference in behavior of TCR heterogeneity in different T-cell subsets of AA patients was further confirmed by analysis of Simpson’s diversity score, which shows how TCR diversity decreased from CD4+ to CD8+ T cells, and even more from both CD4+ and CD8+ T cells to CD8+ CD57+ TEM. In summary, the work performed by Giudice et al. confirms that clonal CD8+ T-cell expansions are common in AA patients, in agreement with the well-established T-cell-mediated immune pathophysiology of AA.1 But for the first time, here the Authors provide evidence that the clonal expansions are not limited to the CD8+ effector T cells, since they can be found even in the TEM compartment. Very interestingly, with the caveat of a limited sample size, Giudice et al. show that abnormalities of the TEM compartment (i.e. increased TEM, with possible clonal expansions) may be associated with a dismal outcome in AA, due to refractory or relapsed disease. This is further supported by the observation of concordance between the expansion of the immunodominant CD8+ CD57+ TEM clone and disease activity (sim747

Editorials

ilar to what has already been shown for expansion within the bulk CD8+ compartment10).

Memory is the hallmark of adaptive immunity; indeed, antigen-driven clonal expansion of effector T cells may generate antigen-specific lymphocytes that may persist life long (reviewed by Sallusto et al.18). These cells, which are known as memory cells, confer immediate and effective protection against pathogens upon antigen rechallenging; indeed, memory immune cells are selected for their higher affinity for cognate antigens, and rapidly activate from their resting state after antigen stimulation.18 Among T cells, three subtypes of memory cells have been described:18 i) TEM, that carry the protective memory because of their ability to deliver effector functions; ii) central memory T cells (TCM), that home to secondary lymphoid organs and exert reactive memory through their cross-differentiation toward TEM and effector T cells (TEFF); and iii) memory stem T cells (TSCM), which have been described as a less differentiated subset that has better self-renewal and the ability to differentiate into distinct subsets of memory T cells. Only limited data are available about memory T cells in AA. In 2009, Hu et al. reported that TEM and TEMRA (a further subset of terminally differentiated TEM characterized by CD45RA expression and by better effector function) are increased in CD4+ and CD8+ T-cell subsets in AA patients, while naĂŻve T cells are decreased.19 More recently, the US National Institutes of Health group has described that even TSCM

are increased in circulating CD8+ T cells of AA patients.20 However, while these observations depict the broad immune derangement which is expected in an autoimmune disease such as AA, they do not provide any clues about the pathogenic role of these cells. In contrast, the demonstration of clonality within the TEM compartment shown by Giudice et al. seems pathogenically more relevant, according to the scenario depicted in Figure 1.

Indeed, a clonal immune response specific for (or crossreactive with) HSC can be elicited by unknown antigens or triggers, eventually causing some impairment of hematopoiesis. If these clonal TEFF do not undergo apoptosis (as usually occurrs in a physiological immune response), they may lead to overt AA.1 Given the plasticity of T cells, some of these clonal, antigen-specific cells may eventually acquire a TEM or a TCM functional phenotype, generating some skewing within the broad TEM repertoire, as found by Giudice et al.2 The presence of these clonal (possibly HSC-specific) TEM would account for continuous damage to the hematopoiesis, since in the presence of persistent antigen spread they may serve as a reservoir for newly-generated TEFF. Thus, irrespective of the regulatory role postulated by Giudice et al.,2 the presence of clonal TEM would represent the signature of a deeper immune derangement, possibly associated with a dismal clinical outcome.

In conclusion, the availability of high-throughput technologies is providing biomarkers which anticipate future

Figure 1. T-cell clonality in aplastic anemia. Unknown antigens and triggers may elicit a clonal immune response specific for (or cross-reactive with) hematopoietic stem cells (HSC). These clonal, activated T cells tend to expand delivering their immune damage over hematopoiesis. At the same time, some of these clonally expanded activated T cells may acquire an effector memory T cell (TEM) functional phenotype, leading to skewing of the TEM cell repertoire. While activated effector T cells (TEFF) may undergo anergy or apoptosis (even as a consequence of immunosuppressive therapies), TEM represent a continuous reservoir for HSC-specific T cells, which may exert their effector function upon rechallenging with the antigen. (This is very likely, given the typical antigen spread seen in autoimmune diseases.) Thus, the presence of clonal TEM, which likely share the same antigen-specificity with large TEFF clones found in bulk CD8+ populations, represent a biomarker of a deep-rooted immune derangement, possibly associated with a dismal disease course.

748

haematologica | 2018; 103(5)

Editorials

applications in the management of AA patients. Indeed, deep whole exome sequencing,15 CyTOF14 and deep TCR analysis2 all help to better describe the pathogenic events underlying bone marrow failure syndromes. Even if none of them translates into immediate therapeutic decisions, they are all useful to confirm the diagnosis, to determine the prognosis and possibly to monitor the clinical course of AA patients. Indeed, this latter application may be useful for early identification of refractory or relapsing patients, paving the way for pre-emptive therapeutic interventions. Moreover, the deep dissection at the clonal and at the functional levels of the immune T-cell compartment (e.g. combining CyTOF and TCR analysis) may also answer some open questions in the field. For example, the differential depletion of some specific T-cell subsets might explain the different outcome seen with different ATG preparations.3 These novel technologies may help identify the specific T-cell subsets which are crucial to the pathophysiology of AA (and possibly differentially depleted by distinct ATG brands), possibly driving the development of future targeted therapies.

References 1. Young NS. Current concepts in the pathophysiology and treatment of aplastic anemia. Hematology Am Soc Hematol Educ Program.

2013;2013:76-81.

2. Giudice V, Feng X, Lin Z, et al. Deep sequencing and flow cytometric characterization of expanded effector memory CD8+CD57+ T cells frequently reveals T-cell receptor Vbeta oligoclonality and CDR3 homology in acquired aplastic anemia. Haematologica. 2018;103(5): 759-769.

3. Scheinberg P, Nunez O, Weinstein B, et al. Horse versus rabbit antithymocyte globulin in acquired aplastic anemia. N Engl J Med. 2011;365(5):430-438.

4. Zoumbos NC, Gascón P, Djeu JY, Trost SR, Young NS. Circulating activated suppressor T lymphocytes in aplastic anemia. N Engl J Med. 1985;312(5):257-265.

5. Zoumbos NC, Gascón P, Djeu JY, Young NS. Interferon is a mediator of hematopoietic suppression in aplastic anemia in vitro and possibly in vivo. Proc Natl Acad Sci USA. 1985;82(1):188-192.

6. Sloand E, Kim S, Maciejewski JP, Tisdale J, Follmann D, Young NS.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Intracellular interferon-gamma in circulating and marrow T cells detected by flow cytometry and the response to immunosuppressive therapy in patients with aplastic anemia. Blood. 2002;100(4):11851191.

Selleri C, Maciejewski JP, Sato T, Young NS. Interferon-gamma constitutively expressed in the stromal microenvironment of human marrow cultures mediates potent hematopoietic inhibition. Blood: 1996;87(10):4149-4157.

Alvarado LJ, Andreoni A, Huntsman HD, et al. Heterodimerization of TPO and IFNγ Impairs Human Hematopoietic Stem/Progenitor Cell Signaling and Survival in Chronic Inflammation Blood.

2017;130(Suppl 1):4.

Zeng W, Nakao S, Takamatsu H, et al. Characterization of T-cell repertoire of the bone marrow in immune-mediated aplastic anemia: evidence for the involvement of antigen-driven T-cell response in cyclosporine-dependent aplastic anemia. Blood. 1999;93(9):30083016.

Risitano AM, Maciejewski JP, Green S, Plasilova M, Zeng W, Young NS. In-vivo dominant immune responses in aplastic anaemia: molecular tracking of putatively pathogenetic T-cell clones by TCR betaCDR3 sequencing. Lancet. 2004;364(9431):355-364. de Latour RP, Visconte V, Takaku T, et al. Th17 immune responses contribute to the pathophysiology of aplastic anemia. Blood.

2010;116(20):4175-4184.

Solomou EE, Rezvani K, Mielke S, et al. Deficient CD4+ CD25+ FOXP3+ T regulatory cells in acquired aplastic anemia. Blood.

2007;110(5):1603-1606.

Kordasti S, Marsh J, Al-Khan S, et al. Functional characterization of CD4+ T cells in aplastic anemia. Blood. 2012;119(9):2033-2043.

Kordasti S, Costantini B, Seidl T, et al. Deep phenotyping of Tregs identifies an immune signature for idiopathic aplastic anemia and predicts response to treatment. Blood. 2016;128(9):1193-1205.

Yoshizato T, Dumitriu B, Hosokawa K, et al. Somatic Mutations and Clonal Hematopoiesis in Aplastic Anemia. N Engl J Med. 2015;373(1):35-47.

Cooper JN, Young NS. Clonality in context: hematopoietic clones in their marrow environment. Blood. 2017;130(22):2363-2372.

Calis JJ, Rosenberg BR. Characterizing immune repertoires by high throughput sequencing: strategies and applications. Trends Immunol. 2014;35(12):581-590.

Sallusto F, Geginat J, Lanzavecchia A. Central memory and effector memory T cell subsets: function, generation, and maintenance. Annu Rev Immunol. 2004;22:745-763.

Hu X, Gu Y, Wang Y, Cong Y, Qu X, Xu C. Increased CD4+ and CD8+ effector memory T cells in patients with aplastic anemia.

Haematologica. 2009;94(3):428-429.

Hosokawa K, Muranski P, Fenx X, et al. Memory Stem T Cells in Autoimmune Disease: High Frequency of Circulating CD8+ Memory Stem Cells in Acquired Aplastic Anemia. J Immunol.

2016;196(4):1568-1578.

Hematopoietic stem cell mobilization with plerixafor in sickle cell disease Matthew M. Hsieh and John F. Tisdale Cellular and Molecular Therapeutics Section, Sickle Cell Branch, National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health (NIH), Bethesda, Maryland E-mail: matthewhs@nhlbi.nih.gov doi:10.3324/haematol.2018.190876

A

After more than a half-century since the molecular basis for sickle cell disease (SCD) was described by Linus Pauling and colleagues, we now possess the molecular tools to contemplate a one-time cure through genetic modification of autologous hematopoietic stem cells (HSC). For these promising gene transfer and gene editing strategies to become a reality, a sufficient number of HSC of high purity must be obtained. Filgrastim, or granulocyte colony-stimulating factor, mobilization and haematologica | 2018; 103(5)

apheresis is the standard method for HSC collection in healthy adult donors, yet this approach is associated with high rates of adverse events requiring hospitalization in SCD, including vaso-occlusive crises, multi-organ failure, and even death, prompting our call for a moratorium on its use for HSC mobilization in SCD.1 Thus, bone marrow harvesting is the default approach, with evidence supporting its utility in both animal models and in vitro studies utilizing patients’ material.2-4 However, bone marrow harvest749

Editorials

ing employed in an ongoing HSC gene therapy trial was recently recognized to result in suboptimal yields of high purity HSC at the end of collection and processing, along with substantial pain after each harvest, and most subjects required two or three harvests to yield sufficient cell doses for manufacturing.5,6 In this issue of the Journal, two groups of investigators report their results using a third approach to HSC collection in SCD through mobilization with an inhibitor of the CXCR4 chemokine receptor, plerixafor. Boulad et al. performed a dose escalation study of plerixafor among a total of 15 SCD patients at steady state.7 Ten of the patients were receiving concomitant treatment with hydroxyurea. Only a minority of patients in each cohort achieved the target of ≥30 CD34+ cells/μL at 12 h after the plerixafor injection: three out six at a dose of 80 μg/kg, one out of three at a dose of 160 μg/kg, and two out of six at a dose of 240 μg/kg. Two patients (15%) experienced a vaso-occlusive crisis during the study period – one each at 80 and 240 μg/kg. None of the patients underwent leukapheresis, thus attribution of these adverse events could be narrowed to plerixafor. On the other hand, LagreslePeyrou et al. reported the outcomes of three patients who received plerixafor at a dose of 240 μg/kg.8 All three patients received at least 2 months of red cell exchange transfusion to target a sickle hemoglobin (HbS) near 30% while hydroxyurea was discontinued. The peak CD34+ cell count reached >75 cells/μL at as early as 3 h after the injection. All three patients also underwent leukapheresis of 15 to 21 L, with a resulting total CD34+ cell yield of 4.5 to 5.8 x 106 cells/kg and a purity of 80% to 95%. No pain, vaso-occlusive crises, or sickle-related events were observed in these three patients.

While the number of patients is relatively small in both studies, important lessons relevant to autologous HSC mobilization and collection in SCD with plerixafor can be gleaned. The first lesson regards preparation of the patients. Specifically, stopping hydroxyurea and utilizing red cell transfusions, simple or exchange, to target a HbS of 30% were likely key factors in the successful mobilization of the series reported by Lagresle-Peyrou et al.

Conversely, the absence of these measures in the study

by Boulad et al. may explain why the majority of their patients failed to reach the target CD34+ concentration.

This is consistent with prior work demonstrating a lower CD34+ cell content in the marrow of SCD patients on hydroxyurea when compared to those not on the drug.3 Discontinuation of hydroxyurea combined with scheduled red cell transfusion to keep the HbS near 30% may also have improved purity, which was 80% to 95% in the study by Lagresle-Peyrou et al., while helping to minimize the risk of sickle cell-related adverse events while hydroxyurea treatment was interrupted. Secondly, leukocyte and neutrophil counts increased 2- to 3-fold just hours after a single injection of plerixafor, even at the lowest dose of 80 μg/kg tested. Although increases of a similar magnitude also occurred with filgrastim, the adverse events seen with filgrastim may have been related to the prolonged duration of 5 to 6 days from filgrastim that led to the high rates of adverse events in the ear-

750

lier reports. We will need more patients to ascertain the contribution of leukocytosis alone and/or the duration of leukocytosis in developing sickle-related complications.

Thirdly, only three patients underwent leukapheresis and though adverse events appear acceptable, expanded accrual could capture additional side effects. Furthermore, if patients with SCD do not meet the goal and need additional mobilization and collection, there could be cumulative side effects. Finally, the peak of mobilization of CD34+ cells appeared to be much earlier, at 3-6 hours.

This observation is distinct from that in healthy donors, in whom the peak is observed at 6-12 hours.9 Perhaps the chronically hyperproliferative marrow in SCD partly explains this early release of HSC; there could be other factors at play. Regardless, this observation suggests that for optimal collection, apheresis should be started within 4-6 hours of dosing.

As clinical applications of gene transfer and gene editing strategies are being implemented in SCD, obtaining adequate numbers of HSC safely from patients could be the ‘bottleneck’, preventing broad dissemination of these exciting approaches. The early results provide optimism that mobilization with plerixafor could be a safer and more efficacious alternative for HSC collection to either filgrastim mobilization or bone marrow harvesting, and provide general confidence for the further development of these promising approaches to a one-time cure for SCD.

©2017 NIH (National Institutes of Health)

References 1. Fitzhugh CD, Hsieh MM, Bolan CD, Saenz C, Tisdale JF. Granulocyte colony-stimulating factor (G-CSF) administration in individuals with sickle cell disease: time for a moratorium? Cytotherapy.

2009;11(4):464-471.

2. Hematti P, Tuchman S, Larochelle A, Metzger ME, Donahue RE, Tisdale JF. Comparison of retroviral transduction efficiency in CD34+ cells derived from bone marrow versus G-CSF-mobilized or G-CSF plus stem cell factor-mobilized peripheral blood in nonhuman primates. Stem Cells. 2004;22(6):1062-1069.

3. Uchida N, Fujita A, Hsieh MM, et al. Bone marrow as a hematopoietic stem cell source for gene therapy in sickle cell disease: evidence from Rhesus and SCD patients. Hum Gene Ther Clin Dev. 2017;28(3):136144.

4. Uchida N, Bonifacino A, Krouse AE, et al. Accelerated lymphocyte reconstitution and long-term recovery after transplantation of lentiviral-transduced rhesus CD34+ cells mobilized by G-CSF and plerixafor.

Exp Hematol. 2011;39(7):795-805.

5. Kanter J, Walters MC, Hsieh MM, et al. Interim results from a phase I/II clinical study of lenitoglobin gene therapy for severe sickle cell disease.

Blood. 2016;128(22):1176.

6. Leonard A, Bonifacino A, Dominical VM, et al. Bone marrow characterization in sickle cell disease: inflammation and stress erythropoiesis lead to suboptimal CD34 recovery compared to normal volunteer bone marrow. Blood. 2017;130(Suppl 1):966.

7. Boulad F, Shore T, van Besien K, et al. Safety and efficacy of plerixafor dose escalation for the mobilization of CD34+ hematopoietic progenitor cells in patients with sickle cell disease: interim results.

Haematologica. 2018;103(5):770-777.

8. Lagresle-Peyrou C, Lefrere F, Magrin E, et al. Plerixafor enables the safe, rapid, efficient mobilization of haematopoietic stem cells in sickle cell disease patients after exchange transfusion. Haematologica.

2018;103(5):778-786.

9. Pantin J, Purev E, Tian X, et al. Effect of high-dose plerixafor on CD34(+) cell mobilization in healthy stem cell donors: results of a randomized crossover trial. Haematologica. 2017;102(3):600-609.

haematologica | 2018; 103(5)

Editorials

Age-related clonal hematopoiesis and monoclonal B-cell lymphocytosis / chronic lymphocytic leukemia: a new association?

Adalgisa Condoluci and Davide Rossi Hematology, Institute of Oncology Research, and Oncology Institute of Southern Switzerland, Bellinzona, Switzerland E-mail: davide.rossi@eoc.ch doi:10.3324/haematol.2018.191098

A

gathangelidis et al.1 compared the mutational landscape of low-count monoclonal B-cell lymphocytosis (MBL), high-count MBL and highly stable chronic lymphocytic leukemia (CLL) with confirmed lack of progression after a very long follow up (>10 years). The Authors also studied the polymorphonuclear (PMN) fraction and germline DNA from buccal swabs of the same individuals. Whole genome sequencing was complemented with deep sequencing of targeted genes.

The sample size, along with the low coverage imposed by whole genome sequencing, are both limitations in efforts to discover yet unknown variants that might actually be recurrent in these conditions. While Agathangelidis et al. acknowledge these limitations, three major findings characterize their manuscript.1 First, lowcount MBL, high-count MBL and highly stable CLL share a similar genetic landscape and mutational signatures, which include the presence of mutations in known drivers associated with poor outcome, such as NOTCH1, SF3B1, POT1,2-4 indicating that these mutations are not sufficient to drive the aggressiveness of the disease by themselves. Second, the mutational landscape of paired PMN suggests that most of these patients carry a clonal hematopoiesis that could possibly be age-related. Third, a number of somatic mutations were found in both the MBL/CLL cells and PMN, supporting the idea that the MBL/CLL clone stemmed from a common ancestral

hematopoietic precursor that was able to participate in both lymphoid and myeloid differentiation. By documenting that the DNA from PMN was free from contamination by MBL/CLL DNA, for example, by using molecular minimal residual disease methods relying on the individual patient’s specific immunoglobulin gene rearrangements, the Authors have provided further evidence of this important finding which, although previously reported,5,6 had remained a subject of debate.

Several hematologic malignancies, including CLL, multiple myeloma (MM) and acute myeloid leukemia (AML), have well-defined precursor states that precede the development of overt cancer. CLL is always preceded by a high MBL count,7 MM is almost always preceded by monoclonal gammopathy of undetermined significance (MGUS),8 and at least a quarter of all patients with myelodysplastic syndromes (MDS) have disease that evolves into AML.9 Deep genomic sequencing of normal subjects revealed that during human aging, the expansion of 1 or more hematopoietic stem and progenitor cells (HSPC) will result in clones that will sustainably contribute more than others to the production of mature blood cells. Accordingly, age-related clonal hematopoiesis (ARCH) is defined as the expansion of HSPC clones, harboring specific, disruptive, and recurrent genetic variants, in individuals without clear diagnosis of hematologic malignancies.10 MDS are frequently preceded by ARCH.11

Figure 1. Hypothetical model of evolution from age-related clonal hematopoiesis to monoclonal B-cell lymphocytosis/chronic lymphocytic leukemia.

haematologica | 2018; 103(5)

751

Editorials

Some ARCH-related mutations can increase the risk for leukemia,12 while others possibly increase the risk for heart disease and diabetes.13 From a pathogenetic standpoint, the study of Agathangelidis et al.1 provides the proof of principle that ARCH may also associate with expansion of B-cell clones with CLL phenotype, and connects ARCH with MBL and CLL in a continuum of evolution from HSCP clones to mature B-cell clones (Figure 1), thus validating in vivo in patients the notion initially reported from mice studies that the propensity to generate clonal B cells has already been acquired at the HSCP stage.14 To robustly establish this association and to gain greater insight into the pathogenetics, larger cohorts of MBL and CLL patients should be investigated with the rigorous approach utilized by Agathangelidis et al.1 One of the long-term complications of chemoimmunotherapy in CLL is the development of treatmentrelated MDS/AML.15 Chemoimmunotherapy poses a strong selection bottleneck to HSCPs, and thus only the fittest HSCPs survive and repopulate after the stress of chemoimmunotherapy.16 HSCP fitness may be sustained by somatic mutations in the context of a preceding ARCH, and it is increasingly recognized as a risk factor for therapy-related MDS/AML.17 Among elderly patients who receive chemotherapy and develop therapy-related MDS/AML, most have ARCH before chemotherapy.

Consistently, ARCH associates with an increased rate of therapy-related AML/MDS.17 Following this line of evidence, the study by Agathangelidis et al.1 prompts investigation into whether the finding of an ARCH in CLL patients who receive chemoimmunotherapy is a risk factor for the development of therapy-related MDS/AML. If this is proved to be the case, given the availability of novel agents for the treatment of CLL that are not stressful for HSCP, ARCH might become a new biomarker for tailoring treatment in CLL.

References 1. Agathangelidis A, Ljungström V, Scarfò L, et al. Highly similar genomic landscapes in monoclonal B-cell lymphocytosis and ultrastable chronic lymphocytic leukemia with low frequency of driver

752

mutations. Haematologica. 2018;103(5):865-873.

2. Rossi D, Rasi S, Fabbri G, et al. Mutations of NOTCH1 are an independent predictor of survival in chronic lymphocytic leukemia.

Blood. 2012;119(2):521-529.

3. Wang L, Lawrence MS, Wan Y, et al. SF3B1 and other novel cancer genes in chronic lymphocytic leukemia. N Engl J Med. 2011;365(26):2497-2506.

4. Herling CD, Klaumünzer M, Rocha CK, et al. Complex karyotypes and KRAS and POT1 mutations impact outcome in CLL after chlorambucil-based chemotherapy or chemoimmunotherapy. Blood.

2016;128(3):395-404.

5. Damm F, Mylonas E, Cosson A, et al. Acquired initiating mutations in early hematopoietic cells of CLL patients. Cancer Discov.

2014;4(9):1088-1101.

6. Quijada-Álamo M, Hernández-Sánchez M, Robledo C, et al. Nextgeneration sequencing and FISH studies reveal the appearance of gene mutations and chromosomal abnormalities in hematopoietic progenitors in chronic lymphocytic leukemia. J Hematol Oncol.

2017;10(1):83.

7. Landgren O, Albitar M, Ma W, et al. B-cell clones as early markers for chronic lymphocytic leukemia. N Engl J Med. 2009;360(7):659667.

8. Landgren O, Kyle RA, Pfeiffer RM, et al. Monoclonal gammopathy of undetermined significance (MGUS) consistently precedes multiple myeloma: a prospective study. Blood. 2009;113(22):5412-5417.

9. Pfeilstöcker M, Tuechler H, Sanz G, et al. Time-dependent changes in mortality and transformation risk in MDS. Blood.

2016;128(7):902-910.

10. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371(26):2488-2498.

11. Malcovati L, Gallì A, Travaglino E, et al. Clinical significance of somatic mutation in unexplained blood cytopenia. Blood.

2017;129(25):3371-3378.

12. Genovese G, K¨ahler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-2487.

13. Jaiswal S, Natarajan P, Silver AJ, et al. Clonal hematopoiesis and risk of atherosclerotic cardiovascular disease. N Engl J Med. 2017;377(2):111-121.

14. Kikushige Y, Ishikawa F, Miyamoto T, et al. Self-renewing hematopoietic stem cell is the primary target in pathogenesis of human chronic lymphocytic leukemia. Cancer Cell. 2011;20(2):246259.

15. Benjamini O, Jain P, Trinh L, et al. Second cancers in patients with chronic lymphocytic leukemia who received frontline fludarabine, cyclophosphamide and rituximab therapy: distribution and clinical outcomes. Leuk Lymphoma. 2015;56(6):1643-1650.

16. 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.

17. Takahashi K, Wang F, Kantarjian H, et al. Preleukaemic clonal haemopoiesis and risk of therapy-related myeloid neoplasms: a casecontrol study. Lancet Oncol. 2017;18(1):100-111.

haematologica | 2018; 103(5)

PERSPECTIVES Denosumab for bone lesions in multiple myeloma – what is its value?

Daniel A. Goldstein Davidoff Cancer Center, Rabin Medical Center, Petach Tikvah, Israel E-mail: danielagoldstein@gmail.com doi:10.3324/haematol.2017.185264

I

n June 2017 the Food and Drug Administration (FDA) accepted a supplemental biologics license application seeking to expand the currently approved indication of denosumab to patients with bone lesions from multiple myeloma. The FDA set a prescription drug user act (PDUFA) action date of February 3, 2018. Denosumab is an inhibitor of receptor activator of nuclear factor κ-B ligand (RANKL) and was previously approved for postmenopausal women at risk of osteoporosis in addition to patients at risk of skeletal-related events due to bone metastases from solid tumors and giant cell tumors of the bone. The application for use in patients with myeloma is based on the findings of the recently presented ‘482 trial.1 This commentary seeks to understand the value of this therapy for patients with multiple myeloma.

Denosumab is a monoclonal antibody and uses a novel mechanism to decrease bone resorption. RANKL is a protein expressed on osteoblastic stromal cells. It binds to receptor activator of nuclear factor-κB (RANK) and thus mediates osteoclastic differentiation, activation, and survival. RANKL therefore controls osteoclast-mediated bone resorption. Osteoprotegerin is a soluble RANKL decoy receptor which binds RANKL and is the key regulator of the RANKL–RANK pathway. Denosumab binds to RANKL thus blocking the interaction of RANKL with RANK, mimicking the endogenous effects of osteoprotegerin. This agent has been shown to lead to a decrease in bone resorption, based on changes in serum and urinary N-telopeptide, which are markers of osteoclastic bone resorption.2 Until recently bisphosphonates had been the standard therapy for strengthening bone in a variety of conditions such as osteoporosis and cancer. Bisphosphonates essentially bind to bone mineral and inhibit the activity of mature osteoclasts. Non-nitrogen containing bisphosphonates achieve this goal by being metabolized to ATP analogs that block osteoclast function and induce osteoclast apoptosis. Nitrogen-containing bisphosphonates inhibit farnesyl pyrophosphate synthase, thus preventing the post-translational modification of guanosine triphosphate binding proteins which are essential for osteoclast function and survival.3 The essential difference between bisphosphonates and denosumab is that bisphosphonates inhibit mature osteoclasts while denosumab inhibits osteoclastic precursors.

Denosumab has already gained FDA approval for multiple indications based on advanced phase clinical trials.

In postmenopausal women with low bone mineral density, it was found to lead to a 3.0% to 6.7% increase in bone mineral density of the lumbar spine.2 Multiple trials have compared zoledronic acid and denosumab in patients with solid tumors. In patients with bone metastases from breast cancer, denosumab was superior to haematologica | 2018; 103(5)

zoledronic acid in delaying or preventing first on-study skeletal-related event [hazard ratio (HR)=0.82; 95% confidence interval (95% CI): 0.71- 0.95; P= 0.01).4 Likewise, denosumab was superior in terms of time to first skeletalrelated event in patients with bone metastases from prostate cancer. The median time to first on-study skeletal-related event was 20.7 months (95% CI: 18.8-24.9) with denosumab compared to 17.1 months (95% CI: 15.0-19.4) with zoledronic acid (HR=0.82, 95% CI: 0.710.95; P=0.008 for superiority).5 Despite the reduction in skeletal-related events with denosumab, there was no associated improvement in overall survival in patients with either breast or prostate cancer.4,5 In patients with giant cell tumors of the bone, an open label study with denosumab demonstrated a high level of efficacy: 96% of patients with surgically unsalvageable giant cell tumors of the bone did not have disease progression after a median follow-up of 13 months.6 The ‘482 trial was an international phase 3, randomized, double-blind trial comparing the safety and efficacy of monthly denosumab to monthly zoledronic acid in patients with multiple myeloma.1 The trial enrolled 1718 patients and the primary endpoint was the time to first on-study skeletal-related event, and was powered to demonstrate non-inferiority. Secondary endpoints were time to first skeletal-related event (powered to superiority), time to first and subsequent skeletal-related events (powered to superiority), and overall survival. The study met the primary endpoint and demonstrated that denosumab was non-inferior to zoledronic acid in terms of skeletal-related events (HR=0.98; 95% CI: 0.85-1.14; P=0.01). The trial failed to meet the secondary endpoints of demonstrating superiority in terms of time to first skeletal-related event or overall survival. The authors performed an unplanned exploratory analysis to evaluate progression-free survival as an endpoint and found a prolonged progression-free survival in the denosumab group (HR=0.82; 95% CI: 0.68-0.99; P=0.036). Although the trial was well conducted with double-blind randomization, this finding should be considered only as hypothesis-generating, as it was an unplanned endpoint analysis and such analyses are known to have a lack of statistical reliability.7 There were no significant differences between the two groups in terms of adverse events apart from hypocalcemia and renal toxicity. In patients with baseline creatinine clearance ≤60 mL/minute, 13% of patients in the denosumab group developed renal toxicity, compared to 26% of patients in the zoledronic acid group (P<0.01).

The rate of creatinine doubling from baseline in the zoledronic acid group was nearly twice as high as in the denosumab group (6.5 versus 3.3%). Conversely, there were higher rates of hypocalcemia in patients receiving deno753

Perspectives

sumab (17%) compared to those receiving zoledronic acid (12%) (P=0.009).

In the USA we can calculate the cost of the drugs to Medicare by using the average sales price (www.cms.gov).

This accounts for discounts and rebates and is a close estimate of the cost to Medicare. The patent for zoledronic acid expired in 2013, at which point the reimbursement cost decreased. The average sales price for 4 mg of zoledronic acid is $48 and that for 120 mg of denosumab is $2044. The annual cost is therefore $576 for zoledronic acid, and $24,528 for denosumab â&#x20AC;&#x201C; a difference of almost $24,000. In addition to this cost there is a mark-up of 4.3% that Medicare reimburses to providers. It should be noted that this mark-up may provide a financial incentive to the physician to prescribe the more expensive medication, despite the higher cost to the patient and insurer.

Finally, treatment centers also charge an infusion cost of approximately $140, billed with code 96413 (www.cms.gov). While these are the costs to Medicare, we must also recognize that the patient often shares a significant proportion of the cost. In 2015, the average annual Medicare beneficiary cost share was $527 for denosumab and $68 for zoledronic acid (www.cms.gov - 2015 Medicare drug spending data). The price of drugs is different in other countries around the world; however, it is clear that everywhere in the world zoledronic acid is significantly cheaper than denosumab. This commentary is not intended to assess what was the most appropriate launch price of these drugs at the very different times of their being launched. The purpose is to discuss the most appropriate choice of therapy in 2018, when the prices are significantly different, due to one of the options being available in the significantly cheaper, generic form.

There is some additional convenience from using denosumab. Firstly, denosumab can be given subcutaneously which may be preferable to the intravenous administration of zoledronic acid. Secondly, denosumab is dosed the same for all patients, and no adjustment is needed according to renal function, whereas dose adjustments are necessary for zoledronic acid. It is doubtful however, that this additional convenience justifies the additional annual cost in the USA of $24,000 per patient.

Recent data for zoledronic acid demonstrate equivalent efficacy in patients with bone metastases secondary to breast cancer, irrespective of whether the drug is given monthly or every 3 months.8 Could these data perhaps be extrapolated to patients with multiple myeloma? There are currently no good quality data regarding the use of denosumab every 3 months in patients with neoplastic bone disease.

In an era of financial challenges for healthcare, we, as physicians, must be careful stewards of finite healthcare resources. There appears to be no benefit from using denosumab instead of zoledronic acid in terms of overall

754

survival or skeletal events. In addition, the safety profile is very similar. There appears to be slightly more renal toxicity with zoledronic acid; however, this is balanced by the higher rates of hypocalcemia with denosumab.

Although there was a demonstration of benefit in terms of progression-free survival, this finding should be treated with caution, as it emerged from a post hoc exploratory analysis. There are, however, significant differences in costs â&#x20AC;&#x201C; both to society and to patients. Denosumab costs approximately $24,000 more per patient per year in the USA. Zoledronic acid is also considerably cheaper than denosumab in Europe as well. Perhaps the most appropriate management would be for all patients to receive zoledronic acid, except those with a contraindication due to a low creatinine clearance. The reason for the high cost of new cancer drugs is complex. Without doubt, one of the many reasons is that the cost of drug development is high, partially related to the many regulatory requirements. However, while cancer is still often an incurable disease, we must strive towards bringing forward new therapies that provide clinically meaningful benefits to our patients.9 In an era of medical bancruptcies and increasing healthcare costs, we owe it to both our patients and society to incorporate costs into clinical decision-making.

References 1. Raje NS, Roodman GD, Willenbacher W, et al. Impact of denosumab (DMB) compared with zoledronic acid (ZA) on renal function in the treatment of myeloma bone disease. J Clin Oncol.

2017;35(15_suppl):8005.

2. McClung MR, Lewiecki EM, Cohen SB, et al. Denosumab in postmenopausal women with low bone mineral density. N Engl J Med. 2006;354(8):821-831.

3. Baron R, Ferrari S, Russell RG. Denosumab and bisphosphonates: different mechanisms of action and effects. Bone. 2011;48(4):677692.

4. Stopeck AT, Lipton A, Body J-J, et al. Denosumab compared with zoledronic acid for the treatment of bone metastases in patients with advanced breast cancer: a randomized, double-blind study. J Clin Oncol. 2010;28(35):5132-5139.

5. Fizazi K, Carducci M, Smith M, et al. Denosumab versus zoledronic acid for treatment of bone metastases in men with castration-resistant prostate cancer: a randomised, double-blind study. Lancet.

2011;377(9768):813-822.

6. Chawla S, Henshaw R, Seeger L, et al. Safety and efficacy of denosumab for adults and skeletally mature adolescents with giant cell tumour of bone: interim analysis of an open-label, parallel-group, phase 2 study. Lancet Oncol. 2013;14(9):901-908.

7. Raghav KP, Mahajan S, Yao JC, et al. From protocols to publications: a study in selective reporting of outcomes in randomized trials in oncology. J Clin Oncol. 2015;33(31):3583-3590.

8. Hortobagyi GN, Van Poznak C, Harker WG, et al. Continued treatment effect of zoledronic acid dosing every 12 vs 4 weeks in women with breast cancer metastatic to bone: the OPTIMIZE-2 randomized clinical trial. JAMA Oncol. 2017;3(7):906-912.

9. Ellis LM, Bernstein DS, Voest EE, et al. American Society of Clinical Oncology perspective: raising the bar for clinical trials by defining clinically meaningful outcomes. J Clin Oncol. 2014;32(12):12771280.

haematologica | 2018; 103(5)

Perspectives

Osteoprotective medication in the era of novel agents: a European perspective on values, risks and future solutions Monika Engelhardt,1,2 Georg W. Herget,3 Giulia Graziani,1,2 Gabriele Ihorst,4 Heike Reinhardt,1,2 Stefanie Ajayi,1,2 Stefan Knop5 and Ralph Wasch1,2 1

Department of Medicine I, Hematology, Oncology & Stem Cell Transplantation, Medical Center, University of Freiburg; Comprehensive Cancer Center Freiburg (CCCF), Medical Center, University of Freiburg; 3Department of Orthopedics and Trauma Surgery, Medical Center, University of Freiburg; 4Clinical Trials Unit, Medical Center, University of Freiburg and 5 Hematology, Oncology, Gastroenterology, University of Würzburg, Germany 2

E-mail: monika.engelhardt@uniklinik-freiburg.de doi:10.3324/haematol.2018.188516

O

steolytic bone disease is one of the most prominent features of multiple myeloma (MM) and is present in up to 80% of patients at diagnosis.1 Bone destruction leads to skeletal-related events (i.e. vertebral and other pathological fractures) and/or spinal cord compression. MM is mainly due to an increase in osteoclastic activity which is accompanied by low osteoblastic function.1 Bisphosphonates and other bone-targeting agents (such as denosumab which inhibits RANKL and osteoclast function and is not renally cleared), effective anti-myeloma treatment, radiotherapy and surgery are the main therapies used for the management of bone disease in MM.1–3 Regarding the definition of MM-defining events, there are important studies which suggest that asymptomatic patients with more than one focal lesion detectable by magnetic resonance imaging have a higher risk of progression to symptomatic MM (>70% within 2 years).1,4–6 These patients have been described by international myeloma experts as having symptomatic disease.5,6 Based on phase 3 studies, the bisphosphonates, pamidronate and zoledronic acid, have been found to reduce skeletal-related events compared to placebo.7–9 Three randomized studies have compared the effect of different bisphosphonates or different dosages of the same bisphosphonate. In the first study, zoledronic acid was as effective as pamidronate in reducing skeletal-related events in the era of conventional chemotherapy.9,10 In the second, two doses of intravenous pamidronate (30 versus 90 mg) showed comparable results regarding time to skeletal-related events and survival time free of such events.11 The limitation of this study was that it was powered to show differences in quality of life and not in skeletal-related events.11 The third study compared intravenous (i.v.) zoledronic acid with oral clodronate and showed that zoledronic acid reduced the risk of skeletalrelated events compared to clodronate in all MM patients, irrespective of the presence of lytic lesions at diagnosis, and improved overall survival by 10 months in patients with lytic lesions at diagnosis.12,13 These effects continued in patients who received zoledronic acid for >2 years.14 There was no sub-analysis according to the response status of the patients, thus it is not clear whether the continuous use of zoledronic acid produces similar results in patients who have achieved excellent responses (≥very good partial response). A meta-analysis was unable to confirm superiority of zoledronic acid over pamidronate, but revealed a survival advantage from zoledronic acid versus placebo.15 This analysis also determined that in haematologica | 2018; 103(5)

order to prevent one skeletal-related event, 6-15 MM patients need to be treated.15 The European Myeloma Network (EMN) and International Myeloma Working Group (IMWG) have therefore recommended that all MM patients with adequate renal function (creatinine clearance >30 mL/min) and osteolytic disease at diagnosis should be treated with zoledronic acid [4 mg i.v. infusion, over at least 15 min, every 4 weeks (Q4W) or pamidronate (90 mg, in a 3-hour infusion, Q4W], in addition to specific anti-myeloma therapy (grade 1A; definition of evidence levels: Online Supplementary Table S1). Symptomatic patients, without bone disease assessed by conventional radiography, can be treated with zoledronic acid (grade 1B). The advantage is not clear for patients without detectable bone involvement on magnetic resonance imaging or positron emission tomography/computed tomography.

Bisphosphonates are not routinely recommended in smoldering MM (grade 1A); but in cases of osteoporosis or vertebral fractures that are not due to the MM, bisphosphonates should be given at the doses given for osteoporosis (5 mg zoledronic acid/year). For high-risk smoldering MM, the treating physician should consider using the bisphosphonate doses and schedules typically used to treat symptomatic MM (grade 1B). Zoledronic acid should be given continuously (grade 1B). It is debatable whether patients who achieve a very good partial response or better have benefits from the continuous use of zoledronic acid. Regarding pamidronate, there are no data to support its continuous use; thus it should be given for 2 years and then at the physician’s discretion (grade 2C).1,2 Of note, bisphosphonates are now available as generic drugs, whereas denosumab has just been approved by the Food and Drug Administration for use in MM (January 2018; likewise anticipated in Europe) and is patent-protected. This approval was based on the results of a large phase III study comparing denosumab with zoledronic acid, in which the efficacy and safety of the drugs were assessed in newly diagnosed MM. Eligible patients were randomized 1:1 to denosumab 120 mg subcutaneously Q4W or zoledronic acid 4 mg (with dose adjustments according to renal function) i.v. Q4W along with anti-myeloma therapy. The primary objective was non-inferiority of denosumab to zoledronic acid with respect to time to first on-study skeletal-related event.

Overall survival was a secondary endpoint; progressionfree survival was an exploratory endpoint. The 1718 patients enrolled were randomized into two arms, each 755

Perspectives

with 859 participants. With regards to delaying time to first on-study skeletal-related event, denosumab was not inferior to zoledronic acid [P=0.01; hazard ratio (HR)=0.98; 95% confidence interval (95% CI): 0.85-1.14].

Fewer adverse events potentially related to renal impairment were reported with denosumab than with zoledronic acid (10.0% versus 17.1%, P<0.001). The HR for progression-free survival was 0.82 (95% CI: 0.68-0.99; P=0.036). The overall survival HR between denosumab and zoledronic acid was 0.9 (95% CI: 0.70-1.16; P=0.41), with fewer deaths in the denosumab arm (n=121; 14.1%)

than in the zoledronic acid group (n=129; 15.0%).

Therefore, denosumab showed non-inferiority to zoledronic acid in delaying time to first on-study skeletalrelated event. Patients on denosumab had a significantly lower rate of renal adverse events compared to those on zoledronic acid. The bone-specific benefits in combination with the renal function results and possible prolongation of progression-free survival with denosumab were promising and have led to invigorating discussions about why progression-free survival data were more favorable with denosumab. The observations definitely need deep-

Table 1. Cost comparison of osteoprotective medications for MM: Germany versus USA.

Drug (original)

Costs Germany [Monthly costs in Euro (€)] Original price Generic price

Dose & mode of administration ®

Denosumab (Xgeva ) Zoledronic acid (Zometa®) Pamidronate

120 mg s.c. bolus 4 mg i.v. over 15 min 90 mg i.v. over 3 hours

440 368 out of trade

279 251

Costs USA [Monthly costs in Euro (€)] Original price Generic price 1890 814 out of trade

49 47

German prices: ATaxx® (Dr. Heni Software GmbH & Co.KG, Freiburg); USA prices: https://www.drugs.com/price-guide/. In Germany, the AMNOG is limiting the cost of new pharmaceutical products; In the USA, a deflation of generic prices has been reported, due to an increasing number of competing companies entering the market (https://www.nytimes.com/2017/08/08/health/generic-drugs-prices-falling.html).

A

C

B

D

Figure 1. Features of the diagnosis of multiple myeloma. (A) Time from onset of symptoms to the diagnosis of MM: patients (%) diagnosed within <3, 3-6, 6-11 or >12 months in the retrospective versus prospective analysis. (B) First suspicion of MM: frequency (%) of patients whose MM was first suspected by different types of physicians (n=176 patients; prospective cohort). (C-D) Prospective analysis: patients‘ satisfaction (n=176).

756

haematologica | 2018; 103(5)

Perspectives

er understanding and statistical evaluation: the relationship between the occurrence of skeletal-related events and progression-free survival-defining events needs to be defined. Furthermore, an assessment of cumulative incidence rates of skeletal-related events with death as a competing event will be helpful, as the slight overall survival disadvantage in the zoledronic acid arm might have led to fewer skeletal-related events. Nevertheless, with full publication of the results16 and with EMA approval, denosumab will be used in Europe in MM patients.1–3 Given these insights and major advances in the understanding of the disease, early MM diagnosis, especially of symptomatic patients, has been advocated. However, since MM is an insidiously developing malignancy and may appear with non-specific symptoms, e.g. bone pain, the diagnosis and therapeutic decisions can be complicated. A German study group (DSMM) and EMN project addressed this aspect with the aim of optimizing the prompt diagnosis and further improving the quality of MM care.17 An initial retrospective analysis of 101 MM patients was followed by a prospective study of 176 patients using a structured MM-specific questionnaire.

The median time from the patients' first symptoms to the final MM diagnosis was 4 months (range, 0.5-120) in the retrospectively studied cohort and very similar to the 6 months (range, 0.5-60) in the prospective cohort. Of interest, the time from onset of symptoms to diagnosis of MM was ≥12 months in 20% of the patients in the retrospective analysis and 35% in the prospective study (Figure 1A). The frequencies of MM-related bone fractures, renal complications and infections occurring before the diagnosis of MM was made were 41%, 35% and 16%, respectively. Moreover, 43% had one, 20% had two and 3% had three of these complications. The most frequent symptom was bone pain, which occurred in 73% of MM patients before the final MM diagnosis was made. In 6% of patients, MM was first suspected by orthopedists, whereas the clinical suspicion was raised by nephrologists in 16% of cases, even though renal impairment was less frequent (Figure 1B). Of interest, 61% of patients were completely or fairly satisfied with the diagnostic process, whereas 39% were less satisfied (Figure 1C).

Fifty-eight percent of the patients believed that their disease could have been diagnosed more expeditiously (Figure 1D). Patients, who criticized the slow diagnostic process had a much longer median time interval from symptom onset to their final MM diagnosis compared to those who were less critical (9 versus 3 months, respectively). These results demonstrate that there is still considerable latency in the diagnosis of MM. However, even with early diagnosis and treatment with novel agents, skeletal-related events continue to occur, in part due to MM responses ("melting-down MM") and relapses, reminding us that progress in MM involves understanding how best to avoid skeletal-related events before the diagnosis of the disease is made and with antimyeloma treatment, because this substantially influences patients' coping and their approval of our MM care.2,3,17 The notion that treatment based on novel agents promotes bonehealing - apart from osteoprotective supportive agents such as bisphosphonates and denosumab - has recently led to the demanding discussion18,19 of whether bonehaematologica | 2018; 103(5)

seeking agents are currently needed. Since skeletal-related events continue to occur in the first months of treatment and with relapse (despite the use of novel agents and osteoprotection1–3), effective prevention and reduction of destructive skeletal-related events remain fundamental.1,20 Recent data from the national registry, Hospital Episode Statistics determined fracture rates and the effect on overall survival in MM patients between 2001 and 2015: expectedly, fracture rates were 18 times higher with MM in the first year after admission than in the general population, and remained elevated for up to 10 years. In line with the data on early diagnosis in MM,2,3,17 the increased fracture risk preceded the first admission with MM and conversely the incidence of MM increased after admission with one or more fractures. Fractures were associated with poorer outcome (HR for overall survival: 1.2), indicating the need for regular use of bone supportive drugs despite novel agent-based treatment.21 In addition, cost analyses in 1028 MM patients (596 with ≥1 skeletalrelated events and 432 without skeletal-related events) demonstrated that a higher frequency of skeletal-related events was associated with greater utilization of healthcare resources, suggesting that bone supportive drugs need to be used diligently to avoid higher healthcare costs due to skeletal complications and patients’ discontent.22 Since bisphosphonates in symptomatic MM have been suggested, but beyond 2 years and with stable MM are left to the discretion of the treating physician, a randomized trial assessed 170 untreated, symptomatic patients using zolendronic acid for 4 versus 2 years.23 All patients were treated with the same induction therapy and stemcell transplantation. The group treated for 4 years had substantially fewer skeletal-related events than the group treated for 2 years (21 versus 43%, respectively; P<0.001).

Actuarial curves at 5 years showed that progression-free survival was 75% (95% CI: 64%-82%) and overall survival 68% (95% CI, 60%-76%) in the group treated for 4 years; these rates were not significantly different from those of the control group treated for 2 years with zoledronic acid (P=0.67); but this trial was underpowered to show differences in survival. The trial did, however, confirm that the continued use of zoledronic acid was useful to reduce skeletal-related events and to preserve a better quality of life.23 With bisphosphonates and denosumab being potent options in MM, Goldstein, in this issue of the Journal, comments on both costs and the fact that novel patentprotected drugs will induce greater expenditure than generically available alternatives.24 While there is an unequivocal need to thoroughly evaluate and measure "real" advances with new drugs, shortcomings of this commentary are the "generalization" regarding patent versus generic medications, the understatement of progression-free survival differences, convenience of subcutaneous versus i.v. medication, and the decreased renal impairment and safer use of denosumab in patients with renal impairment. Moreover, Goldstein’s conclusions only apply to the health system in the USA, whereas reimbursement of medication providers and financial incentives to physicians to prescribe more expensive drugs are different in Europe (Table 1).24 The rising cost of patented cancer medicines in the USA is a known phe757

Perspectives

nomenon: from 2000 until now, there have been 5- to 10fold increases in the cost of new drugs.25 Price differences in Europe can, therefore, be substantially different from those in the USA (Table 1). In Germany, a benefit assessment of pharmaceuticals in accordance with the Act on the Reform of the Market for Medicinal Products (AMNOG) is limiting the cost of new pharmaceutical products. In the USA, a deflation of generic prices has been reported, due to an increasing number of competing companies entering the market. As clinicians and researchers we know that medical advances are needed, and developmental costs for new drugs have steadily increased with regulatory requirements. Goldstein reminds us of the financial burden associated with new cancer agents, which we offer to our patients with the aim to 'never harm but always aid': nevertheless, the judgement regarding bisphosphonate generics versus denosumab is biased and the comparison is cumbersome. His conclusions that generic bisphosphonates have a novel counterpart and that the financial burden with denosumab is higher are, however, worth noting.24 Thus, the good news prevails that treatment options for the prevention of bone complications have increased with the introduction of denosumab, providing a new choice for patients and physicians. Once initiated, bisphosphonates or denosumab should be continued for at least 2 years, after which a suspension of bisphosphonate treatment may be considered in very responsive patients, although this may be associated with skeletal risks, especially in those with prior skeletal complications. For patients in whom bisphosphonates were stopped after 2 years, the drug should be resumed on a monthly basis if the MM recurs and/or new skeletal-related events occur, independently of the use of novel agent-based therapies.1–3 Acknowledgment We are deeply indebted to esteemed experts of the Deutsche Studiengruppe Multiples Myelom, German Multiple Myeloma Study Group, European Myeloma Network Group and International Myeloma Working Group for their valuable discussion.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

References 1. Terpos E, Kleber M, Engelhardt M, et al. European Myeloma Network Guidelines for the management of multiple myeloma-related complications. Haematologica. 2015;100(10):1254-1266.

2. Terpos E, Christoulas D, Gavriatopoulou M. Biology and treatment of myeloma related bone disease. Metabolism. 2018;80:80-90.

3. Yee AJ, Raje NS. Denosumab for the treatment of bone disease in solid tumors and multiple myeloma. Future Oncol. 2018;14(3):195203.

4. Hillengass J, Fechtner K, Weber M-A, et al. Prognostic significance of focal lesions in whole-body magnetic resonance imaging in patients with asymptomatic multiple myeloma. J Clin Oncol.

2010;28(9):1606-1610.

5. 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.

6. Caers J, Fernández de Larrea C, Leleu X, et al. The changing land-

758

21.

22.

23.

24.

25.

scape of smoldering multiple myeloma: a European perspective.

Oncologist. 2016;21(3):333-342.

Berenson JR, Lichtenstein A, Porter L, et al. Efficacy of pamidronate in reducing skeletal events in patients with advanced multiple myeloma. Myeloma Aredia Study Group. N Engl J Med. 1996;334(8):488-493.

Berenson JR, Rosen LS, Howell A, et al. Zoledronic acid reduces skeletal-related events in patients with osteolytic metastases.

Cancer. 2001;91(7):1191-1200.

Rosen LS, Gordon D, Kaminski M, et al. Zoledronic acid versus pamidronate in the treatment of skeletal metastases in patients with breast cancer or osteolytic lesions of multiple myeloma: a phase III, double-blind, comparative trial. Cancer J. 2001;7(5): 377-387.

Rosen LS, Gordon D, Kaminski M, et al. Long-term efficacy and safety of zoledronic acid compared with pamidronate disodium in the treatment of skeletal complications in patients with advanced multiple myeloma or breast carcinoma: a randomized, double-blind, multicenter, comparative trial. Cancer. 2003;98(8):1735-1744.

Gimsing P, Carlson K, Turesson I, et al. Effect of pamidronate 30 mg versus 90 mg on physical function in patients with newly diagnosed multiple myeloma (Nordic Myeloma Study Group): a double-blind, randomised controlled trial. Lancet Oncol. 2010;11(10):973-982.

Morgan GJ, Davies FE, Gregory WM, et al. First-line treatment with zoledronic acid as compared with clodronic acid in multiple myeloma (MRC Myeloma IX): a randomised controlled trial. Lancet.

2010;376(9757):1989-1999.

Morgan GJ, Child JA, Gregory WM, et al. Effects of zoledronic acid versus clodronic acid on skeletal morbidity in patients with newly diagnosed multiple myeloma (MRC Myeloma IX): secondary outcomes from a randomised controlled trial. Lancet Oncol.

2011;12(8):743-752.

Morgan GJ, Davies FE, Gregory WM, et al. Effects of induction and maintenance plus long-term bisphosphonates on bone disease in patients with multiple myeloma: the Medical Research Council Myeloma IX Trial. Blood. 2012;119(23):5374-5383.

Mhaskar R, Redzepovic J, Wheatley K, et al. Bisphosphonates in multiple myeloma: a network meta-analysis. Cochrane Database Syst Rev. 2012;(5):CD003188.

Raje N, Terpos E, Willenbacher W, et al. An international, randomised, double-blind study of denosumab compared to zoledronic acid in bone disease treatment of newly diagnosed multiple myeloma. Lancet Hematol. 2018;19(3):370-381.

Graziani G, Herget G, Ihorst G, et al. Time from first symptom onset to the final diagnosis of multiple myeloma - possible risks and future solutions: large retrospective and conformatory prospective “Deutsche Studiengruppe Multiples Myelom” (DSMM) analysis.

Blood. 2017;130(Suppl 1):4710.

Delforge M, Terpos E, Richardson PG, et al. Fewer bone disease events, improvement in bone remodeling, and evidence of bone healing with bortezomib plus melphalan-prednisone vs. melphalanprednisone in the phase III VISTA trial in multiple myeloma. Eur J Haematol. 2011;86(5):372-384.

Mohty M, Malard F, Mohty B, Savani B, Moreau P, Terpos E. The effects of bortezomib on bone disease in patients with multiple myeloma. Cancer. 2014;120(5):618-623.

Engelhardt M, Terpos E, Kleber M, et al. European Myeloma Network recommendations on the evaluation and treatment of newly diagnosed patients with multiple myeloma. Haematologica.

2014;99(2):232-242.

McIlroy G, Mytton J, Evison F, et al. Increased fracture risk in plasma cell dyscrasias is associated with poorer overall survival. Br J Haematol. 2017;179(1):61-65.

Nash Smyth E, Conti I, Wooldridge JE, et al. Frequency of skeletalrelated events and associated healthcare resource use and costs in US patients with multiple myeloma. J Med Econ. 2016;19(5):477-486.

Avilès A, Nambo M-J, Huerta-Guzmàn J, Cleto S, Neri N. Prolonged use of zoledronic acid (4 Years) did not improve outcome in multiple myeloma patients. Clin Lymphoma Myeloma Leuk. 2017;17(4):207210.

Goldstein D. Denosumab for bone lesions in multiple myeloma what is its value? Haematologica. 2018(5):753-754.

Kantarjian H, Steensma D, Rius Sanjuan J, Elshaug A, Light D. High cancer drug prices in the United States: reasons and proposed solutions. J Oncol Pract. 2014;10(4):e208-211.

haematologica | 2018; 103(5)

ARTICLE

Bone Marrow Failure

Deep sequencing and flow cytometric characterization of expanded effector memory CD8+CD57+ T cells frequently reveals T-cell receptor Vβ oligoclonality and CDR3 homology in acquired aplastic anemia

Ferrata Storti Foundation

Valentina Giudice,1 Xingmin Feng,1 Zenghua Lin,1,2 Wei Hu,3 Fanmao Zhang,3 Wangmin Qiao,3 Maria del Pilar Fernandez Ibanez,1 Olga Rios,1 and Neal S. Young1

Hematology Branch, National Heart, Lung, and Blood Institute (NHLBI), NIH, Bethesda, MD, USA; 2Department of Hematology, Affiliated Hospital of Nantong University, Jiangsu, China and 3BGI Genomics, BGI-Shenzhen, China

1

Haematologica 2018 Volume 103(5):759-769

ABSTRACT

O

ligoclonal expansion of CD8+CD28– lymphocytes has been considered indirect evidence for a pathogenic immune response in acquired aplastic anemia. A subset of CD8+CD28– cells with CD57 expression, termed effector memory cells, is expanded in several immune-mediated diseases and may have a role in immune surveillance.

We hypothesized that effector memory CD8+CD28–CD57+ cells may drive aberrant oligoclonal expansion in aplastic anemia. We found CD8+CD57+ cells frequently expanded in the blood of aplastic anemia patients, with oligoclonal characteristics by flow cytometric Vβ usage analysis: skewing in 1-5 Vβ families and frequencies of immunodominant clones ranging from 1.98% to 66.5%. Oligoclonal characteristics were also observed in total CD8+ cells from aplastic anemia patients with CD8+CD57+ cell expansion by T-cell receptor deep sequencing, as well as the presence of 1-3 immunodominant clones. Oligoclonality was confirmed by T-cell receptor repertoire deep sequencing of enriched CD8+CD57+ cells, which also showed decreased diversity compared to total CD4+ and CD8+ cell pools. From analysis of complementarity-determining region 3 sequences in the CD8+ cell pool, a total of 29 sequences were shared between patients and controls, but these sequences were highly expressed in aplastic anemia subjects and also present in their immunodominant clones. In summary, expansion of effector memory CD8+ T cells is frequent in aplastic anemia and mirrors Vβ oligoclonal expansion. Flow cytometric Vβ usage analysis combined with deep sequencing technologies allows high resolution characterization of the T-cell receptor repertoire, and might represent a useful tool in the diagnosis and periodic evaluation of aplastic anemia patients. (Registered at clinicaltrials.gov identifiers: 00001620, 01623167, 00001397, 00071045, 00081523, 00961064)

fengx2@nhlbi.nih.gov

Introduction

©2018 Ferrata Storti Foundation

Acquired aplastic anemia (AA) is a bone marrow (BM) failure syndrome characterized by peripheral blood (PB) pancytopenia and BM hypocellularity due to hematopoietic stem and progenitor cell destruction.1-7 There is indirect evidence to support the hypothesis of autologous immune attack: clinical response to immunosuppressive therapies (IST),1,2,8-10 the predominant role of activated cytotoxic T cells (CTLs) in BM growth inhibition,1,10-12 identification of putative autoantigens,13,14 and oligoclonal expansion of CD8+ lymphocytes.8,14-17 Oligoclonality of T-cell populations has been defined by flow cytometry and deep sequencing of the T-cell receptor (TCR) Vβ repertoire and by spectratyping of complementarity-determining region 3 (CDR3) length skewing.18-21 The TCR, an aβ or γd heterodimer, is responsible for antigen recognition and T-cell activation.22Haematologica | 2018; 103(5)

Correspondence:

Received: July 21, 2017.

Accepted: December 30, 2017.

Pre-published: February 1, 2018. doi:10.3324/haematol.2017.176701 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/759

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.

759

V. Giudice et al.

26

Each chain is the result of a complex gene locus rearrangement, known as VDJ recombination.23 On a rearranged VDJ segment, a terminal deoxynucleotidyl transferase enzyme increases TCR variability through insertions/deletions within a hypervariable region, the CDR3, generating a unique potential antigen-specific sequence.23 Early in infection, CD28+ CTLs with different antigen affinity are selected and expanded (polyclonal phase).27 In late stages, high antigen-affinity CD28– T cells are in resting state as memory T cells.28-29 The expression of CD57 on CD8+CD28– T cells identifies a subset of memory T cells termed effector memory because of their high antigen-affinity and ability to be rapidly activated after antigen stimulation, as “tissue-guards”.29,30 Direct evidence of their high antigen-affinity is decreased diversity in the CD8+CD57+ TCR repertoire (oligoclonality) due to the presence of only a few selected clones of memory cells.27 In this work, we investigated the frequency and oligoclonal expansion of effector CD28–CD57+ memory cells in CD4+ and CD8+ T cells and the TCR Vβ repertoire in AA patients by flow cytometry and deep sequencing technologies, to provide additional evidence for the immune hypothesis in AA pathophysiology.

Methods Human samples Heparinized whole PB was collected from patients and healthy subjects after informed consent, in accordance with the Declaration of Helsinki31 and protocols approved by the National Heart, Lung, and Blood Institute Institutional Review Board (National Institutes of Health, Bethesda, MD, USA) (see Online Supplementary Table S1 for clinical characteristics). HLA haplotypes are reported in Online Supplementary Table S2. All patients received a diagnosis of severe AA (SAA) according to the International Study of Aplastic Anemia and Agranulocytosis32 and to the criteria of Camitta.33 At the time of blood sampling, none of the patients had received therapy. Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque density gradient centrifugation (MP Biomedicals, LLC, Santa Ana, CA, USA) according to the manufacturer’s instructions.

Flow cytometry A minimum of 4x106 PBMCs from each subject were stained for TCR Vβ repertoire analysis (Online Supplementary Figure S1). The manufacturer’s instructions for the IOTest Beta Mark (Beckman Coulter, Miami, FL, USA) were optimized to avoid incorrect compensation due to the existence of FITC/PE double positive antibodies in the kit, as described in Online Supplementary Methods.

Expansion of CD8+CD57+ cells was defined using as a threshold the mean frequency calculated in healthy subjects. Vβ skewing in SAA patients was described when the frequency of one Vβ family was higher than the mean+3Standard Deviation (SD) in healthy subjects.16 The term “immunodominant clone” was defined as the Vβ+ population by flow cytometry or the DNA sequence by deep sequencing present in the highest abundance.17

TCR repertoire deep sequencing For VDJ combination and CDR3 sequence profiling,34 DNA was isolated from FACS- or beads- (Miltenyi Biotec Inc., San Diego, CA, USA) sorted CD4+ and CD8+ T cells from 12 SAA patients with CD8+CD57+ cell expansion and 9 healthy subjects (mean: 3.2 μg of DNA; range: 0.2-27.4 μg) (Online Supplementary 760

Table S2). DNA was also isolated from beads-sorted CD8+CD57+ cells from 2 of the 12 SAA patients with enough cells for further analysis (mean: 1.6 μg of DNA). TCR repertoire sequencing was performed with an Illumina HiSeq 2000 sequencer (Illumina Inc., San Diego, CA, USA). Detailed information is provided in the Online Supplementary Methods. Data have been deposited in the NCBI GEO database (accession n. GSE101660).

Statistical analysis Data were analyzed using R (RStudio, Boston, MA, USA) and Prism (v.7.02; GraphPad software Inc., La Jolla, CA, USA). MannWhitney U test, Wilcoxon signed rank sum test, pair and unpaired t-tests, or χ2 test were used for data with abnormal distributions. Bonferroni and Dunn’s corrections were used for multiple comparisons. P≤0.05 was considered statistically significant, after adjustment with Bonferroni and false discovery rate (FDR).35 Linear regression was performed for correlations. Log-rank (Mantel-Cox) test was used for progression-free survival data analysis. Simpson's diversity index was calculated according to the following formula:

in which ni represents the clone size as the number of copies of each clonotype (i) or the total number of CDR3 sequences belonging to each i clonotype, and n is the total number of different clonotypes in the sample or the total number of sequences for each sample.

Results Effector memory CD8+CD57+ T cells frequently show oligoclonal expansion of TCR Vβ repertoire by flow cytometry Immunophenotyping and flow-cytometry Vβ usage were performed in 24 SAA patients. Clinical characteristics are reported in Online Supplementary Table S1. A group of 34 healthy subjects was studied in order to define normal ranges of T-cell populations and Vβ family expression.

SAA patients showed higher frequencies of CD8+CD57+ cells (25.6±17.3% vs. 13.3±12.6% in healthy individuals; P=0.003), and decreased frequency of CD8+CD28+ cells (56.8±25.7% vs. 68.8±19.1%; P=0.046). A negative correlation between CD57 and CD28 expression was also present (r2=0.601, P<0.0001). No differences were found for CD28+ and CD57+ cells within the CD4+ subset (P=0.974 and P=0.250, respectively) (Figure 1A).

By Vβ usage, polyclonal expansion was observed in total CD4+, CD4+CD28+, CD4+CD57+ and CD8+CD28+ cells in both healthy subjects and SAA patients (Figure 1B and Online Supplementary Figure S2). Oligoclonal expansion of CD8+CD57+ cells was present in 92% of SAA patients with 1-3 immunodominant clones, while in total CD8+ cells oligoclonality was reported only in 33% of cases (Online Supplementary Figure S3). Patients did not show expansion of a shared Vβ family, as each subject carried a different TCR Vβ rearrangement in effector memory CD8+ T cells (Figure 2A). None of the 7 patients without CD8+CD57+ cell expansion showed Vβ skewing in any subgroup, with mean frequency of the immunodominant clone of 3.8% (range: 0.21-6.01%). Conversely, all 17 patients with effector memory CD8+ cell expansion haematologica | 2018; 103(5)

TCR repertoire of effector memory T cells in AA

showed Vβ skewing in 1-5 Vβ subgroups, and frequencies of the immunodominant clones ranged from 2.1% to 66.5% (mean: 9.9%).

Progression-free survival (PFS) analysis was performed on all SAA patients, divided according to pre-treatment frequencies of effector memory CD8+ T cells (cut-off value 13.3%) (Figure 2B). No patients with low CD8+CD57+ cell

frequency (n=7) experienced relapse (median survival not reached; median follow-up time: 13.6 months), while 7 of the 17 SAA subjects with expanded effector memory T cells relapsed or died (median survival 13.2 months; median follow up: 10.4 months). However, statistical significance between the two curves was not reached (P=0.089).

Vβ usage was analyzed in CD4+ and CD8+ T-cell subsets

A **

B

Figure 1. Immunophenotyping and flow cytometry analysis of Vβ usage in severe aplastic anemia (SAA) patients and healthy subjects. (A) Percentages of CD28+ and CD57+ cells were calculated in both CD4+ and CD8+ compartments for healthy controls and SAA patients. Data are shown as mean±Standard Deviation (SD).

Unpaired t-test was performed. *P<0.05; **P<0.01. (B) Vβ usage was studied in T-cell compartments (by row), and percentages of each Vβ family were reported as total CD4+ or CD8+ cell percentage. For Vβ usage in healthy subjects, data are shown as mean+SD, combining the results from all 34 healthy donors. For SAA patients, 2 representative cases are shown.

haematologica | 2018; 103(5)

761

V. Giudice et al.

in serial samples available for Patients 4, 22, and 34 in order to understand the correlation of Vβ usage with clinical course (Figure 2C and Online Supplementary Figure S4A). For Patient 22, at baseline, Vβ was the immunodominant clone and mostly enriched in CD8+CD57+ cell population, rather than in CD4+CD28+, CD4+CD57+, and CD8+CD28+ cells. After ten days of IST, the size of the CD8+CD57+ clone was greatly reduced. The clone was detected again at six months (3 months before clinical relapse) and further increased at relapse (Figure 2C), suggesting association of Vβ expansion with clinical status.

Patients 4 and 34 were non-responders at three months, and non- and minimal partial-responders at the 6-month time point. However, no significant changes in immunodominant clone size were observed (Online

Supplementary Figure S4A). Vβ usage was also investigated at baseline in the BM of these 2 patients (Online Supplementary Figure S4B), and high concordance with Vβ usage of PB CD8+CD57+ cells was described.

Increased expansion of effector memory CD8+ T cells with age has been reported;36 therefore, we used a pool of age-matched healthy controls to assess the effect of age.

There was no correlation between the size of the immunodominant clone in CD8+CD57+ cells and age in healthy subjects (r2=0.0003, P=0.919) or in SAA patients (r2=0.140, P=0.079) (Online Supplementary Figure S5). However, a correlation was found between CD57 expression and age in SAA patients (r2=0.552, P<0.0001).

We then assessed the effect of transfusion history on oligoclonal expansion of effector memory T cells, as trans-

A

B

C

Figure 2. Vβ usage at diagnosis and during treatment. (A) Percentages of Vβ family in CD8+CD57+ cells were calculated on total CD8+ cells, and Vβ skewing in severe aplastic anemia (SAA) patients was defined using the mean+3Standard Deviation (SD) of a given Vβ group in healthy donors. Relative expansion of each Vβ subgroup is shown in the bar graph. Patients were divided based on the absence or presence of expanded CD8+CD57+ cells, using the mean in healthy donors (13.3%). Skewing of one Vβ family is reported as an orange bar. (B) Progression-free survival rate of SAA patients with CD8+CD57+ cells ≤13.3% (n=7) or >13.3% (n=17) prior to treatment. Log-rank (Mantel-Cox) test was performed. (C) Vβ usage was performed in Patient 22 at diagnosis, at 10 days of treatment, and at 6 and 9 months (relapse).

Perturbations during treatment and relapse are reported as percentage of positive CD8+CD57+ and Vβ2+ cells (left), or absolute lymphocyte and Vβ2 clone count (right).

762

haematologica | 2018; 103(5)

TCR repertoire of effector memory T cells in AA

fusions expose T cells to multiple foreign epitopes.37 A group of 5 pure red cell aplasia (PRCA) patients, 10 sickle cell disease (SCD) subjects, and 8 myelodysplastic (MDS) patients who had been heavily transfused before sampling were studied for CD8+CD57+ cell expansion and Vβ usage (Online Supplementary Table S1). No variations were observed in CD8+CD57+ cell frequencies when PRCA, MDS, and SCD patients were compared to healthy subjects (22.8±25.1% vs. 21.6±9.7% vs. 30.6±28.3% vs.

13.3±12.6%, respectively; P=0.216) (Online Supplementary Figure S6A). Vβ skewing was described in 4 PRCA patients in 1-8 subgroups, in all 8 MDS subjects in 1-4 Vβ families, and in 7 SCD patients in 1-5 subgroups (Online Supplementary Figure S6B). Oligoclonal expansion was described in 91% of cases and subjects without skewing did not present effector memory T-cell expansion.

Subsequently, we combined all subjects and divided them according to their transfusion history and the presence of

A

B

Figure 3. Characterization of Vβ/Jβ plot, CDR3 size and DJ length profiles in healthy donors by deep sequencing. (A) T-cell receptor β variable (TRBV)/T-cell receptor β joining (TRBJ) plots showed a “citylike” landscape for total CD4+ and CD8+ cell populations in healthy subjects (HC). (B) The size of the complementarity region 3 (CDR3) and DJ length profiles were also defined, showing similar features in CD4+ and CD8+ cells.

haematologica | 2018; 103(5)

763

V. Giudice et al.

CD8+CD57+ T-cell expansion. By χ2 test, transfusion history did not correlate to CD8+CD57+ expansion (P=0.255).

Total CD8+ cell TCR repertoires are polyclonal in healthy subjects Deep sequencing of TCR repertoire was performed in CD4+ and CD8+ populations sorted from 9 healthy donors.

The average depth of sequencing was 10,790,646±4,050,138

in CD4+ T cells, and 10,961,961±3,879,596 in CD8+ populations. The mean frequency of the immunodominant clone was 4.1±4.1% in CD4+ cells and 17.3±16.9% in CD8+ cells.

By plotting the number of reads of each Vβ and Jβ matching,17 a “citylike” landscape was described for total CD4+ and CD8+ T-cell populations (Figure 3A and Online Supplementary Figure S7A), because of the presence of a more homogenous distribution of TCR rearrangement fre-

A

B

Figure 4. The TCR repertoire by deep sequencing analysis from total CD8+ cells in severe aplastic anemia (SAA) patients with CD8+CD57+ cell expansion. In contrast to healthy CD8+ profiles, most SAA patients (AA) displayed oligoclonal features in a TRBV/TRBJ rearrangement plot (A) and CDR3 size and DJ length profiles (B). CD4+ and CD8+ profiles are shown for each SAA patient. In AA1 and AA6, only CD8+ cells were sufficient for deep sequencing.

764

haematologica | 2018; 103(5)

TCR repertoire of effector memory T cells in AA

quencies (“midtown”). The normal CDR3 size and DJ length profiles were defined by comparing distributions among healthy donors (Figure 3B and Online Supplementary Figure S7B). In CD4+ and CD8+ cells, profiles were typically distributed in a Gaussian manner with 1012 different size classes of 30-65 nucleotides (nt) sizes at 3 nt intervals, and they completely overlapped. Similarly, DJ length profiles assumed an asymmetric Gaussian distribution with 2-5 predominant length classes without nt intervals in CD4+ and CD8+ subsets.

Deep sequencing allows detailed characterization of oligoclonality in SAA Deep sequencing of TCR repertoire was performed in CD4+ and CD8+ cells from 12 SAA patients who had demonstrated CD8+CD57+ cell expansion by flow cytom-

etry, and also in CD8+CD57+ cells from 2 of these patients.

The average depth of sequencing was T cells, 13,861,048±4,992,677 in CD4+ 13,873,207±5,029,195 in CD8+ T cells, and 21,030,616±1,238,660 in CD8+CD57+ cells. The mean frequency of the immunodominant clone was 3.3±3.4% (range: 0.2-11.8%) in CD4+, 18.2±14.9% (range: 3.354.1%) in CD8+ cells, and 59±28.9% in CD8+CD57+cells.

By plotting TRBV/TRBJ rearrangements from CD4+ cells, the “citylike” landscape was found in 11 out of 12 patients (Figure 4A and Online Supplementary Figure S8A). In the CD8+ cell pool, the “citylike” landscape was found in 3 patients (AA7, AA8, and AA11), whose clones showed very low frequencies (4.3%, 3.4%, and 3.3%, respectively). The remaining patients displayed a “skyscraper” landscape due to the presence of 1-3 immunodominant clones.

A

B

Figure 5. The TCR repertoire by deep sequencing of enriched CD8+CD57+ cells in severe aplastic anemia (SAA) patients. (A) The enrichment of the clone in effector memory CD8+ T cells, comparing TRBV/TRBJ rearrangement in total CD8+ cells (left) with those in CD8+CD57+ cells (right) from the same patients. (B) CDR3 size and DJ length profiles from CD8+CD57+ cells also overlapped with those in CD4+ and CD8+ profiles.

haematologica | 2018; 103(5)

765

V. Giudice et al.

CDR3 size and DJ length profiles were defined in total CD4+ and CD8+ cells and compared within each patient (Figure 4B and Online Supplementary Figure S8B). In CD4+ cells, all patients displayed CDR3 profiles with normal Gaussian distribution, as described in healthy subjects. In CD8+ cells, Patients AA7, AA8, and AA10 to AA12 had CDR3 size profiles with Gaussian distributions. In the

remaining patients, CDR3 profiles showed different shapes with 8-13 different classes and 1 or 2 predominant peaks of various nt sizes (36-57). For DJ length profiles in SAA patients, the asymmetric Gaussian distribution was described in all CD4+ cells, and in 6 (AA5-AA8 and AA11-AA12) CD8+ populations. Similarly, TRBV/TRBJ rearrangement plots from the CD8+CD57+

A

B

Figure 6. Homology assessment. (A) CDR3 sequence pools were analyzed among patients (AA) and healthy subjects (HC) for the presence of homology. Shared and immunodominant sequences were reported as a heatmap based on their relative expression: in the same row, from lowest (gray; <0.01%) to highest (red; >5%) values. (B) Structural analysis was performed comparing the sequences for common pattern, using both alignments at the N-terminal of Vβ gene (left) or at the C-terminal of Jβ gene (right).

766

haematologica | 2018; 103(5)

TCR repertoire of effector memory T cells in AA

cell pool of AA3 and AA4 showed the same “skyscraper” landscape, but more enriched (Figure 5A), as well as CDR3 size and DJ length profiles that overlapped those in total CD8+ cells, although these were again more enriched (Figure 5B).

mean frequency of 0.1±0.2% and 4.7±9.4% in SAA patients, and 0.0003±0.0001% and 0.01±0.002% in healthy subjects. No matches were found in the enriched CD8+CD57+ T-cell pool (Online Supplementary Table S3).

Discussion TCR repertoire diversity and shared CDR3 sequences in SAA patients and healthy subjects Simpson’s index of diversity was calculated for each healthy and each SAA subject. This index is used in ecology for in-depth assessment of the degree of diversity of a system, related to the richness (or number of species present) and evenness (or relative abundance of each).38 Thus, this index assesses the probability that 2 randomly selected individuals from a system belong to the same species.

For a TCR repertoire, the index measures the probability that 2 CDR3 sequences randomly selected from CD4+ or CD8+ pools of one subject could be identical.17,38 A value close to 0 means infinite diversity, and a value close to 1 no diversity.36 In our cohort, Simpson’s indexes were similar in healthy controls and SAA patients for CD4+ (0.990±0.005 vs. 0.991±0.005, respectively; P=0.537) and CD8+ cells (0.983±0.013 vs. 0.927±0.084, respectively; P=0.060). By paired t-test, Simpson’s indexes in SAA subjects were significantly different between CD4+ and CD8+ cells (P=0.047) (Online Supplementary Figure S9A). When compared to those indices in CD8+CD57+ cells, decreased diversity in the CD8+ effector memory compartment was described (total CD4+ cells vs. CD8+CD57+ cells, P<0.0001; total CD8+ cells vs. CD8+CD57+ cells, P=0.0003) (Online Supplementary Figure S9B).

In order to test the hypothesis that an autoreactive clone is triggered by autoantigens, CDR3 amino acid sequences from healthy subjects and patients were screened for homology and then compared to public and private TCR repertoires reported in literature,39 as described in the Online Supplementary Methods. From analysis of sequences in the CD8+ cell pool, a total of 29 CDR3 sequences were shared between patients and controls, but these sequences were highly expressed in SAA patients and also present in their immunodominant clones (Figure 6A). When we searched for these common sequences in the whole CDR3 sequence repertoire from CD8+CD57+ cells in AA subjects, CD8+CD57+ cell pools from AA3 and AA4 showed enrichment in frequencies of the immunodominant clones (from 53.61% to 79.12% for AA3 patient and from 23.45% to 38.14% for AA4 patient). Structural analysis of shared and immunodominant sequences did not show a pattern of common charged residue among patients (Figure 6B and Online Supplementary Methods). Immunodominant and shared CDR3 sequences were compared with CDR3 β repertoires reported in literature for infectious, autoimmune and malignant diseases.39 For confirmation, the VDJdb database, a database of known antigen specific sequences (https://vdjdb.cdr3.net), was also used. No matches were found for infectious diseases, while 2 sequences present in AA9 were reported previously in AA and paroxysmal nocturnal hemoglobinuria (PNH).8,40 Lastly, we sought to perform homology assessment for reported PNH-related clonotypes on the entire TCR repertoire without frequency restriction, as small PNH clones could be present in healthy individuals.40 Two of the reported 12 CDR3 sequences (CATSRTGGETQYF and CATSRVVAGETQYF) were found in our TCR repertoires with a haematologica | 2018; 103(5)

The character of oligoclonal expansion of CD8+CD28– lymphocytes in AA, described by Risitano et al.,8,41 strongly suggests an antigen driven mechanism of T-cell activation, ultimately leading to destruction of hematopoietic stem and progenitor cells. In this study, we focused on a subset of CD8+CD28–CD57+ T lymphocytes, termed effector memory cells because of their high antigen-affinity and their ability to undergo activation after antigen stimulation.27,36 Others have reported the expansion of effector memory CD8+ T cells in the tumor microenvironment and in PB from patients with solid tumors, hematologic malignancies, chronic infections and autoimmune disorders.36,4144 In cancers, effector memory T cells appear to have an important role in immune surveillance; for example, their increase after interferon (IFN)a treatment correlates to better prognosis in melanoma patients,43 and expanded CD8+CD57+ T cells reach normal levels after removal of head and neck cancer.36 Higher frequencies of CD8+CD57+ cells have also been described in SAA and MDS patients, which decrease in responders after anti-thymocyte globulin treatment.44 The expansion of effector memory cells could also occur in older healthy subjects as a result of lifelong exposure to common pathogens, but it is related to reduced overall immune responsiveness to novel antigens.45,46 Consistent with previous studies, the SAA patients in our cohort had higher frequencies of CD8+CD57+ cells at diagnosis (25.6% in SAA patients vs.

13.3% in healthy subjects). Moreover, patients with CD8+ effector memory T-cell expansion at diagnosis experienced shorter PFS. No age-effects were seen for clone size in either patients or healthy subjects.

There is indirect evidence of immune pathophysiology of AA, including oligoclonal expansion of effector CD8+ T lymphocytes.8,41 Clonality of T-cell subsets can be studied by flow cytometry analysis of Vβ usage, CDR3 size spectratyping or deep sequencing of VDJ combinations, and CDR3 nucleotide and amino acid sequences.17-20,47 In SAA patients, expansion of at least one Vβ family in both CD4+ and CD8+ effector CD28dim cells by flow cytometry with polyclonal features in CD4+ and oligoclonal characteristics in CD8+ cells has been described.1,8,15,41,44 The CDR3 size skewing by spectratyping in the CD8+ population but not in CD4+ cells confirmed clonality in CD8+ cells.8,21,41 In our current work, we demonstrate by flow cytometry and deep sequencing that oligoclonal expansion occurs mainly in effector memory CD8+ cell compartment, as the immunodominant clones were highly enriched in CD8+CD57+ cells and had decreased diversity on deep sequencing. Effector memory T cells are a circulating Tcell population that can migrate to the BM under different types of stimulation.30 Vβ usage of peripheral CD8+CD57+ T cells may mirror that in the BM, as suggested by the high concordance between PB and BM in our AA patients.

In our cohort, 75% of SAA patients with effector memory cell expansion also showed oligoclonal features by deep sequencing of the total CD8+ cell compartment: TRBV/TRBJ rearrangement plots with a “skyscraper” pro767

V. Giudice et al.

file due to the presence of 1-3 immunodominant clones, and predominant classes in CDR3 size and DJ length profiles. These findings were confirmed by TCR repertoire sequencing of CD8+CD57+ enriched cells from SAA patients. Similar deep sequencing features have been described in whole blood in T-cell large granular lymphocyte leukemia (T-LGLL).17 T-LGLL, a chronic lymphoproliferation of TCRaβ+CD3+CD5dimCD8+CD57+CD16+ cells with monoclonal TCRγ-chain rearrangement and prevalence in the elderly, is frequently associated with autoimmune diseases.16,17 Thus, similarities between our SAA cohort and T-LGLL patients suggest a common pathophysiology of expanded autoreactive T lymphocytes.

Characterization of long-term Vβ usage has already been proposed as a biomarker of disease progression in T-LGLL and AA,8,16 given that immunodominant clones can remerge during relapse.16 However, high heterogeneity in the TCR repertoire during immunosuppressive therapies has been reported.8,16 In our study, immunodominant clones were longitudinally investigated in 3 AA patients.

In Patient 4, expanded clones slightly decreased during treatment but did not disappear, as no hematologic improvements were observed. In Patient 34, clone 17 remained stable during the course of the disease, while clone 13.6 increased at three months and slightly decreased at six, concomitantly with a minimal partial response. In Patient 22, clone 2 completely disappeared at six months of treatment and the patient achieved hematologic remission; however, at the time of relapse the clone increased again. Our data suggest that Vβ typing of the CD8+ CD57+ T-cell population by flow cytometry might be a useful biomarker to monitor clonal kinetics during the course of AA, but a larger cohort of patients and a more sensitive technique, such as deep sequencing, are needed to validate the clinical usefulness.

Chronic antigen exposure is required to trigger T-cell activation, as in persistent viral infections or with antigen spread during autoimmune and malignant diseases.17 For viral infections, CDR3 homology in public sequence repertoires, and also cross-reactivity between viruses, is limited

References 1. Young NS, Calado RT, Scheinberg P. Current concepts in the pathophysiology and treatment of aplastic anemia. Blood.

2006;108(8):2509-2519.

2. Young NS. Current concepts in the pathophysiology and treatment of aplastic anemia. Hematology Am Soc Hematol Educ Program. 2013;2013:76-81.

3. Maciejewski JP, Selleri C, Sato T, Anderson S, Young NS. A severe and consistent deficit in marrow and circulating hematopoietic cells (long-term culture-initiating cells) in acquired aplastic anemia.

Blood.

1996;88(6):1983-1991.

4. Maciejewski JP, Kim S, Sloand E, Selleri C, Young NS. Sustained long-term hematologic recovery despite a marked quantitative defect in the stem cell compartment of patients with aplastic anemia after immunosuppressive therapy. Am J Hematol.

2000;65(2):123-131.

5. Risitano AM, Maciejewski JP, Selleri C,

768

6.

7.

8.

9.

by the number of possible epitopes.41,48 For autoimmune diseases, including T-LGLL, clonotypes can be private to a specific disease because of the unlimited number of possible epitopes.17,41,49 Despite the large diversity of TCR CDR3 sequences, only 29 shared CDR3 sequences were found in our cohort; these were highly expressed in SAA patients and enriched in the CD8+ CD57+ T-cell population. By using sensitive techniques such as deep sequencing, we detected sequences at very low frequencies. However, the finding that a group of clonotypes is shared between SAA patients and healthy subjects suggests the existence of common epitopes driving activation of T-cell autologous clones (as also described by Gargiulo et al. in PNH patients40). As chronic transfusion could be the source of antigen exposure, we investigated its effect on effector memory T-cell expansion in other hematologic diseases. A comparison between transfused patients and healthy subjects showed no significant variations in CD8+ CD57+ cell frequencies (P=0.216).

Oligoclonal expansion of effector memory CD8+ T cells is frequent in AA and may correlate with prognosis, consistent with a role of effector memory T cells in BM destruction during active disease. Deep sequencing technologies allow in-depth characterization of the TCR repertoire, and flow cytometric analysis of Vβ usage may be useful to determine diagnosis and prognosis of SAA patients, and to monitor their clinical course.

Acknowledgments The authors would like to thank Sachiko Kajigaya and Keyvan Keyvanfar (Hematology Branch, NHLBI), Ying Rao (BGI Genomics), and Swee Lay Thein (Sickle Cell Branch, NHLBI) for assistance; Kinneret Broder (Hematology Branch, NHLBI) for assistance in obtaining healthy volunteer samples; and Barbara Weinstein (Hematology Branch, NHLBI) and Jim Nichols (Sickle Cell Branch, NHLBI) for obtaining patient samples.

Funding This research was supported by the Intramural Research Program of the NIH, National Heart, Lung, and Blood Institute.

Rotoli B. Function and malfunction of hematopoietic stem cells in primary bone marrow failure syndromes. Curr Stem Cell Res Ther. 2007;2(1):39-52.

Maciejewski JP, O’Keefe C, Gondek L, Tiu R. Immune-mediated bone marrow failure syndromes of progenitor and stem cells: molecular analysis of cytotoxic T cell clones.

Folia Histochem Cytobiol. 2007;45(1):5-14.

Feng X, Scheinberg P, Wu CO, et al.

Cytokine signature profiles in acquired aplastic anemia and myelodysplastic syndromes. Haematologica. 2011;96(4):602606.

Risitano AM, Maciejewski JP, Green S, Plasilova M, Zeng W, Young NS. In-vivo dominant immune responses in aplastic anaemia: molecular tracking of putatively pathogenetic T-cell clones by TCR betaCDR3 sequencing.

Lancet.

2004; 364(9431):355-364.

Young NS. Hematopoietic cell destruction by immune mechanisms in aquired aplastic anemia. Semin Hematol. 2000;37(1):3-14.

10. Sloand EM, Kim S, Maciejewski JP, et al.

Intracellular interferon- in circulating and marrow T cells detected by flow cytometry and the response to imunosuppressive therapy in patients with aplastic anemia. Blood.

2002;100:(4)1185-1191.

11. Maciejewski JP, Selleri C, Anderson S, Young NS. Fas antigen expression on CD34+ human marrow cells is induced by interferon-gamma and tumor necrosis factor-alpha and potentiates cytokine-mediated hematopoietic suppression in vitro.

Blood. 1995;85:(11)3183-3190.

12. Selleri C, Maciejewski JP, Sato T, Young NS. Interferon-γ constitutively expressed in the stromal microenviroment of human marrow cultures mediates potent hematopoietic inhibition. Blood. 1996; 87:(10)4149-4157.

13. Hirano N, Butler MO, von Bergwelt-Baildon MS, et al. Autoantibodies frequently detected in patients with aplastic anemia. Blood.

2003;102:(13)4567-4575.

14. Feng X, Chuhjo T, Sugimori C, et al.

Diazepam-binding inhibitor-related protein

haematologica | 2018; 103(5)

TCR repertoire of effector memory T cells in AA

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

1: a candidate autoantigen in acquired aplastic anemia patients harboring a minor populatoin of paroxysmal nocturnal hemoglobinuria-type cells. Blood. 2004; 104:(8)24252431.

Kook H, Risitano AM, Zeng W, et al.

Changes in T-cell receptor VB repertoire in aplastic anemia: effects of different immunosuppressive regimens. Blood.

2002;99(10):3668-3675.

Clemente MJ, Wlodarski MW, Makishima H, et al. Clonal drift demonstrates unexpected dynamics of the T-cell repertoire in Tlarge granular lymphocyte leukemia. Blood.

2011;118(16):4384-4393.

Clemente MJ, Przychodzen B, Jerez A, et al.

Deep sequencing of the T-cell receptor repertoire in CD8+ T-large granular lymphocyte leukemia identifies signature landscapes. Blood. 2013;122(25):4077-4085.

Calis JJ, Rosenberg BR. Characterizing immune repertoires by high throughput sequencing: strategies and applications.

Trends Immunol. 2014;35(12):581-590.

Langerak AW, van Den Beemd R, WolversTettero IL, et al. Molecular and flow cytometric analysis of the Vbeta repertoire for clonality assessment in mature TCRalphabeta T-cell proliferations. Blood.

2001;98(1):165-173.

Wlodarski MW, Gondek LP, Nearman ZP, et al. Molecular strategies for detection and quantitation of clonal cytotoxic T-cell responses in aplastic anemia and myelodysplastic syndrome. Blood. 2006;108(8):26322641.

Schuster FR, Hubner B, FĂźhrer M, et al.

Highly skewed T-cell receptor V-beta chain repertoire in the bone marrow is associated with response to immunosuppressive drug therapy in children with very severe aplastic anemia. Blood Cancer J. 2011;1(3):e8.

Schatz DG, Ji Y. Recombination centres and the orchestration of V(D)J recombination.

Nat. Rev. Immunol. 2011;11:(4)251-263.

Gauss GH, Lieber MR. Mechanistic constraints on diversity in human V(D)J recombination. Mol Cell Biol. 1996;16(1):258-269.

Teng G, Papavasiliou FN. Immunoglobulin somatic hypermutation. Annu Rev Genet.

2007;41:107-120.

Matthews AJ, Zheng S, DiMenna LJ, Chaudhuri J. Regulation of immunoglobulin class-switch recombination: choreography of noncoding transcription, targeted DNA deamination, and long-range DNA repair.

haematologica | 2018; 103(5)

Adv Immunol. 2014;122:1-57.

26. Arstila TP, Casrouge A, Baron V, Even J, Kanellopoulos J, Kourilsky P. A direct estimate of the human alphabeta T cell receptor diversity. Science. 1999;286(5441):958-961.

27. Bretschneider I, Clemente MJ, Meisel C, et al. Discrimination of T-cell subsets and Tcell receptor repertoire distribution.

Immunol Res. 2014;58(1):20-27.

28. Araki K, Youngblood B, Ahmed R. The role of mTOR in memory CD8 T-cell differentiation. Immunol Rev. 2010;235(1):234-243.

29. Appay V, van Lier RA, Sallusto F, Roederer M. Phenotype and function of human T lymphocyte subsets: consensus and issues.

Cytometry A. 2008;73(11):975-983.

30. Farber DL, Yudanin NA, Restifo NP. Human memory T cells: generation, compartmentalization and homeostasis. Nat Rev Immunol. 2014;14(1):24-35.

31. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):21912194.

32. International Agranulocytosis and Aplastic Anaemia Study. Incidence of aplastic anemia: the relevance of diagnostic criteria. By the International Agranulocytosis and Aplastic Anemia Study.

Blood.

1987;70(6):1718-1721.

33. Camitta BM, Thomas ED, Nathan DG, et al.

Severe aplastic anemia: a prospective study of the effect of early marrow transplantation on acute mortality. Blood. 1976;48(1):63-70.

34. Bolotin DA, Shugay M, Mamedov IZ, et al.

MiTCR: software for T-cell receptor sequencing data analysis. Nat Methods.

2013;10(9):813-814.

35. Benjamini Y, Yekutieli D. The control of the false discovery rate in multiple testing under dependency. Ann Stat. 2001; 29:(4)11651188.

36. Strioga M, Pasukoniene V, Characiejus D. CD8+ CD28- and CD8+ CD57+ T cells and their role in health and disease.

Immunology. 2011;134(1):17-32.

37. Nickel RS, Horan JT, Fasano RM, et al.

Immunophenotypic parameters and RBC alloimmunization in children with sickle cell disease on chronic transfusion. Am J Hematol. 2015;90(12):1135-1141.

38. Hill MO. Diversity and evenness: a unifying notation and its consequences. Ecology.

1973;54:(2)427-432.

39. Li H, Ye C, Ji G, Han J. Determinants of pub-

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

lic T cell responses. Cell Res. 2012;22(1):3342.

Gargiulo L, Lastraioli S, Cerruti G, et al.

Highly homologous T-cell receptor beta sequences support a common target for autoreactive T cells in most patients with paroxysmal nocturnal hemoglobinuria.

Blood. 2007;109(11):5036-5042.

Risitano AM, Kook H, Zeng W, Chen G, Young NS, Maciejewski JP. Oligoclonal and polyclonal CD4 and CD8 lymphocytes in aplastic anemia and paroxysmal nocturnal hemoglobinuria measured by V beta CDR3 spectratyping and flow cytometry. Blood.

2002;100(1):178-183.

Filaci G, Fenoglio D, Fravega M et al. CD8+ CD28- T regulatory lymphocytes inhibiting T cell proliferative and cytotoxic functions infiltrate human cancers. J Immunol 2007; 179:(7)4323-4334.

Characiejus D, Pasukoniene V, Jonusauskaite R, et al. Peripheral blood CD8high CD57+ lymphocyte levels may predict outcome in melanoma patients treated with adjuvant interferon-alpha.

Anticancer Res. 2008;28:(2B)1139-1142.

Kook H, Zeng W, Guibin C, Kirby M, Young NS, Maciejewski JP. Increased cytotoxic T cells with effector phenotype in aplastic anemia and myelodysplasia. Exp Hematol.

2001;29:(11)1270-1277.

Boucher N, Dufeu-Duchesne T, Vicaut E, Farge D, Effros RB, Schächter F. CD28 expression in T cell aging and human longevity. Exp Gerontol. 1998;33(3):267282.

Wikby A, Ferguson F, Forsey R, et al. An immune risk phenotype, cognitive impairment, and survival in very late life: impact of allostatic load in Swedish octogenarian and nonagenarian humans. J Gerontol A Biol Sci Med Sci. 2005;60(5):556-565.

Fozza C, Barraqueddu F, Corda G, et al.

Study of the T-cell receptor repertoire by CDR3 spectratyping. J Immunol Methods.

2017;440:1-11.

Zeng W, Maciejewski JP, Chen G, Young NS. Limited heterogeneity of T cell receptor BV usage in aplastic anemia. J Clin Invest.

2001;108(5):765-773.

Le Gal FA, Ayyoub M, Dutoit V, et al.

Distinct structural TCR repertoires in naturally occurring versus vaccine-induced CD8+ T-cell responses to the tumor-specific antigen NY-ESO-1. J Immunother.

2005;28(3):252-257.

769

ARTICLE

Rec Cell Biology & its Disorders

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):770-777

Safety and efficacy of plerixafor dose escalation for the mobilization of CD34+ hematopoietic progenitor cells in patients with sickle cell disease: interim results

Farid Boulad,1,2 Tsiporah Shore,3 Koen van Besien,3 Caterina Minniti,4 Mihaela Barbu-Stevanovic,5 Sylvie Wiener Fedus,6 Fabiana Perna,2 June Greenberg,7 Danielle Guarneri,7 Vijay Nandi,5 Audrey Mauguen,8 Karina Yazdanbakhsh,5 Michel Sadelain2 and Patricia A. Shi4,5 Department of Pediatrics, BMT Service, Memorial Sloan Kettering Cancer Center, New York; 2Center for Cell Engineering, Memorial Sloan Kettering Cancer Center, New York; 3 Bone Marrow and Hematopoietic Stem Cell Transplant Program, Weill Cornell Medicine/ New York Presbyterian Hospital, New York; 4Sickle Cell Program, Division of Hematology, Albert Einstein College of Medicine, Bronx; 5Lindsley F. Kimball Research Institute, New York Blood Center, NY; 6Laboratory Medicine, Memorial Sloan Kettering Cancer Center, New York; 7Division of Hematology and Oncology, Weill Cornell Medicine /New York Presbyterian Hospital, NY and 8Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, USA 1

ABSTRACT

G

Correspondence: pshi@nybc.org

Received: December 22, 2017.

Accepted: January 23, 2018.

Pre-published: February 1, 2018. doi:10.3324/haematol.2017.187047 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/770

ene therapy for sickle cell disease is limited by the yield of hematopoietic progenitor cells that can be harvested for transduction or gene editing. We therefore performed a phase I dose-escalation study of the hematopoietic progenitor cell mobilizing agent plerixafor to evaluate the efficacy and safety of standard dosing on peripheral blood CD34+ cell mobilization. Of 15 patients enrolled to date, only one was chronically transfused and ten were on hydroxyurea. Of eight patients who achieved a CD34+ cell concentration >30 cells/μL, six were on hydroxyurea. There was no clear dose response to increasing plerixafor dosage. There was a low rate of serious adverse events; two patients developed vaso-occlusive crises, at the doses of 80 μg/kg and 240 μg/kg.

Hydroxyurea may have contributed to the limited CD34+ mobilization by affecting baseline peripheral blood CD34 counts, which correlated strongly with peak peripheral blood CD34 counts. Plerixafor administration did not induce significant increases in the fraction of activated neutrophils, monocytes, or platelets. However, increased neutrophils positive for activated β2 integrin and Mac-1 were associated with serious adverse events. In summary, plerixafor was well tolerated but did not achieve consistent CD34+ cell mobilization in this cohort of patients, most of whom were being actively treated with hydroxyurea and only one was chronically transfused. The study will continue with escalation of the dose of plerixafor and modification of hydroxyurea administration. Clinicaltrials.gov identifier: NCT02193191.

Introduction ©2018 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.

770

Autologous gene therapy holds considerable promise for the treatment of patients with sickle cell disease (SCD).1,2 However, its successful application requires an adequate number of hematopoietic progenitor cells (HPC) for gene transfer or gene editing.3 Steady-state bone marrow has been the historical source of HPC for SCD gene therapy, but its harvest requires general anesthesia and has been complicated in current gene therapy trials by the need for repeated bone marrow harvests and a high rate of adverse events.4 Granulocyte colony-stimulating factor (G-CSF) is a standard method of mobilizing HPC but its use in SCD patients has been associated with vaso-occlusive complications and even death.5 The mechanism of action of G-CSF involves activation of neutrophils,6 and is also associated with endothelial cell, platelet, and coagulation system activation,7-11 all of which may play a crucial role in sickle cell vaso-occlusion.12 In contrast to G-CSF, plerixafor is a bicyclam reversible small molecule inhibitor haematologica | 2018; 103(5)

Safety and efficacy of plerixafor in SCD patients Table 1. Patients’ characteristics.

Dose level

80 μg/kg

160 μg/kg

240 μg/kg

Subject ID

Ethnicity

Age yrs

Gender

Genotype

Treatment regimen

1 2 3 (1*) 4 5 6

AfricanAmerican

33 21 29 32 34 25

M F M M M F

SS SS SS; thal(1/4) SS; thal (1/4) SS; thal (1/4) SS; thal (1/4)

HU 16 mg/kg HU 26 mg/kg No HU No HU HU 28 mg/kg HU 16 mg/kg

25 37 32

F M F

SS SS SS

46 38 23 27 31 38

M M F M M M

7 8 (1*) 9 10 12 13 14 3 (2*) 8 (2*)

Hispanic AfricanAmerican AfricanAmerican

AfricanAmerican

HU 25 mg/kg HU 27 mg/kg Chronic transfusion – No HU SS; thal (1/4) No HU SS HU 17 mg/kg SS HU 27 mg/kg SS HU 23 mg/kg SS- thal (1/4) No HU SS HU 27 mg/kg

Clinical complications (in addition to ACS) Avascular necrosis, retinopathy ≥ 3 vaso-occlusive crises per year ≥ 3 vaso-occlusive crises per year Deep venous thrombosis, priapism Leg ulcers, retinopathy Leg ulcers, retinal artery occlusion, retinopathy Cerebral aneurysms (2-3 mm) Avascular necrosis, retinopathy Avascular necrosis, iron overload, leg ulcers Leg ulcers, priapism Avascular necrosis, leg ulcers, priapism Avascular necrosis, retinopathy Avascular necrosis, priapism ≥ 3 vaso-occlusive crises per year Avascular necrosis, retinopathy

*Indicates first and second enrollments for indicated subject. ACS: acute chest syndrome,; HU: hydroxyurea.

of the chemokine receptor CXCR4 and prevents binding of its ligand CXCL12 or stromal cell derived factor-1a to induce HPC mobilization.13 We hypothesized that plerixafor’s mechanism of action would lead to less marked increases in white blood cell (WBC) counts and therefore less cell and coagulation system activation in SCD and supported this with data from a pre-clinical study involving a sickle cell mouse model.14 Nevertheless, the safety of plerixafor in SCD patients remains a matter of concern because of possible activation of WBC and neutrophils which could still lead to vaso-occlusive complications and the risk of early death in SCD.15-19 As CXCR4 is expressed on most WBC and is involved in the retention of these cells in bone marrow, a standard dose of plerixafor of 240 μg/kg increases all major WBC subsets (neutrophils, lymphocytes, monocytes) in normal donors about 3- to 4fold.20 Notably, however, in SCD patients who received GCSF, not all patients who had highly elevated WBC counts experienced vaso-occlusive complications, and conversely, not all patients who experienced vaso-occlusive complications had highly elevated WBC counts,5 suggesting that WBC activation rather than WBC count per se may contribute to vaso-occlusion in SCD. Another issue of concern is whether enough peripheral blood CD34+ cells can be mobilized in SCD patients with plerixafor. The mean and median peak CD34+ counts using plerixafor alone in normal donors are only ~25/μL.13,20 SCD patients might mobilize particularly well, in that SCD patients might have increased circulating HPC even at steady state, although more so during a crisis.21,22 Furthermore, SS and Sβ0 patients tend to be autosplenectomized, and data from patients with thalassemia showed that splenectomized patients mobilized about twice as many peripheral blood CD34 cells with plerixafor alone as non-splenectomized patients.23 Another consideration when using plerixafor is whether to withhold hydroxyurea, the recommended standard of care for most SCD patients. Hydroxyurea may inhibit mobilization and withholding hydroxyurea for 2 weeks leads to a degree of spontaneous mobilization that abets haematologica | 2018; 103(5)

drug-induced mobilization.23,24 However, Richard et al. showed that two of the three SCD patients whose hydroxyurea was withdrawn for 2 weeks developed painful crises following the withdrawal. Given these considerations, we designed a prospective phase I dose escalation study of both the safety and efficacy of plerixafor in patients with SCD in which the patients continued on their standard outpatient treatment used for disease control. We have completed the dosing cohorts through to the standard plerixafor dose of 240 μg/kg and report the interim results here.

Methods Study design This study is conducted under FDA IND 122657, registered in ClinicalTrials.gov as NCT02193191, and approved by the Institutional Review Boards of Memorial Sloan Kettering, Weill Cornell Medical College and the New York Blood Center. The study design is a 3 + 3 dose escalation study with six levels of escalation: doses of 80, 160, 240, 320, 400, and 480 μg/kg. There are two primary endpoints: (i) efficacy, defined by the achievement of a HPC mobilization level of 30 CD34+ cells/μL; and (ii) safety, defined by the occurrence of serious adverse events (≥ grade 3) that are at least possibly plerixafor-related (including vaso-occlusive events).

At any dose level, the occurrence of at least one grade 3 serious adverse event results in the addition of three more patients to the initial three-patient dosing cohort. The occurrence of two grade 3 serious adverse events at a particular dose-level signifies that the maximal tolerated dose has been exceeded and that the previous dose-level is the maximum tolerated dose. The trial will be stopped upon the occurrence of one grade 4 or 5 serious adverse event at least possibly related to plerixafor. Patients are followed for adverse events for 1 month after administration of the plerixafor. This design provides the following probabilities of escalation based on the true chance of a dose-limiting toxicity at a specific dose level: True probability of toxicity 0.10 0.20 0.30 0.40 0.50 0.60 771

F. Boulad et al.

Figure 1. Peripheral blood white blood cell, absolute neutrophil, and CD34 cell counts. There is a trend of increasing white blood cell count (WBC, P=0.05) and absolute neutrophil count (ANC, P=0.03), but not CD34 concentration (P=0.65) with increasing dose of plerixafor. The graphs show the peripheral blood WBC, ANC, and CD34 concentrations of 15 patients with SCD treated whith 80 (circles), 160 (squares) and 240 (triangles) µg/Kg of plerixafor prior to administration of plerixafor (PRE), 6-12 h and 20-24 h after plerixafor. Patients on hydroxyurea are represented by filled circles, squares and triangles, patients off hydroxyurea are represented by open circles, squares and triangles.

Probability of escalation: 0.91 0.71 0.49 0.31 0.17 0.08 For the efficacy endpoint, the dose escalation will continue to 480 μg/kg unless all patients at a preceding dose level achieve a peripheral blood CD34+ concentration of at least 30 cells/μL. In the present dose-escalation phase, no leukapheresis is performed. If and when the efficacy endpoint is safely reached, the study will proceed to a leukapheresis phase (including preclinical transduction and editing) in three patients.

Eligible subjects are adults with SS or Sβ0 disease, normal renal and liver function, hemoglobin concentration ≥6 g/dL, WBC count ≥3,000/μL, absolute neutrophil count (ANC) ≥1,500/μL, and platelet count of ≥150,000/μL. Eligible subjects are admitted to the Clinical Research Center at Weill Cornell Medical College. A single subcutaneous injection of plerixafor (Sanofi-Genzyme) is administered in the evening between 8-9 pm. The protocol calls for peripheral blood sampling at three time points (baseline, 0-2 h prior to plerixafor; peak between 6-12 h after the plerixafor dose; at the presumed return to baseline between 20-24 h after the dose): for reasons of feasibility and patient comfort issues, the peak sample was consistently drawn at a mean of 12 ± 1 h after plerixafor administration and the return to baseline sample at a mean of 20 ± 0.29 h after the dose. Since patients have pre-existing anemia, for reasons of safety no more than a total of 105 mL of blood is drawn at all three time points combined.

Peripheral blood CD34 testing Flow cytometric evaluation of the collected peripheral blood is performed using a FACS Canto flow cytometer (Becton 772

Dickinson Biosciences, San Jose, CA, USA) and FACS Diva software (BD Biosciences). Samples are stained and analyzed within 2-12 h of collection using a modification of the International Society of Hematotherapy and Graft Engineering (ISHAGE) method (see Online Supplementary Methods).

CD34+CD38– enumeration Mononuclear cells are isolated from 2 mL peripheral blood by Ficoll-Hypaque Plus density centrifugation. CD34+ cells are purified by positive selection (MidiMACS™ LS Columns, Miltenyi) and stained with CD34 (BD PharMingen) and CD38 (Invitrogen).

Research cell and coagulation activation studies Peripheral blood samples drawn at baseline and after 12 h are stained within 1 h of collection for activation markers relevant to sickle vaso-occlusion12,25-28 and assessed by flow cytometry (BD FACSCanto™). For CD16b+ (1D3, Beckman Coulter) neutrophils: activated β2 integrin (clone 24, abcam), activated Mac1 (CBRM1/5, eBioscience), E-selectin-Fc chimera (724-ES, R&D Systems), L-selectin (DREG-56, eBioscience), Mac-1/CD11b (ICRF44, BD Pharmingen), and LFA-1/CD11a (HI111, BD Pharmingen). For CD14+ (M5E2, BD Pharmingen) monocytes: tissue factor (HTF-1, BD Pharmingen). For CD41+ (HIP2, BD Pharmingen) platelets: CD16b (1D3, Beckman Coulter) and CD14 (M5E2, BD Pharmingen). The percentages of positive cells and median fluorescent intensity (MFI) are assessed for each haematologica | 2018; 103(5)

Safety and efficacy of plerixafor in SCD patients

A

B

C

Figure 2. Correlation between post-plerixafor CD34+ cell counts and baseline cell counts. The CD34+ level at 12 h correlated positively with the baseline level of CD34+ (P=0.0006) but not baseline levels of absolute neutrophil count (ANC, P=0.66) or white blood cells (WBC, P=0.49). The graphs show the association between the value of peripheral blood CD34 concentration at 12 h after plerixafor and the baseline values of (A) CD34, (B) ANC and (C) WBC in 15 patients with SCD treated with 80 (circles), 160 (squares) and 240 (triangles) μg/Kg of plerixafor. Patients on hydroxyurea are represented by filled circles, squares and triangles, patients off hydroxyurea are represented by open circles, squares and triangles.

patient, with calculation of the absolute number of positive cells by multiplying the percentage of positive cells by the relevant number of cells obtained from a concurrent complete blood count. For plasma studies, whole blood is centrifuged at 2500 rpm for 15 min at 4°C and plasma frozen at ≤ -80°C until batch testing. The following coagulation system activation markers relevant to sickle vaso-occlusion are tested:26,29 prothrombin fragment 1.2 (Enzygnost F1+2, Dade-Behring) and factor VIII (STA®ImmunoDef VIII, Stago).

Statistical analysis Absolute concentrations as well as the fold increases (ratio of peak at 12 h to baseline absolute concentrations) of standard clinical as well as research parameters were analyzed. For patient 8(2), to avoid computing correlations on an infinite value, we replaced the baseline value of 0 by 0.2 in the analyses using CD34+ ratio (10/0.2 = 50). When applying non-parametric statistics, the chosen value does not affect the results, as long as it is close to 0. The presence of a trend in increases of CD34+, ANC, and WBC counts (both at 12 h and fold increases) with increasing dose was tested using the non-parametric Cuzick test for trend. Correlations between baseline values, and values at 12 h or fold increases were estimated using the Kendall tau. A difference in the distribution of 12 h and fold increases according to the administration of hydroxyurea was investigated using a Wilcoxon test.

For analyses of cell activation and coagulation, Wilcoxon signed rank paired testing was performed on the combined dose cohorts for differences between 12 h and baseline values, whereas the presence of significant fold differences between dose levels was examined with the Kruskal-Wallis test. Correlations between values were estimated using the Kendall tau. Twotailed P values <0.05 were considered statistically significant.

Results Patients’ characteristics Fifteen subjects have been recruited to date for the study at the first three dose levels of 80, 160 and 240 μg/kg. Fourteen patients were enrolled from Montefiore haematologica | 2018; 103(5)

Medical Center (New York, USA) and one patient from The Mount Sinai Hospital (New York, USA) (Table 1).

Two patients enrolled at dose levels 1 and 2 were subsequently re-enrolled in the study at a higher plerixafor dose (dose level 3). All subjects had a past history of moderate to severe acute chest syndrome, defined by requiring treatment with simple or exchange transfusion.

Importantly, for safety and feasibility, patients were continued on their standard outpatient treatment being used to control their disease. Ten of 15 patients were on hydroxyurea, with a median HbF level of 12.4% (Online Supplementary Table S1) and median baseline ANC of 4100/μL (Online Supplementary Table S2). Only one of the 15 subjects was receiving chronic transfusion therapy, with a HbF of 1.2% and HbA of 54%; this patient was also on deferasirox for the treatment of transfusion-related iron overload. HbA was absent in all other patients. In the non-transfused patients, HbF levels correlated strongly with hemoglobin concentration, hematocrit, and reticulocyte counts. Of nine patients for whom splenic imaging was available, seven had splenic atrophy (Online Supplementary Table S1).

Efficacy of CD34+ mobilization Absolute WBC counts, neutrophil counts and CD34+ cell concentrations increased from baseline in all patients (Figure 1). Absolute monocyte and lymphocyte counts also increased from baseline (Online Supplementary Table S2). Our target goal of mobilizing at least 30 CD34+ cells/μL was, however, reached in only 50% of patients given the plerixafor dose of 80 μg/kg, 33% of patients given 160 μg/kg, and 33% of patients given 240 μg/kg.

Peak ANC (P=0.03) and WBC count (P=0.05), but not CD34+ cell count (P=0.65), increased with increasing dose level. As previously reported in healthy donors,13,30,31 there was a strong correlation of peak CD34+ count with baseline CD34+ concentration (Kendall tau=0.68, P=0.0006) but no correlation was observed with baseline ANC (Kendall tau=0.09, P=0.66) or baseline WBC count (Kendall tau=0.13, P=0.49) (Figure 2). There was also no correlation, as previously reported,13 with baseline platelet 773

F. Boulad et al.

A

B

C

Figure 3. Association of hydroxyurea treatment with post-plerixafor cell counts. The use or not of hydroxyurea was not associated with different levels of CD34+ (P=0.66), absolute neutrophil count (ANC, P=0.66), or white blood cell count (WBC, P=0.57) 12 h after plerixafor. Peripheral blood (A) CD34 concentrations, (B) ANC, and (C) WBC at 12 h after plerixafor in 15 patients with SCD treated with 80 (circles), 160 (squares) and 240 (triangles) μg/kg of plerixafor. Patients on hydroxyurea are represented by filled circles, squares and triangles, patients off hydroxyurea are represented by open circles, squares and triangles.

count (Kendall tau=0.30, P=0.13) or donor age (Kendall tau=0.14, P=0.49).

There was a significant increase in the median CD34+ fold increase with dose (P=0.01) (Online Supplementary Figure S1). A trend was observed for ANC ratio, although not statistically significant (1.6-, 1.8-, and 2.1-fold increases, P=0.08). No trend was seen for WBC ratio (P=0.13).

With the caveat of statistical adjustment for patient 8(2)'s baseline CD34+ count of 0/μL, there was no correlation between CD34+ fold increase and baseline CD34+ (Kandall tau= -0.32, P=0.11), baseline ANC (Kendall tau=0.19, P=0.32), or baseline WBC (Kendall tau=0.22, P=0.25) (Online Supplementary Figure S2).

Hydroxyurea was not associated with differences in peak absolute CD34+ concentration (P=0.95), peak ANC (P=0.59) or peak WBC (P=0.68) concentrations (Figure 3); or with differences in the fold increases of CD34+ cell (P=0.64), ANC (P=0.12), or WBC (P=0.36) concentrations (Online Supplementary Figure S3).

In a subset of six patients, CD34+CD38– cells were enumerated (Online Supplementary Table S3), showing a median 3-fold increase in CD34+CD38– concentrations at 12 h.

Safety of plerixafor There were no significant changes in hemoglobin concentration, hematocrit, or platelet counts with plerixafor treatment (data not shown, baseline values in Online Supplementary Table S1). Due to the occurrence of one serious adverse event at the 80 μg/kg dose and another one at the 240 μg/kg dose, an additional three patients were enrolled at each of these dose levels. The serious adverse events were both pain crises, possibly related to plerixafor (Online Supplementary Table S4), but also associated with other possibly contributory events. Patient 13 with a serious adverse event had the second highest peak ANC (and third highest peak WBC count) in the study, but high WBC count and ANC were not consistently associated with serious adverse events.

There were no significant differences between dose levels for any of the activation markers of vaso-occlusion tested. With the exception of tissue factor-positive (TF+) mono774

cytes at the 240 µg/kg dose, the median fold-changes in percentage of cells at every dose cohort were ≤1.1, arguing against generalized plerixafor-mediated cell activation (Figure 4A). Median fold increases in absolute numbers of activated β2 integrin-positive (aβ2+) neutrophils, activated Mac-1-positive (aMac-1+) neutrophils, and TF+ monocytes (160 μg/kg, 240 μg/kg) were close to 2 (Figure 4B). There were strong correlations between the fold increase in absolute numbers of neutrophils and fold increases in aβ2+ neutrophils (Kendall tau=0.85, P<0.001) and aMac-1+ neutrophils (Kendall tau=0.46, P=0.02). The absolute numbers of aβ2+ and aMac-1+ neutrophils were significantly increased (Online Supplementary Figure S4A,B), and the two patients with serious adverse events (gray arrows) had relatively high absolute numbers of aβ2+ and aMac-1+ neutrophils, albeit not the highest. There was also a significant increase in plasma prothrombin fragment 1.2 concentrations (Online Supplementary Figure S4C), but the two patients with serious adverse events had absolute concentrations and fold increases at 12 h that were lower than the median and mean for that measure. Both patients with serious adverse events had relatively high fold-increases in L-selectinneg neutrophils and one had a large fold increase in TF+ monocytes (Figure 5), but their absolute numbers of Lselectinneg neutrophils and TF+ monocytes were not particularly high (Online Supplementary Figure S4D,E). There were significant decreases for five parameters: percentage of aβ2+ neutrophils, MFI of aβ2 neutrophils, percentage of TF+ monocytes, and percentage and absolute number of platelet-neutrophil aggregates (Online Supplementary Figure S5). There were no significant changes in the MFI of aMac1+, Mac-1, LFA-1, or L-selectin on neutrophils (data not shown).

Discussion Eight of 15 patients (53%) with SCD treated with plerixafor reached the peripheral blood CD34 cell target count of at least 30 CD34+ cells/μL, including three of six patients treated at a dose of 240 μg/kg. This is in contrast haematologica | 2018; 103(5)

Safety and efficacy of plerixafor in SCD patients

A

Figure 4. Fold changes in the percentage of cells expressing activation markers, according to plerixafor dose. (A) Other than % TF+ monocytes, median fold changes in % cells were ≤1.1. (B) Median fold changes in absolute number of cells for TF+ mono, aβ2+ PMN, and aMac-1+ PMN were closer to 2 than 1.

Fold changes are ratios of 12 h to baseline values. Horizontal lines represent medians. Filled shapes represent hydroxyurea-treated patients, open shapes represent patients not treated with hydroxyurea. Gray arrows point to the values for the two SAE patients who had serious adverse events. TF+ mono: tissue-factor-positive monocytes; E-Sel+ PMN: E-selectin-positive neutrophils; aβ2+ PMN: activated β2 integrin-positive neutrophils, aMac-1+ PMN: activated Mac-1-positive neutrophils; L-Selneg PMN: L-selectin-negative neutrophils; PMA: platelet-monocyte aggregates; PNA: platelet-neutrophil aggregates.

B

with the findings of a recent study by Tisdale et al., in which mobilization was effective in seven of seven SCD subjects (100%) at a dose of 240 μg/kg.4 It should be noted that patients in the National Institutes of Health study were off hydroxyurea and had been transfused for at least 2 months to achieve a HbS <20-30%4,32 while in our study, ten of the 15 patients were on stable doses of hydroxyurea (for at least 1 year) and only one patient was on chronic transfusion. Although hydroxyurea, a ribonucleotide reductase inhibitor, causes myelosuppression and was recently found to reduce CD34 counts in peripheral blood and bone marrow,33 there is no definitive evidence that hydroxyurea negatively affects numbers or quality of cell cycle-quiescent hematopoietic stem cells or immature bone marrow progenitors as opposed to more mature myeloid-erythroid progenitors.33-37 Indeed, in our study, although we did not achieve consistent efficiency in CD34 cell mobilization, no correlation was found between hydroxyurea use, and absolute or fold increases in CD34+ cells/μL. We observed wide inter-donor variability in CD34 mobilization with plerixafor, as previously reported in normal donors (CD34 peaks between 4-157/μL )13 and in patients with SCD (CD34 peaks between 50-200/μL).4 However, we also observed a strong correlation between baseline CD34+ and peak CD34+ concentrations, as previously reported with both G-CSF and plerixafor mobilization in healthy donors (Kendall tau=0.68, P=0.0006).13,30,31 Factors contributing to baseline CD34 count remain unclear, but our data and others’ suggest that baseline CD34+ concentration may be affected by hydroxyureahaematologica | 2018; 103(5)

related myelosuppression.24,33 Patient #8, a subject reenrolled in the study, was particularly instructive regarding this hypothesis. This patient was clinically stable on hydroxyurea at a dose of 27 mg/kg and was enrolled twice at an interval of 13 months. At the time of his second treatment, however, he had a markedly lower baseline ANC (1900/μL down from 6300/μL) and platelet count (217,000/μL down from 400,000/μL), probably related to oscillatory non-toxic hematopoiesis seen in SCD with chronic and dose-intensive treatment with hydroxyurea (ANC oscillations between 2,000-6,000/μL as determined from review of his clinical laboratory records).38 This myelosuppression was associated with a baseline CD34 concentration of 0/μL rather than 1/μL, possibly contributing to the relatively low 12 h CD34+ concentration of 10/μL at the 240 μg/kg dose as compared to 27/μL at the 160 μg/kg dose. In brief, because hydroxyurea can decrease ANC and platelet count,39 hydroxyurea-related myelosuppression may have contributed to the relatively poor CD34+ mobilization obtained in this cohort.

However, avoiding hydroxyurea withdrawal might lower the risk of pain crises;24 we, therefore, plan to explore timing plerixafor administration to the peak rather than nadir of hydroxyurea-related oscillatory hematopoiesis. Finally with regards to hydroxyurea therapy, data from the six patients in whom we enumerated CD34+CD38– cells suggest that hydroxyurea may not adversely affect HSC, given that all patients except one (patient 10) were on hydroxyurea and a median 3-fold increase at 12 h was observed. Only 0.2-2.8% of CD34+ cells were CD38-negative, but this may be consistent with plerixafor’s effect in 775

F. Boulad et al.

normal healthy donors, in that fewer HPC may be mobilized by plerixafor than by G-CSF, where up to 50% of GCSF-mobilized CD34+ cells are CD38-negative.13,40 Because we enrolled only one patient on chronic transfusion, we cannot assess any correlation between transfusion and CD34 mobilization, although notably this patient had the second highest baseline and highest peak CD34+ cell counts in our study. Other studies of plerixafor in SCD4,32 have initiated chronic transfusion based on the hypothesis that the inflammatory nature of SCD affects the bone marrow and transfusion assuages bone marrow inflammation and stress erythropoiesis. Although replicative and oxidative stress of HPC in bone marrow may occur,41-44 there is limited evidence that HPC are damaged in SCD.36 Five of our patients had HbF-associated increases in hemoglobin concentration and hematocrit to more than 10 g/dL and 30%, respectively (similar to values in chronically transfused patients) but HbF levels did not correlate with CD34 cell mobilization.

Based on our data, it is possible that continued dose escalation could result in greater efficacy of mobilization, since we observed a dose-related response in the median CD34+ cell fold increase (P=0.01), as also observed in healthy donors.45 Patient 3, a repeat enrollment who had never been on hydroxyurea and was clinically stable, is instructive in that his 12 h peak CD34+ cell count following a plerixafor dose of 80 μg/kg was only 8/μL whereas at the dose of 240 μg/kg it was 40/μL, even though his baseline CD34+ cell concentrations (1/μL and then 2/μL) were similar, suggesting a dose-response to plerixafor. Notably, his two periods in the study were separated by 19 months, suggesting that, as with healthy donors,46 intra-individual CD34+ cell counts in stable SCD patients not on hydroxyurea may remain stable over time. Based on these data and given the safety and continued dose response between 240 μg/kg and 480 μg/kg observed in healthy donors,20 we plan to continue dose escalation in SCD patients through to the 480 μg/kg dose, barring significant adverse events.

Adding the CXCR2 agonist, GROβ, might be useful.47 Only two of 15 patients (13%) developed serious adverse events as compared to three of seven patients (43%) in the study of plerixafor mobilization in SCD by Tisdale et al.,4 although this must be qualified by the fact that the patients in the study by Tisdale et al. also underwent leukapheresis.

Our low rate of serious adverse events could, however, also be due to chronic hydroxyurea therapy and the subsequent lower WBC and ANC peaks. As the fraction of activated neutrophils did not increase significantly with plerixafor, our low serious adverse event rate may be related to moderation of ANC elevations by hydroxyurea, reducing the absolute number of activated cells. Given the still uncertain risks of morbidity, as seen with G-CSF, the use of plerixafor in SCD requires further evaluation.

In summary, our present data suggest that, with regards the efficacy of CD34 mobilization, red blood cell transfu-

References 1. Mansilla-Soto J, Riviere I, Sadelain M. Genetic strategies for the treatment of sickle cell anaemia. Br J Haematol. 2011;154(6): 715-727.

776

sion may be more effective than continuing standard of care. Whether red blood cell transfusion will remain more effective as we escalate the plerixafor dose (as safety allows) to 480 μg/kg, with protocol revisions for hydroxyurea-treated patients, is unknown. Finally, potential candidates for SCD gene therapy may not be able to receive regular red blood cell transfusions (e.g. if they have red cell alloimmunization or a history of hyperhemolysis) or may not be willing to do so (e.g. Jehovah Witnesses), even for the relatively short duration of 2-3 months.

This study has several limitations. Firstly, despite this study being the largest study to date of plerixafor administration in SCD patients, overall the number of patients involved remains small; thus comparisons, for example, between hydroxyurea-treated and non-hydroxyurea-treated patients, may not be representative of the actual populations. Secondly, we measured WBC, ANC and CD34 mobilization in this study only at ~12 and ~20 h after plerixafor administration. It is possible that an initial peak could have occurred at an earlier time (6-9 h) after plerixafor and could, therefore, have been missed. Nevertheless, CD34 cell concentrations remain at ~70% of peak levels at 12 h.20,45,48 Our current study will be amended to include the addition of earlier post-plerixafor assessments. Thirdly, we determined peripheral blood CD34 cell mobilization in the 15 patients treated with plerixafor, without performing apheresis. However, there is a well-described correlation between peripheral blood CD34 cell concentration and the ultimate CD34 cell dose obtained after apheresis. It is possible that technical adjustments may be required for this equation in the context of SCD. Finally, other than enumerating CD34+CD38- cells, we did not further characterize CD34+ cells to study “stemness”, for example by determining glycophorin A positivity and CD34 dimness.4 CD34+ or CD34+CD38- enumeration is not specific for HSC49 and it is, therefore, unclear whether patients had a true increase in HSC, as opposed to more mature lineagecommitted CD34+ progenitors, which are either mobilized or present in bone marrow.42 We plan to characterize the CD34+ cells further as we move forward in the study, which is currently enrolling at the 320 μg/kg dose level.

Despite mobilization of HSC possibly being less efficient with plerixafor than with G-CSF, plerixafor-mobilized HSC may have an engraftment advantage over G-CSFmobilized HSC with regard to better retention of CXCR4, which facilitates homing.32 Acknowledgments The authors would like to thank our study subjects for their participation; the Doris Duke Charitable Foundation for a 2011 Innovation in Clinical Research Award for trial support (to PAS and MS); Sanofi-Genzyme for provision of plerixafor; Jena Simon for referring one study patient; W. Beau Mitchell for assistance with platelet activation studies; and Henny Billett, Narla Mohandas, and Beth Shaz for departmental support.

2. Ribeil JA, Hacein-Bey-Abina S, Payen E, et al. Gene therapy in a patient with sickle cell disease. N Engl J Med. 2017;376 (9):848-855.

3. Sadelain M, Boulad F, Lisowki L, Moi P, Riviere I. Stem cell engineering for the treatment of severe hemoglobinopathies. Curr Mol Med. 2008;8(7):690-697.

4. Tisdale JF, Pierciey Jr. JF, Kanter J, et al.

Successful Plerixafor-mediated mobilization, apheresis, and lentiviral vector transduction of hematopoietic stem cells in patients with severe sickle cell disease. Blood. 2017;130 (Suppl 1):990.

5. Fitzhugh CD, Hsieh MM, Bolan CD, Saenz C, Tisdale JF. Granulocyte colony-stimulating factor (G-CSF) administration in individ-

haematologica | 2018; 103(5)

Safety and efficacy of plerixafor in SCD patients

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

uals with sickle cell disease: time for a moratorium? Cytotherapy. 2009;11(4):464-471.

Petit I, Szyper-Kravitz M, Nagler A, et al. GCSF induces stem cell mobilization by decreasing bone marrow SDF-1 and up-regulating CXCR4. Nat Immunol. 2002;3(7): 687-694.

Cella G, Marchetti M, Vignoli A, et al. Blood oxidative status and selectins plasma levels in healthy donors receiving granulocytecolony stimulating factor. Leukemia.

2006;20(8):1430-1434. de Haas M, Kerst JM, van der Schoot CE, et al. Granulocyte colony-stimulating factor administration to healthy volunteers: analysis of the immediate activating effects on circulating neutrophils. Blood. 1994;84(11): 3885-3894.

Falanga A, Marchetti M, Evangelista V, et al.

Neutrophil activation and hemostatic changes in healthy donors receiving granulocyte colony-stimulating factor. Blood.

1999;93(8):2506-2514.

Spiel AO, Bartko J, Schwameis M, et al.

Increased platelet aggregation and in vivo platelet activation after granulocyte colonystimulating factor administration. A randomised controlled trial. Thromb Haemost.

2011;105(4):655-662.

Canales MA, Arrieta R, Gomez-Rioja R, Diez J, Jimenez-Yuste V, HernandezNavarro F. Induction of a hypercoagulability state and endothelial cell activation by granulocyte colony-stimulating factor in peripheral blood stem cell donors. J Hematother Stem Cell Res. 2002;11(4):675-681.

Zhang D, Xu C, Manwani D, Frenette PS. Neutrophils, platelets, and inflammatory pathways at the nexus of sickle cell disease pathophysiology. Blood. 2016;127(7):801809.

Schroeder MA, Rettig MP, Lopez S, et al.

Mobilization of allogeneic peripheral blood stem cell donors with intravenous plerixafor mobilizes a unique graft. Blood. 2017;129 (19):2680-2692.

Choi E, Branch C, Cui MH, et al. No evidence for cell activation or brain vaso-occlusion with plerixafor mobilization in sickle cell mice. Blood Cells Mol Dis. 2016;57:67-70.

Charache S. Mechanism of action of hydroxyurea in the management of sickle cell anemia in adults. Semin Hematol.

1997;34(3 Suppl 3):15-21.

Castro O, Brambilla DJ, Thorington B, et al.

The acute chest syndrome in sickle cell disease: incidence and risk factors. The Cooperative Study of Sickle Cell Disease.

Blood. 1994;84(2):643-649.

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.

Litos M, Sarris I, Bewley S, Seed P, Okpala I, Oteng-Ntim E. White blood cell count as a predictor of the severity of sickle cell disease during pregnancy. Eur J Obstet Gynecol Reprod Biol. 2007;133(2):169-172.

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.

Pantin J, Purev E, Tian X, et al. Effect of highdose plerixafor on CD34(+) cell mobilization in healthy stem cell donors: results of a

haematologica | 2018; 103(5)

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

randomized crossover trial. Haematologica.

2017;102(3):600-609.

Lamming CE, Augustin L, Blackstad M, Lund TC, Hebbel RP, Verfaillie CM. Spontaneous circulation of myeloid-lymphoid-initiating cells and SCID-repopulating cells in sickle cell crisis. J Clin Invest.

2003;111(6):811-819.

Croizat H, Ponchio L, Nicolini FE, Nagel RL, Eaves CJ. Primitive haematopoietic progenitors in the blood of patients with sickle cell disease appear to be endogenously mobilized. Br J Haematol. 2000;111 (2):491-497.

Yannaki E, Papayannopoulou T, Jonlin E, et al. Hematopoietic stem cell mobilization for gene therapy of adult patients with severe beta-thalassemia: results of clinical trials using G-CSF or plerixafor in splenectomized and nonsplenectomized subjects. Mol Ther.

2012;20(1):230-238.

Richard RE, Siritanaratkul N, Jonlin E, Skarpidi E, Heimfeld S, Blau CA. Collection of blood stem cells from patients with sickle cell anemia. Blood Cells Mol Dis. 2005;35(3):384-388.

Manwani D, Chen G, Carullo V, et al.

Single-dose intravenous gammaglobulin can stabilize neutrophil Mac-1 activation in sickle cell pain crisis. Am J Hematol.

2015;90(5):381-385.

Jakubowski JA, Zhou C, Jurcevic S, et al. A phase 1 study of prasugrel in patients with sickle cell disease: effects on biomarkers of platelet activation and coagulation. Thromb Res. 2014;133(2):190-195.

Canalli AA, Proenca RF, Franco-Penteado CF, et al. Participation of Mac-1, LFA-1 and VLA4 integrins in the in vitro adhesion of sickle cell disease neutrophils to endothelial layers, and reversal of adhesion by simvastatin.

Haematologica. 2011;96(4):526-533.

Wun T, Styles L, DeCastro L, et al. Phase 1 study of the E-selectin inhibitor GMI 1070 in patients with sickle cell anemia. PLoS One. 2014;9(7):e101301.

Nur E, van Beers EJ, Martina S, et al. Plasma levels of pentraxin-3, an acute phase protein, are increased during sickle cell painful crisis.

Blood Cells Mol Dis. 2011;46(3):189-194. de la Rubia J, Lorenzo JI, Torrabadella M, Marin P, Insunza A, Sanz MA. Basal CD34(+) cell count predicts peripheral blood progenitor cell mobilization and collection in healthy donors after administration of granulocyte colony-stimulating factor.

Haematologica. 2004;89(12):1530-1532.

Martino M, Gori M, Pitino A, et al. Basal CD34(+) Cell count predicts peripheral blood stem cell mobilization in healthy donors after administration of granulocyte colony-stimulating factor: a longitudinal, prospective, observational, single-center, cohort study. Biol Blood Marrow Transplant. 2017;23(7):1215-1220.

Kanter J, Walters MC, Hsieh MM, et al.

Interim results from a phase 1/2 clinical study of lentiglobin gene therapy for severe sickle cell disease. Blood. 2017;128 (22):1176.

Uchida N, Fujita A, Hsieh MM, et al. Bone marrow as a hematopoietic stem cell source for gene therapy in sickle cell disease: evidence from Rhesus and SCD patients. Hum Gene Ther Clin Dev. 2017;28(3):136-144.

Miller ST, Rey K, He J, et al. Massive acci-

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

dental overdose of hydroxyurea in a young child with sickle cell anemia. Pediatr Blood Cancer. 2012;59(1):170-172.

Castro O, Nouraie M, Oneal P. Hydroxycarbamide treatment in sickle cell disease: estimates of possible leukaemia risk and of hospitalization survival benefit. Br J Haematol. 2014;167(5):687-691.

Brunson A, Keegan THM, Bang H, Mahajan A, Paulukonis S, Wun T. Increased risk of leukemia among sickle cell disease patients in California. Blood. 2017;130(13):15971599.

Drasar ER, Jiang J, Gardner K, et al.

Leucocyte telomere length in patients with sickle cell disease. Br J Haematol. 2014;165 (5):725-727.

Baird JH, Minniti CP, Lee JM, et al.

Oscillatory haematopoiesis in adults with sickle cell disease treated with hydroxycarbamide. Br J Haematol. 2015;168(5):737-746.

Zimmerman SA, Schultz WH, Davis JS, et al. Sustained long-term hematologic efficacy of hydroxyurea at maximum tolerated dose in children with sickle cell disease. Blood.

2004;103(6):2039-2045.

Worel N, Frank N, Frech C, Fritsch G. Influence of plerixafor on the mobilization of CD34+ cell subpopulations and lymphocyte subtypes. Transfusion. 2017;57(9): 2206-2215.

Javazon EH, Radhi M, Gangadharan B, Perry J, Archer DR. Hematopoietic stem cell function in a murine model of sickle cell disease.

Anemia. 2012;2012:387385.

Leonard A, Bonifacino A, Dominical VM, et al. Bone marrow characterization in sickle cell disease: inflammation and stress erythropoiesis lead to suboptimal CD34 recovery compared to normal volunteer bone marrow. Blood. 2017;130(Suppl 1):966.

Dallalio G, Brunson CY, Means RT, Jr. Cytokine concentrations in bone marrow of stable sickle cell anemia patients. J Investig Med. 2007;55(2):69-74.

Colella MP, Santana BA, Conran N, et al.

Telomere length correlates with disease severity and inflammation in sickle cell disease. Rev Bras Hematol Hemoter. 2017;39 (2):140-145.

Liles WC, Broxmeyer HE, Rodger E, et al.

Mobilization of hematopoietic progenitor cells in healthy volunteers by AMD3100, a CXCR4 antagonist. Blood. 2003;102(8): 2728-2730.

Eidenschink L, DiZerega G, Rodgers K, Bartlett M, Wells DA, Loken MR. Basal levels of CD34 positive cells in peripheral blood differ between individuals and are stable for 18 months. Cytometry B Clin Cytom.

2012;82(1):18-25.

Hoggatt J, Singh P, Tate TA, et al. Rapid mobilization reveals a highly engraftable hematopoietic stem cell. Cell. 2018;172(12):191-204.e10.

Lemery SJ, Hsieh MM, Smith A, et al. A pilot study evaluating the safety and CD34+ cell mobilizing activity of escalating doses of plerixafor in healthy volunteers. Br J Haematol. 2011;153(1):66-75.

Notta F, Doulatov S, Laurenti E, Poeppl A, Jurisica I, Dick JE. Isolation of single human hematopoietic stem cells capable of longterm multilineage engraftment. Science.

2011;333(6039):218-221.

777

ARTICLE

Red Cell Biology & its Disorders

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):778-786

Plerixafor enables safe, rapid, efficient mobilization of hematopoietic stem cells in sickle cell disease patients after exchange transfusion

Chantal Lagresle-Peyrou,1,2,3* François Lefrère,4* Elisa Magrin,1,4* Jean-Antoine Ribeil,1,4* Oriana Romano,3,5,6 Leslie Weber,2,3,7 Alessandra Magnani,1,4 Hanem Sadek,1,2,3 Clémence Plantier,1,4 Aurélie Gabrion,1,4 Brigitte Ternaux,1,4 Tristan Félix,3,5 Chloé Couzin,1,4 Aurélie Stanislas,1,4 Jean-Marc Tréluyer,8 Lionel Lamhaut,9,10 Laure Joseph,4 Marianne Delville,2,3,4 Annarita Miccio,3,5# Isabelle André-Schmutz1,2,3# and Marina Cavazzana1,2,3,4#

Biotherapy Clinical Investigation Center, Groupe Hospitalier Universitaire Ouest, Assistance Publique-Hôpitaux de Paris, INSERM CIC 1416, France; 2Laboratory of Human Lymphohematopoiesis, INSERM UMR 1163, Imagine Institute, Paris, France: 3 Paris Descartes University – Sorbonne Paris Cité, Imagine Institute, France 4 Department of Biotherapy, Necker Children’s Hospital, Assistance Publique-Hôpitaux de Paris, France; 5Laboratory of Chromatin and Gene Regulation during Development, INSERM UMR1163, Imagine Institute, Paris, France; 6Department of Life Sciences, University of Modena and Reggio Emilia, Modena, Italy; 7Paris Diderot University – Sorbonne Paris Cité, France; 8Mère-Enfant Clinical Investigation Center, Groupe Hospitalier Necker Cochin, Assistance Publique-Hôpitaux de Paris, France; 9Intensive Care Unit, Anaesthesia and SAMU de Paris, Necker Hospital, Assistance PubliqueHôpitaux de Paris, France and 10Paris Descartes University – Sorbonne Paris Cité, France.

1

*CLP, FL and EM and JAR contributed equally to this work, in alphabetical order #

AM, IAS and MC contributed equally to this work

Correspondence: isabelle.andre-schmutz@inserm.fr

Received: November 15, 2017.

Accepted: February 13, 2018.

Pre-published: February 22, 2018. doi:10.3324/haematol.2017.184788 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/778 ©2018 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.

778

ABSTRACT

S

ickle cell disease is characterized by chronic anemia and vaso-occlusive crises, which eventually lead to multi-organ damage and premature death. Hematopoietic stem cell transplantation is the only curative treatment but it is limited by toxicity and poor availability of HLA-compatible donors. A gene therapy approach based on the autologous transplantation of lentiviral-corrected hematopoietic stem and progenitor cells was shown to be efficacious in one patient. However, alterations of the bone marrow environment and properties of the red blood cells hamper the harvesting and immunoselection of patients’ stem cells from bone marrow. The use of Filgrastim to mobilize large numbers of hematopoietic stem and progenitor cells into the circulation has been associated with severe adverse events in sickle cell patients. Thus, broader application of the gene therapy approach requires the development of alternative mobilization methods. We set up a phase I/II clinical trial whose primary objective was to assess the safety of a single injection of Plerixafor in sickle cell patients undergoing red blood cell exchange to decrease the hemoglobin S level to below 30%. The secondary objective was to measure the efficiency of mobilization and isolation of hematopoietic stem and progenitor cells. No adverse events were observed. Large numbers of CD34+ cells were mobilized extremely quickly. Importantly, the mobilized cells contained high numbers of hematopoietic stem cells, expressed high levels of stemness genes, and engrafted very efficiently in immunodeficient mice. Thus, Plerixafor can be safely used to mobilize hematopoietic stem cells in sickle cell patients; this finding opens up new avenues for treatment approaches based on gene addition and genome editing. Clinicaltrials.gov identifier: NCT02212535. haematologica | 2018; 103(5)

Stem cell mobilization in sickle cell patients

Introduction Sickle cell disease (SCD) is caused by a point mutation in the coding region of the HBB (β-globin) gene. As a result, an abnormal β-globin protein is incorporated into hemoglobin tetramers. These mutant tetramers polymerize when the local oxygen tension is low. The sickle hemoglobin (HbS) polymers rigidify red blood cells, change these cells’ shape, and are responsible for structural damage to the red blood cell membrane. In turn, this modifies the cells’ rheological properties, alters their flow in the microcirculation, and thus causes ischemia, stroke, multi-organ damage, severe acute and chronic pain, and chronic hemolytic anemia. Progressive chronic organ complications become the main cause of morbidity and mortality in the third decade of life.1 SCD is endemic in Africa, and the World’s Health Organization considers that 7% of the world population carries the trait.

The only curative treatment for SCD is allogeneic hematopoietic stem cell transplantation (HSCT) from matched sibling donors; the disease-free survival rate 6 years after transplantation is reportedly >90%.2,3 Given the limited availability of suitable donors and the increase in toxicity with age, HSCT is only applied with great caution in adult SCD patients (the main target population for curative treatment).

We recently demonstrated that gene therapy is applicable to SCD patients, and that the associated toxicity and morbidity rates seem to be lower than those for allogeneic HSCT, at least in the first treated patient.4 As is the case with all genetic diseases, the success of gene therapy in SCD relies on several key factors; these include the source, quality and number of transduced cells, the choice of the conditioning regimen, the level of therapeutic transgene expression, and the quality of the bone marrow (BM) microenvironment at the time of harvest and transplantation. It is generally acknowledged that 2 to 3x106 CD34+ hematopoietic stem and progenitor cells (HSPC)/kg are required for a successful outcome in autologous HSCT.5 Considering the typical proportion of HSPC that can be corrected in gene therapy clinical trials (~50% of CD34+ HSPC) and an average recovery of 70% of CD34+ cells post-selection, a minimum harvest of ~6x106 CD34+ cells/kg would be required. For reasons that have not been completely elucidated, as for thalassemic

patients,6-7 the recovery of HSPC from SCD patients’ BM is peculiarly low (M. Cavazzana, unpublished data). In our ongoing gene therapy trial (HGB-205, ClinicalTrials.gov number, NCT02151526), two BM harvests (each requiring an exchange transfusion program before general anesthesia) were needed to obtain enough cells to transplant the three SCD patients enrolled.4 Mobilization with granulocyte colony-stimulating factor (G-CSF, Filgrastim) is widely used to increase the harvest of HSPC (relative to that obtained in a BM aspirate).

However, attempts to mobilize HSPC with Filgrastim in SCD have led to severe adverse events, which hamper the cytokine’s use in this setting. The first report of a severe adverse event following mobilization with low-dose Filgrastim (2.5 μg/kg/day) concerned a SCD patient who developed acute chest syndrome and an elevated white blood cell count (63,400/mm3) as early as 3 days after the first injection.8 The temporal relationship between Filgrastim administration, the rapid rise in the white blood cell count and the severe adverse event were strongly suggestive of a causal link. Two other severe adverse events (multi-organ failure and a death) were reported in 2001.9-10 Between 2003 and 2008, eight other patients requiring autologous HSCT for a malignant hematopoietic disease were mobilized with Filgrastim after additional precautions had been taken: reduction of the HbS level to below 30% via red blood cell exchange, and carefully monitoring of peripheral leukocytosis and the blood ion profile.11-13 Although all eight patients experienced bone pain, hypertension or migraine, no severe adverse events were reported - providing evidence that SCD patients can be mobilized without any major complications. The median [interquartile range] circulating CD34+ cell count was 24.4/μL [21.2-48.6].12-14 As an alternative to Filgrastim, Plerixafor (formerly known as AMD3100) can effectively mobilize HSPC; the CD34+ cell count in the circulation can reach 15 to 40/μL, depending on the dose and the study.15,16 In contrast to Filgrastim (which acts by activating monocytes and neutrophils in both peripheral blood and the BM), Plerixafor directly inhibits the binding of stroma-cell-derived factor1a to its CXC chemokine receptor on HSPC – thus releasing stem cells from the BM niches. This mechanistic difference explains the small increase in the white blood cell count and the short time interval between Plerixafor

Table 1. Clinical parameters of patients treated with Plerixafor.

SCD Pler 1 (19 years) D+1 D+7 D+30

Normal values

D-1

0% 0-17 0-5 125-243 4·0-10·0 1·5-7 0·2-1 22-275 < 36

12·1% 112 8 399 9·9 4·9 2·0 2441

HbS (HPLC G8) Total bilirubin (mol/L) Conjugated bilirubin (mol/L) Lactate dehydrogenase (U/L) White blood cells (109/L) Neutrophils (109/L) Monocytes (109/L) Ferritin (μg/L) Liver quantification (at inclusion period) (μmol/g)

12·6% 15·5% 105 109 7 8 409 361 20·1 11·8 14·9 7·8 2·4 1·7 ND 2674 340 (±50)

36·0% 132 9 429 13·2 8·2 2·3 3111

D-1 20·2% 32 11 566 11·9 7·6 0·9* ND

SCD Pler 2 (20 years) D+1 D+7 D+30 15·8% ND 35 ND 13 ND 584 ND 11·2 ND 9·2 ND 0·9 ND 1009 1473 120 (±30)

37·1% 32 12 605 6·2 2·6 0·7 1962

D-1

SCD Pler 3 (21 years) D+1 D+7 D+30

6·2%*,** 13·2%** 25·2%** 36·8%** 30 30 31 53 10 10 10 6 378 383 426 501 14·0 11·3 9·6 10·4 9·6 9·3 6·4 6·5 1·8* 1·3 1·5 1·8 ND 381 365 446 55 (±30)

HPLC: High-performance-liquid-chromatography; D: day. ND: not determined. * D+3. **Capillary 3 (and not G8).

haematologica | 2018; 103(5)

779

C. Lagresle-Peyrou et al.

administration and HSPC mobilization. Furthermore, Plerixafor synergistically augments Filgrastim-induced mobilization of HSPC,17 and so has been widely combined with Filgrastim (i) in Filgrastim-non-responsive patients, (ii) as an adjunct to increase the overall recovery of CD34+ cells for autologous or allogenic HSCT, and (iii) in gene therapy trials for β-thalassemia and Wiskott-Aldrich syndrome.18-20 In view of all the above data, we initiated a phase I/II clinical trial of Plerixafor as a single mobilizing agent (NCT02212535). Our objective was to demonstrate that SCD patients can be safely mobilized under welldefined, controlled clinical and biological conditions.

Moreover, this clinical trial provided an estimate of the overall recovery of CD34+ HSPC after apheresis in adult Plerixafor-mobilized SCD patients – thus providing a basis for the wider use of this protocol for gene-addition or genome-editing approaches.

Methods Study design and human samples This open-label phase I/II trial (ClinicalTrials.gov number, NCT02212535) was sponsored by the Assistance Publique Hôpitaux de Paris. The protocol was reviewed and approved by the French Drug Agency (Agence Nationale de Sécurité du Médicament) and the local independent ethics committee (Comité de Protection des Personnes Ile-de-France II, Paris, France). The trial was performed in accordance with the Declaration of Helsinki. Adult SCD patients and healthy donors (for control samples) provided their written, informed consent (Online Supplementary Table S1).

The study was designed to demonstrate the safety and efficacy of the mobilization and harvesting of peripheral HSPC following a single injection of 0.24 mg/kg Plerixafor in adult SCD patients, with the inclusion criteria described in the Online Supplementary Methods.

A

B

C

Figure 1. Plerixafor is highly efficient at mobilizing hematopoietic stem and progenitor cells from sickle cell disease patients. (A) Changes in white blood cell (WBC) and (B) CD34+ hematopoietic stem/progenitor cell (HSPC) counts over the 66 h following Plerixafor administration in SCD Pler 1 (red squares), SCD Pler 2 (blue circles) and SCD Pler 3 (green triangles). Arrows indicate time and duration of apheresis. (C) Number of hematopoietic stem cells (HSC, black bars) and multipotent progenitors (MPP, dotted bars) per 1,000 CD34+ cells in samples of various origins. HD: healthy donor, BM: bone marrow, SCD: sickle cell disease, Pler: Plerixafor, Filg: Filgrastim.

780

haematologica | 2018; 103(5)

Stem cell mobilization in sickle cell patients

The patients were carefully monitored in the Intensive Care Unit and a few hours before the procedure started, they were given prophylaxis for vaso-occlusive crisis [hyperhydration 60 mL/kg/day of a 0.9% saline solution and oxygen therapy (2 L/min)], as recommended by French guidelines.21 These guidelines recommend oxygen therapy to prevent sickling phenomena in particularly stressful conditions. As mobilization can be considered stressful, we decided to treat all patients with oxygen therapy. Baseline O2 saturation levels were normal and remained stable during the procedure. The patients received a subcutaneous injection of 0.24 mg/kg of Plerixafor. We then monitored whole blood HbS level, plasma bilirubin, conjugated bilirubin and lactate dehydrogenase levels, and white blood cell, neutrophil and monocyte counts for 30 days after the Plerixafor injection. Plerixafor-mobilized HSPC in peripheral blood were collected by using a COBE® Spectra Optia apheresis system (Terumo BCT, Lakewood, CO, USA) with modifications (Online Supplementary Material and Methods). The apheresis lasted from 4 to 6 h. Product volumes and total blood volumes are indicated in Table 2. Egress of the CD34+ cells was monitored hourly for at least 24 h after the Plerixafor injection: total CD34+/kg collected and CD34+ collection efficiency (CE1) [total CD34+ collected/[L processed x (CD34+ pre + CD34+ post)/2] were evaluated following apheresis (Table 2).

Flow cytometry analysis Cells were stained with specific antibodies (Online Supplementary Table S2) and analyzed on a BD FACSCanto™ II system, with gating on viable, 7AAD-negative cells. The data were processed using FlowJo software (version 10.2, Treestar, Ashland, OR, USA).

RNA-Seq extraction Total RNA was extracted from 0.1-1x106 CD34+ cells using an RNeasy Micro kit (QIAGEN). RNA-Seq libraries were prepared from ~10 ng of total RNA, using the Ovation Human FFPE RNA-Seq Multiplex System kit (Nugen) after DNase treatment (ArcticZymes) and >40 million paired-end reads/sample. Samples were generated on a HiSeq 2500 instrument (Illumina). RNA-Seq analysis was performed as described in the Online Supplementary Methods.

Transplantation into non-obese diabetic severe combined immunodeficiency gamma mice The non-obese diabetic severe combined immunodeficiency gamma (NSG) mice (NOD.CgPrkdcscid Il2rgtm1Wj/SzJ, Charles

River Laboratories) were housed in a pathogen-free facility. All experiments were performed in compliance with the French Ministry of Agriculture’s regulations on animal experiments, and were approved by the regional Animal Care and Use Committee (APAFIS#2101-2015090411495178 v4). Six- to 8-week-old mice were conditioned with intraperitoneally injected busulfan (Sigma, 25 mg/kg body weight/day) 72 h, 48 h and 24 h before transplantation. CD34+ cells (300,000 cells/mouse) from Filgrastim-mobilized healthy donors or Plerixafor-mobilized cells from the three SCD patients were transplanted into the NSG mice via retroorbital sinus injection. Neomycin was added to the animals’ acidified drinking water. Engraftment was analyzed as described in the Online Supplementary Methods.

Statistical analysis Statistical tests are indicated in the Figure legends. The threshold for statistical significance was set at P<0.05.

Results Patients’ characteristics Four SCD homozygous patients followed at NeckerEnfants Malades hospital and meeting our inclusion criteria were enrolled between May 2015 and January 2017.

The first one was excluded from the study during the enrollment stage because of elevated granulocyte counts exceeding the limits established in the protocol (<10x109 granulocytes/L). The other three patients (P1 - P3) had been monitored in a reference center after the diagnosis of SCD within the first 4 years of life. All suffered from severe SCD, with a history of acute chest syndrome and more than two vaso-occlusive crises per year requiring hospitalization. P1 had undergone cholecystectomy and tonsillectomy. P2 had papillary necrosis and osteonecrosis of both femoral heads. P3 had undergone cholecystectomy, and had chronic asthma and a history of osteomyelitis events. P1 and P2 were transfused monthly because years of hydroxyurea treatment had proven to be ineffective.

Hydroxyurea treatment was stopped in P3 3 months before mobilization. P3 was then transfused monthly until mobilization. In view of iron overload caused by the transfusions (Table 1), treatment with deferasirox (P1) and deferiprone (P2) was ongoing.

Table 2. Characteristics of apheresis and CD34+ immunoselection.

Apheresis

SCD Pler 1

SCD Pler 2

SCD Pler 3

Total blood volumes in liters (liters processed by the apheresis device) Product volumes (mL) Hematocrit value (%)* CE1** Total CD34+ cells collected by apheresis x106 Total CD34+ cells collected by apheresis x106/kg of body weight

3.27 (15.8)

4 (21)

4 (17.9)

278 4.8 0.24 354 4.6 (77)

382 5.8 0.30 412 5.8 (71)

322 8.2 0.29 292 4.5 (65)

82% 95%

92% 94.7%

31% 79.5%

CD34+ cell product after immunoselection Recovery Purity post-selection

*Normal value: 2-3%; **CE1 = total CD34+ collected/ [L processed X (CD34+pre+CD34+post)/2]

haematologica | 2018; 103(5)

781

C. Lagresle-Peyrou et al.

Absence of Plerixafor-related toxicity Prior to Plerixafor injection, patients underwent several erythrocytoapheresis sessions in order to decrease their HbS levels to below 30% (Table 1). Each patient received a single, subcutaneous injection of Plerixafor (0.24 mg/kg).

The HbS values remained low until day +1 and increased to pre-treatment levels thereafter (Table 1). No adverse effects were observed, other than moderate hypokalemia related to the anticoagulant citrate infusion during the apheresis; this was corrected by a 2-day course of potassium chloride (Table 1). The white blood cell and neutrophil counts rose to 30.4 ± 2.8x109/μL and 17.8 ± 3.5x109/L, respectively, during the first 3 h, remained stable and then (within 24 to 36 h of Plerixafor injection) returned to pretreatment values (Figure 1A and Online Supplementary Figure S1). Bilirubin, conjugated bilirubin and lactate dehydrogenase values and monocyte counts remained stable up to day +30. Serum levels of inflammatory cytokines, including interleukin-8, which is increased in SCD patients during acute crises and associated with higher

A

numbers of circulating hematopoietic progenitors,22 were comparable to those of healthy controls (data not shown).

Vital parameters and blood O2 saturation remained stable and normal during all the procedure. Monitoring of the patients between discharge and the end of the follow-up period (6 months after treatment) was uneventful.

Efficacy of Plerixafor in mobilizing hematopoietic stem and progenitor cells Baseline CD34+ cell counts were 7, 10 and 10/μL in P1, P2 and P3, respectively. The patients exhibited a very fast, intense increase in peripheral blood CD34+ cell count, exceeding 80 CD34+/μL 3 h after the Plerixafor injection (Figure 1B). Levels greater than 50 CD34+/μL were maintained for 6 h, then decreased and returned to normal pretreatment values (Figure 1B, and data not shown for day 30 and day 60). Apheresis was performed with the technical adjustments described in the Online Supplementary Methods. The quantities of CD34+ cells harvested by apheresis (4.6 x 106, 5.8x106 and 4.5x106/kg body weight

B

C

Figure 2. Analysis of the transcriptomic profiles of hematopoietic stem and progenitor cells from different sources (A) Hierarchical clustering analysis of HD BM, SCD BM, SCD Plerixafor-mobilized (Pler), HD Plerixafor-mobilized and HD Filgrastim-mobilized (Filg) HSPC (cluster method: average; distance: correlation). The color of the sample name indicates the classification. (B) Gene ontology analysis of differentially expressed genes. The most enriched biological process categories are shown on the y-axis. The x-axis shows sample comparisons, as defined in Table 3. The orange and green color gradients correspond to the statistical significance of the enrichment [expressed as –log10 (qvalue)] in up- and downregulated genes, respectively. The first color bar at the top indicates comparisons between HSPC from different types of source (dark red) or the same type of source (light red). The second color bar at the top indicates comparisons between HSPC from different types of donor (dark blue) or the same type of donor (light blue). (C) Heat map of genes involved in HSC and progenitor biology. A proportion of the HSC markers were highly expressed in SCD Plerixafor-mobilized HSPC compared with the other samples. The row Z-score is plotted on a red-blue color scale, where red indicates high expression and blue indicates low expression. The color bar at the top indicates the sample classification. HD: healthy donor; BM: bone marrow; HSPC: hematopoietic stem and progenitor cells; HSC: hematopoietic stem cells; SCD: sickle cell disease; Pler: Plerixafor; Filg: Filgrastim.

782

haematologica | 2018; 103(5)

Stem cell mobilization in sickle cell patients

for P1, P2 and P3, respectively) were high in all three patients despite a limited collection efficiency (Table 2).

This enabled the cryopreservation of 3x106 unselected CD34+ cells/kg as a back-up for the upcoming gene therapy trial (used in the case of absence of engraftment) and the immunoselection and further analyses of mobilized CD34+ cells as detailed below. Following CD34+ selection, CD34+ cell purity was in the normal range (Table 2).

CD34+ recovery was high in P1 and P2 (82% and 92%, respectively) and lower in P3 (31%).

Characterization of mobilized hematopoietic stem and progenitor cells To determine the hematopoietic differentiation capacity and self-renewal potential of the mobilized CD34+ cells, we performed a number of phenotypic, transcriptomic and functional analyses.

Hematopoietic stem cells (HSC) and their immediate progeny (multipotent progenitors) within the CD34+ subset are negative for lineage, CD38, and CD45RA markers and positive for CD13323-25 and can, therefore, be detected using flow cytometry. We compared the numbers of HSC and multipotent progenitors among mobilized SCD CD34+ cells with the values for (i) the BM of healthy donors and SCD patients, and (ii) samples from healthy donors mobilized with either Filgrastim or Plerixafor (Online Supplementary Table S1, Figure 1C and Online Supplementary Figure S2). We estimated the number of

HSC per 1000 CD34+ cells to be >25 in Plerixafor-mobilized SCD samples and <5 in all other samples, suggesting that Plerixafor mobilizes HSC with an unexpectedly high efficacy in SCD patients.

The frequency of erythroid and granulocyte/monocyte colony-forming cells was similar in Plerixafor-mobilized and BM SCD samples and in the range of that observed in Filgastrim-mobilized HSPC (Online Supplementary Figure S3A and data not shown). Upon erythroid differentiation, Plerixafor-mobilized as well as BM SCD HSPC gave rise to >90% of mature GYPA+CD36lowCD71low enucleated red blood cells (Online Supplementary Figure S3B-D).

The transcriptome of highly purified CD34+ HSPC from the different sources was analyzed using RNA-Seq. Unsupervised gene expression analysis showed that the samples clustered into two major groups, based on cell origin: the “BM” group encompassed BM samples from SCD patients and healthy donors, whereas the “mobilized” group encompassed Plerixafor-mobilized HSPC from SCD patients and healthy donors and Filgrastimmobilized HSPC from healthy donors (Figure 2A).

Next, we identified differentially expressed genes among the different populations (false discovery rate <0.05, Table 3). In a comparison of samples from SCD patients and healthy donors, we observed that the genes upregulated in SCD samples are involved in inflammatory and immune responses (e.g. defense response to other organisms, type I interferon signaling pathway, cytokine

A

B

Figure 3. Plerixafor-mobilized CD34+ cells from sickle cell disease patients engraft to the same degree as Filgrastim-mobilized CD34+ cells from healthy donors in NSG mice. NSG mice were sacrificed 3 to 4 months after the injection of SCD (SCD Plerixafor, n=3) or HD (HD Filgrastim, n=2) CD34+ cells. (A) Bone marrow cells and (B) splenocytes were isolated, stained and analyzed by flow cytometry. The chimerism (defined as % human CD45+cells/total CD45+cells) and the numbers of human B lymphocytes (CD19+IgM+), granulocytes (CD11b+CD15+), and monocytes (CD11b+CD14+) were evaluated in each group of mice (red circles and red triangles SCD Pler1; blue circles and blue triangles SCD Pler2; green circles and green triangles SCD Pler3; the two HD Filg control are represented by gray squares/gray diamond and black squares/black diamonds, respectively). Each dot represents an individual mouse. HD: healthy donor; SCD: sickle cell disease; Pler: Plerixafor; Filg: Filgrastim.

haematologica | 2018; 103(5)

783

C. Lagresle-Peyrou et al.

production, leukocyte migration, phagocytosis and adaptive immune response) (Figure 2B). Major differences in gene expression (>900 differentially expressed genes) were observed when comparing mobilized and BM samples (Table 3). The genes downregulated in mobilized versus BM HSPC are involved in cell cycle-related processes (e.g. DNA replication, chromosome segregation, and nuclear division) – confirming that mobilized samples contain more quiescent cells, presumably HSC, than progenitors (Figure 2B, Online Supplementary Figure S4A and Online Supplementary Table S2). Interestingly, mobilized populations poorly expressed genes that are typical of committed hematopoietic progenitors, relative to BM samples (Figure 2C, Online Supplementary Figure S4A and Online Supplementary Table S2). Conversely, a large proportion of the genes involved in HSC biology were strongly expressed in mobilized HSPC (Figure 2C, Online Supplementary Figure S4A and Online Supplementary Table S2). Importantly, some HSC markers (e.g. THY1, HLF and FLT3) were more strongly expressed in Plerixafor-mobilized SCD samples – confirming the latter’s high HSC content – than in BM samples and Filgrastim- or Plerixaformobilized HSPC from healthy donors (Figure 2C).

Fewer than 300 genes were differentially expressed between Filgrastim- or Plerixafor-mobilized samples (Table 3). Of note, genes encoding transcriptional regulators and surface markers of plasmacytoid dendritic cell progenitors are upregulated in Plerixafor-mobilized samples compared to Filgrastim-mobilized samples (Online Supplementary Figure S4B). FACS analyses showed the appearance of a CD133-CD34dim cell population preferentially in Plerixafor-mobilized samples (Online Supplementary Figure S2 and data not shown). This population might contain plasmacytoid dendritic cell progenitors that are mobilized by Plerixafor, as recently described.16 Proof of stemness was further confirmed by transplantation into conditioned, immunodeficient mice. Human chimerism in the BM and spleen was similar in recipients of Filgrastim-mobilized CD34+ cells from healthy donorss and Plerixafor-mobilized CD34+ cells from SCD patients (Figure 3A,B). Similar counts of lymphoid and myeloid subsets were detected in both groups (Figure 3A,B). The numbers of human CD34+ cells, HSC and multipotent progenitors in bone marrow were comparable between the two groups (Online Supplementary Figure S5). After secondary transplantation, all recipients of Plerixafor-mobilized SCD samples and Filgrastim-mobilized CD34+ cells from healthy donors displayed engraftment - further demonstrating the presence of true HSC in HSPC mobilized with Plerixafor in SCD patients (data not shown).

Discussion Successful transplantation of autologous gene-corrected cells primarily depends on the collection and effective genetic modification of a sufficient number of true stem cells. Thus, poor harvesting of BM or mobilized peripheral stem cells limits the success of this procedure. To overcome the need for two or more BM aspirates, we initiated a phase I/II clinical trial with the objective of establishing whether Plerixafor-induced stem cell mobilization in SCD patients can avoid the increased risk of vaso-occlusive crises observed with Filgrastim mobilization and of validating the efficiency of HSPC harvesting with this proce784

Table 3. Number of differentially expressed (up- or downregulated) genes in HSPC from different sources (false discovery rate < 0.05).

Comparison SCD BM vs. HD BM SCD Pler vs. HD Pler SCD Pler vs. HD Filg SCD Pler vs. HD BM SCD Pler vs. SCD BM HD Pler vs. HD BM HD Filg vs. HD BM HD Pler vs. HD Filg HD Filg vs. SCD BM HD Pler vs. SCD BM

Differentially expressed 134 165 265 1685 1544 923 1893 47 1683 1024

Upregulated Downregulated 52 133 188 761 753 315 789 20 800 398

82 32 77 924 791 608 1104 27 883 626

SCD: sickle cell disease. HD: healthy donor. BM: bone marrow. Pler: Plerixafor. Filg: Filgrastim.

dure.26 Because of the risks incurred by the patients and the absence of a direct benefit, the trial was restricted to patients with <10x109 granulocytes/L, knowing their role in vaso-occlusive crises.

We decided to discontinue hydroxyurea treatment 3 months before the mobilization in P3 and to submit all the patients to monthly transfusions. The rationale for this decision was based on the following observations: (i) hydroxyurea has no beneficial role in CD34+ cell mobilization in thalassemic patients;27 (ii) hydroxyurea withdrawal is associated with an increase in the number of circulating CD34+ cells in SCD patients;28 and (iii) in various clinical settings hydroxyurea has been associated with myelosuppression29,30 suggesting BM toxicity and potential impairment of HSC. In order to optimize the safety of the mobilization procedure in SCD patients, we reduced HbS levels to below 30% via erythrocyte exchanges, and we closely monitored white blood cell counts and serum levels of inflammatory cytokines. Furthermore, the use of Plerixafor avoided the adverse events associated with Filgrastim: vascular events and splenic rupture after the administration of this latter have been extensively reported in clinical populations and even in healthy stem cell donors.31 Although these events are usually rare, their frequency may be higher in the presence of other vascular risk factors (such as SCD).31 Thus the benefit/risk ratio of using a hematopoietic growth factor such as Filgrastim (especially at the high doses required in patients without a malignant blood disease) appears to be unacceptably low. Hence, we gave our three patients Plerixafor at the standard dose. No adverse events occurred, serum levels of inflammatory cytokines were in the normal ranges (data not shown) and the white blood cell counts did not exceed 40x109/L (i.e. a value often reported in the literature, and far from the 50 to 75x109/L often observed in healthy donors after 5 days of Filgrastim treatment).32 The rapid mobilization with Plerixafor alone (compared with Filgrastim) is also an important advantage. The current guidelines on mobilization in patients with a malignant disease recommend initiating apheresis 11 h after Plerixafor administration;33 this contrasts with the 5 to 7 days required for collection after Filgrastim administration. Moreover, rapid, transient stem cell mobilization by haematologica | 2018; 103(5)

Stem cell mobilization in sickle cell patients

Filgrastim+Plerixafor (especially in very poor mobilizers) has been observed by various groups (including ours); the rapid decrease in peripheral blood stem cells may cause collection to fail when apheresis is initiated according to conventional guidelines.34 We monitored the egress of CD34+ cells into the blood every hour after Plerixafor injection. The time course of CD34+ mobilization was remarkably similar in our three patients. Peak counts of over 80 CD34+/μL were achieved as early as 2 to 3 h after Plerixafor administration; this confirms that HSPC can be harvested immediately in SCD patients. The decrease in CD34+ cell count observed after 6 h might be the consequence of a short-term mobilization of HSC by Plerixafor and therefore to the reduced egress of CD34+ cells from BM combined with their return to the BM. Another possibility is that the drop in CD34+ cells is due to the leukapheresis procedure. Pantin et al. analyzed the kinetics of CD34+ counts in healthy donors treated with Plerixafor and not subjected to the apheresis procedure.35 In their work, CD34+ cell counts peaked at 6 h at lower values than the ones observed in our study and start to decrease 10 h after Plerixafor administration, with differences in kinetics and range of mobilization probably related to the clinical conditions of healthy donors versus SCD patients.

Overall, the study by Pantin et al. suggests that leukapheresis per se does not cause a decrease in CD34+ cell counts. Additionally, the limited collection efficiency (≤30% of the circulating CD34+ cells) (Table 2) does not support the hypothesis that the drop is due to the leukapheresis procedure. Close monitoring of peripheral blood CD34+ cell counts is therefore a crucial point for efficient apheresis in SCD patients mobilized with Plerixafor.

The leukapheresis product contained significantly more HSC than the other stem cell sources used as controls, i.e.

8- to 10-fold more than in BM from healthy donors or SCD patients and in Filgrastim- or Plerixafor-mobilized cells from healthy donors. Accordingly, HSPC from the patients’ Plerixafor-mobilized samples showed elevated transcription of several HSC-associated genes. We do not have a formal explanation for this result; we can only hypothesize that sickling cycles damage the BM stroma and favor the mobilization of HSC. Genes involved in inflammatory and immune responses were upregulated in Plerixafor-mobilized SCD samples.

The inflammation-related characteristics of HSC and their environment constitute a major obstacle to both allogeneic and autologous transplantation. Significant pathological changes in hematopoiesis have been described by Weisser et al.36 in a setting of murine and human chronic granulomatous disease, an inherited disease characterized by chronic, sterile, granulomatous inflammation). In mice and in humans with chronic granulomatous disease, BM and Filgrastim-mobilized grafts contain a low proportion of HSC; in transplanted mice, this feature is associated with low reconstitution potential.36 Hence, we transplanted mobilized CD34+ cells from SCD patients into

References 1. Ware RE, de Montalembert M, Tshilolo L, Abboud MR. Sickle cell disease. Lancet.

2017;390(10091):311-323.

haematologica | 2018; 103(5)

immunodeficient NSG mice, in order to establish whether an inflammatory expression profile interfered with engraftment and self-renewal capacity. In fact, the cells engrafted as well as Filgrastim-mobilized samples in both primary and secondary transplantation, demonstrating that Plerixafor-mobilized SCD CD34+ cells contain true stem cells that are able to reconstitute human hematopoiesis as well as their Filgrastim-mobilized counterparts. Moreover, it is possible to effectively correct Plerixafor-mobilized SCD HSC by gene addition (manuscript in preparation). Taken as a whole, our results show that the inflammatory characteristics of HSC do not impair self-renewal and engraftment.

In conclusion, the present results show that CD34+ cells can be safely mobilized with Plerixafor in SCD patients under well-defined clinical conditions, including a 3month interruption of hydroxyurea treatment, monthly transfusions and red blood cell exchanges. The proportion of true stem cells in the Plerixafor-mobilized CD34+ population was significantly higher than the proportion from any other source, although their egress must be monitored carefully. After harvesting under specific conditions, the cells can be successfully immunoselected. In the case that the optimal dose of CD34+ cells required for gene therapy (6-9x106/kg) is not reached after one apheresis and in the absence of adverse events, a second mobilization by Plerixafor, 24 h after the first one, will be considered. In our small cohort, the high numbers of cells enabled us to extend our gene-addition therapy project without having to perform several low-yield BM harvesting steps.

Furthermore, the results removed the obstacle of collecting an appropriate graft for genome-editing purposes.

Lastly, our study emphasizes the importance of considering the specific characteristics of a diseased BM; finding a way to put the patient’s hematopoietic system into a steady-state condition may circumvent a lack of true stem cells in the harvested product or poor engraftment of genetically-modified cells.

Acknowledgments The authors would like to thank Jean-Marc Luby for excellent, dedicated technical assistance, Valérie Jolaine, Michaela Semeraro for the logistics of the clinical trial, Michaela Semeraro for care of patients, and Christine Bole and Olivier Alibeu for the RNA sequencing. We also thank Frédéric Galacteros, Pablo Bartolucci and Susanne Matthes-Martin for revision of the clinical protocol.

Funding This work was supported by state funding from the Agence Nationale de la Recherche as part of the Investissements d’Avenir program (ANR-10-IAHU-01 and ANR-16-CE18-0004), the French National Institute of Health and Medical Research (INSERM), and Assistance Publique-Hôpitaux de Paris (APHP). This work was also supported by a grant from the European Research Council (ERC 2015-AdG, GENEFORCURE).

2. Walters MC, De Castro LM, Sullivan KM, et al. Indications and results of HLA-identical sibling hematopoietic cell transplantation for sickle cell disease. Biol Blood Marrow Transplant. 2016;22(2):207-211.

3. Gluckman E, Cappelli B, Bernaudin F, et al.

Sickle cell disease: an international survey of results of HLA-identical sibling hematopoietic stem cell transplantation.

Blood. 2017;129(11):1548-1556.

4. Ribeil JA, Hacein-Bey-Abina S, Payen E, et al. Gene therapy in a patient with sickle

785

C. Lagresle-Peyrou et al.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

786

cell disease. N Engl J Med. 2017;376(9): 848-855.

Cavazzana M, Ribeil JA, Lagresle-Peyrou C, Andre-Schmutz I. Gene therapy with hematopoietic stem cells: the diseased bone marrow's point of view. Stem Cells Dev. 2017;26(2):71-76.

Wang TF, Chen SH, Yang SH, Su YC, Chu SC, Li DK. Poor harvest of peripheral blood stem cell in donors with microcytic red blood cells. Transfusion. 2013;53 (1):91-95.

Constantinou VC, Bouinta A, Karponi G, et al. Poor stem cell harvest may not always be related to poor mobilization: lessons gained from a mobilization study in patients with beta-thalassemia major.

Transfusion. 2017;57(4):1031-1039.

Abboud M, Laver J, Blau CA. Granulocytosis causing sickle-cell crisis.

Lancet. 1998;351(9107):959.

Adler BK, Salzman DE, Carabasi MH, Vaughan WP, Reddy VV, Prchal JT. Fatal sickle cell crisis after granulocyte colonystimulating factor administration. Blood.

2001;97(10):3313-3314.

Grigg AP. Granulocyte colony-stimulating factor-induced sickle cell crisis and multiorgan dysfunction in a patient with compound heterozygous sickle cell/beta+ thalassemia. Blood. 2001;97(12):3998-3999.

Onitilo AA, Lazarchick J, Brunson CY, FreiLahr D, Stuart RK. Autologous bone marrow transplant in a patient with sickle cell disease and diffuse large B-cell lymphoma.

Transplant Proc. 2003;35(8):3089-3092.

Kamble RT, Tin UC, Carrum G. Successful mobilization and transplantation of filgrastim mobilized hematopoietic stem cells in sickle cell-hemoglobin C disease. Bone Marrow Transplant. 2006;37(11):10651066.

Rosenbaum C, Peace D, Rich E, Van Besien K. Granulocyte colony-stimulating factorbased stem cell mobilization in patients with sickle cell disease. Biol Blood Marrow Transplant. 2008;14(6):719-723.

Tormey CA, Snyder EL, Cooper DL. Mobilization, collection, and transplantation of peripheral blood hematopoietic progenitor cells in a patient with multiple myeloma and hemoglobin SC disease.

Transfusion. 2008;48(9):1930-1933.

Liles WC, Broxmeyer HE, Rodger E, et al.

Mobilization of hematopoietic progenitor cells in healthy volunteers by AMD3100, a

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

CXCR4 antagonist. Blood. 2003;102(8): 2728-2730.

Schroeder MA, Rettig MP, Lopez S, et al.

Mobilization of allogeneic peripheral blood stem cell donors with intravenous plerixafor mobilizes a unique graft. Blood.

2017;129(19):2680-2692.

Broxmeyer HE, Orschell CM, Clapp DW, et al. Rapid mobilization of murine and human hematopoietic stem and progenitor cells with AMD3100, a CXCR4 antagonist.

J Exp Med. 2005;201(8):1307-1318.

Devine SM, Flomenberg N, Vesole DH, et al. Rapid mobilization of CD34+ cells following administration of the CXCR4 antagonist AMD3100 to patients with multiple myeloma and non-Hodgkin's lymphoma. J Clin Oncol. 2004;22(6):10951102.

Karponi G, Psatha N, Lederer CW, et al.

Plerixafor+G-CSF-mobilized CD34+ cells represent an optimal graft source for thalassemia gene therapy. Blood. 2015;126(5): 616-619.

Ferrari G, Cavazzana M, Mavilio F. Gene therapy approaches to hemoglobinopathies.

Hematol Oncol Clin North Am. 2017;31(5): 835-852.

Lionnet F, Arlet JB, Bartolucci P, et al.

[Guidelines for management of adult sickle cell disease]. Rev Med Interne. 2009;30 (Suppl 3):S162-223.

Lamming CE, Augustin L, Blackstad M, Lund TC, Hebbel RP, Verfaillie CM. Spontaneous circulation of myeloid-lymphoid-initiating cells and SCID-repopulating cells in sickle cell crisis. J Clin Invest.

2003;111(6):811-819.

Doulatov S, Notta F, Eppert K, Nguyen LT, Ohashi PS, Dick JE. Revised map of the human progenitor hierarchy shows the origin of macrophages and dendritic cells in early lymphoid development. Nat Immunol. 2010;11(7):585-593.

Six EM, Bonhomme D, Monteiro M, et al.

A human postnatal lymphoid progenitor capable of circulating and seeding the thymus. J Exp Med. 2007;204(13):3085-3093.

Handgretinger R, Kuci S. CD133-positive hematopoietic stem cells: from biology to medicine. Adv Exp Med Biol. 2013;777:99111.

Fitzhugh CD, Hsieh MM, Bolan CD, Saenz C, Tisdale JF. Granulocyte colony-stimulating factor (G-CSF) administration in individuals with sickle cell disease: time for a

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

moratorium? Cytotherapy. 2009;11(4):464471.

Yannaki E, Papayannopoulou T, Jonlin E, et al. Hematopoietic stem cell mobilization for gene therapy of adult patients with severe beta-thalassemia: results of clinical trials using G-CSF or plerixafor in splenectomized and nonsplenectomized subjects.

Mol Ther. 2012;20(1):230-238.

Richard RE, Siritanaratkul N, Jonlin E, Skarpidi E, Heimfeld S, Blau CA. Collection of blood stem cells from patients with sickle cell anemia. Blood Cells Mol Dis. 2005;35(3):384-388.

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 J Med. 1995;332(20):1317-1322.

Alvarez-Larran A, Perez-Encinas M, FerrerMarin F, et al. Risk of thrombosis according to need of phlebotomies in patients with polycythemia vera treated with hydroxyurea. Haematologica. 2017;102(1):103109.

Halter J, Kodera Y, Ispizua AU, et al. Severe events in donors after allogeneic hematopoietic stem cell donation.

Haematologica. 2009;94(1):94-101.

Lack NA, Green B, Dale DC, et al. A pharmacokinetic-pharmacodynamic model for the mobilization of CD34+ hematopoietic progenitor cells by AMD3100. Clin Pharmacol Ther. 2005;77(5):427-436.

To LB, Levesque JP, Herbert KE. How I treat patients who mobilize hematopoietic stem cells poorly. Blood. 2011;118(17):45304540.

Lefrere F, Mauge L, Rea D, et al. A specific time course for mobilization of peripheral blood CD34+ cells after plerixafor injection in very poor mobilizer patients: impact on the timing of the apheresis procedure.

Transfusion. 2013;53(3):564-569.

Pantin J, Purev E, Tian X, et al. Effect of high-dose plerixafor on CD34(+) cell mobilization in healthy stem cell donors: results of a randomized crossover trial.

Haematologica. 2017;102(3):600-609.

Weisser M, Demel UM, Stein S, et al.

Hyperinflammation in patients with chronic granulomatous disease leads to impairment of hematopoietic stem cell functions.

J Allergy Clin Immunol. 2016;138(1):219228.e219.

haematologica | 2018; 103(5)

ARTICLE

Red Cell Biology & its Disorders

RON kinase inhibition reduces renal endothelial injury in sickle cell disease mice

Ferrata Storti Foundation

Alfia Khaibullina,1 Elena A. Adjei,1,2 Nowah Afangbedji,1 Andrey Ivanov,1 Namita Kumari,1 Luis E.F. Almeida,3 Zenaide M.N. Quezado,3 Sergei Nekhai1,4,5 and Marina Jerebtsova5

Center for Sickle Cell Disease, College of Medicine, Howard University, Washington, DC; Departments of Genetics and Human Genetics, College of Medicine, Howard University, Washington, DC; 3Department of Perioperative Medicine, NIH Clinical Center, National Institutes of Health, Bethesda, MD; 4Department of Medicine, College of Medicine, Howard University, Washington, DC and 5Department of Microbiology, College of Medicine, Howard University, Washington, DC, USA 1 2

Haematologica 2018 Volume 103(5):787-798

ABSTRACT

S

ickle cell disease patients are at increased risk of developing a chronic kidney disease. Endothelial dysfunction and inflammation associated with hemolysis lead to vasculopathy and contribute to the development of renal disease. Here we used a Townes sickle cell disease mouse model to examine renal endothelial injury. Renal disease in Townes mice was associated with glomerular hypertrophy, capillary dilation and congestion, and significant endothelial injury. We also detected substantial renal macrophage infiltration, and accumulation of macrophage stimulating protein 1 in glomerular capillary. Treatment of human cultured macrophages with hemin or red blood cell lysates significantly increased expression of macrophage membrane-associated protease that might cleave and activate circulating macrophage stimulating protein 1 precursor. Macrophage stimulating protein 1 binds to and activates RON kinase, a cell surface receptor tyrosine kinase. In cultured human renal glomerular endothelial cells, macrophage stimulating protein 1 induced RON downstream signaling, resulting in increased phosphorylation of ERK and AKT kinases, expression of Von Willebrand factor, increased cell motility, and re-organization of F-actin. Specificity of macrophage stimulating protein 1 function was confirmed by treatment with RON kinase inhibitor BMS-777607 that significantly reduced downstream signaling. Moreover, treatment of sickle cell mice with BMS-777607 significantly reduced glomerular hypertrophy, capillary dilation and congestion, and endothelial injury. Taken together, our findings demonstrated that RON kinase is involved in the induction of renal endothelial injury in sickle cell mice. Inhibition of RON kinase activation may provide a novel approach for prevention of the development of renal disease in sickle cell disease.

Introduction Sickle cell disease (SCD) is the most commonly inherited hematologic disorder caused by a single nucleotide mutation in the β-globin gene (HBB) resulting in HbS hemoglobin. HbS polymerization leads to sickling and hemolysis of red blood cells (RBCs), vaso-occlusion and organ damage. SCD patients are at increased risk of developing chronic kidney disease (CKD).1,2 Renal involvement in SCD can be present in childhood, as evidenced in 16-28% of children with clinical manifestations of proteinuria and microalbuminuria.3 Albuminuria and proteinuria are observed in more than 50% of adult SCD patients, and renal failure is developed in about 30%.4,5 SCD-associated nephropathy is characterized by tubular dysfunction, which is manifested by inability to concentrate urine, and consequent hyposthenuria and polyuria, and glomerular damage. Glomerular abnormalities are characterized by glomerular hypertrophy, expansion of mesangium, thrombotic Haematologica | 2018; 103(5)

Correspondence: marina.jerebtsova@howard.edu

Received: September 15, 2017.

Accepted: March 1, 2018.

Pre-published: March 8, 2018. doi:10.3324/haematol.2017.180992 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/787 Š2018 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.

787

A. Khaibullina et al.

microangiopathy, focal segmental glomerulosclerosis (FSGS), membranoproliferative glomerulonephritis, and albuminuria.5-8 Two major disease mechanisms of chronic kidney disease in SCD have been proposed: 1) hemolysisendothelial dysfunction leading to vasculopathy; and 2) inflammation and hyper-viscosity leading to vaso-occlusion.9 Intrarenal RBC hemolysis was suggested to be a trigger for both mechanisms.

Spleen, the physiological site of RBC removal from the circulation, is abnormal in SCD patients. The functional asplenia is likely to increase the rates of intravascular hemolysis. Sickling of RBCs and intra-organ hemolysis stimulate infiltration by circulating monocytes and their differentiation into macrophages. Endocytosis of RBC lysate products affects macrophage phenotypes.10,11 Intravascular RBC hemolysis also releases lysate products that impair endothelial function leading to chronic vasculopathy.

Vascular endothelium and monocytes are activated in SCD patients, and monocyte numbers are increased.12-14 Activated macrophages express matriptase-1 (MT-SP1) which is one of the proteases that cleavages and activates circulating macrophage stimulating protein 1 (MSP1).15,16 MSP1 was shown to accumulate in glomeruli in the rat model of anti-Thy1 glomerular disease; its neutralization by antibodies reduced serum creatinine and proteinuria, and protected rats from glomerular injury.17 MSP1 is a plasma protein secreted by liver and circulated as a single-

A

chain, biologically inactive pro-MSP1. It is activated by proteolytic cleavage of Arg483-Val484 bond by either serum proteases or proteases expressed on the cell surface.18 Pro-MSP1 diffuses into local tissues where it is activated by proteolytic cleavage and plays a role in the tissue injury or repair.18 Activated MSP1 binds to and activates a cell surface receptor tyrosine kinase, Recepteur d'Origine Nantais (RON).19 We hypothesize that endocytosis of RBC lysis products by kidney-infiltrating macrophages stimulates expression of MT-SP1, which then locally activates circulating MSP1. We further hypothesize that MSP1 binds to RON tyrosine kinase receptor and activates glomerular endothelium in SCD. We test these hypotheses using a humanized mouse model of SCD (Townes) which recapitulates several hematologic manifestations of human SCD, including renal vascular occlusion, as well as vascular, tubular and glomerular changes.20,21 Here we showed that glomerular disease in SCD mice was associated with endothelial injury, increase in renal macrophage infiltration, and glomerular MSP1 accumulation. In vitro, treatment of cultured human macrophages with hemin, a breakdown product of hemoglobin, or RBC lysate significantly increased expression of MT-SP1. In cultured human renal glomerular endothelial cells, MSP1 treatment induced phosphorylation of RON downstream signaling ERK and AKT kinases, increased expression of von Willebrand factor (vWF) and cell motility, and induced re-organization of

E P=4x10-17

B

C F P=5x10-9

D

788

Figure 1. Renal disease in sickle cell disease (SCD) mice is characterized by significant glomeruli hypertrophy and capillary dilation.

(A-D) Representative pictures of hematoxylin and eosin staining (H&E) (A and B), and periodic acid–Schiff staining (PAS) (C and D) of renal sections. Squares show enlarged areas (B and D). Bar sizes on microphotographs are 100 μm (A and C) and 40 μm. (B and D). (E and F) Quantification of glomeruli size (E) and capillary size per glomeruli cross section (F) is performed using CellSens Standard software. Five mice per group were used for each staining. For quantification graphs, means are shown. Each dot represents a value obtained from one glomerulus cross-section. Ctrl: control.

haematologica | 2018; 103(5)

Renal accumulation of MSP1 in sickle cell disease

F-actin. RON kinase inhibitor (RONi, BMS-777607) significantly reduced RON signaling. Moreover, injections of RONi in young SCD mice prevented the development of glomerular disease, substantially reducing glomerular hypertrophy, capillary dilation and congestion, and endothelial injury.

Taken together, these data demonstrated that renal glomerular accumulation of MSP1 and the activation of RON kinase were involved in the induction of renal endothelial injury in SCD mice. Inhibition of RON kinase activation is a novel approach to prevent CKD development in SCD.

A

B

C

D

E

P=5x10-16

F

P=8x10-9

Figure 2. Renal disease in sickle cell disease (SCD) mice is associated with significant endothelial injury. (A-D) Representative pictures of von Willebrand factor (vWF) (A and B) and CD34 (C and D) immunostaining (red) of renal sections. Squares show enlarged areas (B and D). Non-specific primary antibodies (Abs) were used as a negative control. Bar sizes on the microphotographs are 100 Îźm (A and C) and 40 Îźm. (B and D). (E and F) Quantification of vWF (E) and CD34 (F) expression in glomeruli cross sections is performed using ImageJ Fiji version. Five mice per group were used for each staining. For quantification graphs, means are shown.

Each dot represents a value obtained from one glomerulus cross-section. Ctrl: control.

haematologica | 2018; 103(5)

789

A. Khaibullina et al.

Methods Mice and RONi injections The animal protocol was approved by the Institutional Animal Care and Use Committee at the Children's National Health

System. Townes mice, here referred to as SCD mice, were obtained from the Jackson Laboratory.

Two-month old mice were injected subcutaneously with 10 mg/kg of body weight of RON inhibitor (RONi, BSM-777607, Santa Cruz Scientific) in 2% DMSO daily for 14 consecutive days.

A

B

C

D

E

F -5

P=4x10

P=3x10-9

Figure 3. Glomerular macrophages infiltration and MSP1 accumulation is increased in sickle cell disease (SCD) mice. (A-D) Representative pictures of macrophages (F4/80) (A and B, red) and macrophage stimulating protein 1 (MSP1) (C and D, brown) immunostaining of renal sections. Squares show enlarged areas immunostaining (B and D). Non-specific primary antibodies (Abs) were used as a negative control. Bar sizes on the microphotographs are 100 Âľm (A and C) and 40 Âľm. (B and D). (E and F) Quantification of F4/80 positive macrophages per glomeruli cross section (E) and MSP1 accumulation in the glomeruli (F) is performed using ImageJ Fiji version software. Five mice per group were used for each staining. For quantification graphs, means are shown. Each dot represents a value obtained from one glomerulus cross-section. Ctrl: control.

790

haematologica | 2018; 103(5)

Renal accumulation of MSP1 in sickle cell disease

Six SCD and 4 control mice were injected with RONi. Similar groups were injected with 2% DMSO (vehicle).

Immunohistochemistry Paraffin embedded tissue sections were used for immunostaining with rat anti-mouse F4/80 (AbD Serotec), rabbit anti-vWF (Dako), rat anti-mouse ICAM (BioLegend), rat anti-mouse CD34 (Cedarline), and mouse anti-MSP1 (R&D Systems). AEC and DAB kits were obtained from Vector Laboratory. Images were acquired with an Olympus 1x51 microscope with an Olympus DP 72 camera. Quantification of positive staining was performed using ImageJ Fiji version and CellSens Standard (Olympus) software.

Cell culture Human glomerular endothelial cell line (HGEC) was generated as described22 and maintained in DMEM media supplemented with 10% fetal bovine serum (FBS) and antibiotics (all from Thermo-Fisher). THP1 human monocytic cell line was purchased from ATCC and maintained in RPMI-1640 media supplemented with 10% FBS, antibiotics and 2 μM β-mercaptoethanol (SigmaAldrich). THP-1 cells were differentiated into macrophages by treatment with 10 nM phorbol 12-myristate 13 acetate (PMA, Sigma-Aldrich) for 48 hours (h).

Treatment of THP-1 cells with hemin and RBC lysate Hemin was obtained from Frontier Scientific. Blood samples were collected from healthy control subjects. RBCs were isolated by centrifugation, frozen at -80°C for 15 min, and then thawed for lysis. Cellular debris was pelleted by centrifugation, and the hemolysates were stored at -80°C. Differentiated THP-1 cells were treated either with different concentrations of hemin or hemolysates for 18 h.

Immunofluorescent staining and western blots Rabbit MT-SP1 (Calbiochem) antibody was used for immunostaining and western blot (WB) of THP1 cells. HGEC were treated with 1 μM of human recombinant MSP1 (R&D Systems) with or without RON inhibitor (200 nM) for varying times. WB analysis was performed with rabbit anti-p44/p42 MAPK (Erk1/2), rabbit anti-phospho-p44/p42 Erk1/2, rabbit anti-pan-Akt, and rabbit anti-phospho-Akt antibodies (all from Cell Signaling Technology).

Mouse anti-β-actin antibodies and phalloidin-FITC conjugate were obtained from Sigma-Aldrich.

Wounding migration assay The HGEC monolayers were wounded by strokes across the diameter of the well with 2.5-mm-wide pipet and medium with MSP1 (1 μM) was added. The width of the wound was visualized with pictures taken at pre-treatment and 5 h post treatment. A total of 20 random measurements were analyzed for each wound at each time point.

Isolation of mouse renal glomeruli and glomeruli permeability assay Mouse renal glomeruli were isolated from control mice using a sieving technique.23 Glomerular permeability was measured by a determination of albumin permeability, as described previously24 with slight modification (Online Supplementary Methods).

Statistical analysis Results were expressed as mean±Standard Deviation.

Differences between two groups were compared by unpaired parametric t-test, between multiple groups by one-way ANOVA (GraphPad software). Differences between groups were considered significant at P<0.05. haematologica | 2018; 103(5)

Results SCD mouse kidneys have increased macrophage infiltration and MSP1 accumulation in the glomeruli Kidneys were collected from 4-month old SCD and control mice (n=5 per group) and renal injury was evaluated.

SCD mice develop RBC sickling, anemia, leukocytosis, behavioral changes, and multi-organ pathology characteristic of SCD patients.21,25,26 Glomerular abnormalities in 4month old SCD mice were characterized by: glomerular hypertrophy [Figure 1A, B, and E, hematoxylin and eosin (H&E) staining]; expansion of mesangium [Figure 1C and D, periodic acid Schiff (PAS) staining, and Online Supplementary Figure S1, PCNA staining]; capillary dilation, and thickening of capillary loops and glomerular basement membrane (Figure 1C, D and F, PAS staining). Glomerular capillaries, peritubular cortical capillaries, and capillaries in the medulla and papilla were markedly congested, and sickling of RBCs was observed especially in the medulla and papilla (Online Supplementary Figure S2, H&E staining).

Capillary congestion and dilation together with hemolysis might induce endothelial cell injury and secretion of von Willebrand Factor (vWF) into the surrounding blood and sub-endothelium.27,28 Indeed, glomerular capillary expression of vWF was significantly increased in SCD mice (Figure 2A, B and E, immunostaining, red). Murine glomerular capillary showed positive immunostaining for endothelial marker CD3429 that was significantly increased in SCD mice (Figure 2C, D and F, immunostaining, red).

Because CD34 is also expressed on hematopoietic progenitors cells which are increased in the circulation in SCD mice30 these cells may also be a source of increased CD34 staining in the congested capillary. We also observed a significant renal glomerular and interstitial infiltration of activated macrophages in SCD mice (Figure 3A, B and E, F4/80 immunostaining, arrows, and Online Supplementary Figure S3). Glomerular infiltrated macrophages were negative for the inducible nitric oxide synthase (iNOS), a marker of M1 pro-inflammatory macrophage (Online Supplementary Figure S4). MSP1 was accumulated in SCD renal glomerular capillary (Figure 3C, D and F, immunohistostaining, brown). High levels of MSP1 accumulation were found in 46±8% of glomeruli in SCD mice. In contrast, less than 20% of glomeruli demonstrated low levels of MSP1 accumulation in control mice. Together, these findings show that renal disease in SCD mice is associated with significant endothelial injury, macrophage infiltration, and accumulation of MSP1 in the glomerular capillaries.

Hypoxia, hemin, and red blood cell lysate stimulate expression of matriptase 1 in human macrophages MSP1 is produced by the liver and secreted into circulation where it is activated by proteolytic cleavage.18 Macrophage membrane-bound matriptase 1 (MT-SP1) is one of the proteases which activate MSP1 protein.16,31 We tested whether renal hypoxia and RBC hemolysis increased expression of MT-SP1. Meta-analysis of the NCBI Geo database demonstrated a 2-fold increase in MTSP1 expression in human monocyte-derived differentiated macrophages compared to non-differentiated monocytes (Figure 4A). Meta analysis also showed that MT-SP1 expression was further increased in macrophages that were cultured under hypoxia (Figure 4B). Non-differentiated human pro-monocytic cell line (THP-1 cells) expressed 791

A. Khaibullina et al.

a low level of MT-SP1, which was increased after THP-1 differentiation into macrophages by PMA treatment (Figure 4C and D, compare lanes 1 and 2). Treatment of THP-1-derived macrophages with hemin or RBC lysate further increased MT-SP1 expression (Figure 4C-F).

Immunofluorescent staining of THP-1-derived macrophages demonstrated membrane and cytoplasmic distribution of MT-SP1 (Figure 4G, green) which was significantly increased upon treatment with hemin (10 μM) or RBC lysate (Figure 4G). Collectively, hemolysis products (tested in this study) and hypoxia (meta-analysis data) demonstrated increased expression of MT-SP1 in human macrophages. Higher levels of MT-SP1 expression might induce local activation and accumulation of circulating MSP1.

MSP1 induces endothelial cell motility and activated downstream RON signaling of ERK and AKT To assess the impact of MSP1 on the endothelium, we examined its effect on cultured human glomerular

A

B P=0.008

C

E

G

792

endothelial cell line (HGEC) recently generated in our laboratory.22 Treatment with 1 μM MSP1 did not induce proliferation of HGEC (Figure 5A). In contrast, MSP1 treatment significantly increased motility of HGEC in a 2D wound assay (Figure 5B and C). F-actin plays a central role in the endothelial cell motility and permeability.32 Inflammatory mediators can alter F-actin formation and distribution.32 In confluent HGEC, F-actin formed mostly a cortical rim with few stress fibers (Figure 5D, green).

MSP1 stimulated re-organization of F-actin and formation of stress fibers. (Figure 5D, green). Treatment cells with specific RON kinase inhibitor (200 nM; RONi) inhibited MSP1-induced formation of stress fibers and vWF expression (Figure 5D, F-actin and vWF staining). MSP1 treatment also increased expression of vWF in HGEC (Figure 5D, red), and this increased expression was inhibited by RONi (Figure 5D). Next, we examined activation of RON kinase signaling in HGEC. Binding of MSP1 to RON leads to phosphorylation of the downstream kinases, ERK and AKT.33 Levels of ERK and AKT phosphorylation in HGEC

P=0.002

D

F

Figure 4. Hemin, red blood cell (RBC) count and hypoxia treatment increase expression of matriptase MT-SP1 in human THP1-derived macrophages. (A and B) Meta-analysis of membrane type serine protease 1 (MTSP1) expression in human primary monocyte-derived macrophages using Geo database NCBI (A) Data Set GDS3203 and (B) Data Set GDS2036. For quantification graphs, mean and Standard Deviation (SD) are shown. (C and E) Western blot of MTSP1 in human THP-1-derived macrophages treated with different concentrations of hemin (C) or RBC lysate (D). β-actin was used for normalization.

(D and F) Quantification of MT-SP1 on Western blot. For quantification graphs, mean and SD are shown. *P<0.05. Results are representative of three independent experiments.

(G) Immunostaining of MTSP1 in THP-1-derived macrophages treated with hemin (10 nM) or RBC lysate (green). DAPI was used for nuclear staining. Bar size 40 μm. PMA: phorbol 12-myristate 13-acetate.

haematologica | 2018; 103(5)

Renal accumulation of MSP1 in sickle cell disease

A

C

B

P=0.003

D

E

G

F

H

I

J

Figure 5. MSP1 treatment of cultured human glomerular endothelial cell line (HGEC) induces cell motility and vWF expression, F–actin re-organization, and RON kinase signaling in HGEC. (A) Cell growth measured by MTT assay. Five wells are used for either control or treatment with 1 μM of recombinant MSP1 in each experiment. Results are representative of three independent experiments. (B) Representative picture of wound migration assay of control cells and cells treated with 1 μM of MSP1. Bar size 300 μm. (C) Quantification of wound migration assay. Three wells are used per treatment in each experiment. Results are representative of three independent experiments. (D) Immunostaining of F-actin (green) and vWF (red) in control and treated MSP1 (1 μM) treated cells with or without RONi (200 nM). DAPI (for F-actin) and Hematoxylin (for vWF) were used for nuclear staining. Non-specific primary antibodies were used for negative control. Bar size 40 μm for F-actin and 100 μm for vWF. (E and F) Western blots of phosphorylated and non-phosphorylated forms of ERK and AKT kinases in HGEC. (E) Cells were treated with MSP1 (1 μM) and collected at different time points after treatment. (F) Cells were treated with MSP1 (1 μM) with or without RON inhibitor (RONi, 200 nM) for 30 min. (G-J) Quantification of pERK and pAKT on Western blot. For quantification graphs, mean and SD are shown. *P<0.05. Results are representative of three independent experiments. β-actin used for loading normalization.

haematologica | 2018; 103(5)

793

A. Khaibullina et al.

A

B P=1x10-5 P=0.008

were significantly increased after MSP1 treatment (Figure 5E-G). Phosphorylation of ERK and AKT was significantly reduced after treatment with RONi (Figure 5H-J). Taken together, MSP1 activated RON receptor signaling in cultured HGEC leading to the re-organization of F-actin and increasing cell motility and vWF expression.

MSP1 increases permeability in SCD mouse glomeruli Glomerular hyperfiltration is an early stage manifestation in the development of SCD glomerulopathy.34,35 To test whether MSP1 increases glomeruli permeability, we utilized a mouse whole glomeruli permeability assay.24 Renal glomerular permeability measurement was based on the determination of albumin permeability. In the absence of capillary filtration, the diffusional loss of albumin from glomeruli capillary is negligible.24 Placement of intact glomeruli into the hypooncotic solution increases glomeruli volume due to water accumulation inside glomeruli according to oncotic gradient.24 Treatment of glomeruli with permeating agents that increase glomerular filtration significantly increases albumin loss and reduces oncotic gradient, leading to a reduced glomerular volume in the hypooncotic solution.24 Renal glomeruli were isolated from control mice and treated with recombinant MSP1 (1 μM) in the presence or absence of RONi (200 nM) for 30 min (see Methods for details). PBS was used as a vehicle control. Placement of vehicle-treated glomeruli in hypooncotic solution significantly increased glomerular volume by more than 3.5-fold (Figure 6A and B). MSP1 treatment reduces glomerular volume, whereas RONi treatment restored the ability of glomeruli to enlarge in the hypooncotic solution (Figure 6A and B). Taken together, pre-treatment of mouse glomeruli with MSP1 significantly 794

Figure 6. MSP1 increases glomerular permeability in whole glomeruli assay. (A) Representative pictures of glomeruli isolated from control mice placed in isooncotic and hypooncotic solutions. Glomeruli were pretreated with vehicle (PBS), or with MSP1 (1 μM) with or without RON inhibitor (RONi 200 nM). Bar size 40 μm. (B) Quantification of glomerular volume change in hypooncotic solution. CellSens Standard software was used for measurement of the glomeruli area and glomeruli volume was calculated. Percent of volume changes in isooncotic solution is shown. For quantification, graphs means are shown. Each dot represents a value obtained from one glomerulus.

increased glomerular permeability for albumin, and RONi reduced MSP1-associated glomerular permeability.

Treatment of SCD mice with RONi significantly ameliorates glomerular endothelial injury To test whether RONi prevents development of glomerular endothelial injury in young SCD mice, 2month old SCD mice were injected with either RONi (10 mg/kg of body weight in 2% DMSO) or vehicle (2% DMSO, n=6 per group) subcutaneously for 14 days.

Control non-SCD mice were also injected with either RONi or vehicle (n=4 per group). We used younger (2month old) mice to test whether RONi prevents development of glomerular disease, because older (4-month old) mice had already developed profound glomerular disease (Figures 1 and 2). Only 20% of 2-month old mice developed microalbuminuria (1 of 5 mice), and 40% of 12-week old non-treated SCD mice developed microalbuminuria (2 of 5 mice) (data not shown). In contrast, we found that all non-treated mice had glomerular endothelia cell injury at 12 weeks of age. Because endothelial injury was detected before the onset of albuminuria, we focused on the role of RON signaling in the development of endothelial injury.

Administration of RONi in SCD mice did not affect body weight (Figure 7A), but significantly reduced kidney size (Figure 7B and C). Moreover, administration of RONi in SCD mice significantly reduced glomerular hypertrophy and capillary congestion (Figure 8A and E and Online Supplementary Figure S5, H&E staining). Capillary dilation and thickening of capillary loops and glomerular basement membrane were significantly reduced in SCD mice treated with RONi compared to mice with vehicle injection (Figure 8B and F and Online Supplementary Figure S5, PAS haematologica | 2018; 103(5)

Renal accumulation of MSP1 in sickle cell disease

staining). RONi treatment did not affect mesangial expansion in SCD mice (Figure 8A and B, H&E and Online Supplementary Figure S6, PAS staining). However, RONi administration in SCD mice significantly reduced capillary expression of vWF (Figure 8C and G and Online Supplementary Figure S5, immunostaining, red), and Intercellular Adhesion Molecule 1 (ICAM1) (Figure 8D and E and Online Supplementary Figure S5, immunostaining, red), which are the markers of endothelial injury.36,37 Thus, inhibiton of RON receptor in SCD mice significantly ameliorated development of endothelial and glomerular injury.

A

B

Discussion Renal disease in SCD patients includes a variety of glomerular and tubular complications, but the mechanism of their development is not fully understood. The most important finding in our study is that MSP1 and its receptor RON kinase may play a role in the activation of renal endothelium and development of glomerular pathology in SCD mice. Moreover, treatment of mice with RONi, an inhibitor of RON kinase, significantly ameliorated glomerular hypertrophy, capillary dilation and congestion, and endothelial injury. These findings reveal a previously unknown mechanism which contributes to glomerular endothelial injury in SCD. Sickle cell disease mice spontaneously develop FSGS that may be directly associated with RBC sickling and chronic hemolysis. The focal nature of glomerulosclerosis in SCD mice and SCD patients apparently excludes the effect of global factors, such as hypoxia, cell-free heme, iron, and other circulating factors that would lead to a global and not focal glomerulosclerosis with involvement of only 50% of glomeruli in SCD mice. Thus, locally-produced factor(s) are more likely to contribute to the development of FSGS. In agreement with a previous report,21 we demonstrate here that macrophage infiltration in renal glomeruli was increased in SCD mice. Sickling and adhesion of RBCs, and accumulation of RBC lysate products within the kidney might stimulate renal infiltration of monocyte-derived macrophages. Monocyte-derived renal macrophages are present in all forms of kidney disease with inflammation, and renal capillary macrophage infiltration is a characteristic pathology of FSGS.38 In many human biopsy studies, number of glomerular or interstitial macrophages correlate with poor outcomes, suggesting their possible role in the disease progression.39,40 However, the role of infiltrating macrophages in the progression of renal disease is not well understood. Phagocytosis of senescent sickled RBCs, RBC exosomes and endocytosis of cell-free hemoglobin increases inflammatory response in the cultured human monocytes/macrophages.12 Activation of proteases is a universal inflammatory response in macrophages.41 We demonstrate here that products of RBC hemolysis significantly increased expression of macrophage membranebound protease, MT-SP1 in cultured human macrophages.

Meta-analysis data also showed that hypoxia and monocyte differentiation increased MT-SP1 expression in primary human macrophages. MT-SP1 is a type II membrane serine protease that plays important roles in cell migration and tumor cell metastasis.42 MT-SP1 is one of the proteases that activate circulating haematologica | 2018; 103(5)

C

P=0.04

Figure 7. Treatment of sickle cell disease (SCD) mice with RON inhibitor reduces kidney hypertrophy. (A) Treatment of SCD mice with RON inhibitor (RONi) does not affect body weight (BW). (B) Representative picture of kidneys of controls and SCD mice treated either with (RONi) or vehicle (2% DMSO). (C) Quantification of kidney weight. Kidney weights (KW) to BW ratios are shown.

MSP1.15 MSP1 expression has been found in the renal tubular cells.43,44 In agreement with previous studies, we did not observe MSP1 expression or accumulation in glomeruli of control mice. In contrast, we found that MSP1 was accumulated in approximately 46Âą8% of renal glomerular capillaries in SCD mice. This accumulation rate was similar to the percentage of injured glomeruli in SCD mice. Glomerular accumulation of MSP1 was previously shown in the rat model of anti-Thy1 glomerular disease, and the neutralization of MSP1 by the injected antibodies reduced serum creatinine and proteinuria, and protected animals from glomerular injury.17 However, the mechanism of MSP1-associated glomerular injury was not clarified. RON is expressed in human renal tubular cells and glomerular mesangial cells.43,44 MSP1 treatment induces growth, motility and collagen invasion of mesangial cells.43 Expression of functional RON in endothelial cells is unknown. MSP1 that is accumulated in glomerular capillary of SCD mice may potentially affect endothelial cells, or leak from capillary to affect mesangial cells or podocytes. The increased glomeruli size associated with dilated glomerular capillary and mesangial proliferation was reported both in SCD patients and mouse model of SCD.20,21,45 In our study, inhibition of RON in SCD mice significantly reduced glomerular hypertrophy, as well as capillary dilation and congestion without reduction of mesangial expansion. It is possible that the short time of treatment was not enough to produce statistically signifi795

A. Khaibullina et al.

cant differences in mesangial cell proliferation, or that factors other than MSP1 could stimulate mesangial growth in SCD mice. Thus, we hypothesized that MSP1 played a role in the activation of endothelial cells. The effect of MSP1 on renal glomerular endothelial cells is still not known.

To test the effect of MSP1 on endothelium, we used cultured HGEC. MSP1 had previously been shown to stimulate motility of murine resident peritoneal macrophages

and kidney epithelial cells.18,46 We demonstrated that MSP1 treatment increased motility of HGECs and induced formation of F-actin stress fibers that is essential for motility and permeability of endothelial cells.32 Further studies are needed to determine whether MSP1/RON signaling induces permeability of cultured endothelial cells.

Interestingly, MSP1 increased vWF expression levels in HGEC. High levels of vWF were also found in SCD murine glomeruli. RONi treatment reduced endothelial

A

B

C

D

E

F P=3x10-8

G P=2x10-7

H P=0.002

P=0.04

Figure 8. Treatment of sickle cell disease (SCD) mice with RON inhibitor (RONi) reduces renal endothelial injury. Representative pictures of renal sections of control and SCD mice treated with either RONi or vehicle (DMSO) are shown. (A) Hematoxylin&Eosin (H&E) staining. (B) Periodic Acid-Schiff (PAS) staining. (C) von Willebrand factor (vWF) immunostaining (red). (D) Intercellular Adhesion Molecule (ICAM) immunostaining (red). Bar sizes 50Âľm. (E and F) Quantification of glomeruli size (E) and capillary size per glomeruli cross section (F) is performed using CellSens Standard software. (G and H) Quantification of vWF (G) and ICAM (H) expression in glomeruli cross sections is performed using ImageJ Fiji version software. Six mice per group were used for each staining. Each dot represents a value obtained from one glomerulus cross-section. For quantification graphs, means are shown.

796

haematologica | 2018; 103(5)

Renal accumulation of MSP1 in sickle cell disease

vWF levels in vivo and in vitro. vWF mediates adhesion of sickle RBCs to endothelial cells, inducing oxidant stress, and increasing expression of ICAM-1, VCAM and Eselectin.47 Indeed, RONi treatment in SCD mice was associated with reduced vWF levels in glomerular capillaries, and decreased RBC adhesion and ICAM-1 levels.

Therefore, the mechanism of RON inhibition may be associated with prevention of glomerular capillary congestion, leading to improvement of renal hemodynamics. A correlation between an alteration of renal hemodynamics and a renal histological injury has been demonstrated in a model of renal ischemia in SCD.48 In addition, our results demonstrate that MSP1 treatment of glomeruli isolated from control mice significantly increased albumin permeability that was effectively prevented by RON inhibition.

Hyperfiltration is associated with glomerular hypertrophy and glomerulosclerosis.49 We demonstrate a significant reduction in glomerular size in SCD mice after treatment with RON inhibitors, suggesting a possible change in hyperfiltration. Reduction of glomerular size after 14-day treatment with RONi is unlikely to be due to structural remodeling or reduced proliferation of endothelial cells.

To the best of our knowledge, this is the first study demonstrating function of MSP1/RON in the glomerular endothelial cells. Future studies will clarify the mechanism of MSP1-associated endothelial cell activation and its role

References 1. Hamideh D, Alvarez O. Sickle cell disease related mortality in the United States (1999-2009). Pediatr Blood Cancer.

2013;60(9):1482-1486.

2. 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.

3. Becton LJ, Kalpatthi RV, Rackoff E, et al.

Prevalence and clinical correlates of microalbuminuria in children with sickle cell disease. Pediatr Nephrol. 2010;25(8): 1505-1511.

4. Saraf SL, Zhang X, Kanias T, et al.

Haemoglobinuria is associated with chronic kidney disease and its progression in patients with sickle cell anaemia. Br J Haematol. 2014;164(5):729-739.

5. Guasch A, Navarrete J, Nass K, Zayas CF. Glomerular involvement in adults with sickle cell hemoglobinopathies: Prevalence and clinical correlates of progressive renal failure. J Am Soc Nephrol. 2006;17(8):22282235.

6. Sharpe CC, Thein SL. Sickle cell nephropathy - a practical approach. Br J Haematol.

2011;155(3):287-297.

7. Scheinman JI. Sickle cell disease and the kidney. Nat Clin Pract Nephrol. 2009;5(2): 78-88.

8. Tejani A, Phadke K, Adamson O, et al.

Renal lesions in sickle cell nephropathy in children. Nephron. 1985;39(4):352-355.

9. Potoka KP, Gladwin MT. Vasculopathy and pulmonary hypertension in sickle cell disease. Am J Physiol Lung Cell Mol Physiol.

2015;308(4):L314-324.

10. Hand WL, King-Thompson NL. Effect of erythrocyte ingestion on macrophage anti-

haematologica | 2018; 103(5)

11.

12.

13.

14.

15.

16.

17.

18.

19.

in the development of glomerular injury.

The present study established for the first time that inhibition of RON kinase significantly ameliorates SCD renal pathology in mice. Short-term inhibition of RON kinase reduced endothelial injury in young SCD mice during the early stage of renal disease before the onset of albuminuria. We do not know how long this effect persists after RONi withdrawal. Whether short-term inhibition of RON kinase will ameliorate already developed renal disease in older mice is currently under investigation.

Future studies will elucidate a role of Ron kinase in renal disease in SCD patients. These findings also highlight a new potential therapeutic target for CKD in SCD. A recent pre-clinical study50 and Phase 1 clinical trial (clinicaltrials.gov identifier: 01721148) of BMS-777607 RON inhibitor for treatment of human cancers demonstrated its safety and good tolerability, providing a proof of principal for the potential use of this pharmacological approach.

Funding This work was supported in part by CHaRM pilot grant awarded to MJ through NIH grant 1P50HL118006 (to SN) and NIH grants 1R01HL125005 (to SN), 5G12MD007597 (to SN), and R41MD008829-01 (to ZMQ). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

bacterial function. Infect Immun. 1983; 40(3):917-923.

Vinchi F, Costa da Silva M, Ingoglia G, et al. Hemopexin therapy reverts hemeinduced proinflammatory phenotypic switching of macrophages in a mouse model of sickle cell disease. Blood.

2016;127(4):473-486.

Belcher JD, Marker PH, Weber JP, Hebbel RP, Vercellotti GM. Activated monocytes in sickle cell disease: potential role in the activation of vascular endothelium and vaso-occlusion. Blood. 2000;96(7):24512459.

Klug PP, Kaye N, Jensen WN. Endothelial cell and vascular damage in the sickle cell disorders. Blood Cells. 1982;8(1):175-184.

Mahoney DH Jr, Fernbach DJ. Monocyte functions in sickle cell disorders. Am J Pediatr Hematol Oncol. 1983;5(4):409-411.

Bhatt AS, Welm A, Farady CJ, et al.

Coordinate expression and functional profiling identify an extracellular proteolytic signaling pathway. Proc Natl Acad Sci USA. 2007;104(14):5771-5776.

Qiu D, Owen K, Gray K, Bass R, Ellis V. Roles and regulation of membrane-associated serine proteases. Biochem Soc Trans.

2007;35(Pt 3):583-587.

Rampino T, Soccio G, Gregorini M, et al.

Neutralization of macrophage-stimulating protein ameliorates renal injury in anti-thy 1 glomerulonephritis. J Am Soc Nephrol.

2007;18(5):1486-1496.

Leonard EJ. Biological aspects of macrophage-stimulating protein (MSP) and its receptor. Ciba Found Symp. 1997;212: 183-191.

Wang MH, Julian FM, Breathnach R, et al.

Macrophage stimulating protein (MSP) binds to its receptor via the MSP beta

20.

21.

22.

23.

24.

25.

26.

27.

28.

chain. J Biol Chem. 1997;272(27):1699917004.

Ryan TM, Ciavatta DJ, Townes TM. Knockout-transgenic mouse model of sickle cell disease. Science. 1997;278(5339): 873-876.

Kasztan M, Fox BM, Speed JS, et al. LongTerm Endothelin-A Receptor Antagonism Provides Robust Renal Protection in Humanized Sickle Cell Disease Mice. J Am Soc Nephrol. 2017;28(8):2443-2458.

Jerebtsova M, Kumari N, Obuhkov Y, Nekhai S. Adenoviral E4 gene stimulates secretion of pigmental epithelium derived factor (PEDF) that maintains long-term survival of human glomerulus-derived endothelial cells. Mol Cell Proteomics.

2012;11(11):1378-1388.

Misra RP. Isolation of glomeruli from mammalian kidneys by graded sieving. Am J Clin Pathol. 1972;58(2):135-139.

Savin VJ, Sharma R, Lovell HB, Welling DJ. Measurement of albumin reflection coefficient with isolated rat glomeruli. J Am Soc Nephrol. 1992;3(6):1260-1269.

Wang L, Almeida LE, de Souza Batista CM, et al. Cognitive and behavior deficits in sickle cell mice are associated with profound neuropathologic changes in hippocampus and cerebellum. Neurobiol Dis. 2015;85:60-72.

Khaibullina A, Almeida LE, Wang L, et al.

Rapamycin increases fetal hemoglobin and ameliorates the nociception phenotype in sickle cell mice. Blood Cells Mol Dis. 2015;55(4):363-372.

Sporn LA, Marder VJ, Wagner DD. Inducible secretion of large, biologically potent von Willebrand factor multimers.

Cell. 1986;46(2):185-190.

Wagner DD, Bonfanti R. von Willebrand

797

A. Khaibullina et al.

29.

30.

31.

32.

33.

34.

35.

36.

798

factor and the endothelium. Mayo Clin Proc. 1991;66(6):621-627.

Kairaitis LK, Wang Y, Gassmann M, Tay YC, Harris DC. HIF-1alpha expression follows microvascular loss in advanced murine adriamycin nephrosis. Am J Physiol Renal Physiol. 2005;288(1):F198-206.

Blouin MJ, De Paepe ME, Trudel M. Altered hematopoiesis in murine sickle cell disease. Blood. 1999;94(4):1451-1459.

Kawaguchi M, Orikawa H, Baba T, Fukushima T, Kataoka H. Hepatocyte growth factor activator is a serum activator of single-chain precursor macrophage-stimulating protein. FEBS J. 2009;276(13):34813490.

Prasain N, Stevens T. The actin cytoskeleton in endothelial cell phenotypes.

Microvasc Res. 2009;77(1):53-63.

Wagh PK, Peace BE, Waltz SE. Met-related receptor tyrosine kinase Ron in tumor growth and metastasis. Adv Cancer Res. 2008;100:1-33.

Haymann JP, Stankovic K, Levy P, et al.

Glomerular hyperfiltration in adult sickle cell anemia: a frequent hemolysis associated feature. Clin J Am Soc Nephrol.

2010;5(5):756-761.

Aygun B, Mortier NA, Smeltzer MP, Hankins JS, Ware RE. Glomerular hyperfiltration and albuminuria in children with sickle cell anemia. Pediatr Nephrol.

2011;26(8):1285-1290.

Pober JS, Cotran RS. Cytokines and endothelial cell biology. Physiol Rev.

1990;70(2):427-451.

37. Wertheimer SJ, Myers CL, Wallace RW, Parks TP. Intercellular adhesion molecule-1 gene expression in human endothelial cells.

Differential regulation by tumor necrosis factor-alpha and phorbol myristate acetate.

J Biol Chem. 1992;267(17):12030-12035.

38. Cao Q, Harris DC, Wang Y. Macrophages in kidney injury, inflammation, and fibrosis. Physiology. 2015;30(3):183-194.

39. Hill GS, Delahousse M, Nochy D, Mandet C, Bariety J. Proteinuria and tubulointerstitial lesions in lupus nephritis. Kidney Int. 2001;60(5):1893-1903.

40. Rollino C, Basolo B, Roccatello D, Menegatti E, Piccoli G. Atypical presence of antimyeloperoxidase antibodies in 2 transplanted patients. Nephron. 1993;63(4):480.

41. Tomlinson ML, Garcia-Morales C, AbuElmagd M, Wheeler GN. Three matrix metalloproteinases are required in vivo for macrophage migration during embryonic development. Mech Dev. 2008;125(1112):1059-1070.

42. Darragh MR, Bhatt AS, Craik CS. MT-SP1 proteolysis and regulation of cell-microenvironment interactions. Front Biosci.

2008;13:528-539.

43. Rampino T, Collesi C, Gregorini M, et al.

Macrophage-stimulating protein is produced by tubular cells and activates mesangial cells. J Am Soc Nephrol. 2002;13(3): 649-657.

44. Cantaluppi V, Biancone L, Romanazzi GM, et al. Macrophage stimulating protein may

45.

46.

47.

48.

49.

50.

promote tubular regeneration after acute injury. J Am Soc Nephrol. 2008;19(10): 1904-1918.

Nath KA, Hebbel RP. Sickle cell disease: renal manifestations and mechanisms. Nat Rev Nephrol. 2015;11(3):161-171.

Zhang K, Yao HP, Wang MH. Activation of RON differentially regulates claudin expression and localization: role of claudin1 in RON-mediated epithelial cell motility.

Carcinogenesis. 2008;29(3):552-559.

Sultana C, Shen Y, Rattan V, Johnson C, Kalra VK. Interaction of sickle erythrocytes with endothelial cells in the presence of endothelial cell conditioned medium induces oxidant stress leading to transendothelial migration of monocytes.

Blood. 1998;92(10):3924-3935.

Juncos JP, Grande JP, Croatt AJ, et al. Early and prominent alterations in hemodynamics, signaling, and gene expression following renal ischemia in sickle cell disease. Am J Physiol Renal Physiol. 2010;298(4):F892899.

Brenner BM, Lawler EV, Mackenzie HS. The hyperfiltration theory: a paradigm shift in nephrology. Kidney Int. 1996;49(6):1774-1777.

Bieniasz M, Radhakrishnan P, Faham N, De La O JP, Welm AL. Preclinical Efficacy of Ron Kinase Inhibitors Alone and in Combination with PI3K Inhibitors for Treatment of sfRon-Expressing Breast Cancer Patient-Derived Xenografts. Clin Cancer Res. 2015;21(24):5588-5600.

haematologica | 2018; 103(5)

ARTICLE

Myeloproliferative Disorders

The KIT and PDGFRA switch-control inhibitor DCC-2618 blocks growth and survival of multiple neoplastic cell types in advanced mastocytosis

Mathias Schneeweiss,1,2 Barbara Peter,1,2 Siham Bibi,3 Gregor Eisenwort,1,2 Dubravka Smiljkovic,1 Katharina Blatt,1,2 Mohamad Jawhar,4 Daniela Berger,2 Gabriele Stefanzl,2 Susanne Herndlhofer,1,2 Georg Greiner,5 Gregor Hoermann,5 Emir Hadzijusufovic,1,2,6 Karoline V. Gleixner,1,2 Peter Bettelheim,7 Klaus Geissler,8 Wolfgang R. Sperr,1,2 Andreas Reiter,4 Michel Arock3 and Peter Valent1,2

Ludwig Boltzmann Cluster Oncology, Medical University of Vienna, Austria; 2Department of Internal Medicine I, Division of Hematology and Hemostaseology, Medical University of Vienna, Austria; 3Laboratoire de Biologie et Pharmacologie Appliquee, CNRS UMR 8113, Ecole Normale Superieure de Cachan, France; 4Department of Hematology and Oncology, University Medical Centre Mannheim and Medical Faculty Mannheim, Heidelberg University, Germany; 5Department of Laboratory Medicine, Medical University of Vienna, Austria; 6Department for Companion Animals and Horses, University Clinic for Small Animals, Internal Medicine Small Animals, University of Veterinary Medicine Vienna, Austria; 7Elisabethinen Hospital Linz, Austria and 8Fifth Medical Department, Hospital Hietzing, Vienna, Austria 1

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):799-809

ABSTRACT

S

ystemic mastocytosis is a complex disease defined by abnormal growth and accumulation of neoplastic mast cells in various organs.

Most patients exhibit a D816V-mutated variant of KIT, which confers resistance against imatinib. Clinical problems in systemic mastocytosis arise from mediator-related symptoms and/or organ destruction caused by malignant expansion of neoplastic mast cells and/or other myeloid cells in various organ systems. DCC-2618 is a spectrum-selective pan KIT and PDGFRA inhibitor which blocks KIT D816V and multiple other kinase targets relevant to systemic mastocytosis. We found that DCC-2618 inhibits the proliferation and survival of various human mast cell lines (HMC-1, ROSA, MCPV-1) as well as primary neoplastic mast cells obtained from patients with advanced systemic mastocytosis (IC50 <1 ÎźM). Moreover, DCC-2618 decreased growth and survival of primary neoplastic eosinophils obtained from patients with systemic mastocytosis or eosinophilic leukemia, leukemic monocytes obtained from patients with chronic myelomonocytic leukemia with or without concomitant systemic mastocytosis, and blast cells obtained from patients with acute myeloid leukemia. Furthermore, DCC-2618 was found to suppress the proliferation of endothelial cells, suggesting additional drug effects on systemic mastocytosis-related angiogenesis. Finally, DCC-2618 was found to downregulate IgE-mediated histamine release from basophils and tryptase release from mast cells. Together, DCC-2618 inhibits growth, survival and activation of multiple cell types relevant to advanced systemic mastocytosis. Whether DCC-2618 is effective in vivo in patients with advanced systemic mastocytosis is currently under investigation in clinical trials.

Introduction Systemic mastocytosis (SM) is a hematopoietic neoplasm with complex biology and pathology, and a variable clinical course.1-7 The disease is characterized by abnormal expansion and accumulation of neoplastic mast cells (MC) in one or more Haematologica | 2018; 103(5)

Correspondence: peter.valent@meduniwien.ac.at

Received: September 1, 2017.

Accepted: January 31, 2018.

Pre-published: February 8, 2018.

doi:10.3324/haematol.2017.179895 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/799 Š2018 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.

799

M. Schneeweiss et al.

internal organs, including the bone marrow.1-3 Various types of SM have been recognized by the World Health Organization (WHO).8-11 The indolent variant of SM is associated with ‘hematologic stability’ and thus with an almost normal life expectancy.12-14 By contrast, the prognosis in patients with advanced SM, including SM with an associated hematologic neoplasm (AHN), aggressive SM (ASM) and MC leukemia (MCL) is unfavorable, with short survival times and poor responses to conventional therapy.1-5,12,13,15 Current research is, therefore, focusing on therapeutic targets and the effects of novel antineoplastic drugs on various cell types relevant to advanced SM.16 Since most patients with SM also suffer from mediatorrelated symptoms that may sometimes be severe or even life-threatening, such drugs are often selected based on their dual effects on MC growth and MC activation.

Most patients with SM express the D816V-mutated variant of the stem cell factor receptor, KIT, which mediates ligand-independent activation and autonomous growth and differentiation of MC.17-22 The D816V KIT point mutation also confers resistance against several tyrosine kinase inhibitors, including imatinib.23-26 Novel kinase blockers acting on KIT D816V have, therefore, been developed. The highlighting example is midostaurin (PKC412).27,28 However, despite superior clinical efficacy seen in a global phase II trial,28 patients with advanced SM often exhibit or acquire resistance.28,29 A number of different mechanisms may underlie resistance against midostaurin. One obvious problem is that the drug does not suppress all clinically relevant sub-clones and cell-types, especially cells lacking KIT D816V.28,29 Such sub-clones are often seen in the context of advanced SM. Over 50% of these patients have or develop an AHN.30-32 Of these patients with an AHN, approximately 80-90% have an associated myeloid neoplasm, the most frequent ones being chronic myelomonocytic leukemia (CMML) and acute myeloid leukemia (AML).8-11,30-32 In these patients, leukemic expansion of monocytes and/or blast cells is typically found. In other patients, an expansion of eosinophils, sometimes resembling chronic eosinophilic leukemia (SM-CEL), is found. In most of these patients,

eosinophils display KIT D816V.33 By contrast, expression of rearranged PDGFR variants is rarely seen in SM, although in some patients with a FIP1L1/PDGFRA fusion gene, the MC expansion has a histopathological picture indistinguishable from that of SM.34 Treatment of SM-AHN represents a clinical challenge because the AHN-component is often resistant.16,32 DCC-2618 is a switch-control type II inhibitor of KIT, which arrests KIT in an inactive state, regardless of activating mutations, such as KIT D816V.35 Moreover, several additional oncogenic kinases, including FLT-3, PDGFRA, PDGFRB, KDR, TIE2 and FMS are recognized by DCC2618.35 Recently, the first clinical trials with DCC-2618 (NCT02571036) were started in patients with kinase-driven malignancies. In addition, first preclinical studies have shown that DCC-2618 may exert antineoplastic effects on neoplastic MC.36 In our current study, we show that DCC-2618 is a potent inhibitor of growth and survival of neoplastic human MC expressing various KIT mutations. Furthermore, we show that DCC-2618 produces growth inhibition and apoptosis in other cell types that play a role in advanced SM. Finally, we show that DCC-2618 inhibits IgE-dependent histamine secretion from basophils and tryptase secretion from MC. All in all, our data suggest that DCC-2618 is a promising, novel drug for the treatment of advanced SM.

Methods Reagents The reagents used in this study are described in the Online Supplement. DCC-2618 and its active metabolite, DP-5439, were kindly provided by Dr. B. Smith (Deciphera Pharmaceuticals LLC, Lawrence, KS, USA).

Isolation of primary neoplastic cells Primary neoplastic cells were isolated from bone marrow samples of 11 patients with SM. The bone marrow cells were obtained during routine diagnostic investigations after written

Table 1. Characteristics of patients with systemic mastocytosis and response of neoplastic cells to DCC-2618 and DP-5439.

Patient n.

Age

m/f

Diagnosis, SM variant

#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11

68 56 49 66 82 57 90 63 65 78 91

m m f m f f m m m m m

ISM ISM ISM ISM-MPN-eo SSM ASM ASM ASM-AML ASM-CMML ASM-MPN-eo MCL

KIT Serum tryptase BM MC D816V ng/mL infiltration % + + + + + + + + + +

83.3 103 22.9 170 284 87.9 125 33.9 220 45.9 330

n.a.

20 10 5 50 50 20 15 50 15 20

% MC in MNC

DCC-2618 IC50

DP-5439 IC50

PKC412 IC50

1 5 n.a. n.a. n.a. n.a.

25 8 30 5 50

114 nM 240 nM 198 nM 394 nM 347 nM 331 nM 386 nM 393 nM 460 nM 83 nM 321 nM

414 nM 390 nM 554 nM 1481 nM 1584 nM 366 nM 679 nM 554 nM 907 nM 64 nM 592 nM

56 nM 35 nM n.a. n.a. n.a. n.a.

360 nM 66 nM n.a.

114 nM 65 nM

Diagnoses were established according to WHO criteria. Primary bone marrow cells were incubated with various concentrations of DCC-2618, DP-5439 or midostaurin (PKC412) at 37°C for 48 h. Proliferation was then determined by measuring uptake of 3H-thymidine and IC50 values were calculated. WHO: World Health Organization; m: male; f: female; SM: systemic mastocytosis; PB, peripheral blood; BM, bone marrow; MC: mast cells; MNC, mononuclear cells, nM, nanomolar; n.a., not available; IC50, half maximal inhibitory concentration; ISM: indolent SM; MPN: myeloproliferative neoplasms; SSM: smoldering SM; ASM: aggressive SM; CMML: chronic monomyelocytic leukemia; MCL: mast cell leukemia.

800

haematologica | 2018; 103(5)

DCC-2618: a new drug against mastocytosis

informed consent had been given. Patients were classified as having indolent SM (ISM; n=3), smoldering SM (SSM; n=1), ASM (n=2), SM-AHN (n=4) and MCL (n=1) according to WHO criteria.8-11 In addition, neoplastic cells were obtained from ten patients suffering from CMML, ten with AML and three with hypereosinophilia. The patients’ characteristics are shown in Table 1 (SM patients) and Online Supplementary Table S1 (other hematologic disorders). Heparinized bone marrow cells were layered over Ficoll to isolate mononuclear cells. The study was approved by the ethics committee of the Medical University of Vienna.

Culture of human cell lines The following human MCL-like cell lines were employed in this study: HMC-1.1 and HMC-1.2,37 three ROSA sub-clones (ROSAKIT WT, ROSAKIT D816V, ROSAKIT K509I)38 and four MCPV-1 subclones (MCPV-1.1, MCPV-1.2, MCPV-1.3, MCPV-1.4).39 In addition, we examined several AML cell lines, the CEL-related cell line EOL-1, the microvascular endothelial cell line HMEC-1, and cultured human umbilical vein endothelial cells (HUVEC). A description of cell lines is provided in the Online Supplement.

Evaluation of growth, survival of neoplastic cells Drug-exposed cells (cell lines or primary cells) were analyzed for proliferation and survival. The bioassays employed are described in the Online Supplementary Methods.

Results DCC-2618 and its metabolite DP-5439 inhibit proliferation of neoplastic mast cells DCC-2618 and its active metabolite, DP-5439 were found to inhibit 3H-thymidine uptake and thus proliferation in a dose-dependent manner in all MC lines tested, with slightly lower IC50 values obtained in HMC-1.1 cells lacking KIT D816V and ROSAKIT WT cells compared to the KIT D816V-positive cell lines HMC-1.2 and ROSAKIT D816V (Figure 1A and Table 2). IC50 values obtained in HMC-1.1 cells with DCC-2618 were also lower than IC50 values obtained with midostaurin.25,26 In addition, DCC-2618 and DP-5439 were found to inhibit proliferation of ROSAKIT K509I cells with lower IC50 values (DCC-2618, IC50: 34±10 nM) compared to ROSAKIT D816V cells (Figure 1A). Unexpectedly, DCC-2618 and its metabolite also produced growth-inhibition in the multi-resistant MC lines MCPV-1.1, MCPV1.2, MCPV-1.3 and MCPV-1.4 (Figure 1B and Table 2).

Finally, we were able to show that DCC-2618 and DP5439 induced dose-dependent inhibition of growth of primary neoplastic bone marrow cells obtained from patients suffering from various forms of SM, including ASM and MCL (Figure 1C, Table 1). Interestingly, the effects of DCC-2618 on primary neoplastic BM cells and the related IC50 values obtained in different SM variants were comparable (Figure 1C, Table 1).

Western blotting For evaluation of KIT and BTK signaling, HMC-1.1, HMC-1.2, ROSAKIT WT and ROSAKIT D816V cells were incubated in control medium or in DCC-2618 (0.5-5 μM) for 4 h at 37°C. Western blotting was performed essentially as described elsewhere.26,40 For evaluation of downstream signaling pathways of KIT, HMC1.1, HMC-1.2, ROSAKIT WT and ROSAKIT D816V cells were first preincubated overnight in Iscove modified Dulbecco medium devoid of fetal calf serum and of stem cell factor. Cells (106) from each line were then treated with DCC-2618 (0.001-10 μM) for 90 min at 37°C. At the end of the treatment, ROSAKIT WT cells were stimulated with stem cell factor-containing supernatants (10%) of Chinese hamster ovary cells transfected with the murine scf (kl) gene (CHO-KL) at room temperature for 10 min.

Thereafter, Western blotting was performed essentially as described previously.26,40 Antibodies against phosphorylated (p)Kit, STAT5, pSTAT5, ERK1/2, pERK1/2 were purchased from Cell Signaling (Danvers, MA, USA), antibodies against pBTK were bought from NovusBiologicals (Littleton, CO, USA) and antibodies against total KIT and total BTK were from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

DCC-2618 inhibits KIT, STAT5, AKT, and ERK activation in neoplastic mast cells As expected, DCC-2618 was found to suppress phosphorylation of KIT in ROSAKIT WT and ROSAKIT D816V cells as well as in both HMC-1 sub-clones (Figure 2A). In addition, DCC-2618 was found to decrease the expression of phosphosphorylated (p)STAT5, pAKT and pERK1/2 in all cell lines tested (Figure 2B). As expression levels of pSTAT5 in

Table 2. Effects of DCC-2618 and DP-5439 on growth of various human cell types.

Cell line / cell type

DCC-2618, IC50

DP-5439, IC50

12.3±3.7 nM 123±36 nM 41±5 nM 168±65 nM 34±10 nM 298±77 nM 1000±932 nM 724±511 nM 670±418 nM 132±95 nM 147±88 nM 2.7±3.7 μM 7.8±5.4 μM 1.8±1.3 nM 3.7±2.2 μM 707±224 nM

13±4 nM 96±23 nM 56±21 nM 188±60 nM 28±13 nM 357±179 nM 913±378 nM 756±435 nM 697±410 nM 73±31 nM 76±24 nM 5.2±4.4 μM 2.5±0.3 μM 1±0.8 nM 2±1.3 μM 605±222 nM

Statistical analysis

HMC-1.1 HMC-1.2 ROSA KIT WT ROSA KIT D816V ROSA KIT K509I MCPV-1.1 MCPV-1.2 MCPV-1.3 MCPV-1.4 MOLM-13 MV4-11 KG-1 U937 EOL-1 HMEC-1 HUVEC

To determine the level of significance the Student t-test was applied. Results were considered to be statistically significantly different when P was <0.05.

Cell lines were incubated with various concentrations of DCC-2618 or DP-5439 at 37°C for 48 h.Then, proliferation was determined by measuring uptake of 3H-thymidine and IC50 values were calculated. Values represent the mean±S.D. from three independent experiments. IC50: half maximal inhibitory concentration.

Measurement of mediator release Drug-exposed cells (blood basophils obtained from healthy individuals and HMC-1 cells) were analyzed for histamine- and tryptase release as described in the Online Supplementary Methods.

Evaluation of apoptosis in basophils Drug-exposed blood basophils obtained from healthy donors by dextran sedimentation were analyzed for cell survival by flow cytometry. Technical details are described in the Online Supplementary Methods.

haematologica | 2018; 103(5)

801

M. Schneeweiss et al.

HMC-1.1 cells were rather low and difficult to quantify by Western blotting, we also performed intracellular flow cytometry-staining experiments using an antibody against pSTAT5. In these experiments, DCC-2618 was found to counteract pSTAT5 expression in HMC-1.1 and HMC-1.2 cells in a dose-dependent manner (Online Supplementary Figure S1). DCC-2618 did not inhibit phosphorylation of BTK, another important target of tyrosine kinase inhibitors expressed by neoplastic MC (Figure 2C).

and Online Supplementary Figure S2A,B). The metabolite DP-5439 was found to be equally effective in producing apoptosis in MC lines compared to DCC-2618 (Figure 3A,B and Online Supplementary Figure S2A,B). Together, these data show that DCC-2618 is a novel potent antineoplastic compound inducing apoptosis and growth arrest in neoplastic MC.

DCC-2618 induces apoptosis in neoplastic mast cells

DCC-2618 produces synergistic growth-inhibitory effects with midostaurin and cladribine (2CdA) in neoplastic mast cells

To explore the mechanism of drug action, we analyzed the effects of DCC-2618 on the survival of neoplastic MC. As assessed by light microscopy, DCC-2618 induced apoptosis in HMC-1.1, HMC-1.2, ROSAKIT WT, ROSAKIT D816V and ROSA KIT K509I cells in a dose-dependent manner (Figure 3A). The effects of DCC-2618 on survival were more pronounced in KIT D816V-negative MC lines than in KIT D816V-positive MC lines (Figure 3A). DCC-2618 was also found to produce apoptosis in the multi-resistant MCPV-1 cell lines (Figure 3B). The apoptosis-inducing effect of DCC-2618 on MC was confirmed by combined annexin V/propidium-iodide staining (Figure 3A,B

In advanced SM, drug combinations may be required to suppress malignant cell growth. We found that DCC-2618 and midostaurin produce clear cooperative (synergistic) growth-inhibitory effects in HMC-1.1 cells (Online Supplementary Figure S3A,C). In HMC-1.2 cells, the drug combination also produced cooperative antineoplastic effects, but these effects were additive rather than synergistic as defined by Calcusyn software (Online Supplementary Figure S3A,C). In addition, we found that DCC-2618 and 2CdA induce clear synergistic growthinhibitory effects on HMC-1.1 and HMC-1.2 cells (Online Supplementary Figure S3B,D).

A

B

C

Figures 1. DCC-2618 and its active metabolite DP-5439 inhibit proliferation of neoplastic mast cells. HMC-1, ROSA (A), MCPV-1 (B) and primary neoplastic mast cells (C) obtained from patients with different variants of systemic mastocytosis (ISM, SSM, ASM, MCL) were incubated in control medium (0 nM) or medium containing various concentrations of DCC-2618 or DP-5439, as indicated, at 37°C for 48 h. Thereafter, 3H-thymidine uptake was measured. Results in (A) and (B) are expressed as percent of control and represent the mean±S.D. from three independent experiments. Results in (C) are expressed as percent of control and represent mean±S.D from triplicates. Asterisk (*): P<0.05 compared to control medium.

802

haematologica | 2018; 103(5)

DCC-2618: a new drug against mastocytosis

DCC-2618 inhibits IgE-dependent histamine release from basophils and spontaneous tryptase release from neoplastic mast cells Since patients with SM often suffer from symptoms caused by mediators released from neoplastic MC and/or basophils, we evaluated the effect of DCC-2618 on antiIgE-induced histamine release. We found that DCC-2618 (0.1-1.0 μM) slightly inhibited anti-IgE mediated histamine release from normal human blood basophils (Figure 4A).

This drug effect was found to be specific in that DCC-2618 did not inhibit C5a- or calcium ionophore-induced histamine release from basophils (Online Supplementary Figure S4A). As expected, DCC-2618 did not affect the viability of basophils between 0.1 and 1.0 μM and did not induce histamine secretion within 30 min of incubation (Online Supplementary Figure S4B). In consecutive experiments, we also found that DCC-2618 suppresses the spontaneous (baseline) secretion of tryptase from HMC-1.1 and HMC1.2 cells during the entire incubation period (days 1 through 6) (Figure 4B).

DCC-2618 counteracts growth and survival of leukemic monocytes and blast cells We next explored the effects of DCC-2618 on AHN celltypes. In a first step, we examined AML cell responses.

DCC-2618 was found to inhibit the proliferation of all AML cell lines tested, with considerably lower IC50 values obtained with the FLT3-mutated cell lines MOLM-13 (132±95 nM) and MV4-11 (147±88 nM) compared to KG1 and U937 cells (Table 2, Figure 5A). Similar effects were seen with DP-5439 (Figure 5A). DCC-2618 was also found to induce apoptosis in MOLM-13, MV4-11 and KG-1 cells (Figure 5B and Online Supplementary Figure S5). Finally, we found that DCC-2618 and DP-5439 produced dose-dependent inhibition of growth in primary leukemia cells obtained from patients with AML or CMML (Online Supplementary Table S1 and Figure 5C). In one patient with ASM-CMML, we isolated mononuclear cells and found that DCC-2618 and DP-5439 induced growth inhibition of these cells in the same way as in mononuclear cells obtained from patients with CMML without SM (Table 1

A

B

C

Figures 2. DCC-2618 inhibits phosphorylation of KIT and other targets in neoplastic mast cells. (A,C) HMC-1 and ROSA cells were incubated in control medium (ROSAKIT : Iscove modified Dulbecco medium (IMDM) with stem cell factor, SCF; ROSAKIT D816V: IMDM without SCF) or medium containing various concentrations of DCC-2618, as indicated, at 37°C for 4 h. Thereafter, cells were harvested and Western blotting was performed as described in the text using antibodies against phosphorylated (p)KIT, total KIT, pBTK and total BTK. (B) HMC-1 and ROSA cells were first pre-incubated overnight in IMDM devoid of fetal calf serum and of SCF. Cells were then treated with DCC-2618 (0.001-10 μM) for 90 min at 37°C. At the end of the treatment, ROSAKIT WT cells were stimulated with SCF (10% CHO-KL) at room temperature for 10 min.

Thereafter, cells were harvested and Western blotting was performed as described in the text using antibodies against pSTAT5, total STAT5, pAKT, total AKT, pERK1/2, total ERK1/2. Western blot experiments were performed at least twice. Western blots in this figure show one representative experiment.

WT

haematologica | 2018; 103(5)

803

M. Schneeweiss et al.

and Figure 5C). Together, these data suggest that DCC2618 counteracts growth of AHN cells, including CMML monocytes and AML blasts.

DCC-2618 inhibits the proliferation of neoplastic eosinophils Advanced SM is often accompanied by eosinophilia. In addition PDGFRA is a known target of DCC-2618. We analyzed the effects of DCC-2618 on proliferation and

survival of the FIP1L1-PDGFRA (F/P) positive EOL-1 cell line. DCC-2618 was found to inhibit proliferation in EOL-1 cells at low nanomolar range of concentrations (IC50: 1.8±1.3 nM) (Online Supplementary Figure S6A).

Similar effects were seen with DP-5439 (Online Supplementary Figure S6A). DCC-2618 also induced apoptosis in EOL-1 cells (Online Supplementary Figure S6C).

Next, we examined the effects of DCC-2618 on growth of primary eosinophils. In these experiments, DCC-2618

A

B

Figures 3. DCC-2618 and DP-5439 induce apoptosis in neoplastic mast cells. HMC-1, ROSA (A) and MCPV-1 (B) were incubated in control medium (0 μM) or medium containing various concentrations of DCC-2618 and DP-5439, as indicated, at 37°C for 48 h. Cells were then harvested and the percentage of apoptotic cells was quantified morphologically on Wright-Giemsa-stained cytospin preparations (left panels) or by flow cytometry (determination of annexinV/PI-positive cells, right panels). Results represent the mean±S.D. of three independent experiments. Asterisk (*): P<0.05 compared to control medium.

804

haematologica | 2018; 103(5)

DCC-2618: a new drug against mastocytosis

was found to inhibit the proliferation of neoplastic bone marrow cells obtained from a patient with ASM (Table 1 and Online Supplementary Figure S6B). In addition, DCC2618 was found to block the growth of bone marrow cells obtained from patients with secondary hypereosinophilic syndromes (Online Supplementary Table S1 and Online Supplementary Figure S6B).

DCC-2618 inhibits growth of human endothelial cells Increased bone marrow angiogenesis has been implicated in the pathogenesis of SM.41 To investigate potential effects of DCC-2618 on angiogenesis, we explored drug effects on growth of HUVEC and the microvascular endothelial cell line HMEC-1. As assessed by 3H-thymidine uptake, DCC-2618 and its metabolite were found to inhibit the proliferation of HUVEC and HMEC-1 cells in a dose-dependent manner (Online Supplementary Figure S7). DCC-2618 exerted stronger effects on HUVEC (707±224 nM) than on HMEC-1 cells (3.7±2.2 µM).

Discussion Due to the poor response to conventional drugs, treatment of patients with advanced SM is still a major challenge in clinical practice. Despite the availability of new

drugs the prognosis of these patients remains poor with short survival times.10,16,28,29 Research is, therefore, seeking new effective drugs and novel treatment concepts. DCC2618 is a novel switch-control type II blocker that exerts inhibitory effects on KIT D816V, other KIT mutants, and several other critical target kinases, such as FLT3, PDGFRA and KDR.35 We here describe that DCC-2618 inhibits the proliferation of nine different human MC lines, with lower IC50 values obtained in HMC-1.1 cells and ROSAKIT WT cells than in KIT D816V-positive HMC-1.2 and ROSAKIT D816V cells. In addition, DCC-2618 was found to block the proliferation of primary neoplastic MC obtained from patients with ASM or MCL. Moreover, DCC-2618 exerted major antineoplastic effects on AHN cells and endothelial cells, all of which may be relevant in the pathogenesis of advanced SM. Based on these observations DCC-2618 is a novel emerging drug candidate for advanced SM. Indeed, clinical trials with DCC-2618 have been started recently.

The multi-kinase inhibitor midostaurin (PKC412) is effective against the D816V-mutated variant of KIT and has shown promising results in patients with advanced SM in a global phase II trial, with an overall response rate of 60%.28 In addition, midostaurin was found to suppress mediator-related symptoms and IgE-dependent histamine release from basophils.28,42 However, despite clinical effi-

A

Figures 4. Effects of DCC-2618 on anti-IgE-induced histamine release from normal basophils. (A) Primary blood basophils from healthy donors were incubated in control medium (0 μM) or in various concentrations of DCC-2618, as indicated, at 37°C for 30 min.

Thereafter, cells were incubated in control buffer or in buffer containing anti-IgE antibody E-124.2.8 (1 μg/mL) at 37°C for 30 min. After incubation, cells were centrifuged at 4°C, and cell-free supernatants and cell suspensions recovered and examined for histaminecontent by radioimmunoassay. Histamine release was calculated as percent of total histamine and is expressed as percent of control. Results represent the mean±S.D. of four independent experiments. Asterisk (*): P<0.05 compared to control medium. (B) HMC-1 cells were cultured in the presence or absence of DCC-2618 (HMC-1.1: 25 nM; HMC-1.2: 500 nM) over a 6-day period. Spontaneous release of tryptase from HMC-1.1 and HMC-1.2 cells was measured by determining tryptase concentrations in cell-free supernatants and lysates. Tryptase release is expressed as percent of total (intra- and extra-cellular) tryptase.

Results represent the mean±S.D. of three independent experiments. Asterisk (*): P<0.05.

B

haematologica | 2018; 103(5)

805

M. Schneeweiss et al.

A

B

C

Figures 5. Effects of DCC-2618 and DP-5439 on proliferation and survival of acute myeloid leukemia (AML) and chronic myelomonocytic leukemia (CMML). (A,C) MOLM-13, MV4-11, KG-1, U937 and primary leukemic cells were incubated in control medium (0 μM) or medium containing various concentrations of DCC-2618 or DP-5439, as indicated at 37°C for 48 h. Thereafter, 3H-thymidine uptake was determined. Results in (A) are expressed as percent of control and represent the mean±S.D. from three independent experiments. Asterisk (*): P<0.05 compared to control medium. Results of (C) are expressed as percent of control and represent the mean±S.D. from triplicates. (B) MOLM-13, MV4-11, KG-1 and U937 cells were incubated with control medium (0 nM) or various concentrations of DCC-2618 and DP-5439, as indicated, for 48 h. Thereafter cells were harvested and the percentage of apoptotic cells was quantified morphologically on Wright-Giemsa-stained cytospin preparations (left panels) or by flow cytometry (determination of annexinV/PI-positive cells, right panels). Results represent the mean±S.D. of three independent experiments. Asterisk (*): P<0.05 compared to control medium.

806

haematologica | 2018; 103(5)

DCC-2618: a new drug against mastocytosis

cacy, midostaurin is unable to produce long-lasting complete remission in all patients.28 Therefore, new drugs and drug-combinations are currently being tested in the context of advanced SM. DCC-2618 might be a promising candidate for several reasons. First, DCC-2618 exhibits a broad target profile and is able to block growth of various neoplastic cells.36 In the current study, DCC-2618 was found to block growth of neoplastic cells obtained from patients with ASM and MCL. In addition, the drug produced growth inhibition in all MCL-like cell lines tested, including KIT-mutated cells and cell lines in which other oncogenic pathways (such as the RAS pathway) trigger malignant cell growth. Moreover, unlike other KIT-targeting drugs, DCC-2618 is able to suppress the growth and survival of other cell types relevant to advanced SM and AHN, including monocytes, blast cells, neoplastic eosinophils and endothelial cells. The concentrations required to mediate these cellular inhibitory effects are readily achievable based on the recent report of clinical exposure of 5 ÎźM or higher in patients with gastrointestinal stroma tumors.43 After intake, DCC-2618 is considered to be converted to one active metabolite, DP-5439. We therefore investigated whether DP-5439 is also able to counteract growth and survival of neoplastic cells. In these experiments, we were able to show that DP-5439 is able to suppress growth and survival of neoplastic MC and of other leukemic (non-MC-lineage) cells in the same way (and with comparable IC50 values) as DCC-2618. These data suggest that DCC-2618 treatment should be effective even if the DP-5439 metabolite may accumulate over time.

It is well known that about one-third of all patients with advanced SM have an AHN at diagnosis. Of these patients, most have a myeloid neoplasm, often in the form of CMML or AML.2-8,13,32 The treatment of these SMAHN patients is a clinical challenge because the AHN is often drug-resistant. In fact patients with SM-AHN still have a poor prognosis with an overall survival time of about 24 months.13,32 Because of its broad activity profile, we asked whether DCC-2618 might be a promising agent for patients with SM-AHN. In a first step, we found that DCC-2618 is a potent inhibitor of proliferation and survival of the FLT3-mutated AML cell lines MOLM-13 and MV4-11. DCC-2618 also inhibits the growth of other AML cell lines examined (KG-1 and U937), but at IC50 values considerably higher than those for MOLM-13 or MV4-11 cells. We also found that DCC-2618 counteracts proliferation of primary leukemic cells obtained from patients with SM-AHN, AML or CMML (Table 1 and Online Supplementary Table S1). These findings suggest that DCC-2618 may be a promising agent for SM-AHN. In SM patients, disease progression is often accompanied by expansion of neoplastic eosinophils, sometimes even resembling (chronic) eosinophilic leukemia. In most cases the eosinophils are of clonal origin as they express KIT D816V.33 In rare cases, neoplastic eosinophils display the FIP1L1/PDGFRA fusion gene.34 However, this fusion gene is usually detectable only in eosinophilic neoplasms, such as CEL. Since DCC-2618 is known to exert inhibitory effects against PDGFRA35 we examined its effects on EOL-1 cells harboring FIP1L1-PDGFRA. DCC-2618 was found to exert strong anti-proliferative and apoptosisinducing effects in EOL-1 cells, with IC50 values in the low nanomolar range. In addition, DCC-2618 was found to haematologica | 2018; 103(5)

inhibit growth of primary eosinophils obtained from patients with KIT D816V-positive SM or reactive hypereosinophilia. Together, these data suggest that DCC-2618 inhibits multiple AHN-related cell types, which may be relevant clinically as progression of SM is often accompanied by multilineage expansion of various sub-clones, including cells harboring or lacking KIT D816V.15,28,29 A number of different pro-oncogenic pathways and targets may be involved in KIT D816V-dependent expansion and accumulation of MC in advanced SM.25,40,44-50 Several of these target pathways may be sensitive to therapy with tyrosine kinase inhibitors. We studied whether key target pathways in neoplastic MC can be disrupted by DCC-2618. As assessed by Western blotting, DCC-2618 was found to block the phosphorylation and thus activation of wild-type KIT and KIT D816V. In addition, we were able to show that DCC-2618 blocks the activation of AKT, ERK and STAT5, suggesting that multiple target pathways are accessible to this drug. By contrast, however, the drug did not disrupt activation of BTK, another important target displayed by neoplastic MC.46 Since the target spectrum of midostaurin (PKC412) and DCC-2618 is not identical, we were also interested to learn whether DCC-2618 and midostaurin can produce synergistic antineoplastic effects on neoplastic MC. Indeed, we found that both drugs induce cooperative or even synergistic growth-inhibitory effects on HMC-1.1 and HMC-1.2 cells.

Specific alterations in the microenvironment, including increased angiogenesis, are frequently detectable in advanced bone marrow neoplasms and are often considered to play an important role in disease progression. A typical finding in the affected bone marrow of patients with advanced SM is increased microvessel density.41 We found that DCC-2618 inhibits the proliferation of human endothelial cells, including HUVEC and a microvascular endothelial cell line, HMEC-1. These data suggest that DCC-2618 also acts as an anti-angiogenic agent.

Interestingly, the IC50 values obtained for HMEC-1 cells were higher than those for HUVEC, which may be explained by the fact that HMEC-1 is a cell line, whereas HUVEC are primary cells. An alternative explanation would be the lack of key targets in HMEC-1 cells.

Indeed, it is well known, that KDR, a key target of DCC2618, is only expressed in HUVEC but not in HMEC-1 cells.

Patients with SM frequently suffer from symptoms produced by MC-derived mediators.4-6,16 These mediators are released on IgE-dependent activation of MC and may cause severe problems or even lead to life-threatening anaphylaxis. Concomitant (IgE-dependent) allergies are, therefore, relevant comorbidities in the context of SM. We found that DCC-2618 counteracts IgE-dependent secretion of histamine from basophils obtained from healthy donors. In addition, we were able to show that DCC-2618 blocks IgE-independent, spontaneous release of tryptase from HMC-1.1 and HMC-1.2 cells in vitro.

These results suggest that, apart from its antineoplastic effects, DCC-2618 might also have an impact on mediator release (and probably on the resulting symptoms) in patients with SM with concomitant allergies. Whether these data can be reproduced in vivo and whether the drug is able to suppress mediator symptoms in patients with advanced SM or SM with concomitant allergies remains to be determined. In fact, whereas the concentrations of 807

M. Schneeweiss et al.

DCC-2618 required to block tryptase secretion in MC were rather low (<1.0 μM), the concentrations required to block IgE-dependent histamine release were rather high (1 μM).

Collectively, our data indicate that DCC-2618 is a novel promising agent that counteracts growth and survival of various cell types relevant to the pathogenesis of advanced SM. Whether DCC-2618 is able to block

References 1. Metcalfe DD. Classification and diagnosis of mastocytosis: current status. J Invest Dermatol. 1991;96(3):2S-4S. 2. Valent P. Biology, classification and treatment of human mastocytosis. Wien Klin Wochenschr. 1996;108(13):385-397.

3. Horny HP, Valent P. Diagnosis of mastocytosis: general histopathological aspects, morphological criteria, and immunohistochemical findings. Leuk Res. 2001;25(7): 543-551.

4. Escribano L, Akin C, Castells M, et al.

Mastocytosis: current concepts in diagnosis and treatment. Ann Hematol. 2002;81(12): 677-690.

5. Valent P, Akin C, Sperr WR, et al. Diagnosis and treatment of systemic mastocytosis: state of the art. Br J Haematol. 2003;122(5): 695-717.

6. Akin C, Metcalfe DD. Systemic mastocytosis. Annu Rev Med. 2004;55:419-432.

7. Arock M, Valent P. Pathogenesis, classification and treatment of mastocytosis: state of the art in 2010 and future perspectives.

Expert Rev Hematol. 2010;3(4):497-516.

8. Valent P, Horny H-P, Escribano L, et al.

Diagnostic criteria and classification of mastocytosis: a consensus proposal. Leuk Res. 2001;25(27):603-625.

9. Valent P, Horny H-P, Li CY, et al.

Mastocytosis (mast cell disease). World Health Organization (WHO) Classification of Tumours. Pathology & Genetics.

Tumours of Haematopoietic and Lymphoid Tissues. Eds: Jaffe ES, Harris NL, Stein H, Vardiman JW 2001;1:291-302.

10. Valent P, Akin C, Metcalfe DD. Mastocytosis: 2016 updated WHO classification and novel emerging treatment concepts. Blood. 2017;129(11):1420-1427.

11. Horny HP, Akin C, Arber D, et al.

Mastocytosis. In: WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Eds: Swerdlow SH, Campo E, Harris NL et al. IARC Press Lyon, France, vol 3, pp 62-69, 2017(2):62-69.

12. Lim KH, Tefferi A, Lasho TL, et al. Systemic mastocytosis in 342 consecutive adults: survival studies and prognostic factors. Blood.

2009;113(23):5727-5736.

13. Pardanani A, Lim KH, Lasho TL, et al.

Prognostically relevant breakdown of 123 patients with systemic mastocytosis associated with other myeloid malignancies.

Blood. 2009;114(18):3769-3772.

14. Escribano L, Alvarez-Twose I, SánchezMuñoz L, et al. Prognosis in adult indolent systemic mastocytosis: a long-term study of the Spanish Network on Mastocytosis in a series of 145 patients. J Allergy Clin Immunol. 2009;124(3):514-521.

15. Valent P, Akin C, Sperr WR, et al. Aggressive

808

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

growth of neoplastic MC in patients with advanced SM is currently being explored in a clinical trial (NCT02571036).

Acknowledgments This study was supported by a grant from Deciphera Pharmaceuticals LLC. We like to thank Dr. Dan Flynn, Dr. Bryan Smith and Dr. Oliver Rosen for helpful discussion.

systemic mastocytosis and related mast cell disorders: current treatment options and proposed response criteria. Leuk Res. 2003;27(7):635-641.

Arock M, Akin C, Hermine O, et al. Current treatment options in patients with mastocytosis: status in 2015 and future perspectives.

Eur J Haematol. 2015;94(6):474-490.

Furitsu T, Tsujimura T, Tono T, et al.

Identification of mutations in the coding sequence of the proto-oncogene c-kit in a human mast cell leukemia cell line causing ligand-independent activation of the c-kit product. J Clin Invest. 1993;92(4):17361744.

Nagata H, Worobec AS, Oh CK, et al.

Identification of a point mutation in the catalytic domain of the protooncogene c-kit in peripheral blood mononuclear cells of patients who have mastocytosis with an associated hematologic disorder. Proc Natl Acad Sci. 1995;92(23):10560-10564.

Feger F, Ribadeau Dumas A, et al. Kit and ckit mutations in mastocytosis: a short overview with special reference to novel molecular and diagnostic concepts. Int Arch Allergy Immunol. 2002;127(2):110-114.

Longley BJ, Tyrrell L, Lu SZ, et al. Somatic ckit activating mutation in urticaria pigmentosa and aggressive mastocytosis: establishment of clonality in a human mast cell neoplasm. Nat Genet. 1996;12(3):312-314.

Fritsche-Polanz R, Jordan JH, Feix A, et al.

Mutation analysis of C-KIT in patients with myelodysplastic syndromes without mastocytosis and cases of systemic mastocytosis. Br J Haematol. 2001;113(2):357-364.

Arock M, Sotlar K, Akin C, et al. KIT mutation analysis in mast cell neoplasms: recommendations of the European Competence Network on Mastocytosis. Leukemia.

2015;29(6):1223-1232.

Shah NP, Lee FY, Luo R, et al. Dasatinib (BMS-354825) inhibits KITD816V, an imatinib-resistant activating mutation that triggers neoplastic growth in most patients with systemic mastocytosis. Blood.

2006;108(1):286-291.

Akin C, Brockow K, D'Ambrosio C, et al.

Effects of tyrosine kinase inhibitor STI571 on human mast cells bearing wild-type or mutated c-kit.

Exp Hematol.

2003;31(8):686-692.

Gleixner KV, Mayerhofer M, Sonneck K, et al. Synergistic growth-inhibitory effects of two tyrosine kinase inhibitors, dasatinib and PKC412, on neoplastic mast cells expressing the D816V-mutated oncogenic variant of KIT. Haematologica.

2007;92(11):1451-1459.

Gleixner KV, Mayerhofer M, Aichberger KJ, et al. PKC412 inhibits in vitro growth of neoplastic human mast cells expressing the D816V-mutated variant of KIT: comparison with AMN107, imatinib, and cladribine

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

(2CdA) and evaluation of cooperative drug effects. Blood. 2006;107(2):752-759.

Fabbro D, Ruetz S, Bodis S, et al. PKC412 — a protein kinase inhibitor with a broad therapeutic potential. Anticancer Drug Des. 2000;15(1):17-28.

Gotlib J, Kluin-Nelemans HC, George TI, et al. Efficacy and safety of midostaurin in advanced systemic mastocytosis. N Engl J Med. 2016;374(26):2530-2541.

Gotlib J, Berubé C, Growney JD, et al.

Activity of the tyrosine kinase inhibitor PKC412 in a patient with mast cell leukemia with the D816V KIT mutation. Blood.

2005;106(8):2865-2870.

Sperr WR, Horny HP, Valent P. Spectrum of associated clonal hematologic non-mast cell lineage disorders occurring in patients with systemic mastocytosis. Int Arch Allergy Immunol. 2002;127(2):140-142.

Horny HP, Sotlar K, Sperr WR, et al.

Systemic mastocytosis with associated clonal haematological non-mast cell lineage diseases: a histopathological challenge. J Clin Pathol. 2004;57(6):604-608.

Pardanani A. Systemic mastocytosis in adults: 2017 update on diagnosis, risk stratification and management. Am J Hematol.

2016;91(11):1146-1159.

Pardanani A, Reeder T, Li CY, et al.

Eosinophils are derived from the neoplastic clone in patients with systemic mastocytosis and eosinophilia. Leuk Res. 2003;27(10): 883-885.

Pardanani A, Ketterling RP, Brockman SR, et al. CHIC2 deletion, a surrogate for FIP1L1PDGFRA fusion, occurs in systemic mastocytosis associated with eosinophilia and predicts response to imatinib mesylate therapy. Blood. 2003;102(9):3093-3096.

Smith BD, Hood MM, Wise SC, et al. DCC2618 is a potent inhibitor of wild-type and mutant KIT, including refractory exon 17 D816 KIT mutations, and exhibits efficacy in refractory GIST and AML xenograft models. Cancer Res. 2015;75(15): abstract 2690.

Bai Y, Bandara G, Ching Chan E, et al, Targeting the KIT activating switch control pocket: a novel mechanism to inhibit neoplastic mast cell proliferation and mast cell activation. Leukemia. 2013; 27(2):278-285.

Butterfield JH, Weiler D, Dewald G, Gleich GJ. Establishment of an immature mast cell line from a patient with mast cell leukemia.

Leuk Res. 1988;12(4):345-355.

Saleh R, Wedeh G, Herrmann H, et al., A new human mast cell line expressing a functional IgE receptor converts to factor-independence and tumorigenicity by KIT D816V-transfection. Blood. 2014;124(1): 111-120.

Hoermann G, Blatt K, Greiner G, et al., CD52 is a molecular target in advanced systemic mastocytosis. FASEB J. 2014;28(8):

haematologica | 2018; 103(5)

DCC-2618: a new drug against mastocytosis

3540-3551.

40. Peter B, Gleixner KV, Cerny-Reiterer S, et al.

Polo-like kinase-1 as novel target in neoplastic mast cells: demonstration of growthinhibitory effects of siRNA and the Pololike kinase-1 targeting drug BI 2536.

Haematologica. 2011;96(5):672-680.

41. Wimazal F, Jordan JH, Sperr WR, et al.

Increased angiogenesis in the bone marrow of patients with systemic mastocytosis. Am J Pathol. 2002;160(5):1639-1645.

42. Peter B, Winter GE, Blatt K, et al. Target interaction profiling of midostaurin and its metabolites in neoplastic mast cells predicts distinct effects on activation and growth.

Leukemia. 2016;30(2):464-472.

43. Janku F, Razak ARA, Gordon MS, et al.

Pharmacokinetic-driven phase I study of DCC-2618, a pan-KIT and PDGFRA

haematologica | 2018; 103(5)

inhibitor, in patients with gastrointestinal stroma tumor (GIST) and other solid tumors. J Clin Oncol. 2017;35:15(S)2515.

44. Ustun C, Arock M, Kluin-Nelemans HC, et al. Advanced systemic mastocytosis: from molecular and genetic progress to clinical practice. Haematologica. 2016;101(10): 1133-1143.

45. Jawhar M, Schwaab J, Schnittger S, et al.

Additional mutations in SRSF2, ASXL1 and/or RUNX1 identify a high-risk group of patients with KIT D816V(+) advanced systemic mastocytosis. Leukemia. 2016;30(1): 136-143.

46. Gleixner KV, Mayerhofer M, Cerny-Reiterer S, et al. KIT-D816V-independent oncogenic signaling in neoplastic cells in systemic mastocytosis: role of Lyn and BTK activation and disruption by dasatinib and bosutinib.

Blood. 2011;118(7):1885-1898.

47. Wilson TM, Maric I, Simakova O, et al.

Clonal analysis of NRAS activating mutations in KIT-D816V systemic mastocytosis.

Haematologica. 2011;96(3):459-463.

48. Schwaab J, Schnittger S, Sotlar K, et al.

Comprehensive mutational profiling in advanced systemic mastocytosis. Blood.

2013;122(14):2460-2466.

49. Bibi S, Langenfeld F, Jeanningros S, et al.

Molecular defects in mastocytosis: KIT and beyond KIT. Immunol Allergy Clin North Am. 2014;34(2):239-262.

50. Damaj G, Joris M, Chandesris O, et al., ASXL1 but not TET2 mutations adversely impact overall survival of patients suffering systemic mastocytosis with associated clonal hematologic non-mast-cell diseases.

PLoS One. 2014;9(1):e85362.

809

ARTICLE

Acute Myeloid Leukemia

Ferrata Storti Foundation

Distinct protein signatures of acute myeloid leukemia bone marrow-derived stromal cells are prognostic for patient survival

Steven M. Kornblau,1* Peter P. Ruvolo,1* Rui-Yu Wang,1 V. Lokesh Battula,1 Elizabeth J. Shpall,2 Vivian R. Ruvolo,1 Teresa McQueen,1 YiHua Qui,1 Zhihong Zeng,1 Sherry Pierce,1 Rodrigo Jacamo,1 Suk-Young Yoo,3 Phuong M. Le,1 Jeffrey Sun,1 Numsen Hail Jr,1 Marina Konopleva1 and Michael Andreeff1

Haematologica 2018 Volume 103(5):810-821

Section of Molecular Hematology and Therapy, Department of Leukemia; 2Department of Stem Cell Transplantation and 3Bioinformatics and Computational Biology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, TX, USA 1

*SMK and PPR contributed equally to this work.

ABSTRACT

M

Correspondence: skornblau@mdanderson.org or mandreef@mdanderson.org Received: May 9, 2017.

Accepted: February 1, 2018.

Pre-published: March 15, 2018.

esenchymal stromal cells (MSC) support acute myeloid leukemia (AML) cell survival in the bone marrow (BM) microenvironment. Protein expression profiles of AML-derived MSC are unknown. Reverse phase protein array analysis was performed to compare expression of 151 proteins from AML-MSC (n=106) with MSC from healthy donors (n=71). Protein expression differed significantly between the two groups with 19 proteins over-expressed in leukemia stromal cells and 9 over-expressed in normal stromal cells.

Unbiased hierarchical clustering analysis of the samples using these 28 proteins revealed three protein constellations whose variation in expression defined four MSC protein expression signatures: Class 1, Class 2, Class 3, and Class 4. These cell populations appear to have clinical relevance. Specifically, patients with Class 3 cells have longer survival and remission duration compared to other groups. Comparison of leukemia MSC at first diagnosis with those obtained at salvage (i.e. relapse/refractory) showed differential expression of 9 proteins reflecting a shift toward osteogenic differentiation. Leukemia MSC are more senescent compared to their normal counterparts, possibly due to the overexpressed p53/p21 axis as confirmed by high β-galactosidase staining. In addition, overexpression of BCL-XL in leukemia MSC might give survival advantage under conditions of senescence or stress and overexpressed galectin-3 exerts profound immunosuppression. Together, our findings suggest that the identification of specific populations of MSC in AML patients may be an important determinant of therapeutic response.

doi:10.3324/haematol.2017.172429 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/810 Š2018 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.

810

Introduction There is growing evidence to support the importance of the leukemia bone marrow (BM) niche in the process of acute myeloid leukemia (AML) chemoresistance.1,2 Hence, optimal therapeutic strategies should also address neighboring cells in the tumor microenvironment. The critical support cells in the leukemia BM microenvironment are mesenchymal stromal cells (MSC).3-8 Depending on the type, MSC can act either to support or suppress tumors.4,8-15 Our group and others have found that MSC support leukemia cell survival by diverse mechanisms that include secretion of cytokines and chemokines, activation of survival signaling in tumor cells, and blocking immune surveillance by suppressing natural killer (NK) and T cells.2-5,13 Mesenchymal stromal cells are essential for human hematopoiesis, particularly as a source of SDF-1, which regulates homing, proliferation, and differentiation.6,9,10,16-18 Moreover, studies from our group and others have demonstrated that MSC protect leukemia cells from chemotherapy.6,19-23 We have recently found that there is reciprocal activation of NFkB signaling between MSC and AML and acute lymphoblastic leukemia (ALL) cells that likely contribute to the effectiveness of the microenvironment to protect malignant cells.7 Medyouf et al. recently demonstrathaematologica | 2018; 103(5)

Proteomic profiling reveals complexity of AML MSC

ed that blast cells from myelodysplastic syndrome (MDS) patients induce changes in MSC reflecting reprograming of the stromal cells.24 MSC may also influence hematopoietic precursors to promote leukemogenesis as evidenced by the development of AML and MDS in mice where the MSC osteo-progenitors were engineered to lack Dicer, a key regulator of microRNA (miR) processing.2 Furthermore, a recent study from Zhao et al. reported that p21 could be critical for inducing senescence in MSC from MDS patients with concomitant induction of interleukin6 (IL6) and transforming growth factor β (TGF-β).25 This study is consistent with findings that support the role of Dicer in regulating MSC biology and also establish a possible mechanism of aberrant survival functions in malignant MSC that may be associated with p21 and senescence. The ability of malignant MSC to withstand senescence may depend on the expression of the anti-apoptotic molecule BCL-XL.26,27 The cellular composition of stromal cells in a cancer microenvironment, such as the leukemic BM niche, is likely markedly different from that of the normal BM. We, therefore, set out to study the protein expression and activation in leukemic MSC (AML-MSC) and compared and contracted these to normal MSC (NL-MSC) to determine if and how they are functionally different. Reverse phase protein array analysis (RPPA, pioneered in our laboratory)28-32 was used to examine expression of 151 proteins in MSC derived from AML BM (n=106) with those derived from healthy donors (n=71). The results presented here identify 28 that were differentially expressed between the two. Importantly, the 28 proteins identified as differentially expressed in the AML versus normal MSC could be grouped into four protein constellation (PC) expression signatures with different biological properties and clinical implications regarding patient response to therapy.

Methods Patients’ samples Bone marrow was obtained from AML patients (n=106) undergoing diagnostic BM aspiration and from healthy donors (n=71) who were undergoing BM harvest for use in allogeneic BM transplantation. Samples were acquired in accordance with the regulations and protocols approved by the Investigational Review Board of MD Anderson Cancer Center. Informed consent was obtained in accordance with the Declaration of Helsinki. Samples were analyzed under an Institutional Review Board-approved laboratory protocol. Patients' characteristics are presented in Table 1. Details of isolation of MSC are available in the Online Supplementary Methods.

RPPA Proteomic profiling was carried out on MSC samples from patients with AML and healthy donors using RPPA. The RPPA method and sample validation technique are described fully elsewhere.28-32 Antibodies against 151 proteins were used for analysis.

(A list and the source of the antibodies and the concentrations utilized is provided in the Online Supplementary Table S1). The sources of antibodies have been been reported previously.30 An IgG subtype-specific secondary antibody was used to amplify the signal, and finally a stable dye was precipitated. The stained slides were analyzed using the Microvigene software (version 3.0, Vigene Tech, Carlise, MA, USA) to produce quantified data. Statistical analyses are described in the Online Supplementary Methods. haematologica | 2018; 103(5)

Cell senescence assessment

Microscopy assessment of β-galactosidase staining was used to detect cell senescence using a detection kit from Cell Signal Technology (Boston, MA, USA). Early passage cells (passage 2) were imaged using a Nikon Coolpix 950 camera attached to a Nikon TMS light microscope (Nikon Instruments Inc.). AML-MSC (n=4) and NL-MSC (n=5) were lysed in kit buffer. Measurement of β-galactosidase was performed using an in vitro fluorometric assay with fluorescein di-β-D-galactopyranoside (FDG) as substrate.

Incubation time was 2 hours (h). Fluorescence was measured using an Optima Fluorometer (Durham, NC, USA). Activity is presented as fluorescence units/1000 cells/minute.

Pathway analysis String software (String 10.1; available from: http://string-db.org)33 was used to determine protein associations. Pathway analysis to identify canonical pathways, upstream regulators, and protein networks was performed using Ingenuity Pathway software (Qiagen).

Results Proteins are differentially expressed in AML versus healthy MSC We have routinely utilized RPPA to analyze protein expression from clinical samples from many hematologic malignancies.28-32 We examined protein expression in blasts from newly diagnosed AML patients (n=85), CD34+ cells from normal donors (n=10), MSC from healthy donors (n=71), and MSC from newly diagnosed AML patients (n=54). Both normal MSC and AML-MSC expressed MSC defining lineage markers CD73, CD90, and CD 105 as determined by flow cytometry (Online Supplementary Figure S1). MSC from salvage samples (i.e. relapse/refractory) were also studied (n=46). The RPPA was probed with 151 antibodies targeting 119 different proteins (114 targeting total protein with 32 paired antibodies targeting phosphoepitopes on 26 proteins, and 5 with only a phosphoepitope but not total protein epitope) covering a wide variety of cellular functions and pathways (Online Supplementary Table S1). Protein expression in AML-MSC, NL-MSC, AML blasts and normal CD34+ cells was compared using principle component analysis (Online Supplementary Figure S1) and unbiased hierarchical clustering (Online Supplementary Figure S2B). NL-MSC and AMLMSC formed a cluster distinct from AML blasts and NLCD34+ cells with the vast majority of the 151 proteins tested showing statistically significant differential expression (143 of 151, P=0.01; 124 of 151, P≤10-6) between MSC and the blast/CD34+ cells. This unsurprising observation is consistent with a previous report that gene expression profiles are distinct between blood cells and MSC.34 Principal component analysis (PCA) also shows that protein expression in NL-CD34+ cells is distinct from that of AML blasts. These findings were identical to those observed when the analysis was restricted to samples from newly diagnosed patients alone.

Next we investigated whether protein expression in AML-MSC was different from that of NL-MSC. In PCA, the NL-MSC occupied a distinct space from that of the AML-MSC (Figure 1A). Unbiased hierarchical clustering comparing AML-MSC and NL-MSC revealed differential expression of 28 of those proteins (P<0.001; Q=0.0059).

The Q-value, a measure of the false discovery rate determined by a β-uniform mixture model highlights that these 811

S.M. Kornblau et al.

differences are very unlikely to be random and suggest that there are significant differences in the protein expression patterns of AML-MSC relative to those of NL-MSC.

The proteins clustered into three PCs (from top to bottom in Figure 1B). Nineteen of these proteins had generally higher expression in AML-MSC, including P1 [STAT1, p-

Table 1. Patientsâ&#x20AC;&#x2122; demographics (see Figures 2A and 3A).

Variable Cohort size Status Sex Race

FAB

PS

AHD Cyto

FLT3

D835

NPM1

Response

Relapse Vital

Category

Total

Class 2

Class 1

Class 3

Class 4

Number Percentage New Salvage M F Asian Black White Hispanic Unknown M0 M1 M2 M4 M5 M6 RAEB-T Unk Asymptomatic Symptoms In bed <50 In bed >50 100% bedridden N Y Favorable Intermediate Unfavorable Negative Positive Not Done Negative Positive Not Done Negative Positive Not Done CR PR HI No response FAIL Yes Alive Dead

101 100.00% 53.50% 45.50% 61.40% 38.60% 6.90% 7.90% 73.30% 9.90% 3.00% 4.00% 16.80% 24.80% 26.70% 12.90% 3.00% 4.00% 7.90% 15.80% 62.40% 13.90% 2.00% 1.00% 75.20% 24.80% 2.00% 54.50% 43.60% 65.30% 20.80% 8.90% 79.20% 6.90% 8.90% 46.50% 10.90% 5.00% 58.80% 29.70% 1.00% 3.00% 6.90% 70.00% 21.60% 78.40%

6 5.90% 66.70% 33.30% 33.30% 66.70% 0.00% 16.70% 50.00% 33.30% 0.00% 0.00% 16.70% 16.70% 66.70% 0.00% 0.00% 0.00% 0.00% 16.70% 50.00% 33.30% 0.00% 0.00% 50.00% 50.00% 16.70% 83.30% 0.00% 33.30% 16.70% 50.00% 50.00% 0.00% 50.00% 50.00% 0.00% 50.00% 66.70% 0.00% 0.00% 33.30% 0.00% 100.00% 0.00% 100.00%

13 12.90% 53.80% 46.20% 30.80% 69.20% 7.70% 15.40% 69.20% 7.70% 7.70% 0.00% 38.50% 23.10% 23.10% 7.70% 0.00% 0.00% 7.70% 15.40% 46.20% 23.10% 7.70% 0.00% 84.60% 15.40% 7.70% 53.80% 38.50% 69.20% 23.10% 7.70% 92.30% 0.00% 7.70% 69.20% 7.70% 23.10% 71.40% 0.00% 0.00% 14.30% 14.30% 80.00% 28.60% 71.40%

18 17.80% 66.70% 33.30% 72.20% 27.80% 11.10% 0.00% 83.30% 5.60% 0.00% 5.60% 5.60% 22.20% 27.80% 16.70% 11.10% 0.00% 11.10% 11.10% 77.80% 5.60% 0.00% 5.60% 83.30% 16.70% 0.00% 50.00% 50.00% 66.70% 16.70% 16.70% 77.80% 5.60% 16.70% 72.20% 0.00% 27.80% 63.60% 0.00% 9.10% 9.10% 18.20% 28.60% 36.40% 63.60%

64 63.40% 48.40% 50.00% 67.20% 32.80% 6.30% 7.80% 73.40% 9.40% 3.10% 4.70% 15.60% 26.60% 23.40% 14.10% 1.60% 6.30% 7.80% 17.20% 62.50% 12.50% 1.60% 0.00% 73.40% 26.60% 0.00% 53.10% 46.90% 64.10% 21.90% 14.10% 76.60% 9.40% 14.10% 45.30% 17.20% 37.50% 53.30% 3.30% 6.70% 13.30% 23.30% 81.30% 16.70% 83.30%

P

0.55 0.03 0.5

0.21

0.44

0.33 0.026

0.0979

0.005

0.08

0.8

0.076 0.41

M: male; F: female; FAB: French-American-British Classification; PS: propensity score; AHD: antecedent hematologic disease; N: no; Y: yes; Cyto: cytogenetic profile; CR: complete response; PR: partial response; HI: hematologic improvement;*Statistically significant (P) sets are in bold.

812

haematologica | 2018; 103(5)

Proteomic profiling reveals complexity of AML MSC

PDK1 (S241)], CCND1, CDKN1A (p21), ITGA2, PARP1, PPP2R2A/B/C/D, the PP2A B regulatory subunit family B55, BAK1, CSNK2A1, CDK4, GSK3A/B) and PC2 [STAT5A/B, BCL2L1 (BCL-XL)], DIABLO, TP53 (p53), NOTCH 1 (cleaved 1744), SPP1, p-EGFR (Y992), and ERBB2). Expression of 17 of the 19 proteins was validated by immunoblot analysis (Online Supplementary Figure S3).

Although expression of EGFR and ERBB2 expression could not be confirmed by western blot analysis, this may reflect the enhanced sensitivity of RPPA over standard immunoblot technology. The remaining nine proteins in PC3 were elevated in healthy donor MSC compared to AML-MSC: SMAD1, CREB1.p133 STMN1, SIRT1, CREB1 SMAD4, p-Foxo1/3 (S32), HSP90AA1/B1, and EIF2S1.

A

B

Figure 1. Mesenchymal stromal cell (MSC) protein expression signatures. Protein expression is distinct between normal MSC and acute myeloid leukemia (AML)MSC. (A) Principal component analysis (PCA) of 151 proteins examined in Class 1 (yellow), Class 2 (light blue), Class 3 (orange) and Class 4 (dark blue). (B) Unbiased hierarchical clustering identifies 3 protein signature groups: Group 1 (11 members), Group 2 (8 members) and Group 3 (9 members) in Class 1 (yellow), Class 2 (light blue), Class 3 (orange) and Class 4 (dark blue) groups, identified in top row as “MSC protein type”. MSC derived from normal donor (light blue) or AML patient (dark blue) is shown in the second row marked “cell type”.

haematologica | 2018; 103(5)

813

S.M. Kornblau et al.

Table 2. List of proteins with reverse correlation to one or more other proteins in acute myeloid leukemia (AML) mesenchymal stromal cells (MSC) versus normal MSC.

Protein

AML-MSC negative normal MSC positive

AKT1S1 AKT3 ARC ATF3 BCL2 BCL2L11 BECN1 BID BIRC5 CAV1 CCNB1 CTNNB1

STAT3 BCL2L11, ERG, SFN, SRCp416 STAT3 p727 BCL2, KIT, SFN, TP53 p15 ATF3, CTNNB1 ATF3, CTNNB1 PSMD9, YWHAE GAPDH MS4A1, FN1, SRC p416

CDK1 CDK4 CREB1 CREB2 p133 DIABLO EGLN1 EIF2S1 p51 ELK1 p383 ERBB2 p1248 ERG FN1 FOXO1.3/FOXO3 p24.p32 FOXO3 GAB2 p452 GAPDH GSK3A.B HDAC3 HNRNPK HSP90AA1.B1 HSPB1 IRS1 p1101 ITGAL JMJD6 JUN p73 KIT LEF1 MAP2K1 MAPK1.3 p202.p204 MAPK9 MAPK14 MS4A1 NPM1 NR4A1 NRP1 PECAM1

MAPK1.3 p202.p204 BCL2, BCL2L11, FOXO1.3 p24.p32, FOXO3, KIT, ITGAL, PSMD9 p10, RPS6KB1 p389, SFN, SRC, SRC p416 CREB1 GAPDH, MAPK1.3 p202.p204, YWHAE CDK1, JUN p73 MAPK9 HSPB1

SMAD1 TCF4 AKT3 BIRC5 CTNNB1 CTNNB1, IRS1 p1101

AML-MSC positive normal MSC negative

CTNNB1

BIRC5, GAB2 p452 BID, EIF2S1 p51, IRS1 p1101, RPS6KB1 PTK2 ARC, HSP90AA1.B1, MAPK9, PTGS2, YWHAZ

DIABLO, PECAM1, SFN TP53 CREB1 FOXO1.3 p24.p32, FOXO3, MS4A1, SRC p416, TCF4, YWHAE BIRC5, TGM2

EGLN1 EGLN1 BID

BID, CDK4, GSK3A.B GAPDH, STK11 PSMD9 STAT5A.B CTNNB1 DIABLO FOXO3 CTNNB1 STAT3 p727 CREB1 ATF3, CTNNB1 PTK2 CCNB1, CDK4 CREB2 p133 STAT1 BIRC5 PECAM1 JUN p73 NR4A1

BIRC5

NRP1

XIAP SRC p527 CTNNB1 EGLN1 SMAD1

CREB1 continued on the next page

814

haematologica | 2018; 103(5)

Proteomic profiling reveals complexity of AML MSC continued from the previous page

PPARG PRKAA1.2 p172 PSMD9 p10 PSMD9 PTGS2 PTK2 RPS6KB1 p389 RPS6KB1 SFN SMAD1 SRC SRC p416 SRC p527 STAT1 STAT3 p727 STAT3 p727 STAT5A.B STK11 STMN1 TCF4 TGM2 TP53 TP53 p15 XIAP YWHAE YWHAZ

SMAD1 SRC p527 CTNNB1 BECN1, HDAC3 LEF1 CTNNB1 AKT3, ATF3, CTNNB1 ELK1 p383 CTNNB1 AKT3, BIRC5, CTNNB1, XIAP MAPK14, STAT5A.B AKT1S1 ARC, JMJD6 STAT1 GSK3A.B ERBB2 p1248

ATF3 SRC p416 BECN1, CDK4

Based on various combinations of proteins expressed within these three groups, we observed that normal or leukemic MSC samples clustered into four distinct PCs. The first (aqua in Figure 1A; top row of annotations above the heat map in Figure 1B), was comprised predominantly of NL-MSC (42 of 46 members), characterized by lower expression of proteins in constellations 1 and 2, and higher expression of constellation 3 proteins. This was considered to represent the protein expression pattern of NLMSC. AML cases in this signature are hereafter called Class 2.

A second signature (blue in Figure 1A and B) was comprised predominantly of AML-MSC (62 of 72 members), and was characterized by the opposite protein expression of the NL-MSC signature, with high expression of PC1 and 2 proteins and lower expression of PC3 proteins. This signature is called Class 4. Two groups with expression signatures between that of the Class 2 MSC and Class 4 MSC were also recognized. One signature (yellow in Figure 1A and top row of Figure 1B), comprised of 12 AML-MSC and 11 NL-MSC, mirrored the expression of Class 2 for protein constellation 1, but that of Class 4 for PC2 and 3. AML cases in this group are hereafter referred to as Class 1. The final signature (orange in Figure 1A and top row of Figure 1B), consisting of 16 AML-MSC and 8 NL-MSC, mirrored the expression of Class 4 for PC1 and 3 but had low expression of PC2, similar to Class 2. AML cases from this signature are referred to as Class 3. haematologica | 2018; 103(5)

CTNNB1 CAV1 BIRC5 CREB1, STMN1 NPM1, PPARG, EGLN1 PRKAA1.2 p172, MAPK1.3 p202.p204

HNRNPK BCL2L11, SFN EGLN1 EIF2S1 p51 CREB2 p133 MAP2K1 EGLN1 CTNNB1

Clinical correlation Among the AML cases, there were very few clinical or laboratory features that showed a statistically significant correlation with the four MSC PCs. Specifically there was no association for continuous variables: age, percent BM or blood blasts, hemoglobin, platelet count, serum lactate dehydrogenase, albumin, creatinine, bilirubin fibrinogen, CD7, CD10, CD13, CD14, CD20, CD33 or CD34, or discrete variables disease status (new vs. salvage), race, French-American-British (FAB) Classification, performance status, history of an antecedent hematologic disorder, presence of an FLT3-ITD or presence of an NPM1 mutation (Table 1). However, a few notable associations between MSC protein signature and clinical features were observed. There was a statistical difference in presence of FLT3 D835 mutation (P=0.05). While no Class 1 or Class 2 samples contained FLT3 D835 mutation, 9.4% of Class 4 MSC and 5.6% of Class 3 MSC had the mutation (Table 1). Interestingly, there was a statistical difference in MSC population types between men and women (P=0.03) (Table 1). Men had a higher percentage of Class 4 and Class 3 MSC while women had a higher percentage of Class 1 and Class 2 MSC. At present it is unclear if sex influences MSC biology in the leukemic niche. Sex-specific effects in the microenvironment in leukemia is not unprecedented as integrin-mediated adhesion-triggered activation of GSK3B occurs exclusively in leukemia progenitor cells derived from female AML patients.35 Total GSK3 and ITGA2 are present in protein constellation 1 815

S.M. Kornblau et al.

and the proteins in this constellation are elevated in Class 3 and Class 4 while they are lower in Class 1 and Class 2.

Perhaps the integrin/GSK3 axis is also important in AMLMSC. Finally, there were statistically significant biases with different cytogenetic groups associating with different MSC protein signatures (P=0.026) (Table 1). Specifically, although few patients had favorable cytogenetics in this dataset, these were found exclusively in the Class 1 and Class 2 signatures. Conversely, patients with unfavorable cytogenetics were not observed in the Class 2 signature, but they were found in relatively equal proportions in the three AML specific signatures (i.e. Classes 1, 3, and 4).

Having determined that leukemic MSC divide AML patients into discrete protein signatures, we next examined whether signature membership affected outcomes.

There was no significant difference in median overall survival between the four MSC populations, although the median survival of 105 weeks (wk) in the Class 3 signature was longer than the other three groups at 25, 48 and 58 wk (Figure 2A); this suggested that patients with Class 3 MSC may have a better disease prognosis. Despite the small sample size, relapse rates among the newly diagnosed cases were different, with 4 of 5 Class 1 MSC patients and 13 of 16 Class 4 MSC patients relapsing compared to only 2 of 7 Class 3 MSC cases (P=0.076).

Furthermore the median remission duration was different between patients with Class 3 MSC and the three other signatures (101 wk for Class 3 MSC vs. 42.2 wk for Class 4 MSC; P=0.05) (Figure 2B). Other comparisons were not made due to the small sample size. These findings suggest that MSC can influence patient survival, although at present the mechanism is unknown.

These findings suggest that the disease state dictates expression of at least some of the identified proteins in MSC.

Evidence for dysregulated signaling in AML-MSC The observed differences in protein expression between AML-MSC and NL-MSC suggest that pathway utilization may be dysregulated or non-canonical in the AML-MSC. To look for evidence of abnormal utilization, we searched for statistically significant (R>0.2; P< 0.0001) protein-protein correlations that were reversed in the direction of correlation in the AML-MSC compared to the pattern in the NL-MSC. A representative analysis of STAT5 expression with the other proteins in NL-MSC revealed that this transcription factor is negatively correlated with hnRPK and positively correlated with STAT1 (Online Supplementary Figure S5B). However, in AML-MSC the correlations are reversed (Online Supplementary Figure S5B).

A list of all protein correlations that are reversed in AML-MSC compared to NL-MSC is provided in Table 2.

Of note, there are significant changes in protein correlations involving β-catenin. As shown in Table 2, BCL2, BCL2L11, FOXO1.3 p24.p32, FOXO3, KIT, ITGAL, PSMD9 p10, RPS6KB1 p389, SFN, SRC, and SRC p416 are negatively correlated with β-catenin in NL-MSC but

A

P=0.26

P=0.6

Effect of age on protein expression in NL-MSC and AML-MSC As AML is a disease of the elderly, many of the AMLMSC samples were obtained from individuals aged over 60 years (y). Of the 106 AML-MSC samples, 50 samples were from individuals under 60 y. Of the 71 MSC samples from healthy donors, 68 samples were from individuals under 60 y. To assess the impact of age on the differences in protein expression between the AML-MSC and NLMSC groups, the protein expression of 24 of the differentially-expressed proteins was reassessed in age-matched sets. The three groups were: a) under 30 y (AML n=13; NL n=37); b) 30-49 y (AML n=17; NL n=25); and c) 50 to 59 y (AML n=20; NL n=6). As shown in Online Supplementary Table S2 and Online Supplementary Figure S4A, of the 19 proteins elevated in AML-MSC compared to NL-MSC, only two (i.e. BCL-XL and PPP2R2A/B/C/D) were significantly higher in all age groups. This suggests that increased expression of these 2 proteins is dependent on the disease state and not on age. Six other proteins including p53 and CDKN1A (p21) were also elevated in two of the three age groups suggesting expression of these proteins is also dependent on the disease state rather than on age (Online Supplementary Table S2).

Analysis of p53 and CDKN1A (p21) protein expression was performed by western blot analysis on age-matched AML and NL samples (age 40-49 y; n=3 for both groups).

Protein expression of both p53 and CDKN1A (p21) was at least 2- to 3-fold higher in AML-MSC compared to the healthy donor samples (Online Supplementary Figure 4B).

816

B

P=0.05

P=0.12

Figure 2. Survival and remission duration differ in patients based on mesenchymal stromal cell (MSC) population. Kaplan-Meir curves showing overall survival (A) and remission duration (B). N: number.

haematologica | 2018; 103(5)

Proteomic profiling reveals complexity of AML MSC

are positively correlated with the protein in AML-MSC. ARC (NOL3), HSP90AA1/B1, MAPK9, PTGS2, and YWHAZ were positively correlated with β-catenin in NL-MSC but are negatively correlated with the protein in AML-MSC. Ingenuity Pathway Analysis (IPA) was performed using software on the proteins identified as differentially correlated with β-catenin in the MSC sets and β-catenin itself (i.e. CTNNB1; BCL2; BCL2L11; FOXO1; FOXO3; KIT; ITGAL; PSMD9; RPS6KB1; SFN; SRC; NOL3; HSP90AA1; HSP90B1; MAPK9; PTGS2; YWHAZ). IPA revealed these proteins were highly associated with PI3K/AKT signaling (top canonical pathway; P=7.35E-19) (Online Supplementary Figure S5). IPA also identified the top upstream regulator of this set of proteins as p53 (P=1.04E-11).

Proteins differentially expressed in AML-MSC share interactomes To assess the relationship among the proteins identified in the RPPA analysis, protein association network analysis was performed using STRING 10.533 on proteins identified as significantly different in the AML-MSC and NLMSC (Figure 1B). Blalock et al. used a previous version of String software to map the nuclear interactome.36 In cases where a family of proteins was identified (e.g. PPP2R2 set and HSP90 set), a representative member was included in the analysis. With the exception of PDK-1 all proteins are interconnected at least through one association (Figure 3).

This finding suggests that there is an interconnection between the various proteins that are distinctly expressed between the NL-MSC and AML-MSC groups.

Activation Inhibition Binding Phenotype Catalysis Post-translational regulation Reaction Transcriptional regulation Figure 3. Proteins differentially expressed in acute myeloid leukemia (AML) and normal mesenchymal stromal cell (MSC) are highly interactive. (A) String analysis was performed by String 10.0 using interactions based on “action”; available from: http://string-db.org. (B) Model of involvement of Group 1 and 2 proteins in AKT signaling. Red: proteins are members of Group 1 or 2. Yellow; proteins are non-members but possible links.

haematologica | 2018; 103(5)

817

S.M. Kornblau et al.

Pathway analysis suggests NL-MSC have prominent adipogenic signaling while all AML-MSC populations have prominent PI3K/AKT signaling Pathway analysis was performed using IPA. Two separate data sets were created: one with proteins from PC3 (elevated in normal MSC) and one with proteins from PC1 and 2 (elevated in AML-MSC). Proteins in group 3 are associated with adipogenesis (ninth top canonical pathway; P=2.19E-05) (Online Supplementy Figure S6). PC3 proteins are elevated in MSC that were presumably normal but reduced in AML-MSC suggesting differences in differentiation potential of MSC between NL-MSC and AMLMSC. Of the 7 proteins identified, SIRT1, FOXO1, and SMAD1 each can activate PPARβ which is a critical regulator of adipocyte differentiation.37-39 PC3 also displayed the strongest association with AMPK signaling (P=6.84E07). The top upstream regulator identified in the set of PC3 proteins was PDGFB (P=1.01E-06). PI3K/AKT pathway was highly associated with PC1 and 2 proteins, which are elevated in AML-MSC (top canonical pathways; P=7.16E-17) (Online Supplementary Figure S7). AKT

was identified as one of the top three upstream regulators (P=1.22E-14), suggesting that signaling mediated by this survival kinase is prominent in AML-MSC.

AML-MSC are senescent compared to NL-MSC The p21 protein appears to be critical for senescence in myelodysplastic syndrome (MDS)-MSC22 and AMLMSC similar to MDS-MSC have elevated p21 (Figure 1B).

This finding suggests that AML-MSC might also be more senescent than NL-MSC, as was the case in MDS-MSC.25 Senescence was observed in normal donor MSC- and AML-derived MSC using β-galactosidase staining. AMLMSC were more senescent than MSC derived from healthy donors in this representative example (Figure 4A).

To account for age effects on senescence, MSC were taken from donors under 58 y. Average age of the AML patients (n=4) was 52 y and the average age of normal donors (n=5) was 47 y. Also, MSC of similar cell passage (passage 2 or 3) were used, so effects of cell passage were unlikely. As shown in Figure 4B, β-galactosidase activity was significantly higher (almost 2-fold) in AML-MSC compared to

A

B

Figure 4. Acute myeloid leukemia (AML) mesenchymal stromal cell (AML-MSC) are more senescent than normal MSC (NL-MSC). (A) Representative microscopy of an AML-MSC and a normal MSC with two different slide areas. (B). Level of β-galactosidase was assessed by enzymatic assay in normal MSC (n=5) and AML-MSC (n=4). Statistical significance determined by Student t-test; *P=0.027.

818

haematologica | 2018; 103(5)

Proteomic profiling reveals complexity of AML MSC

NL-MSC (P=0.032). These findings suggest that AMLMSC tend toward senescence.

Therapy alters proteins expression in AML-MSC Protein expression in AML-MSC might change between diagnosis and relapse, perhaps as a result of relapse, or in response to acquired changes in AML blasts. To determine if AML-MSC protein expression was changing between diagnosis and relapse we compared expression between the samples obtained at first diagnosis (n=53) to those of AML-MSC collected from primary refractory or relapsed patients (n=54). Nine proteins are differentially expressed between the two MSC groups (Figure 5).

Phosphorylated β-catenin, phosphorylated RPS6, and galectin-3 are expressed at higher levels in MSC in the salvage set. SMAD6, TCF4, LYN, integrin-β3, phosphorylated EIF4BP1, and phosphorylated ELK1 are higher in MSC at first diagnosis compared to MSC from salvage AML patients. IPA reveals that, for canonical pathways, a set associated with osteoblast differentiation was found to be the pathway most associated with the proteins differentially expressed at diagnosis compared to salvage (Online Supplementary Figure S8).

Discussion This study presents the first systematic study of protein expression differences between NL-MSC and AML-MSC. There were several notable observations. There were clear differences between the protein expression of MSC (whether from healthy donor or AML patient) and AML blast cells. This result was not surprising as one would predict different proteins would be prominent in mesenchymal cells and cells of hematopoietic/myeloid lineage. The major observation of this research was the dis-

covery that AML-MSC have significantly different protein expression patterns compared to normal MSC, with 28 of 151 analyzed proteins being highly significantly different (FDR<0.006). These changes assumed four signatures in AML, 81% of which were very different from that of normal MSC, while 6% had an identical signature to NLMSC, and another 13% were more like the normal signature than the leukemia patterns. Signature membership showed an association with cytogenetics, with 'favorable' cytogenetics being limited to the more NL-MSC-like pattern and 'unfavorable' cytogenetics not occurring in AML with an NL-MSC-like pattern. There was a difference in the distribution of the MSC population between men and women. The significance of these differences is not clear, but women tended to have higher percentages of Class 1 and Class 2 MSC compared to men. In leukemia progenitor cells, GSK3B is activated via an integrin-mediated mechanism in response to adhesion to a stromal cell exclusively in women patients.35 As ITGA2 and GSK3 are members of a protein constellation (i.e. constellation 1) that is differentially expressed in Class 1 and Class 2 (lower levels) compared to Class 3 and Class 4 (higher levels), it is tempting to speculate that integrin/GSK3 axis may contribute to sex-specific effects in MSC. Furthermore, these signatures influence outcomes including response rates, remission duration and, perhaps, survival. Patients with Class 3 MSC fare much better than patients with Class 4 MSC as demonstrated by significantly longer remission duration and a trend toward longer OS. Changes in protein expression were often characterized by protein-protein correlations that were reversed from those seen in normal MSC, providing insight into the nature of this dysregulation and potentially providing therapeutic targets. In NL-MSC, the signature proteins were associated with adipocyte differentiation. That normal MSC, but not AML-MSC, possess protein pathways impor-

Figure 5. A distinct set of proteins is associated with acute myeloid leukemia (AML) patient salvage status. (A) Reverse phase protein arrays (RPPA) reveals protein expression in AML mesenchymal stromal cells (MSC) differs between diagnosis and the salvage setting for 9 proteins (P=0.05; false discovery rate=0.68).

haematologica | 2018; 103(5)

819

S.M. Kornblau et al.

tant in adipogenesis is consistent with other studies from our group and another group that found AML-MSC are primed toward osteoblastic differentiation and not adipocytic differentiation.9,40,41 We reported that AML-MSC express higher levels of osteogenic markers, including Tissue Non-specific Alkaline Phosphatase (TNAP), RUNX2, Osterix, and Ostepontin, compared to MSC from agematched heathy donors.40 In addition, in that study we found that AML-MSC readily differentiate along the osteogenic lineage pathway but are unable to differentiate into adipocytes. This differentiation potential of the MSC may be influenced by the leukemia cells themselves, as exposure of healthy donor MSC to AML cell lines such as OCI-AML3 induces gene expression of RUNX2, TNAP, and other osteogenic genes, and induces osteogenic differentiation of the stromal cells.40 The protein networks prevalent in NL-MSC that we identify here are consistent with signaling that skews toward adipocytic differentiation in the normal cells. Diaz de la Guardia et al. found that MSC from a highrisk AML group failed to differentiate into adipcocytes.41 The IPA data on canonical pathways in the AML-MSC compared at first diagnosis with refractory and relapse samples suggests the importance of MSC of the osteoblastic lineage in the AML niche, as many of these proteins are associated with osteoblast survival and differentiation.

PI3K/AKT is very prevalent in group 1 proteins which are associated with Class 4 MSC by IPA. We have previously reported that leukemia cells in co-culture with MCS induce activation of AKT and other survival kinases.42 It is plausible that the observed activation of AKT signaling in AML-MSC is due to the presence of leukemia cells in the niche or at least that the malignant cells may contribute to activation. In MSC, AKT has been implicated in positive regulation of Cyclin D1 (CCND1) and CDK4.43,44 The presence of the PP2A B55 a subunit in group 1 suggests that either the protein phosphatase is not active against AKT in those cells or that AKT is less active in the AML-Like MSC compared to Normal-Like MSC. Despite its tumor suppressor role in suppression of AKT signaling, the PP2A subunit does support β-catenin expression by dephosphorylating serine and threonine residues that are required for destruction of the transcription factor (e.g. serine 33, threonine 41).45-47 It is interesting that when expression of the 151 proteins surveyed by RPPA are correlated in the normal MSC and in the AML-MSC, β-catenin exhibits the greatest difference in correlation pattern between normal MSC and AML-MSC compared to the other proteins. It is plausible that the PP2A B55 a subunit may be a factor in this phenomenon, though further investigation will be required to verify this potential mechanism. It is interesting to note that suppression of PP2A (albeit via the catalytic core subunit Ca) in MSC cell line or pre-adipocyte cells

References 1. Konopleva M, Tabe Y, Zeng Z, Andreeff M. Therapeutic targeting of microenvironmental interactions in leukemia: mechanisms and approaches. Drug Resist Updat.

2009;12(4-5):103-113.

2. Tabe Y, Konopleva M. Advances in understanding the leukaemia microenvironment.

Br J Haematol. 2014;164(6):767-778.

820

results in differentiation to adipocytes via a mechanism involving loss of β-catenin.48 As pathway analysis of proteins associated with normal MSC (i.e. Constellation 3) pathways identifies adipogenesis as a key pathway, it will be important to determine if AML-MSC would be less likely to skew toward adipocyte differentiation because of a PP2A B55 a subunit/β-catenin axis.

BCL-XL is necessary for MSC survival during differentiation.27 AML-MSC expressed higher BCL-XL and Cyclin D1 protein. These findings may be attributed to STAT5 as this transcription factor is a regulator of both BCL-XL and Cyclin D1.49 We found by qRT-PCR that gene expression of both BCL-XL and Cyclin D1 was higher in AML-MSC (n=9) compared to normal MSC (n=10) (Online Supplementary Figure S9). These findings suggest that elevation of BCL-XL and Cyclin D1 protein in AML-MSC can be attributed at least in part to a transcriptional mechanism possibly involving STAT5.

Two prominent proteins identified as elevated in the AML-MSC group are p21 and p53 (Figure 1B). Elevated expression of p21 and increased senescence of MSC is consistent with the study on MDS-MSC that showed a role for p21 in IL-6 and TGF-β production.25 However, a recent study from Desbourdes et al. found p21 and p53 levels were similar between AML-MSC and healthy donor-derived MSC.50 The reason for the difference between our results and their results is not clear. It should be noted that the Desbourdes et al. study used less than 5 samples each of AML-derived MSC and healthy donorderived MSC to determine p21 and p53 levels, so perhaps Class 1 or Class 2 MSC (which would have lower levels of p21 and p53) are over-represented in their samples.50 In addition, the average age of the AML patients in the Desbourdes et al. study was 49 y while the average age of the healthy donors was near 60 y, so perhaps the p21 and p53 levels in the healthy donor MSC are skewed higher as the donors are older than the patients. For p21 and p53 expression, an age match comparison of p21 and p53 in the AML-MSC and NL-MSC shows levels of these proteins are higher in the AML-MSC in at least 2 ages for each (Online Supplementary Table S2).

In conclusion, proteomic analysis identified a distinct set of proteins that distinguish normal MSC from AMLMSC. Our RPPA studies identified four major signatures of MSC in AML patients that may impact their function in the tumor microenvironment.

Funding This work was supported in part by the research funding from the National Institutes of Health (P01CA49639 and MD Anderson Cancer Center Support Grant CA016672) and the Paul and Mary Haas Chair in Genetics (to MA).

3. Barcellos-de-Souza P, Gori V, Bambi F, Chiarugi P. Tumor microenvironment: bone marrow-mesenchymal stem cells as key players.

Biochim Biophys Acta.

2013;1836(2):321-335.

4. da Silva Meirelles L, Chagastelles PC, Nardi NB. Mesenchymal stem cells reside in virtually all post-natal organs and tissues. J Cell Sci. 2006;119(Pt 11):2204-2213.

5. Eggenhofer E, Luk F, Dahlke MH, Hoogduijn

MJ. The life and fate of mesenchymal stem cells. Front Immunol. 2014;5:148.

6. Greenbaum A, Hsu YM, Day RB, et al.

CXCL12 in early mesenchymal progenitors is required for haematopoietic stem-cell maintenance. Nature. 2013;495(7440):227230.

7. Jacamo R, Chen Y, Wang Z, et al. Reciprocal leukemia-stroma VCAM-1/VLA-4-dependent activation of NF-kappaB mediates

haematologica | 2018; 103(5)

Proteomic profiling reveals complexity of AML MSC

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

chemoresistance. Blood. 2014;123(17):26912702.

Studeny M, Marini FC, Dembinski JL, et al.

Mesenchymal stem cells: potential precursors for tumor stroma and targeted-delivery vehicles for anticancer agents. J Natl Cancer Inst. 2004;96(21):1593-1603.

Battula VL, Chen Y, Cabreira Mda G, et al.

Connective tissue growth factor regulates adipocyte differentiation of mesenchymal stromal cells and facilitates leukemia bone marrow engraftment. Blood. 2013; 122(3): 357-366.

Chen Y, Jacamo R, Shi YX, et al. Human extramedullary bone marrow in mice: a novel in vivo model of genetically controlled hematopoietic microenvironment. Blood.

2012;119(21):4971-4980.

Kidd S, Spaeth E, Watson K, et al. Origins of the tumor microenvironment: quantitative assessment of adipose-derived and bone marrow-derived stroma. PLoS One. 2012;7(2):e30563.

Ling X, Marini F, Konopleva M, et al.

Mesenchymal Stem Cells Overexpressing IFN-beta Inhibit Breast Cancer Growth and Metastases through Stat3 Signaling in a Syngeneic Tumor Model.

Cancer Microenviron. 2010;3(1):83-95.

Poggi A, Giuliani M. Mesenchymal Stromal Cells Can Regulate the Immune Response in the Tumor Microenvironment. Vaccines (Basel). 2016;4(4).

Spaeth E, Klopp A, Dembinski J, Andreeff M, Marini F. Inflammation and tumor microenvironments: defining the migratory itinerary of mesenchymal stem cells. Gene Ther. 2008;15(10):730-738.

Spaeth EL, Labaff AM, Toole BP, Klopp A, Andreeff M, Marini FC. Mesenchymal CD44 expression contributes to the acquisition of an activated fibroblast phenotype via TWIST activation in the tumor microenvironment. Cancer Res. 2013; 73(17):53475359.

Kim JA, Shim JS, Lee GY, et al.

Microenvironmental remodeling as a parameter and prognostic factor of heterogeneous leukemogenesis in acute myelogenous leukemia. Cancer Res. 2015; 75(11): 2222-2231.

Raaijmakers MH, Mukherjee S, Guo S, et al.

Bone progenitor dysfunction induces myelodysplasia and secondary leukaemia.

Nature. 2010;464(7290):852-857.

Shi C. Recent progress toward understanding the physiological function of bone marrow mesenchymal stem cells. Immunology.

2012;136(2):133-138.

Balakrishnan K, Burger JA, Quiroga MP, et al. Influence of bone marrow stromal microenvironment on forodesine-induced responses in CLL primary cells. Blood.

2010;116(7):1083-1091.

Kojima K, McQueen T, Chen Y, et al. p53 activation of mesenchymal stromal cells partially abrogates microenvironment-mediated resistance to FLT3 inhibition in AML through HIF-1alpha-mediated down-regulation of CXCL12. Blood. 2011; 118(16):44314439.

Marquez ME, Hernandez-Uzcategui O, Cornejo A, Vargas P, Da Costa O. Bone marrow stromal mesenchymal cells induce down regulation of CD20 expression on B-

haematologica | 2018; 103(5)

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

CLL: implications for rituximab resistance in CLL. Br J Haematol. 2015;169(2):211-218.

Xia B, Tian C, Guo S, et al. c-Myc plays part in drug resistance mediated by bone marrow stromal cells in acute myeloid leukemia.

Leuk Res. 2015;39(1):92-99.

Zeng Z, Shi YX, Samudio IJ, et al. Targeting the leukemia microenvironment by CXCR4 inhibition overcomes resistance to kinase inhibitors and chemotherapy in AML. Blood. 2009;113(24):6215-6224.

Medyouf H, Mossner M, Jann JC, et al.

Myelodysplastic cells in patients reprogram mesenchymal stromal cells to establish a transplantable stem cell niche disease unit.

Cell Stem Cell. 2014;14(6):824-837.

Zhao Y, Wu D, Fei C, et al. Down-regulation of Dicer1 promotes cellular senescence and decreases the differentiation and stem cellsupporting capacities of mesenchymal stromal cells in patients with myelodysplastic syndrome.

Haematologica.

2015; 100(2):194-204.

Oliver L, Hue E, Priault M, Vallette FM. Basal autophagy decreased during the differentiation of human adult mesenchymal stem cells. Stem Cells Dev. 2012; 21(15):27792788.

Oliver L, Hue E, Rossignol J, et al. Distinct roles of Bcl-2 and Bcl-Xl in the apoptosis of human bone marrow mesenchymal stem cells during differentiation. PLoS One. 2011;6(5):e19820.

Kornblau SM, Singh N, Qiu Y, Chen W, Zhang N, Coombes KR. Highly phosphorylated FOXO3A is an adverse prognostic factor in acute myeloid leukemia. Clin Cancer Res. 2010;16(6):1865-1874.

Kornblau SM, Womble M, Qiu YH, et al.

Simultaneous activation of multiple signal transduction pathways confers poor prognosis in acute myelogenous leukemia.

Blood. 2006;108(7):2358-2365.

Ruvolo PP, Qiu Y, Coombes KR, et al.

Phosphorylation of GSK3alpha/beta correlates with activation of AKT and is prognostic for poor overall survival in acute myeloid leukemia patients. BBA Clin. 2015;4:59-68.

Ruvolo PP, Qui YH, Coombes KR, et al. Low expression of PP2A regulatory subunit B55alpha is associated with T308 phosphorylation of AKT and shorter complete remission duration in acute myeloid leukemia patients. Leukemia. 2011; 25(11):1711-1717.

Tibes R, Qiu Y, Lu Y, et al. Reverse phase protein array: validation of a novel proteomic technology and utility for analysis of primary leukemia specimens and hematopoietic stem cells. Mol Cancer Ther.

2006;5(10):2512-2521.

Szklarczyk D, Franceschini A, Wyder S, et al. STRING v10: protein-protein interaction networks, integrated over the tree of life.

Nucleic Acids Res. 2015;43(Database issue):D447-452.

Huang JC, Basu SK, Zhao X, et al.

Mesenchymal stromal cells derived from acute myeloid leukemia bone marrow exhibit aberrant cytogenetics and cytokine elaboration. Blood Cancer J. 2015;5:e302.

Bertrand J, Despeaux M, Joly S, et al. Sex differences in the GSK3beta-mediated survival of adherent leukemic progenitors.

Oncogene. 2012;31(6):694-705.

Blalock WL, Piazzi M, Bavelloni A, et al.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

Identification of the PKR nuclear interactome reveals roles in ribosome biogenesis, mRNA processing and cell division. J Cell Physiol. 2014;229(8):1047-1060.

Ali AT, Hochfeld WE, Myburgh R, Pepper MS. Adipocyte and adipogenesis. Eur J Cell Biol. 2013;92(6-7):229-236.

Jin W, Takagi T, Kanesashi SN, et al.

Schnurri-2 controls BMP-dependent adipogenesis via interaction with Smad proteins.

Dev Cell. 2006;10(4):461-471.

Kauppinen A, Suuronen T, Ojala J, Kaarniranta K, Salminen A. Antagonistic crosstalk between NF-kappaB and SIRT1 in the regulation of inflammation and metabolic disorders. Cell Signal. 2013;25 (10):19391948.

Battula VL, Le PM, Sun JC, et al. AMLinduced osteogenic differentiation in mesenchymal stromal cells supports leukemia growth. JCI Insight. 2017;2(13).

Diaz de la Guardia R, Lopez-Millan B, Lavoie JR, et al. Detailed Characterization of Mesenchymal Stem/Stromal Cells from a Large Cohort of AML Patients Demonstrates a Definitive Link to Treatment Outcomes. Stem Cell Reports.

2017;8(6):1573-1586.

Tabe Y, Jin L, Tsutsumi-Ishii Y, et al.

Activation of integrin-linked kinase is a critical prosurvival pathway induced in leukemic cells by bone marrow-derived stromal cells. Cancer Res. 2007;67(2):684694.

Gharibi B, Ghuman MS, Hughes FJ. Aktand Erk-mediated regulation of proliferation and differentiation during PDGFRbetainduced MSC self-renewal. J Cell Mol Med. 2012;16(11):2789-2801.

Lai VK, Ashraf M, Jiang S, Haider K. MicroRNA-143 is a critical regulator of cell cycle activity in stem cells with co-overexpression of Akt and angiopoietin-1 via transcriptional regulation of Erk5/cyclin D1 signaling. Cell Cycle. 2012;11(4):767-777.

Hein AL, Seshacharyulu P, Rachagani S, et al.

PR55alpha Subunit of Protein Phosphatase 2A Supports the Tumorigenic and Metastatic Potential of Pancreatic Cancer Cells by Sustaining Hyperactive Oncogenic Signaling. Cancer Res. 2016; 76(8):22432253.

Stamos J, Weis W. The β-catenin destruction complex. Cold Spring Harb Perspect Biol 2013:5: a007898.

Zhang W, Yang J, Liu Y, et al. PR55 alpha, a regulatory subunit of PP2A, specifically regulates PP2A-mediated beta-catenin dephosphorylation. J Biol Chem. 2009; 284(34): 22649-22656.

Okamura H, Yang D, Yoshida K, Teramachi J, Haneji T. Reduction of PP2A Calpha stimulates adipogenesis by regulating the Wnt/GSK-3beta/beta-catenin pathway and PPARgamma expression. Biochim Biophys Acta. 2014;1843(11):2376-2384.

Grad JM, Zeng XR, Boise LH. Regulation of Bcl-xL: a little bit of this and a little bit of STAT. Curr Opin Oncol. 2000;12(6):543-549.

Desbourdes L, Javary J, Charbonnier T, et al.

Alteration Analysis of Bone Marrow Mesenchymal Stromal Cells from De Novo Acute Myeloid Leukemia Patients at Diagnosis. Stem Cells Dev. 2017;26(10): 709-722.

821

ARTICLE

Acute Myeloid Leukemia

Ferrata Storti Foundation

Clinical relevance of IDH1/2 mutant allele burden during follow-up in acute myeloid leukemia. A study by the French ALFA group

Yann Ferret,1,2,3 Nicolas Boissel,4,5 Nathalie Helevaut,1 Jordan Madic,6 Olivier Nibourel,1,2,3 Alice Marceau-Renaut,1,2,3 Maxime Bucci,1 Sandrine Geffroy,1 Karine Celli-Lebras,7 Sylvie Castaigne,8 Xavier Thomas,9 Christine Terré,10 Hervé Dombret,4,5 Claude Preudhomme1,2,3 and Aline Renneville1,2,3

Haematologica 2018 Volume 103(5):822-829

Hematology Laboratory, Biology and Pathology Center, CHRU of Lille; 2INSERM, UMR-S 1172, Cancer Research Institute of Lille; 3University of Lille, F-59000; 4Hematology Department, Saint-Louis Hospital, AP-HP, Paris; 5EA3518, Institut Universitaire d’Hématologie (IUH), University 7 Paris Diderot; 6Circulating Biomarkers Laboratory, Curie Institute, Paris; 7Acute Leukemia French Association (ALFA) coordination, Saint-Louis Hospital, AP-HP, Paris; 8Hematology Department, Versailles Hospital, Le Chesnay; 9 Hematology Department, Lyon Sud Hospital, Pierre Benite and 10Cytogenetic Laboratory, Versailles Hospital, Le Chesnay, France 1

ABSTRACT

A Correspondence: aline.renneville@gmail.com

Received: October 29, 2017.

Accepted: February 16, 2018.

Pre-published: February 22, 2018.

doi:10.3324/haematol.2017.183525 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/822

ssessment of minimal residual disease has emerged as a powerful prognostic factor in acute myeloid leukemia. In this study, we investigated the potential of IDH1/2 mutations as targets for minimal residual disease assessment in acute myeloid leukemia, since these mutations collectively occur in 15-20% of cases of acute myeloid leukemia and now represent druggable targets. We employed droplet digital polymerase chain reaction assays to quantify IDH1R132, IDH2R140, and IDH2R172 mutations on genomic DNA in 322 samples from 103 adult patients with primary IDH1/2 mutant acute myeloid leukemia and enrolled on Acute Leukemia French Association (ALFA) 0701 or -0702 clinical trials. The median IDH1/2 mutant allele fraction in bone marrow samples was 42.3% (range, 8.2 - 49.9%) at diagnosis of acute myeloid leukemia, and below the detection limit of 0.2% (range, <0.2 - 39.3%) in complete remission after induction therapy. In univariate analysis, the presence of a normal karyotype, a NPM1 mutation, and an IDH1/2 mutant allele fraction <0.2% in bone marrow after induction therapy were statistically significant predictors of longer disease-free survival. In multivariate analysis, these three variables remained significantly predictive of disease-free survival. In 7/103 (7%) patients, IDH1/2 mutations persisted at high levels in complete remission, consistent with the presence of an IDH1/2 mutation in pre-leukemic hematopoietic stem cells. Five out of these seven patients subsequently relapsed or progressed toward myelodysplastic syndrome, suggesting that patients carrying the IDH1/2 mutation in a pre-leukemic clone may be at high risk of hematologic evolution.

©2018 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.

822

Introduction Assessment of minimal residual disease (MRD) has emerged as a powerful prognostic factor in acute myeloid leukemia (AML).1–6 Many studies have shown that MRD detection using multiparameter flow cytometry or real-time quantitative polymerase chain reaction (qPCR) provides powerful independent prognostic information in AML.7,8 Chimeric fusion genes, such as PML-RARA, CBFB-MYH11 or RUNX1-RUNX1T1, NPM1 mutations, as well as WT1 expression are well-established molecular markers for MRD monitoring in AML. However, sensitive and leukemia-specific MRD markers are lacking in approximately 40% of AML patients. This prompted us to investigate the potential of other recurrent molecular haematologica | 2018; 103(5)

IDH1/2 mutant allele burden in AML

abnormalities as targets for MRD assessment in AML, such as mutations in isocitrate dehydrogenase (IDH) 1 and 2.

IDH1/2 mutations affecting IDH1R132, IDH2R140, and IDH2R172 residues are single-nucleotide mutations that collectively occur in 15-20% of AML and represent driver mutations in leukemogenesis.9 Mutant IDH1/2 enzymes have neomorphic activity and catalyze the reduction of aketoglutarate to an oncometabolite, the R-enantiomer of 2-hydroxyglutarate (2-HG), which promotes DNA and histone hypermethylation, altered gene expression, and impaired hematopoietic differentiation.10,11 Quantification of single-nucleotide mutations by qPCR can be challenging because of problems with background amplification from the wild-type allele. Recently, the development of digital PCR has enabled absolute quantification of various genomic targets with high precision and sensitivity and has, therefore, turned out to be a promising technique for MRD monitoring, especially for gene mutations.7,12 The clinical significance of residual IDH1/2 mutations in bone marrow in complete remission after chemotherapy is currently unknown. In this study, we employed digital PCR assays to quantify IDH1/2 mutant allele fraction at AML diagnosis and during follow-up in a large cohort of AML patients intensively treated in the Acute French Leukemia Association (ALFA) trials to investigate whether IDH1/2 mutations are suitable MRD markers that could predict clinical outcome in AML patients and provide further information for risk-adapted therapy.

Methods Patients and treatment This study was performed in 103 adult patients (18-70 years) with previously untreated primary IDH1/2 mutated AML and enrolled on the prospective ALFA-0701 (Eudra-CT 2007-002933-

36; ClinicalTrials.gov NCT00927498 or ALFA-0702 (Eudra-CT 2008-000668-18; ClinicalTrials.gov NCT00932412) trials.

Treatment schemes have been previously reported for both trials.13,14 These studies were approved by the ethics committee of Saint-Germain en Laye and Sud Est IV, France, respectively, and the institutional review board of the French Regulatory Agency.

Bone marrow or peripheral blood samples collected at the time of diagnosis of AML and during follow-up were obtained from the tissue bank Tumorothèque du Centre de Référence Régional en Cancérologie de Lille (CRRC)” and approval for this study was obtained from the institutional review board of CHRU of Lille (CSTMT089). All patients provided written informed consent to both treatment and genetic analysis before inclusion in the study, in accordance with the declaration of Helsinki. Among all patients included in the ALFA-0701 (n=278) or ALFA-0702 (n=704) trials, we selected patients meeting the following criteria: (i) the presence of an IDH1R132 or an IDH2R140/R172 mutation at AML diagnosis (n=160), (ii) achievement of complete remission after induction therapy (n=130), and (iii) one or more bone marrow follow-up sample available for IDH1/2 variant allele fraction (IDH1/2-VAF) assessment (n=103) (Figure 1).

Molecular analysis Droplet DigitalTM PCR (ddPCR) assays were used to quantify the IDH1/2 mutant allele and its wild-type counterpart in diagnostic and follow-up samples. During complete remission, only bone marrow samples were analyzed for IDH1/2-VAF assessment.

IDH1/2-VAF was quantified on genomic DNA using Bio-RadTM reagents, primers and probes (HEX-labeled wild-type allele; FAMlabeled mutant alleles). All samples were tested in duplicate wells, using 90 ng of DNA per well. The PCR product from each well was then subjected to the QX100 droplet reader (Bio-RadTM), which measures the fluorescence of each droplet individually using a two-color detection system. Raw data were analyzed using QuantaSoft software, version 1.7.4.0917 (Bio-RadTM).

Representative two-dimensional plots of droplet fluorescence for

Figure 1. Patient flow chart. VAF, variant allele fraction.

haematologica | 2018; 103(5)

823

Y. Ferret et al.

IDH1/2 wild-type controls and IDH1/2 mutant samples are shown in Online Supplementary Figure S1. The mutant allele frequency was then estimated using a Poisson distribution model as the fraction of positive droplets divided by total droplets containing a target.

The limit of detection was defined for each mutation as the mean value of IDH1/2 wild-type controls plus three standard deviations (Online Supplementary Table S1). The upper detection limit of these ddPCR assays (rounded to 0.2% of mutant allele frequency) was further considered as the threshold for statistical analysis. An IDH1/2-VAF level below 0.2% was hereafter considered as negative MRD. Gene mutation analysis and next-generation sequencing assays are described in the Online Supplementary Methods and Online Supplementary Tables S2-S4.

mutations in AML.9,17 In our cohort, IDH2R172K mutations were mutually exclusive with NPM1 and FLT3 mutations, but co-occurred with DNMT3A mutations. An isolated trisomy 11 was identified in 5/21 (24%) patients with the IDH2R172K mutation, while this cytogenetic abnormality was not found in any patient with other types of IDH1/2 mutations (24% versus 0%; P<0.001) (Figure 2). In the subgroup of IDH2R172K mutant AML (n=21), single-nucleotide polymorphism array analysis revealed an additional genomic lesion involving chromosome 11, consisting of a 11p11.2-q12.1 uniparental disomy, in one patient with normal karyotype AML. No MLL partial tandem duplication, known to be strongly associated with trisomy 11,18 was found by reverse transcriptase

Statistical analysis Group comparison for categorical and continuous variables was performed with the Fisher exact and Mann-Whitney test, respectively. Overall survival was calculated from the date of AML diagnosis to the last follow-up date by censoring patients alive at that date. Disease-free survival was calculated from the date of complete remission to the date of relapse or death, censoring patients alive without an event at the last follow-up date. In some analyses, data were censored at the time of allogeneic stem cell transplantation. Univariate and multivariate analyses assessing the impact of categorical and continuous variables were performed with a Cox model.15 The proportional-hazards assumption was checked before conducting multivariate analyses.16 Covariates with a P-value <0.1 in univariate analysis were included in the multivariable models.

Statistical analyses were performed with STATA software (STATA 12.0 Corporation, College Station, TX, USA). P-values were twosided, with P<0.05 denoting statistical significance.

Results Baseline characteristics of the patients and acute myeloid leukemias The patientsâ&#x20AC;&#x2122; median age was 54 years (range, 22-70).

The median follow-up was 2.7 years (95% CI: 2.3-3.0).

Results of conventional cytogenetic studies were available for 98/103 (95%) patients, of whom 72% had normal karyotype AML. A concomitant NPM1 mutation was found in 50/103 (48%) patients. Only 4/103 (4%) patients harbored a concomitant TET2 mutation (Table 1), in accordance with the fact that IDH1/2 and TET2 mutations tend to be mutually exclusive.10 As opposed to IDH1R132 and IDH2R140 mutations, IDH2R172 mutations are less likely to be accompanied by additional frequently recurring

Table 1. Baseline characteristics of the patients and acute myeloid leukemias.

Number of patients (%) ALFA-0701 ALFA-0702 Total Gender Male 10 Female 16 Median age (range), years 62 (51-70) Median white blood cell 18 (1-157) count (range), x 109/L Cytogenetics Normal 21 Abnormal 4 Failure 1 IDH1/2 mutation IDH1 p.R132H/C/G 10 IDH2 p.R140Q 10 IDH2 p.R172K 6 Other gene mutations NPM1 mutation 16 FLT3 internal tandem duplication 3 FLT3-tyrosine kinase domain mutation2 DNMT3A mutation 6 TET2 mutation 2 CEBPA mutation 1 (1 sm) European LeukemiaNet 2008 risk-group Favorable 13 Non-favorable 12 Not defined 1

38 39 50 (22-60) 5 (1-377)

48 (47) 55 (53) 54 (22-70) 7 (1-377)

50 23 4

71 (69) 27 (26) 5 (5)

26 36 15

36 (35) 46 (45) 21 (20)

34 50 (48) 16 19 (19) 7 9 (9) 23 29 (35) 2 4 (4) 3 (2 sm, 1 dm) 4 (4) 20 52 5

33 (32) 64 (62) 6 (6)

sm: single mutation; dm: double mutation.

Figure 2. Barcoding representing the co-occurrence of gene mutations and cytogenetic alterations in our cohort of 103 patients with IDH1/2 mutant acute myeloid leukemia. ITD: internal tandem duplication; TKD: tyrosine kinase domain.

824

haematologica | 2018; 103(5)

IDH1/2 mutant allele burden in AML

PCR in this subgroup (data not shown). The association between IDH2R172 mutation and trisomy 11 observed in our cohort is consistent with results from a previous study,19 and suggests a potential cooperation between these two genetic alterations in leukemogenesis.

IDH1/2 mutation level at diagnosis of acute myeloid leukemia and during follow-up At AML diagnosis, IDH1/2-VAF could be assessed by next-generation sequencing in 80/103 patients (Online Supplementary Table S5). The median IDH1/2-VAF value was 41% (range, 16-53%) in bone marrow and 39.5% (range, 6-50%) in peripheral blood samples. In the subset of NPM1-mutated AML, IDH1/2-VAF was systematically higher than NPM1-VAF, except in one patient with similar VAF for both mutations [n=34 comparisons; median difference IDH1/2-VAF - NPM1-VAF, 10.5% (range, 0-25%); P<0.001] (Online Supplementary Figure S2). This finding supports the notion that IDH1/2 mutations were present in pre-existing clones that subsequently acquired NPM1 mutations.

We also performed ddPCR assays in diagnostic and follow-up samples to quantify the IDH1/2-VAF. A total of 322 samples from 103 patients with IDH1/2 mutations were analyzed by ddPCR at diagnosis (n=97, of which 69 were bone marrow and 28 peripheral blood samples), during hematologic remission (n=211 bone marrow samples), and at relapse (n=14 bone marrow samples). At AML diagnosis, the median IDH1/2-VAF assessed by ddPCR was 42.3% (range, 8.2-49.9%) in bone marrow and 40.6% (range, 5.5-53%) in peripheral blood samples, consistent with our next-generation sequencing data. After induction therapy, the IDH1/2 mutant allele fraction in bone marrow samples decreased significantly compared to the pretreatment levels (P<0.001) with a median value below 0.2% (range, <0.2-39.3%). At AML relapse, the median IDH1/2VAF was 21.3% (range, 0.2-38.5%). Among the 14 patients for whom a bone marrow sample was available for molecular analysis at AML relapse, only one lost the mutation during disease evolution (Figure 3A).

Persistent clonal hematopoiesis with IDH1/2 mutations IDH1/2 mutations persisted at high levels during hematologic remission in 7/103 (7%) patients, including four

A

with an IDH2R140 mutation and three with an IDH1R132 mutation, but none with an IDH2R172 mutation. The main characteristics of these seven patients are summarized in Table 2 and their IDH1/2-VAF profiles are shown in Figure 3B. The only common characteristic identified in these patients was age over 50 years. In this subgroup, the median IDH1/2-VAF was 8% (range, 0.8-28.5%) after induction and 40% (range, 26-43.5%) after consolidation therapy. Of these seven patients, only one is still alive in first complete remission, one died from transplant-related mortality, three relapsed, and two developed overt myelodysplastic syndrome. Altogether, 5/7 (71%) patients with persistent clonal hematopoiesis with IDH1/2 mutations relapsed or progressed toward myelodysplastic syndrome within 1 to 4 years after AML diagnosis.

Univariate and multivariate prognostic analyses The prognostic impact of IDH1/2 mutations in AML remains controversial.9 In the present cohort composed exclusively of IDH1/2-mutated AML, the presence of an IDH2R172 mutation was associated with a shorter disease-free survival compared to other IDH1/2 mutation types, but without the difference reaching statistical significance (P=0.088). No difference according to the type of IDH1/2 mutation was observed regarding overall survival (Table 3; Figure 4).

The prognostic impact of IDH1/2-VAF was evaluated in complete remission after induction therapy in a subset of 95 patients for whom a post-induction bone marrow sample was available for IDH1/2-VAF assessment (Figure 1).

We were not able to perform statistical analysis at later follow-up time-points, such as post-consolidation, because of the lack of available DNA samples for many patients. Variables considered for univariate and multivariate analyses were age, white blood cell count, cytogenetics, mutational status of five genes, and IDH1/2-VAF after induction therapy. In univariate analysis for disease-free survival, the presence of a normal karyotype, a NPM1 mutation, and a IDH1/2-VAF <0.2% were significantly associated with a longer disease-free survival. In multivariate analysis, these three variables remained significantly predictive of disease-free survival. Factors significantly associated with overall survival were age, the presence of a normal karyotype, the presence of a NPM1 mutation or a TET2 mutation. Other molecular abnormalities studied,

B

Figure 3. IDH1/2 mutant allele fraction assessed by droplet digital polymerase chain reaction at diagnosis of acute myeloid leukemia and during follow-up (A) in the whole cohort and (B) for the seven patients with persistent clonal hematopoiesis with IDH1/2 mutations. The plain lines in the dot plot indicate the median values.

haematologica | 2018; 103(5)

825

Y. Ferret et al.

as well as IDH1/2-VAF, had no impact on overall survival (Table 3; Figure 5).

Discussion In this study including 103 adult patients with primary IDH1/2 mutant AML who were intensively treated, we showed the feasibility of IDH1/2-VAF monitoring using ddPCR and its prognostic relevance after induction therapy, independently of pretreatment risk factors. Our findings also suggest that patients with persistent IDH1/2mutated clonal hematopoiesis may be at high risk of dismal hematologic evolution.

The prognostic value of IDH1/2 mutations is still a matter of debate9 and may be influenced by the type of mutations, as we previously reported,20,21 or the profile of con-

comitant mutations, such as NPM1 or DNMT3A mutations.17,22 The present study, which only included patients with IDH1/2 mutations, was not designed to explore the prognostic significance of IDH1/2 mutations.

The role of MRD in the management of AML patients is growing. Because of the marked heterogeneity of AML, no single MRD marker can be applied to all patients.

Additionally, the optimal method for measuring clearance of leukemia cells after chemotherapy remains to be determined. Here, we focused on IDH1/2 mutations because they are recurrent genetic events in AML, mostly in normal karyotype AML, and now represent druggable targets.

The digital PCR technique had been previously shown to allow absolute quantification of a nucleic acid target with high precision and sensitivity.12 Our data provide evidence that measurement of IDH1/2-VAF by ddPCR is feasible.

Table 2. Clinical and biological characteristics of the seven patients with persistent clonal hematopoiesis with IDH1/2 mutations.

UPN 1

UPN 2

UPN 3

UPN 4

Age (years) 50 55 55 50 Gender F M M M 9 WBC count, x 10 /L 28 2.4 4.7 43 Cytogenetics Normal Trisomy 8 Normal Normal NPM1 mutation Pos. Neg. Pos. Neg. FLT3-ITD Pos. Neg. Neg. Pos. FLT3-TKD mutation Pos. Neg. Neg. Neg. CEBPA mutation Neg. NA Neg. Neg. DNMT3A mutation NA p.R882H (VAF 26%) NA p.R882H (VAF 48%) TET2 mutation NA Neg. NA Neg. IDH1/2 mutation IDH2 p.R140Q IDH2 p.R140Q IDH2 p.R140Q IDH2 p.R140Q (VAF at diagnosis) (44%) (43%) (39%) (47%) IDH1/2-VAF in CR 28.5% 0.76% 4.87% 28.1% after induction IDH1/2-VAF in CR 28.1% 7.2% 42.9% 39.9% after consolidation Clinical outcome Alive in CR1 2 years Relapse 4 years MDS 1 year after Relapse 1.5 year after AML after AML AML diagnosis after AML diagnosis diagnosis diagnosis

UPN 5

UPN 6

UPN 7

68 F 34 Failure Pos. Pos. Neg. Neg. Neg. Neg. IDH1 p.R132G (44%) NA

60 F 100 Normal Pos. Neg. Neg. Neg. Neg. Neg. IDH1 p.R132C (48%) 8.2%

63 F 3.2 Normal Neg. Neg. Neg. Neg. Neg. Neg. IDH1 p.R132C (43%) NA

43.5%

NA

27.1%

Death after allo-SCT

Relapse 1.5 year after AML diagnosis

MDS 2.5 years after AML diagnosis

UPN: unique patient number; F: female; M: male; WBC: white blood cell; Pos.: positive; Neg.: negative; ITD: internal tandem duplication; TKD: tyrosine kinase domain; NA: not available; VAF: variant allele fraction; CR: complete remission; AML: acute myeloid leukemia; MDS: myelodysplastic syndrome; allo-SCT: allogeneic stem cell transplantation.

Table 3. Prognostic analysis for disease-free survival and overall survival.

Disease-free survival Variable Age* Log10 (white blood cell count)* NPM1 mutation Normal karyotype FLT3 internal tandem duplication FLT3 tyrosine kinase domain mutation DNMT3A mutation TET2 mutation IDH2 p.R172K mutation IDH1/2-VAF after induction <0.2%

HR 1.04 1.00 0.23 0.26 1.11 0.20 1.42 2.66 2.04 0.32

Univariate 95% CI 0.98 0.99 0.11 0.12 0.33 0.03 0.61 0.58 0.90 0.15

1.10 1.01 0.50 0.59 3.70 1.48 3.31 12.30 4.61 0.69

P

HR

0.202 0.537 <0.001 0.001 0.865 0.115 0.413 0.209 0.088 0.004

0.32 0.41 0.85 0.40

Multivariate 95% CI 0.12 0.17 0.31 0.18

0.88 0.99 2.32 0.90

P

HR

Overall survival Univariate 95% CI

0.027 0.046 0.751 0.026

1.13 1.00 0.19 0.24 1.00 2.42 12.59 1.44 0.46

1.00 0.99 0.05 0.07 0.12 0.64 1.68 0.39 0.14

1.27 1.01 0.72 0.76 8.08 9.15 94.62 5.34 1.54

HR: hazard ratio; CI: confidence interval; VAF: variant allele fraction; *: continuous variable.

826

haematologica | 2018; 103(5)

P 0.047 0.698 0.014 0.016 1.000 0.078 0.192 0.014 0.586 0.208

IDH1/2 mutant allele burden in AML

However, despite technical optimizations, we were not able to reach the 0.01% or even 0.1% threshold that we would expect as the quantitative detection limit with the specific ddPCR assays. This problem was due to a relatively high background observed in negative controls, which always consisted of double-positive (actually false-positive) droplets. Polymerase errors occurring during the PCR amplification step seem to be responsible for the generation of these false-positive signals.

The present study is the first to quantify IDH1/2 mutation levels in a large cohort of AML patients. Previous studies using Sanger sequencing,23 qPCR,24 or next-generation sequencing technology25 suggested that the presence or the level of IDH1/2 mutations was correlated to disease status in most patients with AML, but the small number of IDH1/2-mutated patients included in these studies precluded statistical analysis. Our study revealed that a positive IDH1/2-VAF after induction chemotherapy was associated with a shorter disease-free survival. Whether patients with residual IDH1/2 mutations in complete remission may benefit from allogeneic stem cell transplantation remains to be addressed by future studies. In clinical

practice, IDH1/2-VAF assessment during and after treatment could be especially valuable in AML patients without recurrent fusion genes or NPM1 mutations, which are both leukemia-specific and more sensitive MRD markers.

Keeping in mind the caveat that IDH1/2 mutations can be present in the pre-leukemic clone in some cases, one could argue that these mutations could be good MRD markers for those patients in whom MRD becomes undetectable after induction or at early follow-up time-points.

However, sequential monitoring of IDH1/2-VAF after consolidation therapy or allogeneic stem cell transplantation could still help to detect disease persistence and guide preemptive therapy to prevent hematologic relapse, as suggested in a recent study.26 An alternative approach to MRD monitoring in IDH1/2-mutated patients is to quantify the oncometabolite 2-HG.27,28 A previous study from the ALFA group showed that total 2-HG serum levels <2 Îźmol/L after induction were associated with better disease-free survival and overall survival.29 We were not able to correlate IDH1/2-VAF and 2-HG levels in this study because of the lack of serum samples.

We found that 7/103 (7%) patients had an IDH1/2

A

B

Figure 4. Kaplan-Meier estimates of (A) disease-free survival and (B) overall survival according to the type of IDH1/2 mutation.

A

B

Figure 5. Prognostic analysis according to post-induction IDH1/2 mutant allele fraction. Kaplan-Meier estimates of (A) disease-free survival and (B) overall survival according to IDH1/2 variant allele fraction (IDH1/2-VAF). MRD+ denotes IDH1/2-VAF â&#x2030;Ľ0.2% and MRD- denotes IDH1/2-VAF <0.2%.

haematologica | 2018; 103(5)

827

Y. Ferret et al.

mutation that persisted at high levels in hematologic remission, consistent with the presence of this mutation in pre-leukemic hematopoietic stem cells. Unlike AML blasts, these hematopoietic stem cells survive chemotherapy and persist in remission bone marrow, providing a potential reservoir for leukemic progression.30 In our study, 5/7 (71%) patients with persistent clonal hematopoiesis with IDH1/2 mutations relapsed or progressed toward myelodysplastic syndrome, suggesting that these patients may be at high risk of hematologic evolution and should probably be monitored more closely. Klco et al. showed that initiating mutations, such as DNMT3A, TET2, and IDH1/2 mutations, are less likely to be cleared after chemotherapy than cooperating mutations,31 in accordance with our own and previous data.25,32 Furthermore, the prognostic value of persisting somatic mutations in complete remission appears to vary depending on the gene involved. Recent studies suggested that the presence of persistent mutations in DNMT3A, TET2 or ASXL1 lacks prognostic impact in terms of AML relapse or survival,33,34 in contrast with what we observed for IDH1/2 mutations.

Patients with IDH1/2 mutations are candidates for targeted therapies. Small-molecule inhibitors of mutant IDH1 such as ivosidenib or IDH2 such as the recently approved enasidenib are currently under clinical investigation and, when used as single agents, have shown promising results in patients with AML or myelodysplastic syndrome as a first-line treatment or in relapsed or refractory diseases.9 These molecules have been shown to induce differentiation of primary leukemic cells in vitro35,36 and in vivo37 to promote clinical responses. Future studies should determine whether patients with high levels of IDH1/2-VAF after induction therapy could benefit from a consolidation or maintenance therapy including IDH1/2 inhibitors.

Ultimately, one could imagine that the use of these small

References 1. Schnittger S, Weisser M, Schoch C, Hiddemann W, Haferlach T, Kern W. New score predicting for prognosis in PMLRARA+, AML1-ETO+, or CBFBMYH11+ acute myeloid leukemia based on quantification of fusion transcripts. Blood.

2003;102(8):2746–2755.

2. Stentoft J, Hokland P, Ostergaard M, Hasle H, Nyvold CG. Minimal residual core binding factor AMLs by real time quantitative PCR--initial response to chemotherapy predicts event free survival and close monitoring of peripheral blood unravels the kinetics of relapse. Leuk Res. 2006;30(4):389– 395.

3. Cilloni D, Renneville A, Hermitte F, et al.

Real-time quantitative polymerase chain reaction detection of minimal residual disease by standardized WT1 assay to enhance risk stratification in acute myeloid leukemia: a European LeukemiaNet study. J Clin Oncol. 2009;27(31):5195–5201.

4. Freeman SD, Virgo P, Couzens S, et al.

Prognostic relevance of treatment response measured by flow cytometric residual disease detection in older patients with acute myeloid leukemia. J Clin Oncol.

2013;31(32):4123–4131.

5. Terwijn M, van Putten WLJ, Kelder A, et al.

High prognostic impact of flow cytometric minimal residual disease detection in acute myeloid leukemia: data from the

828

6.

7.

8.

9.

10.

11.

12.

molecules might also be considered in patients with persistence of clonal hematopoiesis with IDH1/2 mutations, although clearance of the clone carrying the drug targets seems to occur only in a small subset of treated patients, even with the most potent inhibitors.38 Additionally, preclinical and clinical data indicate that IDH1/2 mutations may identify patients likely to respond to pharmacological BCL-2 inhibition.39,40 The use of IDH1/2-VAF monitoring in patients treated with an IDH1/2 or BCL-2 inhibitor, such as venetoclax, could therefore contribute to the evaluation of treatment efficacy.

In conclusion, our study is the first to show that IDH1/2 mutant allele fraction in complete remission after induction therapy significantly correlates with disease-free survival, independently of pretreatment prognostic factors.

However, this difference did not translate into distinct overall survival rates in our cohort. Our data provide evidence that IDH1/2 mutant allele fraction has the potential to become a useful tool for the management of AML patients as a biomarker of treatment response, in addition to being a molecular predictor of response to targeted therapies. Further studies based on larger cohorts of patients are required to confirm and extend our findings, and to address the question of whether the residual level of IDH1/2 mutation may help to refine the assignment into distinct risk groups and guide the decision of whether to perform allogeneic stem cell transplantation or give targeted therapies.

Acknowledgments This work was supported by the Association Laurette Fugain, the Ligue Contre le Cancer (North Center), the SIRIC ONCOLille, the North-West Canceropole (GIRCI AAP-AE 2015_53), and the Institut National du Cancer - Direction Generale de l’Offre de Soin (INCA-DGOS_9967).

HOVON/SAKK AML 42A study. J Clin Oncol. 2013;31(31):3889–3897.

Krönke J, Schlenk RF, Jensen K-O, et al.

Monitoring of minimal residual disease in NPM1-mutated acute myeloid leukemia: a study from the German-Austrian acute myeloid leukemia study group. J Clin Oncol. 2011;29(19):2709–2716.

Grimwade D, Ivey A, Huntly BJP. Molecular landscape of acute myeloid leukemia in younger adults and its clinical relevance. Blood. 2016;127(1):29–41.

Ommen HB. Monitoring minimal residual disease in acute myeloid leukaemia: a review of the current evolving strategies.

Ther Adv Hematol. 2016;7(1):3–16.

Medeiros BC, Fathi AT, DiNardo CD, Pollyea DA, Chan SM, Swords R. Isocitrate dehydrogenase mutations in myeloid malignancies. Leukemia. 2017;31(2):272–281.

Figueroa ME, Abdel-Wahab O, Lu C, et al.

Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell. 2010;18(6): 553–567.

Losman J-A, Looper RE, Koivunen P, et al.

(R)-2-hydroxyglutarate is sufficient to promote leukemogenesis and its effects are reversible. Science. 2013;339(6127):1621– 1625.

Hindson BJ, Ness KD, Masquelier DA, et al.

High-throughput droplet digital PCR system for absolute quantitation of DNA copy

13.

14.

15.

16.

17.

18.

19.

number. Anal Chem. 2011;83(22):8604– 8610.

Castaigne S, Pautas C, Terré 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.

Thomas X, de Botton S, Chevret S, et al.

Randomized phase II study of clofarabinebased consolidation for younger adults with acute myeloid leukemia in first remission. J Clin Oncol. 2017;35(11):1223-1230.

Cox D. Regression models and life tables. J R Stat Soc B. 1972;34:187-220.

Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81(3): 515–526.

Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209–2221.

Caligiuri MA, Strout MP, Schichman SA, et al. Partial tandem duplication of ALL1 as a recurrent molecular defect in acute myeloid leukemia with trisomy 11. Cancer Res. 1996;56(6):1418–1425.

Eisfeld A-K, Kohlschmidt J, Mrózek K, et al.

Adult acute myeloid leukemia with trisomy 11 as the sole abnormality is characterized by the presence of five distinct gene mutations: MLL-PTD, DNMT3A, U2AF1, FLT3-ITD and IDH2.

Leukemia.

haematologica | 2018; 103(5)

IDH1/2 mutant allele burden in AML

2016;30(11):2254–2258.

20. Boissel N, Nibourel O, Renneville A, et al.

Prognostic impact of isocitrate dehydrogenase enzyme isoforms 1 and 2 mutations in acute myeloid leukemia: a study by the Acute Leukemia French Association group.

J Clin Oncol. 2010;28(23):3717–3723.

21. Boissel N, Nibourel O, Renneville A, Huchette P, Dombret H, Preudhomme C. Differential prognosis impact of IDH2 mutations in cytogenetically normal acute myeloid leukemia. Blood. 2011;117(13): 3696–3697.

22. Patel JP, Gönen M, Figueroa ME, et al.

Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med. 2012;366(12):1079–1089.

23. Chou W-C, Peng K-Y, Lei W-C, et al.

Persistence of mutant isocitrate dehydrogenase in patients with acute myeloid leukemia in remission. Leukemia.

2012;26(3):527–529.

24. Jeziskova I, Razga F, Toskova M, et al.

Quantitative detection of IDH2 mutation for minimal residual disease monitoring in patients with acute myeloid leukemia and its comparison with mutations in NPM1 gene. Leuk Lymphoma. 2013;54(4):867–870.

25. Debarri H, Lebon D, Roumier C, et al.

IDH1/2 but not DNMT3A mutations are suitable targets for minimal residual disease monitoring in acute myeloid leukemia patients: a study by the Acute Leukemia French Association. Oncotarget. 2015;6 (39):42345–42353.

26. Brambati C, Galbiati S, Xue E, et al. Droplet digital polymerase chain reaction for DNMT3A and IDH1/2 mutations to

haematologica | 2018; 103(5)

27.

28.

29.

30.

31.

32.

33.

improve early detection of acute myeloid leukemia relapse after allogeneic hematopoietic stem cell transplantation.

Haematologica. 2016;101(4):e157-161.

Ward PS, Patel J, Wise DR, et al. The common feature of leukemia-associated IDH1 and IDH2 mutations is a neomorphic enzyme activity converting a-ketoglutarate to 2-hydroxyglutarate. Cancer Cell.

2010;17(3):225–234.

Pollyea DA, Kohrt HE, Zhang B, et al. 2hydroxyglutarate in IDH mutant acute myeloid leukemia: predicting patient responses, minimal residual disease and correlations with methylcytosine and hydroxymethylcytosine levels. Leuk Lymphoma. 2013;54(2):408–410.

Janin M, Mylonas E, Saada V, et al. Serum 2-hydroxyglutarate production in IDH1and IDH2-mutated de novo acute myeloid leukemia: a study by the Acute Leukemia French Association group. J Clin Oncol.

2014;32(4):297–305.

Shlush LI, Zandi S, Mitchell A, et al.

Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature. 2014;506(7488):328– 333.

Klco JM, Miller CA, Griffith M, et al.

Association between mutation clearance after induction therapy and outcomes in acute myeloid leukemia.

JAMA. 2015;314(8):811–822.

Pløen GG, Nederby L, Guldberg P, et al.

Persistence of DNMT3A mutations at longterm remission in adult patients with AML. Br J Haematol. 2014;167(4):478–486.

Gaidzik VI, Weber D, Paschka P, et al.

34.

35.

36.

37.

38.

39.

40.

DNMT3A mutant transcript levels persist in remission and do not predict outcome in patients with acute myeloid leukemia.

Leukemia. 2018;32(1):30-37.

Jongen-Lavrencic M, Grob T, Kavelaars FG, et al. Prospective molecular MRD detection

by NGS: a powerful independent predictor for relapse and survival in adults with newly diagnosed AML. Blood. 2017;130 (Suppl 1):LBA-5.

Chaturvedi A, Araujo Cruz MM, Jyotsana N, et al. Mutant IDH1 promotes leukemogenesis in vivo and can be specifically targeted in human AML. Blood. 2013;122(16): 2877–2887.

Wang F, Travins J, DeLaBarre B, et al.

Targeted inhibition of mutant IDH2 in leukemia cells induces cellular differentiation. Science. 2013;340(6132):622–626.

Amatangelo MD, Quek L, Shih A, et al.

Enasidenib induces acute myeloid leukemia cell differentiation to promote clinical response. Blood. 2017;130(6):732–741.

Perl AE. The role of targeted therapy in the management of patients with AML. Hematol Am Soc Hematol Educ Program.

2017;2017(1):54–65.

Chan SM, Thomas D, Corces-Zimmerman MR, et al. Isocitrate dehydrogenase 1 and 2 mutations induce BCL-2 dependence in acute myeloid leukemia. Nat Med. 2015;21(2):178–184.

Konopleva M, Pollyea DA, Potluri J, et al.

Efficacy and biological correlates of response in a phase II study of venetoclax monotherapy in patients with acute myelogenous leukemia. Cancer Discov. 2016;6 (10):1106–1117.

829

ARTICLE

Acute Lymphoblastic Leukemia

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):830-839

MSH6 haploinsufficiency at relapse contributes to the development of thiopurine resistance in pediatric B-lymphoblastic leukemia Nikki A. Evensen,1 P. Pallavi Madhusoodhan,1 Julia Meyer,2 Jason Saliba,1 Ashfiyah Chowdhury,1 David J. Araten,3 Jacob Nersting,4 Teena Bhatla,1 Tiffaney L. Vincent,5 David Teachey,5 Stephen P. Hunger,5 Jun Yang,6 Kjeld Schmiegelow4 and William L. Carroll1

Departments of Pediatrics and Pathology, Perlmutter Cancer Center, NYU-Langone Medical Center, New York, NY, USA; 2Huntsman Cancer Institute, University of Utah Medical Center, Salt Lake City, USA; 3Department of Medicine, Perlmutter Cancer Center, NYU-Langone Medical Center, New York NY, USA; 4Department of Pediatrics and Adolescent Medicine, The University Hospital Rigshospitalet, Copenhagen, Denmark; 5 Department of Pediatrics and the Center for Childhood Cancer Research, Children’s Hospital of Philadelphia and The Perelman School of Medicine at The University of Pennsylvania, Philadelphia, PA, USA and 6St. Jude Children’s Research Hospital, Memphis, TN, USA 1

ABSTRACT

S

Correspondence: william.carroll@nyumc.org

Received: July 13, 2017.

Accepted: February 7, 2018.

Pre-published: February 15, 2018. doi:10.3324/haematol.2017.176362 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/830

urvival of children with relapsed acute lymphoblastic leukemia is poor, and understanding mechanisms underlying resistance is essential to developing new therapy. Relapse-specific heterozygous deletions in MSH6, a crucial part of DNA mismatch repair, are frequently detected. Our aim was to determine whether MSH6 deletion results in a hypermutator phenotype associated with generation of secondary mutations involved in drug resistance, or if it leads to a failure to initiate apoptosis directly in response to chemotherapeutic agents. We knocked down MSH6 in mismatch repair proficient cell lines (697 and UOCB1) and showed significant increases in IC50s to 6-thioguanine and 6-mercaptopurine (697: 26- and 9-fold; UOCB1: 5- and 8-fold) in vitro, as well as increased resistance to 6-mercaptopurine treatment in vivo. No shift in IC50 was observed in deficient cells (Reh and RS4;11). 697 MSH6 knockdown resulted in increased DNA thioguanine nucleotide levels compared to non-targeted cells (3070 vs. 1722 fmol/μg DNA) with no difference observed in mismatch repair deficient cells. Loss of MSH6 did not give rise to microsatellite instability in cell lines or clinical samples, nor did it significantly increase mutation rate, but rather resulted in a defect in cell cycle arrest upon thiopurine exposure. MSH6 knockdown cells showed minimal activation of checkpoint regulator CHK1, γH2AX (DNA damage marker) and p53 levels upon treatment with thiopurines, consistent with intrinsic chemoresistance due to failure to recognize thioguanine nucleotide mismatching and initiate mismatch repair.

Aberrant MSH6 adds to the list of alterations/mutations associated with acquired resistance to purine analogs emphasizing the importance of thiopurine therapy.

©2018 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.

830

Introduction Relapsed B-precursor acute lymphoblastic leukemia (B-ALL) is a leading cause of cancer mortality amongst children. Development of chemoresistance is a crucial factor contributing to relapse, therefore understanding the biological mechanisms underlying this resistance is imperative for discovering innovative treatment strategies.1 Recent work has begun to highlight the direct role of relapse specific/enriched genetic alterations in the emergence of clones that have gained a selective advantage under the pressure of specific chemotherapeutics, such as NT5C2, TBL1XR1, PRPS1, and CREBBP.1-5 Many of these mutations cause resistance specifically to thiopurines, which are the backbone of maintenance therapy haematologica | 2018; 103(5)

MSH6 haploinsufficiency contributes to resistance

and have proven vital for achieving cures.6 Our analysis of copy number alterations (CNAs) in diagnosis/relapse pairs revealed a relapse specific hemizygous deletion on chromosome 2p16.3 involving MSH6 in 4-10% of patients.7,8 MutS homolog 6 (MSH6) is a major component of the mismatch repair (MMR) system, which is a highly conserved biological process that recognizes and repairs errors in nascent DNA strands during replication to maintain genomic integrity. Initial recognition of replicative errors is carried out by protein heterodimers consisting of either MSH6 and MSH2 (hMutSa), or MSH3 and MSH2 (hMutSβ). Upon recognition of a mismatch, hMutSa recruits MutLa (MLH1-PMS2) which engages downstream proteins and enzymes involved in DNA repair.9,10 Constitutional defects in MMR, including monoallelic mutations in Lynch syndrome and biallelic loss in constitutional mismatch repair deficiency (CMMRD), are strongly linked to carcinogenesis, where loss of MMR functionality causes increased mutability and predisposition to malignancy.11-15 Previous work has linked defects in MMR to drug resistance, including thiopurines, in various cancers.16-19 However, it is uncertain if resistance in MMR defective clones occurs through the acquisition of secondary mutations as a consequence of mutagenic therapy, or the outgrowth of clones that have intrinsic drug resistance. Our lab previously demonstrated that lower expression of MSH6 in patient samples was associated with increased ex vivo resistance to 6-mercaptopurine and prednisone,7 highlighting the clinical importance of understanding the role of this genetic alteration in B-ALL. The mechanism of action of thiopurines is based upon the insertion of a false nucleotide, namely a thioguanine (TGN), into DNA that when thiomethylated pairs with a thymine instead of a cytosine.18 Cytotoxicity is thought to be dependent on the MMR machinery recognizing the mismatch and attempting to match the TGN on the parental strand with an appropriate base on the daughter strand.19,20 Whether the DNA damage induced by the repetitive, futile cycles of DNA excision and repair, or simply the recognition of mismatches by hMutSa is enough to initiate a signaling cascade culminating in cell cycle arrest and apoptosis is not entirely understood.

We sought to delineate whether reduced expression of MSH6 could give rise to chemoresistance in B-precursor ALL and elucidate the mechanism responsible for the resistance. Our data here support the view that reduced MSH6 directly results in an increased tolerance to incorporated TGN and subsequent mismatches through a failure to initiate MMR, thus allowing cells to proliferate and survive under thiopurine treatment both in vitro and in vivo. We demonstrate that ALL cell lines with a functional MMR trigger a CHK1-mediated cell cycle arrest in response to thiopurines that is followed by DNA damage and apoptosis. In contrast, upon reduction of MSH6, the MMR signaling cascade is not fully activated and cells do not undergo apoptosis.

Methods Cells and reagents The B-lineage leukemia cell lines RS4;11 (ATCC, Manassas, VA, USA), Reh (ATCC), 697 (DSMZ, Braunschweig, Germany), and UOCB1 (a kind gift from Dr. Terzah Horton at Texas Children’s Cancer Center/Baylor College of Medicine) were grown in haematologica | 2018; 103(5)

RPMI1640 medium. HEK293T (ATCC) cells were grown in DMEM medium. All media were supplemented with 10% FBS, 1% penicillin/streptomycin under 5% CO2 at 37°C.

Drug preparation, viral preparation, immunoblotting, apoptosis assays, and cell cycle Standard protocols were followed and have been previously described.3,21 More detailed information is provided in the Online Supplementary Appendix.

Patients’ samples Cryopreserved pediatric B-ALL specimens were obtained from the Children's Oncology Group (COG) ALL cell bank. All patients were treated on COG protocols for newly diagnosed ALL. All subjects provided consent for banking and future research use of these specimens in accordance with the regulations of the institutional review boards of all participating institutions.

Microsatellite instability analysis Microsatellite instability (MSI) analysis was performed using MSI Analysis System, v.1.2 (Promega, Madison, WI, USA) following the manufacturer’s protocol. Detailed information is provided in the Online Supplementary Appendix.

Measurement of mutation rate Spontaneous mutation rate was measured using a flow cytometry assay previously described by Araten et al.22 that detects the presence of numerous glycosylphosphatidylinositol-linked (GPI) membrane proteins (see Online Supplementary Appendix). Briefly, GPI(+) isolated clones from the NT and MSH6-KD cell lines were expanded either untreated or treated with 6-TG (0.040 μg/mL and 0.100 μg/mL, respectively, based on IC50 values determined for clones). Cells were then stained for GPI-dependent markers including FLAER-Alexa 488 (Pinewood), CD48, CD52, and CD59 (Serotec),23 and analyzed by flow cytometry. The mutant frequency (f) was calculated as the number of GPI(-) events divided by the total number of live events, and mutation rate (μ) was calculated as f divided by cell divisions.22

Thioguanine quantification assay Cells were treated with 6-thioguanine (6TG) and collected every day for four days. DNA was extracted using Puregene Core Kit A (QIAGEN). DNA TGN levels were measured using liquid chromatography-tandem mass spectrometry as described previously.24

In vivo mouse model of chemoresistance All experiments were conducted on protocols approved by the Institutional Animal Care and Use Committee and Institutional Review Board of the Children's Hospital of Philadelphia. Briefly, 1 million UOCB1 NT GFP-CBG or MSH6 shRNA1 GFP-CBR cells were injected into NSG mice via tail vein on day 0 (total 20 mice; 10 per cell line). On day 6 leukemic burden was confirmed via bioluminescence imaging (BLI) (IVIS Spectrum imaging system, Perkin Elmer) and animals were randomized to treatment groups [PBS vehicle or Purixan (50 mg/kg) diluted in PBS]. Mice were treated on day 7 by gavage (0.2 mL/mouse). For BLI, 3 mg of luciferin was injected intraperitoneally and mice were imaged ten minutes post injection.

Quantification of total flux was determined by analyzing the BLI images using Living Image Software (Perkin Elmer) (see Online Supplementary Appendix).

Statistical analysis Statistical significance was calculated using unpaired t-test for IC50s, paired t-test for mutation rates, one-way ANOVA for cell 831

N.A. Evensen et al.

A

B

C

Figure 1. Knockdown of MSH6 in mismatch repair (MMR) proficient cells lead to decreased sensitivity to thiopurines. (A-C, left) Western blot analysis of whole cell lysates from 697 (A), UOCB1 (B), and Reh and RS4;11 (C). (A-C, right) Apoptotic cells measured by Annexin V and 7AAD staining followed by flow cytometry after 5 days of treatment. Graphs represent 3 experiments each performed with duplicates. Bars indicate mean+Standard Deviation.

832

haematologica | 2018; 103(5)

MSH6 haploinsufficiency contributes to resistance

cycle analysis, and two-way ANOVA for mutation rates with and without treatment as well as in vivo studies.

Results Previously we noted relapse-specific heterozygous deletions in MSH6 in 4 out of 76 patients that were near identical and deleted MSH6 only for 3 patients while one harbored a larger deletion involving more genes within the region (Online Supplementary Figure S1). To begin to elucidate the impact of MSH6 deletion on the development of relapsed disease, we knocked down expression of MSH6 using shRNA in 697 cells, a B-ALL cell line that expresses all four MMR proteins (Figure 1A) and is MMR proficient,25 and tested for changes in chemosensitivity. We observed approximately 80-90% (shRNA1) and 50-60% (shRNA2) knockdown of MSH6 expression, as well as decreased expression of MSH2, compared to non-targeting (NT) control cells (Figure 1A), which is consistent with literature on the loss of protein stability of MSH2 and MSH6 when not dimerized.17,26 Knockdown of MSH6 with both shRNA1 and shRNA2 leads to a significant decrease in apoptotic cells when treated with thiopurines for five days (Figure 1A). A 26-fold increase in IC50 with 6-TG (NT: 0.027 vs. shRNA1: 0.716 μg/mL; P=0.007) and 8.5-fold increase for 6-MP (NT: 0.340 vs. shRNA1: 2.89 μg/mL; P=0.006) was observed for shRNA1 (Online Supplementary Figure S2). A 1.7-fold (NT: 0.015 vs. shRNA2: 0.025 μg/mL; P=0.015) and a 2.6-fold (NT: 0.143 vs. shRNA2: 0.373 μg/mL; P=0.032) increase in IC50 for 6TG and 6-MP, respectively, were observed for shRNA2 cells compared to NT cells (Online Supplementary Figure

A

S2). However, no significant differences were observed when cells were treated with prednisolone (Pred), doxorubicin (Doxo), cytarabine (Ara-c), or methotrexate (MTX) (Online Supplementary Figure S3A). Interestingly, we found that knockdown of MSH6 also resulted in decreased sensitivity to temozolomide (TMZ), an alkylating agent used to treat glioblastomas, as reported previously (Online Supplementary Figure S3B).27,28 To further support the role of MSH6 in chemoresistance, we knocked down expression in UOCB1 cells, another B-ALL cell line that expresses all four MMR proteins (Figure 1B). Similar to the effect observed in 697 cells, depletion of MSH6 with either shRNA significantly reduced the induction of apoptosis upon treatment with thiopurines (Figure 1B) [fold increase in IC50 as compared to NT with 6TG: 4.8 for shRNA1 (P=0.007) and 3 for shRNA2 (P<0.001); 6MP: 8.3 for shRNA1 (P<0.001) and 9.2 shRNA2 (P<0.001)] (Online Supplementary Figure S4).

Additionally, a similar impact on TMZ resistance was observed with UOCB1 MSH6 shRNA1 expressing cells compared to NT control cells, although shRNA2 did not show the same effect, possibly due to less depletion by shRNA2 (Online Supplementary Figure S3B).

To determine the specificity of the phenotype observed for MSH6 depletion versus defects in other MMR proteins, we assessed the effect of MSH6 knockdown in MMR deficient B-ALL cell lines Reh and RS4;11.25,29 Both Reh and RS4;11 have minimal to no expression of MLH1 and PMS2 (Figure 1C). Knockdown of MSH6 expression had no effect on the sensitivity of either Reh or RS4;11 to 6-TG or 6-MP (Figure 1C and Online Supplementary Figure S5).

To begin to elucidate the mechanism of resistance, we

B

C Figure 2. Knockdown of MSH6 leads to increased tolerance of incorporated thioguanine (TGN). (A and B) TGN incorporation into DNA was measured over time after treatment with 0.1 μg/mL of 6-thioguanine (6-TG) in 697 (A) and Reh (B) cells using liquid chromatography-tandem mass spectrometry. A representative graph from 3 independent experiments is shown. (C) Combined ratio of TGN fmole/μg of DNA in MSH6 shRNA1 knockdown (KD) cells compared to non-targeting (NT) cells from 3 experiments. Bars indicate mean+Standard Deviation.

haematologica | 2018; 103(5)

833

N.A. Evensen et al.

measured the level of TGN incorporation into DNA upon treatment with 6-TG. 697 MSH6 shRNA1 cells accumulated more TGN/μg DNA over time than NT cells (NT 1722 and KD 3070 fmol/μg DNA) (Figure 2A and C). In contrast, no difference in TGN levels was observed in Reh cells (Figure 2B and C). Additionally, Reh cells had approximately 10-fold higher TGN levels compared to 697 cells (Figure 2A and B), highlighting the difference between MMR deficient and proficient cells in their ability to

respond to and survive thiopurine exposure. Thus, MMR proficient cells with high TGN succumb to the damage and therefore display less TGN/μg DNA over time, meanwhile deficient cells tolerate higher levels of TGN. We next tested whether or not a change occurs in cell cycle progression upon treatment. 697 NT cells slowed their growth and had a significantly higher proportion of cells in S phase and less cells in G1 beginning at 96 hours (h) (6-TG, P=0.014; 6-MP, P=0.051) and progressing

A

B

Figure 3. Thiopurine treatment resulted in an S phase arrest, which was abrogated upon knockdown of MSH6. 697 (A) and UOCB1 (B) NT and MSH6 shRNA1 and 2 expressing cells were treated with indicated drug for 5 days. Cells were fixed with 70% ethanol, treated with RNAse, and then stained with propidium iodide. DNA content was analyzed by flow cytometry. Representative images from 3 individual experiments are shown. A one-way ANOVA was performed to determine statistical significance of the increase in % of cells in S phase at each time point.

834

haematologica | 2018; 103(5)

MSH6 haploinsufficiency contributes to resistance

through 120 h of thiopurine treatment (6-TG, P=0.013; 6MP, P=0.001) compared to the MSH6 shRNA lines by oneway ANOVA (Figure 3A and Online Supplementary Figure S6A). MSH6 shRNA1 cells had only a modest decrease in growth with no clear S phase arrest (6-TG, P=0.011, and 6-MP, P=0.001, for percent of cells in S phase compared to NT at 120 h using Tukey’s multiple comparison test), even at higher concentrations of 6-TG (Figure 3A and Online Supplementary Figure S6A and B). MSH6 shRNA2 cells had a more moderate accumulation of cells in S phase and drop of cells in G1 (Figure 3B and Online Supplementary Figure S6A), which is consistent with the modest levels of knockdown and apoptosis. Similar trends were observed with UOCB1 cells (6-TG, P=0.31; and 6-MP, P=0.34 at 120 h) (Figure 3B). This more moderate effect observed with the UOCB1 cells is consistent with the degree of impact MSH6 knockdown had on chemoresistance compared to the 697 cells. Neither NT nor MSH6 shRNA1 Reh cells showed alterations in cell cycle upon exposure (Online Supplementary Figure S6B).

To gain a more complete understanding of the mechanism leading to apoptosis following TGN incorporation, we analyzed downstream pathways in 697 NT and MSH6 shRNA1 cells after treatment with 6-TG. Based on the observed S phase arrest and previous research demonstrating activation of the ataxia telangiectasia and Rad3-related (ATR)-Chk1 pathway downstream of MMR,30,31 we first assessed the level of activation of Chk1 by probing for phosphorylation of serine 317 (pChk1). 697 NT cells had a low level of pChk1 at 48 h with a significant increase through 96 h of exposure. 697 MSH6 shRNA1 cells had minimal to low levels of pChk1 at 72 h with minimal increase over time (Figure 4A). The 697 cells expressing shRNA2 had a similar pattern but slightly lower levels of pChk1 compared to NT cells, which is consistent with the cell cycle data. We next assessed the level of phosphorylated H2AX (γH2AX), a marker of DNA damage that is phosphorylated downstream of the ATR/ATM pathways following drug treatment,32 as well as levels of the apopto-

A

sis marker p53. There was a very modest level of γH2AX starting at 72 h that increased to a higher level at 96 h after treatment in 697 NT cells compared to very modest levels in the MSH6 shRNA1-2 cells (Figure 4A), suggesting that the functional MMR system in the NT cells was attempting to repair the DNA, leading to nicks. Additionally, the levels of phosphorylated and total p53 were higher in NT cells at 72 and 96 h compared to MSH6 shRNA1-2 cells (Figure 4A) and the shRNA2 cells had higher levels than the shRNA1 cells. Similar results were found in UOCB1 cells (Figure 4B).

We next examined the impact of MSH6 knockdown on mutation rate by performing two assays that measure genomic instability and mutation burden. Microsatellite instability (MSI) is a marker for genomic instability and has been observed in cases where expression of MLH1 or MSH2 is lost.25,33 We investigated MSI on 2 patient sample pairs that we previously found to have deletions of MSH6 at relapse, as well as on 697 NT and MSH6 shRNA1 cells treated with 6-TG for 120 h. No MSI was observed in the patient samples comparing diagnosis to relapse or in the 697 cells comparing either untreated to 6-TG treated or NT to MSH6 shRNA1 cells (Figure 5A). These data are consistent with previous literature that found alterations in MSH6 expression alone do not lead to high MSI.34 To investigate the effect of MSH6 disruption on the rate of spontaneous mutations in PIG-A, which is required for expression of GPI, we used a flow cytometry-based assay that measures surface expression of several GPI-dependent markers (CD48, CD52, and CD59).22,35 Although there was a trend to suggest that 697 MSH6 shRNA1 cells had a slightly higher mutation rate, statistical significance was not achieved (Figure 5B). Furthermore, treatment of the clones from each cell line with 6-TG did not lead to an increased mutation rate (Figure 5B).

To investigate the clinical relevance of reduced MSH6 expression and drug resistance, we utilized an in vivo mouse model. We injected mice with either UOCB1 NT or UOCB1 MSH6 shRNA1 cell lines (knockdown con-

B

Figure 4. Thiopurine treatment leads to activation of cell cycle regulator Chk1 and DNA repair that ultimately resulted in DNA damage and cell death. Western blot analysis of whole cell lysates from 697 (A) and UOCB1 (B) non-targeting (NT), MSH6 shRNA1, and shRNA2 cells after treatment with 6-thioguanine (0.1 μg/mL, and 0.025 μg/mL, respectively). (C) Untreated cells; numbers are hours after treatment. Blots were probed for Chk1 activation, γH2AX for DNA damage, and apoptosis marker p53. Total Chk1, actin, and total H2AX were used as loading controls. Images are representative of 3 individual experiments.

haematologica | 2018; 103(5)

835

N.A. Evensen et al.

firmed day of injection; Figure 6C) and, following confirmation of leukemic burden on day 6, the mice were treated with PBS (control) or purixan (an oral suspension form of 6-MP). Following the 10-day course of purixan treatment, the leukemic burden was significantly diminished in the mice harboring NT cells compared to that observed in the NT PBS treated mice (P=0.0001), suggesting that these cells were unable to survive and expand under the selective pressure of the purixan (Figure 6A and B). In contrast, the MSH6 shRNA1 mice treated with purixan were not significantly different from the NT PBS group (P=0.828). Although purixan also had a statistically significant impact on MSH6 shRNA1 cells compared to MSH6 shRNA1 PBS control (P=0.0005), these cells were able to continue proliferating under the selective pressure, unlike the NT cells (Figure 6A and B). Finally, a comparison between PBS MSH6 shRNA1 and PBS control NT cells at day 17 showed that MSH6 depleted cells also had a growth advantage in vivo (P<0.0001).

Discussion In recent years, there has been an abundance of evidence demonstrating the outgrowth of clones at relapse in ALL that are associated with unique or enriched relapse

A

specific mutations that confer drug resistance. Some of the most common relapse specific mutations found thus far occur in NT5C2 and PRPS1 and lead to the outgrowth of thiopurine resistant clones.2,4 Our data presented here demonstrate that reduction of MSH6 in ALL also leads to decreased sensitivity to purine analogs due to a failure to initiate the apoptotic cascade directly in response to nucleotide mismatches. Even with only 50-60% reduced expression, which potentially mimics levels in patients with heterozygous loss, we demonstrate a significant decrease in sensitivity to thiopurines. Our data are consistent with the recent work of Diouf et al. who showed that lower levels of MSH2 in cell lines were associated with resistance to 6-TG and 6-MP. They found 11% of ALL samples showed decreased protein levels of MSH2 through copy number loss of genes controlling MSH2 degradation.17 Thus defects in MMR, including heterozygous deletion of MSH6, can be added to the list of genetic alterations that result in the development of resistance to purine analogs, the foundation of maintenance therapy.

The variety of mutations that lead to selective outgrowth of such clones in a substantial number of patients underscores the selective pressure of thiopurines on tumor cells.

The outgrowth of MSH6 deleted/mutated clones not found at diagnosis has been observed at relapse in malignant gliomas following treatment with temozolo-

B

C

Figure 5. Knockdown of MSH6 did not lead to a mutator phenotype or increased mutation rate. (A) Microsatellite instability (MSI) was measured in diagnosis/relapse pairs that had relapse specific, heterozygous MSH6 deletions and in 697 non-targeting (NT) and MSH6 shRNA1 cells left untreated or treated with 0.05 Îźg/mL of 6-thioguanine (6-TG) for 5 days. (B) Mutation rate in the PIC-A gene was measured in 697 NT and MSH6 shRNA1 clones that were expanded for 2-3 weeks with or without 6-TG. The cells were analyzed for loss of GPI-dependent cell surface markers, including FLAER, CD48, CD52, and CD59 using flow cytometry. (Left) Individual mutation rates/cell divisions for each clone; the line represents mean+Standard Deviation. (Right) Mutation rates/cell divisions for three clones with and without 6TG treatment.

836

haematologica | 2018; 103(5)

MSH6 haploinsufficiency contributes to resistance

mide,28,36,37 which produces DNA O6-methyguanine, a lesion structurally similar to 6-TG.38 Our data demonstrating decreased sensitivity of MSH6 knockdown cells to temozolomide support the hypothesis that MMR deficient clones gain an advantage under this selective pressure leading to resistant recurrences.27,39 The difference in TMZ sensitivity between the two UOCB1 shRNA knockdown cell lines could be due to the interplay between MSH6 and SETD2 protein levels since UOCB1 cells have a copy number loss of SETD2 (NA Evensen et al., 2018, unpublished data) and there is a greater reduction of MSH6 with shRNA1 compared to shRNA2. SETD2, the gene that codes for the methyltransferase responsible for the trimethylation of H3K36 that serves as the docking site for MSH6,40 is among the epigenetic regulators commonly found mutated in relapse patients.41 Ongoing studies in our lab are focused on identifying the relationship of epigenetic readers, writers, and erasers, such as SETD2, MSH6, and WHSC1 in chemoresistance.

Mechanistically, our in vitro and in vivo data support the hypothesis that the delayed cytotoxic response to thiopurines is due to the MMR system recognizing a mismatch and initiating futile, damaging DNA repair that ultimately

leads to apoptosis.20,42,43 This pathway is not fully activated in cells with reduced MSH6 because the mismatch goes undetected, allowing these cells to tolerate excess TGN mismatches and, ultimately, to continue to survive and proliferate while under treatment. Our data provide evidence that, upon recognition of mismatches, NT cells slow their progression through S phase by activating Chk1 as they begin to repair their DNA. Due to the mismatch being on the daughter strand, the excision/repair process is unsuccessful, and over time nicks build up in the DNA, demonstrated by increased levels of ÎłH2AX. Eventually, the damage becomes overwhelming and cells initiate apoptosis, as shown by increased p53. MSH6 shRNA1 cells exhibited minimal to no change in cell cycle, activation of Chk1, or increased ÎłH2AX and p53. The moderate changes observed with the MSH6 shRNA2 cells highlight the idea that even a more modest reduction in MSH6 expression could lead to subtle changes that have a significant impact on chemoresistance. The MMR deficient Reh cells also had no alteration in their cell cycle, suggesting that recognition of mismatches by MutSa is not sufficient for full activation of this cascade, but rather damage induced by the repair, which is orchestrated by

A

B

C

Figure 6. Knockdown of MSH6 leads to decreased sensitivity to purixan in vivo. (A) Bioluminescence imaging (BLI) of mice injected with UOCB1 non-targeting (NT) or MSH6 shRNA1 cells. Six days after injection mice were imaged and then randomized to treatment. Treatment was started on day 7 and images were taken again on days 13 and 17. C: PBS control treatment; T: purixan treatment. (B) Quantification of total flux was determined by analyzing the BLI images using Living Image software. (C) Western blot to confirm knockdown of MSH6 in cells used to inject mice. Actin was used as loading control.

haematologica | 2018; 103(5)

837

N.A. Evensen et al.

MutLa, is what initiates apoptosis. Our data demonstrating the involvement of the ATR-Chk1-H2AX signaling cascade is supported by the work of Eich et al. which demonstrated activation of this pathway upon treatment with temozolomide.32 Interestingly, understanding how MMR deficient cells respond to thiopurines in terms of TGN incorporation could prove essential given the emerging idea of measuring these parameters in patients on maintenance therapy.44 The data presented here do not support the hypothesis of increased mutation burden, genomic instability, or MSI when MSH6 is reduced. Our inability to demonstrate MSI-high, which is considered a standard method for clinical testing of MMR deficiencies in tumors,45 in MSH6 depleted cell lines and clinical samples is consistent with the lack of MSI in glioma samples with MSH6 deletions or mutations.46,47 Haploinsufficiency of MSH6 or compensation by MSH2/MSH3 may account for this observation.25,48 In addition, ALL clonal evolution from diagnosis to relapse is not associated with increased mutation burden supporting our mutation rate analysis, although Ma et al. reported a subset of hypermutated relapse cases.49 Of these, one had a bialleic mutation of PMS2, another had multiple damaging MSH6 mutations as well as an MLH1 splice site mutation, while the others harbored no MMR mutations.49 Furthermore, one case demonstrated that a heterozygous deletion of MSH6 at diagnosis was not sufficient to cause a hypermutator phenotype, but the acquisition of a second hit in the WT allele at relapse was.49 Likewise, the majority of hypermutated gliomas at relapse show defects in multiple MMR genes or loss of heterozy-

References 9.

1. Pui CH, Evans WE. Treatment of acute lymphoblastic leukemia. N Engl J Med. 2006;354(2):166-178.

2. Meyer JA, Wang J, Hogan LE, et al. Relapsespecific mutations in NT5C2 in childhood acute lymphoblastic leukemia. Nat Genet.

2013;45(3):290-294.

3. Jones CL, Bhatla T, Blum R, et al. Loss of TBL1XR1 disrupts glucocorticoid receptor recruitment to chromatin and results in glucocorticoid resistance in a B-lymphoblastic leukemia model. J Biol Chem. 2014; 289(30):20502-20515.

4. Li B, Li H, Bai Y, et al. Negative feedbackdefective PRPS1 mutants drive thiopurine resistance in relapsed childhood ALL. Nat Med. 2015;21(6):563-571.

5. Mullighan CG, Zhang J, Kasper LH, et al.

CREBBP mutations in relapsed acute lymphoblastic leukaemia. Nature. 2011; 471(7337):235-239.

6. Nielsen SN, Grell K, Nersting J, et al. DNAthioguanine nucleotide concentration and relapse-free survival during maintenance therapy of childhood acute lymphoblastic leukaemia (NOPHO ALL2008): a prospective substudy of a phase 3 trial. Lancet Oncol. 2017;18(4):515-524.

7. Yang JJ, Bhojwani D, Yang W, et al.

Genome-wide copy number profiling reveals molecular evolution from diagnosis to relapse in childhood acute lymphoblastic leukemia. Blood. 2008;112(10):4178-4183.

8. Hogan LE, Meyer JA, Yang J, et al.

Integrated genomic analysis of relapsed childhood acute lymphoblastic leukemia

838

10.

11.

12.

13.

14.

15.

16.

gosity.50 Thus, our work supports a model whereby haploinsufficiency of MSH6 results in TGN tolerance and resistance directly rather than by generation of secondary mutations. However, it does not rule out the possibility that haploinsufficiency, along with other defects in the MMR pathway, may result in a mutator phenotype.

Overall, it has become increasingly evident that the genetic and epigenetic landscape of cancer cells is vital to the overall effectiveness of treatment. These studies illustrate yet another example of a mutation/deletion found at relapse that directly influences the response to a therapeutic agent that is currently heavily relied on. Continuous efforts to elucidate the potential functions and mechanisms of genes found mutated at relapse will help lead us to novel treatment strategies.

Funding This work was supported by the Leukemia and Lymphoma Society SCOR grant: 7010-14 (WLC, JY, DT, SPH), the US National Institutes of Health (NIH) funded grant RO1 CA140729 (WLC), and the Perlmutter Cancer Center Support Grant: P30 DA016087.

Acknowledgments We gratefully acknowledge the Childrenâ&#x20AC;&#x2122;s Oncology Group (COG) Specimen Bank for samples. Support for flow cytometry was provided by NYU School of Medicineâ&#x20AC;&#x2122;s Cytometry and Cell Sorting Laboratory, which is supported in part by grant P30CA016087 from the NIH/NCI, and the CHOP Flow Cytometry Core. We acknowledge the VA-Mertid award 1I01BX-000670, which helped support this work.

reveals therapeutic strategies. Blood.

2011;118(19):5218-5226.

Li GM. Mechanisms and functions of DNA mismatch repair. Cell Res. 2008;18 (1):85-98.

Edelbrock MA, Kaliyaperumal S, Williams KJ. Structural, molecular and cellular functions of MSH2 and MSH6 during DNA mismatch repair, damage signaling and other noncanonical activities. Mutat Res. 2013;743-744:53-66.

Dunlop MG, Farrington SM, Carothers AD, et al. Cancer risk associated with germline DNA mismatch repair gene mutations.

Hum Mol Genet. 1997;6(1):105-110.

Tiwari AK, Roy HK, Lynch HT. Lynch syndrome in the 21st century: clinical perspectives. QJM. 2016;109(3):151-158.

Goecke T, Schulmann K, Engel C, et al.

Genotype-phenotype comparison of German MLH1 and MSH2 mutation carriers clinically affected with Lynch syndrome: a report by the German HNPCC Consortium. J Clin Oncol. 2006; 24(26):4285-4292.

Ripperger T, Schlegelberger B. Acute lymphoblastic leukemia and lymphoma in the context of constitutional mismatch repair deficiency syndrome. Eur J Med Genet.

2016;59(3):133-142.

Ripperger T, Beger C, Rahner N, et al.

Constitutional mismatch repair deficiency and childhood leukemia/lymphoma-report on a novel biallelic MSH6 mutation.

Haematologica. 2010;95(5):841-844.

Fink D, Aebi S, Howell SB. The role of DNA mismatch repair in drug resistance.

Clin Cancer Res. 1998;4(1):1-6.

17. Diouf B, Cheng Q, Krynetskaia NF, et al.

Somatic deletions of genes regulating MSH2 protein stability cause DNA mismatch repair deficiency and drug resistance in human leukemia cells. Nat Med. 2011;17(10):1298-1303.

18. Swann PF, Waters TR, Moulton DC, et al.

Role of postreplicative DNA mismatch repair in the cytotoxic action of thioguanine. Science. 1996;273(5278):1109-1111.

19. Waters TR, Swann PF. Cytotoxic mechanism of 6-thioguanine: hMutSalpha, the human mismatch binding heterodimer, binds to DNA containing S6-methylthioguanine. Biochemistry. 1997;36(9):25012506.

20. Karran P, Attard N. Thiopurines in current medical practice: molecular mechanisms and contributions to therapy-related cancer.

Nat Rev Cancer. 2008;8(1):24-36.

21. Morrison DJ, Hogan LE, Condos G, et al.

Endogenous knockdown of survivin improves chemotherapeutic response in ALL models. Leukemia. 2012;26(2):271279.

22. Araten DJ, Golde DW, Zhang RH, et al. A quantitative measurement of the human somatic mutation rate. Cancer Res. 2005;65(18):8111-8117.

23. Araten DJ, Sanders KJ, Anscher D, Zamechek L, Hunger SP, Ibrahim S. Leukemic blasts with the paroxysmal nocturnal hemoglobinuria phenotype in children with acute lymphoblastic leukemia.

Am J Pathol. 2012;181(5):1862-1869.

24. Jacobsen JH, Schmiegelow K, Nersting J. Liquid chromatography-tandem mass spectrometry quantification of 6-thioguanine in

haematologica | 2018; 103(5)

MSH6 haploinsufficiency contributes to resistance

25.

26.

27.

28.

29.

30.

31.

32.

DNA using endogenous guanine as internal standard. J Chromatogr B Analyt Technol Biomed Life Sci. 2012;881-882:115-118.

Gu L, Cline-Brown B, Zhang F, Qiu L, Li GM. Mismatch repair deficiency in hematological malignancies with microsatellite instability. Oncogene. 2002;21(37):57585764.

Marra G, Iaccarino I, Lettieri T, Roscilli G, Delmastro P, Jiricny J. Mismatch repair deficiency associated with overexpression of the MSH3 gene. Proc Natl Acad Sci USA. 1998;95(15):8568-8573.

McFaline-Figueroa JL, Braun CJ, Stanciu M, et al. Minor Changes in Expression of the Mismatch Repair Protein MSH2 Exert a Major Impact on Glioblastoma Response to Temozolomide. Cancer Res. 2015; 75(15):3127-3138.

Xie C, Sheng H, Zhang N, Li S, Wei X, Zheng X. Association of MSH6 mutation with glioma susceptibility, drug resistance and progression. Mol Clin Oncol. 2016; 5(2):236-240.

Matheson EC, Hall AG. Assessment of mismatch repair function in leukaemic cell lines and blasts from children with acute lymphoblastic leukaemia. Carcinogenesis.

2003;24(1):31-38.

Yoshioka K, Yoshioka Y, Hsieh P. ATR kinase activation mediated by MutSalpha and MutLalpha in response to cytotoxic O6-methylguanine adducts. Mol Cell.

2006;22(4):501-510.

Yan T, Desai AB, Jacobberger JW, Sramkoski RM, Loh T, Kinsella TJ. CHK1 and CHK2 are differentially involved in mismatch repair-mediated 6-thioguanineinduced cell cycle checkpoint responses.

Mol Cancer Ther. 2004;3(9):1147-1157.

Eich M, Roos WP, Nikolova T, Kaina B. Contribution of ATM and ATR to the resistance of glioblastoma and malignant melanoma cells to the methylating anticancer drug temozolomide. Mol Cancer Ther. 2013;12(11):2529-2540.

haematologica | 2018; 103(5)

33. Verma L, Kane MF, Brassett C, et al.

Mononucleotide microsatellite instability and germline MSH6 mutation analysis in early onset colorectal cancer. J Med Genet.

1999;36(9):678-682.

34. Bodo S, Colas C, Buhard O, et al. Diagnosis of Constitutional Mismatch RepairDeficiency Syndrome Based on Microsatellite Instability and Lymphocyte Tolerance to Methylating Agents.

Gastroenterology. 2015;149(4):1017-1029. e1013.

35. Araten DJ, Krejci O, Ditata K, et al. The rate of spontaneous mutations in human myeloid cells. Mutat Res. 2013;749(1-2):4957.

36. Cahill DP, Codd PJ, Batchelor TT, Curry WT, Louis DN. MSH6 inactivation and emergent temozolomide resistance in human glioblastomas. Clin Neurosurg.

2008;55:165-171.

37. Cahill DP, Levine KK, Betensky RA, et al.

Loss of the mismatch repair protein MSH6 in human glioblastomas is associated with tumor progression during temozolomide treatment. Clin Cancer Res. 2007; 13(7):2038-2045.

38. Zhang J, Stevens MF, Laughton CA, Madhusudan S, Bradshaw TD. Acquired resistance to temozolomide in glioma cell lines: molecular mechanisms and potential translational applications. Oncology. 2010; 78(2):103-114.

39. Hunter C, Smith R, Cahill DP, et al. A hypermutation phenotype and somatic MSH6 mutations in recurrent human malignant gliomas after alkylator chemotherapy. Cancer Res. 2006; 66(8):3987-3991.

40. Li F, Mao G, Tong D, et al. The histone mark H3K36me3 regulates human DNA mismatch repair through its interaction with MutSalpha. Cell. 2013;153(3):590-600.

41. Mar BG, Bullinger LB, McLean KM, et al.

Mutations in epigenetic regulators including SETD2 are gained during relapse in pae-

42.

43.

44.

45.

46.

47.

48.

49.

50.

diatric acute lymphoblastic leukaemia. Nat Commun. 2014;5:3469.

Yan T, Berry SE, Desai AB, Kinsella TJ. DNA mismatch repair (MMR) mediates 6thioguanine genotoxicity by introducing single-strand breaks to signal a G2-M arrest in MMR-proficient RKO cells. Clin Cancer Res. 2003;9(6):2327-2334.

Tidd DM, Paterson AR. Distinction between inhibition of purine nucleotide synthesis and the delayed cytotoxic reaction of 6-mercaptopurine. Cancer Res. 1974;34(4):733-737.

Nielsen SN, Grell K, Nersting J, Frandsen TL, Hjalgrim LL, Schmiegelow K. Measures of 6-mercaptopurine and methotrexate maintenance therapy intensity in childhood acute lymphoblastic leukemia.

Cancer Chemother Pharmacol. 2016; 78(5):983-994.

Buza N, Ziai J, Hui P. Mismatch repair deficiency testing in clinical practice. Expert Rev Mol Diagn. 2016;16(5):591-604.

Yip S, Miao J, Cahill DP, et al. MSH6 mutations arise in glioblastomas during temozolomide therapy and mediate temozolomide resistance. Clin Cancer Res. 2009;15(14):4622-4629.

Maxwell JA, Johnson SP, McLendon RE, et al. Mismatch repair deficiency does not mediate clinical resistance to temozolomide in malignant glioma. Clin Cancer Res. 2008;14(15):4859-4868.

Umar A, Risinger JI, Glaab WE, Tindall KR, Barrett JC, Kunkel TA. Functional overlap in mismatch repair by human MSH3 and MSH6. Genetics. 1998;148(4):1637-1646.

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. van Thuijl HF, Mazor T, Johnson BE, et al.

Evolution of DNA repair defects during malignant progression of low-grade gliomas after temozolomide treatment.

Acta Neuropathol. 2015;129(4):597-607.

839

ARTICLE

Hodgkin Lymphoma

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):840-848

A phase II study of the oral JAK1/JAK2 inhibitor ruxolitinib in advanced relapsed/refractory Hodgkin lymphoma

Eric Van Den Neste,1 Marc André,2 Thomas Gastinne,3 Aspasia Stamatoullas,4 Corinne Haioun,5 Amine Belhabri,6 Oumedaly Reman, 7Olivier Casasnovas,8 Hervé Ghesquieres,9 Gregor Verhoef,10 Marie-José Claessen,11 Hélène A. Poirel,12 Marie-Christine Copin,13 Romain Dubois,13 Peter Vandenberghe,14 Ioanna-Andrea Stoian,15 Anne S. Cottereau,16 Sarah Bailly,1 Laurent Knoops17 and Franck Morschhauser18

Department of Hematology, Cliniques Universitaires Saint-Luc, UCL Brussels, Belgium; Hematology Department, CHU UCL Namur, Yvoir, Belgium; 3Hematology, CHU Nantes, France; 4Clinical Hematology, Centre Henri Becquerel, Rouen, France; 5Lymphoid Malignancies Unit, AP-HP, Groupe Hospitalier Mondor, Créteil, France; 6Onco-hematology, Centre Leon Berard, University Claude Bernard Lyon 1, France; 7Hematology, Centre Hospitalier Universitaire, Caen, France; 8Hematology Department, Hopital Le Bocage, CHU Dijon, France; 9Hospices Civils de Lyon, Université Claude Bernard, Centre Hospitalier LyonSud, Pierre Bénite, France; 10Department of Hematology, University Hospitals Leuven, Belgium; 11Erasmus MC, Rotterdam, the Netherlands; 12Center for Human Genetics, Cliniques universitaires Saint-Luc, Université Catholique de Louvain, Brussels, Belgium; 13 CHRU de Lille, France; 14Center for Human Genetics, Katholieke Universiteit - Leuven, Belgium; 15Nuclear Medicine, Cliniques Universitaires Saint-Luc, UCL Brussels, Belgium; 16 Nuclear Medicine, Hôpital Tenon, Paris, France; 17Cliniques Universitaires Saint-Luc and de Duve Institute, Université Catholique de Louvain, Brussels, Belgium and 18CHU Lille, Hematology Department, and Université de Lille, GRITA, France 1 2

Preliminary results were presented at the 58th Annual Meeting of the American Society of Hematology, held on December 3 – 6, 2016, in San Diego, USA.

ABSTRACT

Correspondence: franck.morschhauser@chru-lille.fr

Received: September 12, 2017.

Accepted: January 10, 2018.

Pre-published: January 19, 2018.

doi:10.3324/haematol.2017.180554 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/840 ©2018 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.

840

J

AK2 constitutive activation/overexpression is common in classical Hodgkin lymphoma, and several cytokines stimulate Hodgkin lymphoma cells by recognizing JAK1-/JAK2-bound receptors. JAK blockade may thus be therapeutically beneficial in Hodgkin lymphoma. In this phase II study we assessed the safety and efficacy of ruxolitinib, an oral JAK1/2 inhibitor, in patients with relapsed/refractory Hodgkin lymphoma. The primary objective was overall response rate according to the International Harmonization Project 2007 criteria. Thirty-three patients with advanced disease (median number of prior lines of treatment: 5; refractory: 82%) were included; nine (27.3%) received at least six cycles of ruxolitinib and six (18.2%) received more than six cycles. The overall response rate after six cycles was 9.4% (3/32 patients). All three responders had partial responses; another 11 patients had transient stable disease. Best overall response rate was 18.8% (6/32 patients). Rapid alleviation of B-symptoms was common. The median duration of response was 7.7 months, median progression-free survival 3.5 months (95% CI: 1.94.6), and the median overall survival 27.1 months (95% CI: 14.4-27.1).

Forty adverse events were reported in 14/33 patients (42.4%). One event led to treatment discontinuation, while 87.5% of patients recovered without sequelae. Twenty-five adverse events were grade 3 or higher.

These events were mostly anemia (n=11), all considered related to ruxolitinib. Other main causes of grade 3 or higher adverse events included lymphopenia and infections. Of note, no cases of grade 4 neutropenia or thrombocytopenia were observed. Ruxolitinib shows signs of activity, albeit short-lived, beyond a simple anti-inflammatory effect. Its limited toxicity suggests that it has the potential to be combined with other therapeutic modalities. ClinicalTrials.gov: NCT01877005

haematologica | 2018; 103(5)

Ruxolitinib in advanced relapsed/refractory HL

Introduction

Methods

Hodgkin lymphoma (HL) is regarded as a curable malignancy in most cases, yet treatment failure still occurs in about 10% of patients with early-stage disease.1 In advanced-stage disease, up to 10% of cases do not reach complete remission and are thus considered to have primary refractory HL,2 while 20-30% of primary responders eventually relapse following first-line treatment.3 For most patients with relapsed or refractory HL (R/R HL), the standard of care consists of high-dose salvage chemotherapy followed by autologous stem-cell transplantation (SCT). For patients who develop R/R HL within 1 year of autologous SCT, the prognosis proves extremely poor, since they have a median survival of 1.2 years.4 For patients in whom all classical approaches have failed, new strategies, including checkpoint inhibitors targeting PD-1 and antibody-drug conjugates targeting CD30, have become part of the therapeutic armamentarium against R/R HL.5-8 However, patients with multiple relapses or those who develop refractory disease remain in medical need, especially those in whom treatment with brentuximab-vedotin (BV) and PD-1 blockers fails.

Classical HL is characterized by the presence of Hodgkin and Reed-Sternberg (HRS) cells and their variants.9 HRS cells were demonstrated to shape their environment by secreting immunosuppressive cytokines and chemokines.10 With this in mind, the Janus kinase (JAK) – signal transducer and activator of transcription (STAT) pathway appears to be a relevant cytokine-induced signal transduction pathway that has been shown to transfer signals directly from cell surface cytokine receptors to the cell nucleus. Given that enhanced JAK-mediated signaling has been demonstrated in a significant number of HL patients,11 this signaling pathway has become a focus for developing novel therapeutic agents for the disease. Van Roosbroeck et al. reported the translocation of JAK2 in several cases of HL,12 and JAK inhibition was shown to decrease the proliferation of cell lines. Although such translocations are relatively rare, 9p24.1 genomic amplification including the JAK2 locus appears common in HL, along with increased protein expression and activity, resulting in the constitutive activation of STAT6, an essential messenger of tumor cell growth.13-15 In corollary, JAK 1/2 inhibition may be suitable to target the constitutive activation caused by either JAK2 translocation or JAK2 amplification and to modify the reactive microenvironment which contributes to HL growth via aberrant cytokine production.16 Ruxolitinib is the first potent, selective, and oral inhibitor of JAK1/2 being developed for clinical use.17 Its major effects include inhibition of proliferation, induction of apoptosis, and reduction in cytokine plasma levels, all mediated by the drug's ability to inhibit JAK-induced phosphorylation of STAT.18 Used in the treatment of myelofibrosis, ruxolitinib had durable efficacy in reducing splenomegaly and alleviating constitutional symptoms, the patients gained weight and their general physical condition improved.19 The dose-limiting toxicity was thrombocytopenia, which was fairly well managed by dose reductions or brief interruptions of treatment. In the present phase II study, we sought to investigate the safety and efficacy of ruxolitinib in patients with R/R HL. Exploratory biomarker analyses pertaining to plasma cytokine profiles and aberrations of JAK2 were also carried out.

Patients’ eligibility

haematologica | 2018; 103(5)

Patients aged 18 years or older with a diagnosis of R/R HL for whom no treatment with proven efficacy was available were eligible to enter the trial after having receiving at least one prior therapy provided that they had measurable nodal disease at baseline (≥1 cm in the longest transverse diameter, clearly measurable in at least two perpendicular dimensions) on computed tomography or magnetic resonance imaging, as well as an Eastern Cooperative Oncology Group performance score of ≤3.

Additional inclusion criteria were an absolute neutrophil count ≥1.0 x 109/L, platelet count ≥75 x 109/L, serum creatinine ≤1.5 x upper limit of normal, serum bilirubin ≤1.5 x upper limit of normal, and ALT and AST levels ≤2.5 or ≤5.0 x upper limit of normal in the event the transaminase increase was due to HL-related liver disease. Pregnant or lactating patients were not allowed to enter the trial, and men and women of childbearing potential had to agree to employ an adequate contraceptive method during the study treatment. Patients were permitted to have received an undefined number of prior lines of therapy, and a previous allogeneic SCT was likewise allowed provided that patients had not received any immunosuppressive therapy within the 90 days prior to starting the screening procedures. Patients were required to have a life expectancy of ≥3 months.

Study design and treatment This multicenter, open-label, phase II study (HIJAK, NCT01877005) was conducted at ten LYSA centers in France and Belgium, with patients recruited from July 2013 through December 2014. Its primary efficacy endpoint was overall response rate (ORR), defined as the proportion of patients with a complete response or partial response at 6 months of treatment by investigator assessment based on the revised 2007 International Harmonization Project response criteria for malignant lymphoma.20 Secondary objectives included relief of B symptoms, best ORR (occurring at any time during study), duration of response, progression-free survival, overall survival, as well as the incidence and severity of adverse events.

The study was carried out in line with the ethical principles of the Helsinki Declaration and in compliance with the International Conference on Harmonization Guideline for Good Clinical Practice. The protocol was approved by the institutional review board of each study site and written informed consent was obtained from all patients.

The starting dose of ruxolitinib was 20 mg given twice daily during six 28-day cycles for the induction period if the platelet count was >200 x 109/L. The ruxolitinib dose was decreased to 15 mg twice daily in patients with platelet counts between 75 x 109/L and 200 x 109/L. Patients who achieved at least stable disease at the end of cycle 6 and who had, in the investigator's opinion, a clinical benefit were eligible to continue ruxolitinib (15 mg or 20 mg), which was defined as “maintenance” therapy. Treatment could be continued for up to 2 years or until progressive disease, intolerability, or as long as the investigator thought that there was clinical benefit.

Administration of the study drug could be stopped for any grade ≥3 non-hematologic toxicity, with the exception of deep venous thrombosis and alopecia. Following event resolution to grade ≤1, ruxolitinib could be resumed, with a 5 mg dose reduction and a maximum delay of 4 weeks. Mandatory dose decreases or interruptions for hematologic toxicity as well as the rules for permanent discontinuation are detailed in the Online Supplementary Appendix. Growth factors were allowed as per American Society of Clinical Oncology guidelines and infectious prophylaxis as per the guidelines of Heine et al.21 841

E. Van Den Neste et al.

Study assessments Baseline assessments comprised documentation of diseaserelated symptoms, physical examination, laboratory tests, and imaging studies of the neck, chest, abdomen, and pelvis, using computed tomography or magnetic resonance imaging. Biopsy prior to inclusion was recommended, but not mandatory. Tumors were measured at baseline, at the end of every two cycles of ruxolitinib, and following the six-cycle induction, as well as during maintenance therapy. Given the exploratory nature of the study, there was no centralized review of computed tomography response. However, positron emission tomographic images of the responders were all centrally reviewed by a nuclear physician (ASC) to confirm partial or complete metabolic response based on the Deauville five-point scale. The evaluable study population for efficacy was restricted to patients who had received at least 28 days of the study drug.

Safety was monitored for up to 1 month after treatment.

Adverse events were summarized by means of the Medical Dictionary for Regulatory Activities, and graded using the National Cancer Institute’s Common Terminology Criteria for Adverse Events (NCI-CTCAE), version 3.0. Laboratory abnormalities were assessed according to NCI-CTCAE version 4.0. Only grade 3 or 4 toxicities and grade 2 infections were to be reported.

All patients were included in the toxicity analysis.

safety analysis comprised all patients who received at least one dose of the study drug. All statistical analyses were performed using SAS software, version 9.2. P-values <0.05 were considered statistically significant. All available data were included in data listings and tabulations, with no imputations of values for missing data. An interim analysis was neither planned nor performed.

Results Patients’ disposition and characteristics The patients’ characteristics are listed in Table 1. From July 2013 to December 2014, a total of 33 patients with R/R HL were recruited. Their median age was 37 years (range, 19-80). Most of the patients had advanced HL (stage III/IV) and had been heavily pretreated, with a median number of five prior regimens including autologous SCT (54%), allogeneic SCT(15%), and BV (82%).

Of the 33 patients recruited, 27 (82%) had refractory HL and 22 had biopsy-confirmed relapse of HL. Among the six patients displaying a response, a biopsy was performed in five of them at relapse [8 days, 12 days, 6 weeks (n=2) and 14 months prior to inclusion in the study].

Exploratory biomarker analysis Blood samples (5 mL) were taken at baseline prior to drug administration and on day 1 of cycle 2 for the measurement of 27 cytokines related to the immune system using bead-based immunoassays. JAK2 gains, amplifications, and gene rearrangements were also investigated using fluorescent in situ hybridization with two tri-color sets of probes associating JAK2/9p24 break-apart probes with a control centromeric probe (CEP9/9q21): the already prepared probes from Empire genomics on the one hand, and the association of the JAK2 B/A probe from Kreatech with the CEP9 probe from Vysis on the other hand. The CD274/PDL1 and PDCD1LG2/PDL2 loci at 9p24 were studied with home-made prepared bacterial artificial chromosome probes purchased from the Chori BACPAC Resources Center (Oakland, CA, USA). Extraction, labeling and hybridization were performed on paraffin-embedded tissue, as previously reported.22

Statistical methods The sample size for this phase II study was calculated using an exact single-stage phase II design.23 A two-stage design with interim analysis for activity or toxicity was not planned given the very advanced stage of the patients, the relative paucity of alternative options, and the potential toxicity of ruxolitinib that was expected to be in the low range, based on myelofibrosis data. The treatment was considered ineffective if the ORR was ≤15%, and effective if the ORR was ≥35%. Under the assumption of an alpha first-order risk error set at 5% and beta at 20% with a one-sided test, it was deemed necessary to include a total of 28 evaluable patients with a cut-off number of eight. If at least eight patients had a response, the hypothesis of an ORR ≤15% was rejected with both a target error rate and an actual error rate of 0.05. If seven or fewer patients had a response, the hypothesis of an ORR ≥35% was rejected with a target error rate of 0.2 and an actual error rate of 0.187. The ORR estimate and its 90% confidence intervals (CI) were calculated for all patients who completed at least one cycle of the study drug.

The Kaplan-Meier method was employed to estimate the median value and its 95% CI for time to response, duration of response, progression-free survival and overall survival. The 842

Table 1. Patient’s demographics and characteristics.

Patients’ demographics and characteristics

All patients (n=33)

Gender, n (%) Male 21 (63.6%) Female 12 (36.4%) Age in years, median (range) 37.0 (19.0-80.0) ECOG score 0 11 (33.3%) 1 15 (45.5%) 2 5 (15.2%) 3 2 (6.1%) Ann Arbor stage I 1 (3.0%) II 7 (21.2%) III 3 (9.1%) IV 22 (66.7%) B symptoms Yes 16 (48.5%) No 17 (51.5%) Extranodal involvement Bone 13 (39.4%) Liver 6 (18.2%) Lung 12 (36.4%) Soft tissues 4 (12.1%) Time since initial diagnosis in months, median (range) 55.4 (8.7 – 216.1) Prior therapies Prior lines, median (range) 5 (1 – 16) Chemotherapy 33 (100%) Radiotherapy 18 (54.5%) Brentuximab vedotin 27 (82%) Autologous SCT 18 (54.5%) Allogeneic SCT 5 (15.2%) Interval since last treatment in months, median (range) 6 (1.1 – 75.0) Disease status at inclusion Relapse 6 Refractory 27 (81.8%) SCT: stem cell transplantation; ECOG: Eastern Cooperative Oncology Group.

haematologica | 2018; 103(5)

Ruxolitinib in advanced relapsed/refractory HL

Patients’ exposure to the study drug The median number of ruxolitinib cycles administered was four (range, 1 to 12) (Table 2). Nine patients received all six of the planned cycles of ruxolitinib and six of these patients continued on maintenance therapy with the JAK inhibitor. The remainder discontinued ruxolitinib therapy, most because of progressive disease and in one case due to adverse events.

Responses and outcomes The patients’ disposition through the study is illustrated in Figure 1. Among the 33 HL patients included in the trial, one patient did not complete the first cycle of treatment because of progressive disease and was not, therefore, included in the efficacy analysis. At the end of the ruxolitinib induction period (6 months) three of 32 patients had a response, for an ORR of 9.4% (90% CI: 2.6-22.5%); the response in all three was partial. At some point during induction six of the 32 patients had a response, which was, in all six cases a partial response, for a best ORR of 18.8% (95% CI: 7.2-36.4%). A detailed analysis of the responders’ characteristics is provided in Table 3. Figures 2 and 3 illustrate metabolic evolution in two patients.

Interestingly, UPN 611001, who had achieved a partial

response after six cycles of treatment, eventually entered complete remission during the follow-up, beyond the six cycles. Achievement of complete metabolic response was confirmed by central review. At the time of writing, two patients (UPN 611001 and 881001) are still taking ruxolitinib. Figure 4 illustrates changes in target tumor measurements in individual patients. The best reduction, if any, at any time throughout treatment is shown.

In addition, during the 6-month induction, transient stable disease was recorded in 11 patients, albeit of limited duration. Overall, the disease control rate (including stable disease with complete and partial responses) was 53.1% (17/32 patients) (95% CI: 34.7-70.9%) with a median duration of 1.9 months.

The alleviating effect on systemic symptoms, such as pruritus, fever, and sweating, was noteworthy, starting within the first month of drug administration and commonly lasting. The impact was most remarkable on the

Table 2. Treatment exposure and modifications.

Treatment exposure and modifications

All patients (n=33)

Cycles given, number, median (range) Received full induction (6 cycles), % Received maintenance (> 6 cycles), % Percentage of planned dosea <75% 75%-90% 90%-110% 110%-125% Dose modification Yes Type of modificationb Dose reduction Dose interruption Dose increase Number of days of interruption if any, median (range)

4 (1-22) 9 (27.3%) 6 (18.2%) 3 (9.1%) 4 (12.1%) 25 (75.8%) 1 (3.0%) 20 (60.6%) 2 (10.0%) 18 (90.0%) 3 (15.0%) 4 (1 – 28)

Defined as follows: (total number of tablets taken/total expected number of tablets) *100, taking into account protocol-defined dose reduction. bThe total sum of the percentages for the type and the reasons of modification may be greater than 100.0% as a patient may have had several types of modification and reasons for treatment modification. a

Figure 1. Patients’ disposition. *N months on maintenance therapy: 4, 6, 6, 21, 16, 22; PD: progressive disease.

Table 3. Characteristics of responders (best response achieved during the 6-month ruxolitinib induction).

UPN 611001 211004 601001 601004 641001 881001

Prior treatment N

Type

9 8 5 5 1 5

ABVD, BEACOPP, MINE, IGEV, GVD, CAELYX, GVD, RT, BV ABVD, RT, IVA, transplantation, MINE, GVD, BV, ASHAP BEACOPP, DHAP, IGEV, transplantation-RT, BV ABVD, DHAP, RT, RT, BV ABVD2 ABVD, transplantation, MINE, BV, GEMOX

Extranodal involvement

Response (Cheson 2007)20

Liver Breast Liver, bone, lung None None Lung

PR*,1 PR PR PR PR PR1

*Patient eventually achieved a complete response during maintenance therapy. 1Patients still under treatment with ruxolitinib at the time of writing; 2Patient with morbid obesity not eligible for standard approaches with chemo/immunotherapy. UPN:unique patient number; RT: radiotherapy; BV: brentuximab vedotin; PR: partial response.

haematologica | 2018; 103(5)

843

E. Van Den Neste et al.

control of pruritus, which affected 35.5% of patients prior to initiating therapy but only 6.6% after the first cycle of ruxolitinib. Sweating, which was present in 32.2% of the patients at inclusion, was reduced, with 20% still having this symptom after one cycle of treatment. Fever was abolished in 3/4 patients with this symptom at inclusion.

The median follow-up was 17.5 months. As illustrated in Figure 5, the median progression-free survival was 3.5 months (95% CI: 1.9-4.6). The median duration of response was 7.7 months (95% CI: 1.8-NA) for the six patients who eventually a achieved response (data not shown). Overall, 30 patients had progressive disease, with 97% at the initial site and/or 60% at new sites. Following ruxolitinib discontinuation, 25 (83.3%) patients were given further treatments, consisting of chemotherapy in 19, and immunotherapy in nine, the latter comprising rituximab in four, BV in three, and nivolumab in two.

Transplantation was eventually carried out in five patients, and was allogeneic in four cases and autologous in the remaining one. Among the 25 patients prescribed further therapy, the observed complete and partial response rates were 10% and 15%, respectively. Overall, 12 patients died on account of lymphoma progression (83.3%), toxicity of other treatments (8.3%), or other reasons (8.3%). The median overall survival was 27.1 months (95% CI: 14.4-27.1).

A

Safety All patients enrolled in the study received at least one dose of study medication and were, therefore, included in the safety analysis. Overall, 40 adverse events were observed in 14/33 patients (42.4%). In six patients, the adverse events were related to the ruxolitinib therapy. In eight of them, the events were of grade ≥3 (Table 4A).

Among the 40 adverse events recorded, 30 (75%) occurred during induction and 18 (45%) were related to ruxolitinib.

No drug-related deaths were recorded. One adverse event resulted in permanent drug discontinuation, while 87.5% of the adverse events resolved without sequelae. The characteristics and grade of the adverse events, listed by system organ class and preferred terms, are displayed in Table 4B. Twenty-five (62.5%) were of grade ≥3. These were mostly anemia (n=11), all considered related to ruxolitinib. Other main causes of grade ≥3 adverse events included lymphopenia (n=4), infections (n=3) and miscellaneous causes. Of note, no cases of grade 4 neutropenia or thrombocytopenia were observed.

Eight serious adverse events were reported in four patients, during induction (n=5), maintenance (n=1) or after the end of treatment (n=2). These serious adverse events consisted of infection in three patients (devicerelated sepsis, gastroenteritis, and lung infection). The other serious adverse events were anemia, diarrhea, subdural hematoma, bone pain, and pulmonary embolism.

Two serious adverse events (anemia, lung infection) were deemed drug-related and six were considered grade ≥3: infection (n=3), anemia, subdural hematoma, and pulmonary embolism. Of the eight serious adverse events, six resolved without sequelae, while the device-related sepsis and pulmonary embolism, observed in the same patient, persisted until the patient died due to progressive disease and were thus not considered as the cause of death.

Among the 33 patients, one second primary malignancy was observed (adenocarcinoma of the colon in an 80-year old male patient).

Biomarker analysis Using bead-based immunoassays, plasma levels of 27 cytokines related to the immune system were measured at

B

Figure 2. Response after ruxolitinib. Illustrative patient (UPN 601004). (A) Positron emission tomography (PET)-computer tomography (CT) frontal view. (B) PET-CT sagittal view. Partial response with allievation of B symptoms and blood inflammation was achieved 2 months after starting ruxolitinib. At month 6, the patient had slowly progressive disease but refused to stop ruxolitinib. CRP: Creactive protein.

844

Figure 3. Response after ruxolitinib. Illustrative patient (UPN 601001).

Comparison of frontal positron emission tomography (PET)-scan prior to inclusion and after 2 months of ruxolitinib. There was a rapid improvement of constitutional symptoms after a few days on ruxolitinib. PET after 2 months showed metabolic partial response with a total volume reduction of tumor lung lesions of 64%.

haematologica | 2018; 103(5)

Ruxolitinib in advanced relapsed/refractory HL

baseline and after the first cycle of treatment. At baseline, there was no difference in cytokine levels between responders and non-responders. In responders, the only cytokine that decreased significantly was CX-CL10 (P=0.01). In patients presenting with pruritus (n=11), the levels of platelet-derived growth factor-BB (PDGF-BB) (Online Supplementary Appendix), interleukin (IL)-5, IL-10, IL-12, IL-13, IL-17, eotaxin, fibroblast growth factor basic (FGF basic), macrophage inflammatory protein 1b (MIP1b), regulated on activation, normal T-cell expressed and secreted (RANTES), and vascular endothelial growth factor (VEGF) were significantly increased. In the latter patients, ruxolitinib treatment significantly decreased the levels of PDGF-BB, IL-10, IL-12, IL-13, IL-17, FGF basic and VEGF. Among the patients who could be analyzed for JAK2 amplification in HRS cells (n=12), polysomy (suggesting hyperdiploidy) was detected in all of them, and specific JAK2 amplification in only one. This latter patient

achieved a partial response as determined by computed tomography criteria and also a positron emission tomography-determined response lasting 4 months. It is noteworthy that the PDL1 and PDL2 loci (which are in the vicinity of the JAK2 locus at 9p24), analyzed by fluorescent in situ hybridization with bacterial artificial chromosome probes, showed the same pattern of gains as for the JAK2 locus.

Discussion JAK/STAT activation, driven by an aberrant network of cytokines and chemokines in the HL microenvironment, is critical for the proliferation and survival of neoplastic HRS cells.24,25 The JAK/STAT pathway also plays a role in immune evasion by HL cells via the secretion of chemokines leading to Th2 homing or via the regulation

Figure 4. Waterfall plot demonstrating percent change from baseline in target tumor dimensions (best response, n=27). Of note, among the 32 patients evaluable for disease response, five had no end-of-treatment SPD measurements by computer tomography (CT) scan as planned by protocol because there were obvious signs of disease progression. *Persisting positive positron emission tomography scan, considered as partial response.

Figure 5. Kaplan-Meier estimate of progression-free survival in 32 evaluable patients with Hodgkin lymphoma receiving ruxolitinib.

haematologica | 2018; 103(5)

845

E. Van Den Neste et al.

of PD-L1/L2 expression, which confers an immune privilege to HRS cells. Chromosome 9p24.1/PD-L1/PD-L2 alterations increase the abundance of the PD-1 ligands, PD-L1 and PD-L2, and their further induction through JAK/STAT signaling.26-28 This complex crosstalk between malignant HRS cells and the reactive microenvironment could be targeted to overcome chemoresistance. Based on this rationale, we explored JAK 1/2 inhibition in a phase II study of fixed dose ruxolitinib in patients with advanced HL patients before the onset of the era of PD-1 blockers. With an ORR of 9.4% at the end of the 6-month induction period, this study did not reach its primary efficacy goal. Nevertheless, when including transient responses seen before the 6-month evaluation, the ORR was 18.8% in some heavily pretreated patients, most of whom were refractory and had failed treatment with BV. These responses were sometimes durable (median=7.7 months). Some other patients had disease control, but with uncertain clinical benefit. A notable finding to be highlighted was the relief of B symptoms and pruritus, which was quick and long-lasting, resulting in a number of patients being reluctant to discontinue the compound, despite progressive disease. The latter effect should not be interpreted as a proven surrogate of anti-lymphoma activity.

These results tend to lend some support to the concept of JAK1/2 inhibition as a potential therapeutic means in HL. There are presently only scarce data available on the use of ruxolitinib in HL. In a preliminary report of an ongoing study, Kim et al. described rapid achievement of disease control (1 complete response, 5 partial responses, 1 stable disease) in 13 patients with advanced HL treated with ruxolitinib at a dose of 20 mg bid.29 Younes et al. reported changes in tumor measurements in HL patients treated in a phase I study with SB1518, an inhibitor of JAK2 and FLT-3.30 In vitro, AZD1480, an inhibitor of JAK1 and JAK2, could regulate proliferation in HL cell lines.27 The multikinase inhibitor lestaurtinib also inhibited growth and increased apoptosis of HL cell lines and HL cells from lymph nodes.31 Finally, a clinical grade JAK2 inhibitor, fedratinib, inhibited the proliferation of classical HL cell lines in a JAK2 copy number-dependent manner implying decreased phosphorylation of STAT and expression of downstream targets including PD-L1 showing immunomodulation by JAK inhibitors.32 If JAK2 is actually an appropriate target, questions arise as to why the study outcome was not more convincing.

Could the drug's limited activity be attributed to insufficient dosage? Given that we observed unambiguous cytokine profile changes and frequent improvements in B symptoms, it would seem that the dosage of 20 mg twice daily, a dosage at which target inhibition occurs in myelofibrosis,33 was appropriate. Another factor possibly influencing the outcome was our patientsâ&#x20AC;&#x2122; disease stage, represented by a high percentage of refractoriness. At this late stage, the genetic changes would be so complex that selective inhibition of JAK is insufficient in cells dependent on other signaling pathways to promote their survival, thus further curbing the study's potential. It is known that genomic aberrations, such as chromosome breakpoints, are more numerous in later clinical stages of HL.34 Mechanisms of resistance to JAK/STAT inhibition have been reported such as a feedback loop of paradoxically activated extracellular signal-regulated kinases 1 and 2 (ERK1/2).27 Aberrations of the 9p24.1 amplicon, which 846

Table 4. Treatment-emergent adverse events.

A. Patients with an adverse event.

Treatment-emergent adverse events1

All patients (n=33)

Patients with > 1 AE N. of AE/patient, median (range) N. of patients with AE > grade 3 Patients with AE related to ruxolitinib Patients with AE leading to drug discontinuation Patients with AE leading to death

14 (42.4%) 2 (1-11) 8 (24.2%) 6 (18.2%) 1 (3%) 0 (0%)

AE: adverse event. 1Total number of AE, 40.

B. Characteristics of the adverse event (N = 40) by system organ class and preferred terms

Adverse event Any adverse event Infections and infestations Anemia Lymphopenia Thrombocytopenia Weight decrease Respiratory and thoracic disorders Diarrhea Infuenza-like illness Subdural hematoma Bone pain Epilepsy

Any grade

Grade 2

Grades > 3

40 13 11 4 2 1 3 1 1 1 1 2

15 10 0 0 0 0 2 1 1 0 1 0

25 31 11 4 2 1 1 0 0 1 0 2

Implantable device infection, gastro-enteritis, lung infection.

1

contains the JAK2 gene, are more frequent in advanced disease.28 Surprisingly, in our patients, a low incidence of JAK2 amplification was seen, suggesting a low proportion of patients harboring the target of ruxolitinib, although this inference should be considered with caution since not all patients could be analyzed.

With respect to safety, ruxolitinib was by and large well-tolerated, with no drug-related mortality reported.

The most prominent toxicities included drug-related anemia and manageable infectious events with no specific pattern. The relative lack of hematologic toxicity suggests that it could be feasible to combine ruxolitinib treatment with genotoxic compounds. For patients who discontinued ruxolitinib therapy, a switch to chemotherapy and/or immunotherapy was feasible, suggesting that the compound does not jeopardize further treatment.

The question now remains as to how this compound can best be utilized in the future. The exploratory nature of our study did not allow identification of the best candidates on the basis of clinical stage or biomarkers. The cytokine profile showed some changes in patients with pruritus, but these changes were not correlated with clinical response. Although JAK2 status was explored in a minority of patients, the only patient with JAK2 amplification achieved a response. It will be important to focus on biomarker results in ongoing studies of JAK inhibition in HL. Given ruxolitinibâ&#x20AC;&#x2122;s limited benefits as monotherapy, use in combination with other drugs may possibly enhance its therapeutic potential. Ruxolitinib, which has no overlapping toxicity with chemotherapy, has been combined with hypomethylating agents, lenalidomide, haematologica | 2018; 103(5)

Ruxolitinib in advanced relapsed/refractory HL

and even intensive chemotherapy.35-38 In vitro data have shown that ruxolitinib could restore the sensitivity of cisplatin-resistant cell lines with higher Jak2 expression.39 Interestingly, the combination of BV with ruxolitinib resulted in additive and synergistic killing in a xenograft mouse model of HL through a mechanism involving mitochondrial control of apoptosis.40 Another means to boost ruxolitinibâ&#x20AC;&#x2122;s potential would be to combine it with agents blocking other signaling pathways. Interestingly, the combination of ruxolitinib with a Bcl2/Bcl-xL inhibitor displayed dramatic synergy in an adult T-cell leukemia cell line via a mechanism implying BAX activation.41 Finally, the effect of combining chemical JAK blockade and an anti-PD1/L1 strategy should be analyzed in HL, keeping in mind, however, that a potential antagonism may be encountered due to these two drugs acting on the same target, given that PD1-L1 expression is dependent on JAK2 activity.

In conclusion, based on a strong biological rationale for clinical evaluation of JAK2 blockade in HL, we initiated a

References 1. Meyer RM, Hoppe RT. Point/counterpoint: early-stage Hodgkin lymphoma and the role of radiation therapy. Hematology Am Soc Hematol Educ Program. 2012;2012:313-321.

2. Horning SJ. Primary refractory Hodgkin's disease. Ann Oncol. 1998; 9 (Suppl 5): S97101.

3. Kuruvilla J, Keating A, Crump M. How I treat relapsed and refractory Hodgkin lymphoma. Blood. 2011;117(16):4208-4217.

4. Arai S, Fanale M, DeVos S, et al. Defining a Hodgkin lymphoma population for novel therapeutics after relapse from autologous hematopoietic cell transplant. Leuk Lymphoma. 2013;54(11):2531-2533.

5. Ansell SM. Hodgkin lymphoma: MOPP chemotherapy to PD-1 blockade and beyond. Am J Hematol. 2016;91(1):109-112.

6. Armand P, Shipp MA, Ribrag V, et al.

Programmed death-1 blockade with pembrolizumab in patients with classical hodgkin lymphoma after brentuximab vedotin failure. J Clin Oncol. 2016 Jun 27.

[Epub ahead of print].

7. Younes A, Bartlett NL, Leonard JP, et al.

Brentuximab vedotin (SGN-35) for relapsed CD30-positive lymphomas. N Engl J Med. 2010;363(19):1812-1821.

8. Moskowitz C. Novel agents and strategies in transplant-eligible patients with relapsed and refractory Hodgkin lymphoma.

Hematology Am Soc Hematol Educ Program. 2016;2016(1):331-338.

9. Slovak ML, Bedell V, Hsu YH, et al.

Molecular karyotypes of Hodgkin and ReedSternberg cells at disease onset reveal distinct copy number alterations in chemosensitive versus refractory Hodgkin lymphoma.

Clin Cancer Res. 2011;17(10):3443-3454.

10. Aldinucci D, Gloghini A, Pinto A, De Filippi R, Carbone A. The classical Hodgkin's lymphoma microenvironment and its role in promoting tumour growth and immune escape. J Pathol. 2010;221(3):248-263.

11. Navarro A, Diaz T, Martinez A, et al.

Regulation of JAK2 by miR-135a: prognostic impact in classic Hodgkin lymphoma.

haematologica | 2018; 103(5)

phase II study of ruxolitinib in R/R HL patients. The study failed to fulfill the efficacy criteria for further development of the drug as monotherapy. Nonetheless, in patients with very advanced disease ruxolitinib showed hints of activity that surpassed solely an anti-inflammatory effect. This may suggest that further improvements will come from a more complete inhibition of signaling pathways involved in HRS cell survival or from combination with chemotherapy, such as BV. Acknowledgments The authors would like to thank the patients, their families and their caregivers who made this study possible. We also thank all the study investigators and study staff at each of the clinical sites.

For the LYSARC, we acknowledge the project manager and all members of the data monitoring committee. We express our thanks to LoĂŻc Chartier and Sami Boussetta, biostatisticians at the LYSARC, who contributed to the statistical design and analysis of the study. Editorial assistance was provided by Cremer Consulting.

Blood. 2009;114(14):2945-2951.

12. Van Roosbroeck K, Cox L, Tousseyn T, et al.

JAK2 rearrangements, including the novel SEC31A-JAK2 fusion, are recurrent in classical Hodgkin lymphoma. Blood. 2011;117 (15):4056-4064.

13. Meier C, Hoeller S, Bourgau C, et al.

Recurrent numerical aberrations of JAK2 and deregulation of the JAK2-STAT cascade in lymphomas. Mod Pathol. 2009;22(3):476487.

14. Hartmann S, Martin-Subero JI, Gesk S, et al.

Detection of genomic imbalances in microdissected Hodgkin and ReedSternberg cells of classical Hodgkin's lymphoma by array-based comparative genomic hybridization. Haematologica. 2008;93(9): 1318-1326.

15. Joos S, Granzow M, Holtgreve-Grez H, et al.

Hodgkin's lymphoma cell lines are characterized by frequent aberrations on chromosomes 2p and 9p including REL and JAK2.

Int J Cancer. 2003;103(4):489-495.

16. Aldinucci D, Celegato M, Casagrande N. Microenvironmental interactions in classical Hodgkin lymphoma and their role in promoting tumor growth, immune escape and drug resistance. Cancer Lett. 2016;380(1): 243-252.

17. Assi R, Verstovsek S, Daver N. 'JAK-ing' up the treatment of primary myelofibrosis: building better combination strategies. Curr Opin Hematol. 2017;24(2):115-124.

18. Vannucchi AM, Harrison CN. Emerging treatments for classical myeloproliferative neoplasms. Blood. 2017;129(6):693-703.

19. Massaro F, Molica M, Breccia M. How ruxolitinib modified the outcome in myelofibrosis: focus on overall survival, allele burden reduction and fibrosis changes. Expert Rev Hematol. 2017;10(2):155-159.

20. Cheson BD, Pfistner B, Juweid ME, et al.

Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579-586.

21. Heine A, Brossart P, Wolf D. Ruxolitinib is a potent immunosuppressive compound: is it time for anti-infective prophylaxis? Blood.

2013;122(23):3843-3844.

22. Duhoux FP, Ameye G, Lambot V, et al.

Refinement of 1p36 alterations not involv-

23.

24.

25.

26.

27.

28.

29.

30.

31.

ing PRDM16 in myeloid and lymphoid malignancies. PLoS One. 2011;6(10):e26311.

A'Hern RP. Sample size tables for exact single-stage phase II designs. Stat Med. 2001;20(6):859-866.

Aldinucci D, Pinto A, Gloghini A, Carbone A. Chemokine receptors as therapeutic tools in Hodgkin lymphoma: CCR4 and beyond.

Blood. 2010;115(3):746-747; author reply 748.

Carbone A, Gloghini A, Castagna L, Santoro A, Carlo-Stella C. Primary refractory and early-relapsed Hodgkin's lymphoma: strategies for therapeutic targeting based on the tumour microenvironment. J Pathol.

2015;237(1):4-13.

Green MR, Monti S, Rodig SJ, et al.

Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma.

Blood. 2010;116(17):3268-3277.

Derenzini E, Lemoine M, Buglio D, et al.

The JAK inhibitor AZD1480 regulates proliferation and immunity in Hodgkin lymphoma. Blood Cancer J. 2011;1(12):e46.

Roemer MG, Advani RH, Ligon AH, et al.

PD-L1 and PD-L2 genetic alterations define classical Hodgkin lymphoma and predict outcome. J Clin Oncol. 2016;34(23):26902697.

Kim SJ, Kang HJ, Dong-Yeop S, et al. The efficacy of JAK2 inhibitor in heavily pretreated classical Hodgkin lymphoma: a prospective pilot study of ruxolitinib in relapsed or refractory classical Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood. 2016;128(22): 1820.

Younes A, Romaguera J, Fanale M, et al.

Phase I study of a novel oral Janus kinase 2 inhibitor, SB1518, in patients with relapsed lymphoma: evidence of clinical and biologic activity in multiple lymphoma subtypes. J Clin Oncol. 2012;30(33):4161-4167.

Diaz T, Navarro A, Ferrer G, et al.

Lestaurtinib inhibition of the Jak/STAT signaling pathway in Hodgkin lymphoma inhibits proliferation and induces apoptosis.

847

E. Van Den Neste et al.

PLoS One. 2011;6(4):e18856.

32. Hao Y, Chapuy B, Monti S, Sun HH, Rodig SJ, Shipp MA. Selective JAK2 inhibition specifically decreases Hodgkin lymphoma and mediastinal large B-cell lymphoma growth in vitro and in vivo. Clin Cancer Res. 2014;20(10):2674-2683.

33. Cervantes F, Vannucchi AM, Kiladjian JJ, et al. Three-year efficacy, safety, and survival findings from COMFORT-II, a phase 3 study comparing ruxolitinib with best available therapy for myelofibrosis. Blood.

2013;122(25):4047-4053.

34. Falzetti D, Crescenzi B, Matteuci C, et al.

Genomic instability and recurrent breakpoints are main cytogenetic findings in Hodgkin's disease. Haematologica. 1999;84 (4):298-305.

35. Naqvi K, Daver N, Pemmaraju N, et al.

848

Clinical use of ruxolitinib in an academic medical center in unselected patients with myeloproliferative neoplasms not on clinical study. Leuk Lymphoma. 2017;58(4): 866-871.

36. Daver N, Cortes J, Newberry K, et al.

Ruxolitinib in combination with lenalidomide as therapy for patients with myelofibrosis. Haematologica. 2015;100(8):10581063.

37. Devillier R, Raffoux E, Rey J, et al.

Combination therapy with ruxolitinib plus intensive treatment strategy is feasible in patients with blast-phase myeloproliferative neoplasms. Br J Haematol. 2016;172(4):628630.

38. Mayfield JR, Czuchlewski DR, Gale JM, et al. Integration of ruxolitinib into dose-intensified therapy targeted against a novel JAK2

F694L mutation in B-precursor acute lymphoblastic leukemia. Pediatr Blood Cancer.

2017;64(5).

39. Hu Y, Hong Y, Xu Y, Liu P, Guo DH, Chen Y. Inhibition of the JAK/STAT pathway with ruxolitinib overcomes cisplatin resistance in non-small-cell lung cancer NSCLC. Apoptosis. 2014;19(11):1627-1636.

40. Ju W, Zhang M, Wilson KM, et al.

Augmented efficacy of brentuximab vedotin combined with ruxolitinib and/or Navitoclax in a murine model of human Hodgkin's lymphoma. Proc Natl Acad Sci USA. 2016;113(6):1624-1629.

41. Zhang M, Mathews Griner LA, Ju W, et al.

Selective targeting of JAK/STAT signaling is potentiated by Bcl-xL blockade in IL-2dependent adult T-cell leukemia. Proc Natl Acad Sci USA. 2015;112(40):12480-12485.

haematologica | 2018; 103(5)

ARTICLE

Non-Hodgkin Lymphoma

A B-cell receptor-related gene signature predicts survival in mantle cell lymphoma: results from the Fondazione Italiana Linfomi MCL-0208 trial

Riccardo Bomben,1* Simone Ferrero,2,3* Tiziana D'Agaro,1 Michele Dal Bo,1 Alessandro Re,4 Andrea Evangelista,5 Angelo Michele Carella,6 Alberto Zamò,7 Umberto Vitolo,8 Paola Omedè,3 Chiara Rusconi,9 Luca Arcaini,10 Luigi Rigacci,11 Stefano Luminari,12,13 Andrea Piccin,14 Delong Liu,15 Adrian Wiestner,15 Gianluca Gaidano,16 Sergio Cortelazzo,17 Marco Ladetto2,18 and Valter Gattei1**

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):849-856

Clinical and Experimental Onco-Hematology Unit, Centro di Riferimento Oncologico, IRCCS, Aviano (PN), Italy; 2Department of Molecular Biotechnologies and Health Sciences, Hematology Division 1, University of Torino, Italy; 3Hematology Division 1, AOU “Città della Salute e della Scienza di Torino” University-Hospital, Italy; 4Hematology, AO “Spedali Civili di Brescia”, Italy; 5Unit of Cancer Epidemiology, AOU “Città della Salute e della Scienza di Torino” University-Hospital, Italy; 6Hematology Division 1, IRCCS AOU San Martino IST, Genova, Italy; 7Department of Diagnostics and Public Health, University of Verona, Italy; 8Hematology Division 1, AOU “Città della Salute e della Scienza di Torino” University-Hospital, Italy; 9Hematology Division, “Niguarda Ca’ Granda” Hospital, Milano, Italy; 10Hematology Division, Department of Molecular Medicine, IRCCS Fondazione San Matteo, Pavia, Italy; 11Hematology Division, AOU “Careggi”, University of Firenze, Italy; 12Hematology, Azienda Sanitaria Locale IRCCS, Reggio Emilia, Italy; 13 Department of Diagnostic, Clinical and Public Health Medicine, University of Modena and Reggio Emilia, Reggio Emilia, Italy; 14Department of Hematology and BMT Unit, Bolzano/Bozen Regional Hospital, Italy; 15Hematology Branch, National Heart, Lung, and Blood Institute, NIH, Bethesda, MD, USA; 16Division of Haematology, Department of Translational Medicine –University of Eastern Piedmont, Novara, Italy; 17Hematology, Medical Oncology and Hematology Division, “Istituto Clinico Humanitas Gavazzeni”, Bergamo, Italy and 18SC Ematologia Azienda Ospedaliera Nazionale SS. Antonio e Biagio e Cesare Arrigo, Alessandria, Italy 1

*RB and SF contributed equally to this work as first authors: **ML and VG contributed equally to this work as senior authors

Correspondence: ABSTRACT

M

antle cell lymphoma patients have variable clinical courses, ranging from indolent cases that do not require immediate treatment to aggressive, rapidly progressing diseases. Thus, diagnostic tools capable of stratifying patients according to their risk of relapse and death are needed. This study included 83 samples from the Fondazione Italiana Linfomi MCL-0208 clinical trial. Through gene expression profiling and quantitative real-time PCR we analyzed 46 peripheral blood and 43 formalin-fixed paraffin-embedded lymph node samples. A prediction model to classify patients was developed. By analyzing the transcriptome of 27 peripheral blood samples, two subgroups characterized by a differential expression of genes from the B-cell receptor pathway (B-cell receptorlow and B-cell receptorhigh) were identified.

The prediction model based on the quantitative real-time PCR values of six representative genes (AKT3, BCL2, BTK, CD79B, PIK3CD, and SYK), was used to classify the 83 cases (43 B-cell receptorlow and 40 B-cell receptorhigh). The B-cell receptorhigh signature associated with shorter progression-free survival (P=0.0074), selected the mantle cell lymphoma subgroup with the shortest progression-free survival and overall survival (P=0.0014 and P=0.029, respectively) in combination with high (>30%) Ki-67 staining, and was an independent predictor of short progressionfree survival along with the Mantle Cell Lymphoma International Prognostic Index-combined score. Moreover, the clinical impact of the 6gene signature related to the B-cell receptor pathway identified a mantle cell lymphoma subset with shorter progression-free survival intervals also in an external independent mantle cell lymphoma cohort homogenously treated with different schedules. In conclusion, this 6-gene signature associates with a poor clinical response in the context of the MCL0208 clinical trial. (clinicaltrials.gov identifier: 02354313).

Haematologica | 2018; 103(5)

rbomben@cro.it

Received: November 10, 2017.

Accepted: February 14, 2018.

Pre-published: February 22, 2018. doi:10.3324/haematol.2017.184325 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/849 ©2018 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.

849

R. Bomben et al.

Introduction Mantle cell lymphoma (MCL) is a distinctive B-cell malignancy accounting for 5-10% of all lymphomas,1-3 whose molecular hallmark and initiating oncogenic event, the t(11;14)(q13;q32) translocation, leads to constitutive overexpression of the proto-oncogene cyclin D1 (CCND1).2,4 Once considered as uniformly characterized by a poor prognosis, MCL has been demonstrated to have unexpectedly variable clinical courses, ranging from indolent cases that do not require immediate treatment to aggressive, rapidly progressing disease.2,5-10 Even among patients requiring treatment, prognosis is highly heterogeneous, with patients experiencing prolonged remissions and others rapidly relapsing even after cytarabine-containing induction regimens followed by autologous transplantation. Thus, diagnostic tools capable of stratifying MCL patients in different risk classes are warranted in order to direct treatment strategies.11 For this reason, many attempts have been made to identify clinical, histological, and molecular markers that can stratify patients according to their risk of relapse and death.12-25 In addition to the clinical MCL prognostic score (MCLInternational Prognostic Index, MIPI)12,14 capable of stratifying patients into risk groups with different overall survival (OS),14 the Ki-67 proliferation index has been proposed as one of the most powerful and independent predictors of survival in MCL even in the context of prospective trials and modern therapies,5,13,26 and for these reasons has been integrated into the so-called MIPI-combined (MIPI-c) score.13,26 Moreover, effective prognostic discrimination is achieved by post-treatment response monitoring by positron emission tomography (PET)-scan and minimal residual disease (MRD). Furthermore, a seminal study identified a specific signature associated with proliferation

Table 1. Characteristics of 83 mantle cell lymphoma cases entering the study.

Samples Number of cases Median age, years (range) Ratio male/female (%) Abnormal LDH (%) Median WBC (x109/L) Typical morphology Blastoid morphology Median Proliferation Index (Ki-67 staining), % MIPI-c class Low Low/intermediate High/intermediate High na Median survival, months (range) Median progression-free survival, months (range)

83 56 (28-65) 57/26 (68) 45/29 (39) 13.7 74 5 20.0 (0-99) 38 (46%) 23 (28%) 11 (13%) 6 (7%) 5 (6%) 34.7 (1.4-73.4) 31.3 (1.4-73.4)

LDH: lactate dehydrogenase; WBC: white blood count; MIPI-c: MCL:-International Prognostic Index-combined; na: not available.

850

as the strongest predictor of OS in a large MCL series.20 In this context, a cohort of 20 proliferation-associated genes constructed on the basis of gene expression analysis was demonstrated to be superior to other molecular markers.20 Since approaches based on microarray technology have not yet been incorporated into routine clinical practice, a PCR-based surrogate method investigating expression of five genes has been proposed and applied to paraffinembedded tissues.18 Recent evidence suggests that the B-cell receptor (BCR) pathway may contribute to the pathogenesis of several histological types of B-cell non-Hodgkin lymphomas, including MCL.27-30 The importance of BCR signaling pathway in B-cell malignancy pathogenesis has driven interest in the use of small-molecule inhibitors of BCR-associated kinases, potentially preventing the activation of one or more of the distal BCR signaling pathway proteins.28,31 In the present study, we developed a survival predictive model for younger patients with advanced MCL treated in the context of the Fondazione Italiana Linfomi (FIL) MCL0208 Phase III randomized clinical trial. This model is based upon the quantitative evaluation of six genes, mostly from the BCR pathway, selected from a gene expression profile (GEP) of peripheral blood (PB) MCL cells and was applied to formalin-fixed paraffin-embedded (FFPE) tissue specimens. Notably, the model predicts poor response in the context of the FIL-MCL-0208 trial.

Methods Primary MCL cases The study included 83 out of 300 samples of adult patients under 66 years of age with advanced stage MCL, enrolled in the FIL-MCL-0208 prospective, multicenter, Phase III randomized clinical trial (clinicaltrials.gov identifier: 02354313),32 divided as follows: i) a panel of 27 PB samples utilized for GEP upon positive sorting of the clonal CD5+/CD19+ MCL cells; ii) an additional panel of 19 PB samples utilized for quantitative real-time PCR (qRT-PCR) of the identified gene signature in the purified MCL cell component; iii) a panel of 43 lymph node (LN) samples utilized for qRT-PCR of the identified gene signature (in this LN panel 6 samples had a matched PB sample). The clinical and histopathological details of the 83 MCL cases used in this study are reported in Table 1. No significant differences were found between the 83 cases entering the study versus the 217 remaining cases enrolled in the clinical trial in terms of median age, MIPI score, Ki-67 index and PFS intervals (Online Supplementary Table S1 and Online Supplementary Figure S1). No differences in clinical and biological parameters were observed between PB and LN MCL samples (data not shown). All patients were treated according to the FIL-MCL-0208 clinical trial, as reported in Online Supplementary Figure S2.

Mantle cell lymphoma diagnosis was prospectively confirmed by centralized histological review according to the World Health Organization (WHO) 2008 criteria.3,33 All patients provided informed consent in accordance with Institutional Review Board requirements (0016331-BZ 09/02/2010) and the Declaration of Helsinki, and protocol consent included use of MRD sample leftovers for the study.

All the procedures employed for RNA extraction, GEP and downstream analyses, qRT-PCR, analyses and qRT-PCR validations were carried out according to standard protocols, as reported previously.34-37 (See the Online Supplementary Appendix for details). Microarray data are available in Gene Expression haematologica | 2018; 103(5)

B-cell receptor signature in mantle cell lymphoma

Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE89447. Cases used for these procedures are reported in Online Supplementary Table S2.

2017.32 Investigators are still blinded to the investigation arm as the primary study end point has still not been met.

Results Validation procedures The 6-gene signature was tested in the MCL cohort described

by Saba et al.,30 enrolled in another clinical trial (clinicaltrials.gov identifier: 00114738), by using the sum of the array gene expression values, as reported.30 Gene signatures related to MCL outcome were retrieved from previous papers,30,38 and imported in the GeneSpring GX and tested in the present cohort with GEP data available.

Statistical analysis Overall survival was computed from trial registration to death as a result of any cause, censored at the latest follow up in patients who were still alive. Progression-free survival (PFS) was computed from trial registration to progression or death as a result of any cause, censored at the latest tumor assessment if no progression was observed. Clinical correlations, performed with the MedCalc v.9.5 software, were made using Kaplan-Meier plots and log-rank test. The Cox proportional model was chosen for multivariable analysis. Clinical outcome results were up-dated as of January

A

GEP identifies MCL patients with distinct expression of genes belonging to the BCR pathway Global GEP was performed in purified MCL cells from 27 PB samples. An unsupervised analysis performed by principal component analysis (PCA) divided the cohort into two groups of 14 cases and 13 cases, respectively (Figure 1A). Consistently, a hierarchical clustering, which was run with all the GEP features, split MCL cases into two major groups perfectly resembling the PCA groups (Figure 1B).

Supervised analysis according to the PCA classification defined a gene expression signature composed of 922 probes, 713 up-regulated and 209 down-regulated in group-2 versus group-1 samples (Figure 1C and Online Supplementary Table S3).

Pathway analysis revealed that “Antigen processing and presentation” and “B-cell receptor signaling pathway” were among the top ranked pathways enriched in the

B

C

Figure 1. Gene Expression Profile (GEP) analysis of 27 mantle cell lymphoma samples. (A) Principal Component Analysis (PCA) scores represented in a 3D scatter plot. One point per array/sample is shown. Black line indicates separation between PCA classes. (B) Hierarchical clustering of 14 group-1 cases and 13 group-2 cases, using 50,739 probes. (C) Hierarchical clustering of 14 group1 cases and 13 group2 cases, using the 922 differentially expressed probes. Color codes for gene expression values refer to mean centered log-ratio values.

haematologica | 2018; 103(5)

851

R. Bomben et al.

group-2 category (Online Supplementary Table S4). Similar results were obtained by GSEA which highlighted a constitutive overexpression of genes related to the BCR signaling pathways in the context of group-2 patients (Figure 2A and Online Supplementary Table S5). Therefore, hereafter the two PCA groups were identified as BCRlow (group-1) and BCRhigh (group-2).

A

A 6-gene signature identifies BCRlow and BCRhigh MCL samples Having identified two different groups of MCL patients at diagnosis with a different expression of genes related to the BCR pathway, we overlapped the genes included in the gene sets related to the BCR pathway (115 probes) and the differentially expressed genes (922 probes) to cre-

B

C

D

E

Figure 2. 6-gene signature and Decision Tree (DT) prediction model. (A) Gene Expression Profile data of BCRlow and BCRhigh MCL samples were tested using Gene set enrichment analysis (GSEA). Reported are the significant gene sets differentially expressed and related to the B-Cell Receptor (BCR) pathway. (B) Venn diagram derived by merging the differentially expressed probes and the genes belonging to the BCR related gene sets. In bold genes selected as the 6-gene signature. (C) Hierarchical clustering of 14 BCRlow cases and 13 BCRhigh cases, using the six gene values. Color codes for gene expression values refer to mean centered log-ratio values. (D) Hierarchical clustering of 8 BCRlow cases and 9 BCRhigh cases belonging to the training set of DT prediction model, using the six gene qRT-PCR values. (E) Hierarchical clustering of 6 BCRlow cases and 4 BCRhigh cases belonging to the validation set of DT prediction model, using the six gene qRT-PCR values. Bar under the heat-map refers to prediction generated by the DT prediction model. Color codes for gene expression values refer to mean centered log-ratio values.

852

haematologica | 2018; 103(5)

B-cell receptor signature in mantle cell lymphoma

ate a reduced signature (Figure 2B). In this way, 18 probes corresponding to 15 genes, all over-expressed in BCRhigh cases were identified (Figure 2B). Among these genes, a subgroup of six genes (AKT3, BCL2, BTK, CD79B, PIK3CD, and SYK) was selected for further validations due to their direct involvement in the BCR pathway and/or the existence of drugs targeting the related proteins. A hierarchical cluster using only these six genes was able to discriminate patients belonging to the BCRlow or BCRhigh groups (Figure 2C).

Development of a qRT-PCR-based predictor for BCRlow and BCRhigh in MCL samples By analyzing the expression levels of the selected six genes in the same 27 MCL PB samples by qRT-PCR approach, a strict correlation with GEP data was found (Online Supplementary Figure S3B). Moreover, the 27 MCL cases were randomly divided into a training set (17 cases; 8 BCRlow and 9 BCRhigh samples) and a validation set (10 cases; 6 BCRlow and 4 BCRhigh samples) to develop and test a decision tree (DT) model based on qRT-PCR data capable of categorizing patients into one of the two categories.

The DT model based on qRT-PCR data correctly classified 16 of 17 cases belonging to the training set and 10 of 10 cases of the validation cohort, and allowed the classification of 19 additional PB samples screened with qRT-PCR (9 BCRlow and 10 BCRhigh (Figure 2D and E and Online Supplementary Table S2).

(median PFS: 21.6 months vs. not reached; P=0.0375) (Figure 3).

Application of the 6-gene signature to LN samples from MCL patients To evaluate the capability of the 6-gene signature to identify different subgroups also in the context of MCL LN cases, we tested our qRT-PCR approach in a series of 43 LN samples preserved as FFPE LN specimens. Thirtyfive (81%) out of 43 samples were amplifiable for all six genes, and using a DT model based on qRT-PCR values from FFPE, 23 cases were classified as BCRlow and 20 classified as BCRhigh (Online Supplementary Table S2). Notably, for 6 out of 43 LN samples, a PB matched sample was available, and by comparing qRT-PCR results performed on PB samples and LN FFPE samples from these cases, a good concordance was overall observed, although FFPE samples generally amplified at higher Ct values (Online Supplementary Figure S5). Of note, 5 out of 6 these MCL cases were consistently classified. The misclassified case

Association between BCR categories and biological and clinical parameters Collectively, the 6-gene signature was re-evaluated by setting up a validated qRT-PCR approach (see Online Supplementary Appendix and Online Supplementary Table S6) in PB samples from 46 MCL cases, 23 were identified as BCRlow and 23 as BCRhigh. By correlating the BCR groups with the available biological parameters, no association was found between the 6-gene signature and IGHV gene status (P=0.93) (Online Supplementary Table S2 and Online Supplementary Figure S4A), Ki-67 expression, white blood cells, hemoglobin, lymphocytes, platelets, and neutrophil count (data not shown). The only significant difference was between the BCR classification and lactate dehydrogenase (LDH) levels; BCRhigh cases showed a higher level of LDH with respect to BCRlow MCL (416.6Âą191.6 vs. 292.2Âą127.4; P=0.023) (Online Supplementary Figure S4B).

Clinically, MCL patients classified as BCRhigh experienced shorter PFS with respect to BCRlow MCL cases

Figure 3. BCRhigh mantle cell lymphoma (MCL) group is associated with a worse clinical outcome. Kaplan-Meier curves obtained by comparing progression-free survival intervals of 23 BCRlow MCL cases with 23 BCRhigh MCL cases. The number of patients in each group is reported under relative categories; P refers to logrank test.

Table 2. Cox regression analysis on mantle cell lymphoma cases.

Variable BCR signature BCR high MIPI-c Low/intermediate High/intermediate High

Univariable HR (95%CI)

P

Multivariable HR (95%CI)

P

2.81 (1.28-6.19)

0.01

3.48 (1.47-8.25)

0.005

1.5 (0.6-3.73) 1.41 (0.45-4.43) 3.46 (1.1-10.91)

0.384 0.557 0.034

4.17 (1.3-13.34)

0.016

Multivariable Cox regression analysis of progression-free survival was performed by including the 6-gene BCR categorization and the mantle cell lymphoma International Prognostic Index-combined (MIPI-c) score as defined by Hoster et al.13 HR: Hazard Ratio; CI: Confidence Interval; BCR: B-cell receptor.

haematologica | 2018; 103(5)

853

R. Bomben et al.

was considered as BCRlow according to GEP data. Also in the context of LN samples, no correlation was found between the different biological parameters and BCR groups (data not shown).

By merging the MCL cases analyzed either in PB or in LN, a total of 83 cases were collected, 43 BCRlow and 40 BCRhigh. BCRhigh patients had a shorter PFS with respect to BCRlow patients (median PFS: 42.1 months vs. not reached; P=0.0074) (Figure 4A). Since Ki-67 is a well-known prognosticator in MCL,26 we combined the BCR groups with the prognostic groups defined by Ki-67 score. Cases with high Ki-67 (â&#x2030;Ľ30% of Ki-67 expressing cells) and classified in the BCRhigh group experienced the shortest PFS, while

A

C

854

cases classified as BCRlow had similar longer PFS intervals irrespective of the high or low Ki-67 score (median PFS: 20.5 months vs. not reached for all the other combinations; P=0.0014) (Figure 4B). Consistently, multivariable analysis carried out by including the BCR signature and the MIPI-c categories selected the BCRhigh and the high risk MIPI-c category as independent predictors of PFS (Table 2). Regarding OS, while the BCR readout failed to identify groups with different OS intervals, possibly due to the low rate of events and short follow up (Figure 4C), the combination of high Ki-67 score and a BCRhigh 6-gene signature was able again to select the MCL subgroup with the shortest OS (46.7 vs. not reached; P=0.029) (Figure 4D).

B

D

Figure 4. BCRhigh mantle cell lymphoma (MCL) group is associated with a worse clinical outcome (overall series).

(A) Kaplan-Meier curves obtained by comparing progression-free survival (PFS) intervals of 43 BCRlow MCL cases with 40 BCRhigh MCL cases. (B) Kaplan-Meier curves obtained by comparing PFS intervals of 19 BCRlow and low Ki-67 MCL cases, with 20 BCRlow and high Ki-67 MCL cases, with 21 BCRhigh and Ki-67 low MCL cases, with 10 BCRhigh and Ki-67 high MCL cases. (C) Kaplan-Meier curves obtained by comparing overall survival (OS) intervals of 43 BCRlow MCL cases with 40 BCRhigh MCL cases.

(D) Kaplan-Meier curves obtained by comparing OS intervals of 19 BCRlow and low Ki-67 MCL cases, with 20 BCRlow and high Ki-67 MCL cases, with 21 BCRhigh and Ki67 low MCL cases, with 10 BCRhigh and Ki-67 high MCL cases. The number of patients in each group is reported under relative categories; P refers to log-rank test.

haematologica | 2018; 103(5)

B-cell receptor signature in mantle cell lymphoma

Validations of BCR signature To verify whether the BCR signature maintained its prognostic impact in an independent set of patients, we used the gene expression data of MCL LN biopsies reported by Saba et al.30 Also in this different setting, a high expression of the 6-gene signature, as in the context of BCRhigh cases, identified an MCL patient subset with inferior PFS (P=0.049) (Online Supplementary Figure S6).

In another set of analyses, by taking advantage of our 27 MCL cases with GEP data available, we correlated our BCR signature with other MCL signatures with proven clinical impact.30,38 As reported in Online Supplementary Figure S7A, the BCR signature reported in Saba et al.30 divided MCL cases in two groups that corresponded exactly to our BCR definition (Online Supplementary Figure S7B).30 Similarly, the 17 genes of the proliferation signature reported by Scott et al.38 split our MCL cases in 3 different groups resembling the 3 different groups originally defined (Online Supplementary Figure S8A). In this context, the shortest PFS and OS intervals were observed in the third group characterized by a higher expression of genes related to proliferation and a BCRhigh phenotype in keeping with our findings (Online Supplementary Figure S8A-C).

Discussion In this study, we demonstrated that a BCR-derived signature based on the differential expression of six genes correlated with shorter PFS intervals in the context of a Phase III prospective clinical trial (FIL-MCL-0208) for younger MCL patients receiving R-CHOP induction, followed by high-dose cytarabine and autologous stem cell transplantation (clinicaltrials.gov identifier: 023541313).32 Notably, the BCR-related 6-gene signature reported here was able to identify an MCL subset with shorter PFS intervals also in the context of an external independent MCL cohort homogenously treated with different schedules.30 On the other hand, when the signature described by Saba et al.30 and Scott et al.38 was applied to our MCL cases, the patient subsets with the worse prognosis turned out to be particularly enriched in BCRhigh cases, even though these signatures did not include any gene from our signature.

Therefore, although composed of genes located upstream of the BCR machinery, our signature was able to identify cases with an active BCR pathway as defined by other signatures.30 In this regard, however, experiments with primary MCL cases and/or MCL cell lines combining BCR stimulation with the use of specific BCR inhibitors should be performed to investigate the contribution of the 6-gene signatures described here to the actual activation of the BCR pathway.

Again in agreement with this line of reasoning, BCRhigh samples presented a significant upregulation of PAX5 (see GEP data in Online Supplementary Table S3), a gene whose product is known to prevent plasma cell differentiation thus preserving the capacity to respond to antigeninduced activation and proliferation.39 Taken together these data corroborate recent findings of ongoing active BCR signaling in MCL cell in vivo,29,30 and further underline the role of antigen stimulation in the ontogeny of MCL, as suggested by the skewed IGVH gene repertoire found in MCL cells.40 In order to discriminate between BCRlow and BCRhigh MCL samples, we developed a DT model based on the haematologica | 2018; 103(5)

expression of the selected six genes.28 This DT model was applied in an independent cohort of PB samples and then to a further series of FFPE LN samples, thus demonstrating that two MCL subsets with different expression levels of BCR-related genes could also be recognized in the LN compartment, mirroring PB. Taken together, by combining data from the PB and LN compartments, MCL cases classified as BCRhigh showed higher LDH levels and shorter PFS with respect to BCRlow patients, suggesting that activation of BCR signaling drives tumor proliferation and determines clinical outcome of MCL patients, which is in keeping with recent findings.30 By combining the predictive capacity of the 6-gene BCR signature with the Ki-67 index, we identified a particularly unfavorable category (BCRhigh and high Ki-67) with a substantially shorter PFS and OS than the other groups.

Consistently, the BCRhigh signature turned out to be an independent prognosticator along with the high-risk MIPI-c category for short PFS by multivariate analysis.

There is no indication that the validity of the model may be affected by the different recruitment site (PB vs. LN), or by different sample storage (frozen vs. FFPE) because the main clinical parameters were equally distributed between the different series (PB/frozen vs. LN/FFPE) (R Bomben et al., 2018, unpublished observation). In this regard, an important feature of this model/assay is its applicability to both PB and LN FFPE samples, having, therefore, the chance to combine results of qRT-PCR with Ki-67 staining in all the cases.

Our data underscore the increasing importance of BCRrelated genes in the pathogenesis and development of MCL, further underlined by the clinical significance of drugs specifically targeting genes belonging to this pathway. In particular, therapeutic targeting of BTK41 can be rationally exploited in lymphoid malignancies that have been proved to be dependent on an antigen-dependent BCR-mediated active signaling. However, despite the relatively high response rate to single agent ibrutinib in relapsed/refractory MCL, it remained unclear as to why some patients showed clear responses, while others received little therapeutic benefit.31,42 The BCR-related signature described here may provide insights into molecular factors that explain the divergent responses of MCL patients to ibrutinib, although other causes of primary resistance might be related to gene mutations in the other pathways, e.g. NF-ÎşB pathway and epigenetic modifiers, as recently reported.43,44 In conclusion, in the present study we developed a survival model for patients with MCL composed of six genes (AKT3, BTK, CD79B, PIK3CD, SYK, BCL2) whose expression can easily be investigated by qRT-PCR and also in FFPE specimens. The signature was associated with a poor clinical response in the context of a high-dose chemoimmunotherapy regimen, and might, therefore, be considered for validation and application in future clinical trials.

Acknowledgments The authors would like to thank Progetto Giovani Ricercatori GR-2011-02347441, GR-2009-1475467, and GR-201102351370, Ministero della Salute, Rome, Italy; Progetto Ricerca Finalizzata RF-2009-1469205, and RF-2010-2307262, Ministero della Salute, Rome, Italy; Associazione Italiana contro le Leucemie, linfomi e mielomi (AIL), Venezia Section, Pramaggiore Group, Italy; Associazione Italiana Ricerca Cancro 855

R. Bomben et al.

(AIRC), Investigator Grant IG-2015 (17622); “5x1000 Intramural Program”, Centro di Riferimento Oncologico, Aviano, Italy; Provincia Autonoma di Bolzano/Bozen, Italy; A.O. S. Maurizio, Bolzano/Bozen, Italy; Progetto di Rilevante Interesse Nazionale (PRIN2009) 7.07.02.60 AE01, Ministero Italiano dell'Università e della Ricerca (MIUR), Roma, Italy; Fondi di

References 1. Dreyling M, Ferrero S, Hermine O. How to manage mantle cell lymphoma. Leukemia.

2014;28(11):2117-2130.

2. Cheah CY, Seymour JF, Wang ML. Mantle Cell Lymphoma. J Clin Oncol. 2016; 34(11):1256-1269.

3. Campo E, Swerdlow SH, Harris NL, Pileri S, Stein H, Jaffe ES. The 2008 WHO classification of lymphoid neoplasms and beyond: evolving concepts and practical applications.

Blood. 2011;117(19):5019-5032.

4. Jares P, Colomer D, Campo E. Molecular pathogenesis of mantle cell lymphoma. J Clin Invest. 2012;122(10):3416-3423.

5. Dreyling M, Ferrero S, Vogt N, Klapper W. New paradigms in mantle cell lymphoma: is it time to risk-stratify treatment based on the proliferative signature? Clin Cancer Res. 2014;20(20):5194-5206.

6. Ghielmini M, Zucca E. How I treat mantle cell lymphoma. Blood. 2009;114(8):14691476.

7. Herrmann A, Hoster E, Zwingers T, et al.

Improvement of overall survival in advanced stage mantle cell lymphoma. J Clin Oncol.

2009;27(4):511-518.

8. Zucca E, Roggero E, Pinotti G, et al. Patterns of survival in mantle cell lymphoma. Ann Oncol. 1995;6(3):257-262.

9. Barista I, Romaguera JE, Cabanillas F. Mantle-cell lymphoma. Lancet Oncol.

2001;2(3):141-148.

10. Martin P, Chadburn A, Christos P, et al.

Outcome of deferred initial therapy in mantle-cell lymphoma. J Clin Oncol. 2009; 27(8):1209-1213.

11. Dreyling M, Ferrero S. The role of targeted treatment in mantle cell lymphoma: is transplant dead or alive? Haematologica.

2016;101(2):104-114.

12. Geisler CH, Kolstad A, Laurell A, et al. The Mantle Cell Lymphoma International Prognostic Index (MIPI) is superior to the International Prognostic Index (IPI) in predicting survival following intensive first-line immunochemotherapy and autologous stem cell transplantation (ASCT). Blood.

2010;115(8):1530-1533.

13. Hoster E, Rosenwald A, Berger F, et al.

Prognostic Value of Ki-67 Index, Cytology, and Growth Pattern in Mantle-Cell Lymphoma: Results From Randomized Trials of the European Mantle Cell Lymphoma Network. J Clin Oncol. 2016; 34(12):1386-1396.

14. Hoster E, Dreyling M, Klapper W, et al. A new prognostic index (MIPI) for patients with advanced-stage mantle cell lymphoma.

Blood. 2008;111(2):558-565.

15. Bea S, Valdes-Mas R, Navarro A, et al.

Landscape of somatic mutations and clonal evolution in mantle cell lymphoma. Proc Natl Acad Sci USA. 2013;110(45):1825018255.

16. Zhang J, Jima D, Moffitt AB, et al. The genomic landscape of mantle cell lymphoma is related to the epigenetically determined chromatin state of normal B cells. Blood.

2014;123(19):2988-2996.

856

Ricerca Locale, Università degli Studi di Torino, Italy; Fondazione Neoplasie Del Sangue (Fo.Ne.Sa), Torino, Italy; CRT 2015.1044, Fondazione CRT, Torino, Italy. We are grateful to all the Clinical Investigators, to the Pathologists, to Luigia Monitillo, Daniela Barbero, Marina Ruggeri, Paola Ghione, and Gian Maria Zaccaria.

17. Katzenberger T, Petzoldt C, Holler S, et al.

The Ki67 proliferation index is a quantitative indicator of clinical risk in mantle cell lymphoma. Blood. 2006;107(8):3407.

18. Hartmann E, Fernandez V, Moreno V, et al.

Five-gene model to predict survival in mantle-cell lymphoma using frozen or formalinfixed, paraffin-embedded tissue. J Clin Oncol. 2008;26(30):4966-4972.

19. Ek S, Bjorck E, Porwit-MacDonald A, Nordenskjold M, Borrebaeck CA. Increased expression of Ki-67 in mantle cell lymphoma is associated with de-regulation of several cell cycle regulatory components, as identified by global gene expression analysis. Haematologica. 2004;89(6):686-695.

20. Rosenwald A, Wright G, Wiestner A, et al.

The proliferation gene expression signature is a quantitative integrator of oncogenic events that predicts survival in mantle cell lymphoma. Cancer Cell. 2003;3(2):185-197.

21. Mozos A, Royo C, Hartmann E, et al.

SOX11 expression is highly specific for mantle cell lymphoma and identifies the cyclin D1-negative subtype. Haematologica.

2009;94(11):1555-1562.

22. Navarro A, Clot G, Royo C, et al. Molecular subsets of mantle cell lymphoma defined by the IGHV mutational status and SOX11 expression have distinct biologic and clinical features. Cancer Res. 2012;72(20):53075316.

23. Majlis A, Pugh WC, Rodriguez MA, Benedict WF, Cabanillas F. Mantle cell lymphoma: correlation of clinical outcome and biologic features with three histologic variants. J Clin Oncol. 1997;15(4):1664-1671.

24. Wiestner A, Tehrani M, Chiorazzi M, et al.

Point mutations and genomic deletions in CCND1 create stable truncated cyclin D1 mRNAs that are associated with increased proliferation rate and shorter survival.

Blood. 2007;109(11):4599-4606.

25. Jares P, Colomer D, Campo E. Genetic and molecular pathogenesis of mantle cell lymphoma: perspectives for new targeted therapeutics. Nat Rev Cancer. 2007;7(10):750762.

26. Determann O, Hoster E, Ott G, et al. Ki-67 predicts outcome in advanced-stage mantle cell lymphoma patients treated with antiCD20 immunochemotherapy: results from randomized trials of the European MCL Network and the German Low Grade Lymphoma Study Group. Blood. 2008; 111(4):2385-2387.

27. Perez-Galan P, Dreyling M, Wiestner A. Mantle cell lymphoma: biology, pathogenesis, and the molecular basis of treatment in the genomic era. Blood. 2011;117(1):26-38.

28. Young RM, Staudt LM. Targeting pathological B cell receptor signalling in lymphoid malignancies. Nat Rev Drug Discov. 2013; 12(3):229-243.

29. Akhter A, Street L, Ghosh S, et al.

Concomitant high expression of Toll-like receptor (TLR) and B-cell receptor (BCR) signalling molecules has clinical implications in mantle cell lymphoma. Hematol Oncol.

2015;35(1):79-86.

30. Saba NS, Liu D, Herman SE, et al.

Pathogenic role of B-cell receptor signaling

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

and canonical NF-kappaB activation in mantle cell lymphoma. Blood. 2016; 128(1):8292.

Wang ML, Rule S, Martin P, et al. Targeting BTK with ibrutinib in relapsed or refractory mantle-cell lymphoma. N Engl J Med. 2013;369(6):507-516.

Cortellazzo S, Martelli M, Ladetto M, Ferrero S, Ciccone G, Evangelista A, et al.

High dose sequential chemotherapy with rituximab and ASCT as first line therapy in adult MCL patients: clinical and molecular response of the MCL-0208 trial, a FIL study.

Haematologica. 2015;100(s1):3.

Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization (WHO) classification of lymphoid neoplasms.

Blood.

2016; 127(20):2375-2390.

Bomben R, Gobessi S, Dal BM, et al. The miR-17-92 family regulates the response to Toll-like receptor 9 triggering of CLL cells with unmutated IGHV genes. Leukemia.

2012;26(7):1584-1593.

Dal Bo M, D'Agaro T, Gobessi S, et al. The SIRT1/TP53 axis is activated upon B-cell receptor triggering via miR-132 up-regulation in chronic lymphocytic leukemia cells.

Oncotarget. 2015;6(22):19102-19117.

Draghici S, Khatri P, Martins RP, Ostermeier GC, Krawetz SA. Global functional profiling of gene expression.

Genomics.

2003;81(2):98-104.

Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA. 2005;102(43):15545-15550.

Scott DW, Abrisqueta P, Wright GW et al.

New Molecular Assay for the Proliferation Signature in Mantle Cell Lymphoma Applicable to Formalin-Fixed ParaffinEmbedded Biopsies. J Clin Oncol. 2017; 35(15):1668-1677.

Nera KP, Kohonen P, Narvi E, et al. Loss of Pax5 promotes plasma cell differentiation.

Immunity. 2006;24(3):283-293.

Hadzidimitriou A, Agathangelidis A, Darzentas N, et al. Is there a role for antigen selection in mantle cell lymphoma?

Immunogenetic support from a series of 807 cases. Blood. 2011;118(11):3088-3095.

Rickert RC. New insights into pre-BCR and BCR signalling with relevance to B cell malignancies. Nat Rev Immunol. 2013; 13(8):578-591.

Wang ML, Blum KA, Martin P, et al. Longterm follow-up of MCL patients treated with single-agent ibrutinib: updated safety and efficacy results. Blood. 2015; 126(6):739745.

Rahal R, Frick M, Romero R, et al.

Pharmacological and genomic profiling identifies NF-kappaB-targeted treatment strategies for mantle cell lymphoma. Nat Med. 2014;20(1):87-92.

Lenz G, Balasubramanian S, Goldberg J, Rizo A, Schaffer M, Phelps C, et al.

Sequence variants in patients with primary and acquired resistance to ibrutinib in the phase 3 MCL3001 (RAY) trial.

Haematologica. 2016;101(s1):155.

haematologica | 2018; 103(5)

ARTICLE

Non-Hodgkin Lymphoma

Incidence and risk factors for relapses in HIV-associated non-Hodgkin lymphoma as observed in the German HIV-related lymphoma cohort study Philipp Schommers,1,2,* Daniel Gillor,1,* Marcus Hentrich,3 Christoph Wyen,1,4 Timo Wolf,5 Mark Oette,6 Alexander Zoufaly,7 Jan-Christian Wasmuth,8 Johannes R. Bogner,9 Markus Müller,10 Stefan Esser,11 Alisa Schleicher,12 Björn Jensen,13 Albrecht Stoehr,14 Georg Behrens,15,16 Alexander Schultze,17 Jan Siehl,18 Jan Thoden,19 Ninon Taylor20 and Christian Hoffmann12,21

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):857-864

1 Department I of Internal Medicine, University Hospital Cologne, Germany; 2German Center for Infection Research (DZIF), Partner Site Bonn-Cologne, Germany; 3Department of Medicine III, Red Cross Hospital Munich, Germany; 4Praxis am Ebertplatz, Cologne, Germany; 5Department of Medicine II, University of Frankfurt, Germany; 6Department of General Medicine, Gastroenterology and Infectious Diseases, Augustinerinnen Hospital, Cologne, Germany; 7Department of Medicine IV, Kaiser Franz Josef Hospital, Vienna, Austria; 8Department of Internal Medicine I, University of Bonn, Germany; 9Department of Medicine IV, University of Munich, Munich, Germany; 10Department of Infectious Diseases, Vivantes Auguste-Viktoria- Hospital, Berlin, Germany; 11Department of Dermatology, University Hospital Essen, Germany; 12University of Schleswig Holstein, Campus Kiel, Kiel, Germany; 13Department of Gastroenterology, Hepatology and Infectious Diseases, Düsseldorf University Hospital, Germany; 14Ifi-Institute for Interdisciplinary Medicine, Hamburg, Germany; 15Department of Clinical Immunology and Rheumatology, Hannover Medical School, Germany; 16German Center for Infection Research (DZIF), Hannover, Germany; 17Department of Emergency Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 18Ärzteforum Seestraße, Berlin, Germany; 19Medical Group Practice for Internal Medicine and Rheumatology, Freiburg, Germany; 20Department of Internal Medicine III with Hematology, Medical Oncology, Hemostaseology, Infectious Diseases, Rheumatology, Oncologic Center, Laboratory of Immunological and Molecular Cancer Research, Paracelsus Medical University Salzburg, Austria and 21IPM Study Center, Hamburg, Germany

*PS and DG contributed equally to this work.

Correspondence: philipp.schommers@uk-koeln.de ABSTRACT

O

utcome of HIV-infected patients with AIDS-related lymphomas has improved during recent years. However, data on incidence, risk factors, and outcome of relapses in AIDS-related lymphomas after achieving complete remission are still limited. This prospective observational multicenter study includes HIV-infected patients with biopsy- or cytology-proven malignant lymphomas since 2005. Data on HIV infection and lymphoma characteristics, treatment and outcome were recorded. For this analysis, AIDS-related lymphomas patients in complete remission were analyzed in terms of their relapsefree survival and potential risk factors for relapses. In total, 254 of 399 (63.7%) patients with AIDS-related lymphomas reached a complete remission with their first-line chemotherapy. After a median follow up of 4.6 years, 5-year overall survival of the 254 patients was 87.8% (Standard Error 3.1%). Twenty-nine patients relapsed (11.4%). Several factors were independently associated with a higher relapse rate, including an unclassifiable histology, a stage III or IV according to the Ann Arbor Staging System, no concomitant combined antiretroviral therapy during chemotherapy and R-CHOP-based compared to more intensive chemotherapy regimens in Burkitt lymphomas. In conclusion, complete remission and relapse rates observed in our study are similar to those reported in HIV-negative non-Hodgkin lymphomas. These data provide further evidence for the use of concomitant combined antiretroviral therapy during chemotherapy and a benefit from more intensive chemotherapy regimens in Burkitt lymphomas. Modifications to the chemotherapy regimen appear to have only a limited impact on relapse rate.

Haematologica | 2018; 103(5)

Received: September 17, 2017.

Accepted: February 2, 2018.

Pre-published: February 8, 2018. doi:10.3324/haematol.2017.180893 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/857 ©2018 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.

857

P. Schommers et al.

Introduction Over the last two decades, the incidence of AIDS-related lymphomas (ARL) has markedly declined due to the introduction of combination antiretroviral therapy (cART). However, ARL remain a major cause of morbidity and mortality, and represent the highest proportion of all AIDS-related deaths.1 Patients with ARL are usually treated with the same chemotherapy protocols established in the HIV-negative setting,2 and the rates of complete remission (CR) achieved are comparable to those reported in their HIV-negative counterparts.3,4 However, available data on the incidence and potential risk factors of recurrent disease in ARL are scarce, and treatment of disease relapse remains challenging.4,5 In recent studies on HIV-negative patients with diffuse large B-cell lymphoma (DLBCL), RCHOP-based regimens (rituximab, cyclophosphamide, adriamycin, vincristine, and prednisone) resulted in CR rates of around 65-80%. Differences in response rates largely depend on the pre-treatment International Prognostic Index (IPI) for aggressive lymphomas.6,7 In patients who had achieved a CR, relapse rates ranged from 6-10%.6,8 The second most common ARL are Burkitt or Burkitt-like lymphomas (BL).9,10 In the HIV-negative setting, CR and overall survival (OS) rates of around 80-90% were reported by different groups.10-12 In a large prospective trial on short-intensive chemotherapy combined with rituximab for patients with BL, the relapse rate was 12%.10 Although this approach also proved feasible in HIV-related BL,9 it remains unclear whether relapse rates reported in HIV-negative DLBCL and BL are different to those in ARL. Thus, we investigated the risk factors and incidence of relapse in a large cohort of ARL patients who had achieved a CR after first-line treatment.

Methods Study design The German HIV Lymphoma Cohort is an ongoing, prospective observational multicenter study including all adult HIV infected patients who are diagnosed with biopsy- or cytology-proven malignant lymphoma in 33 participating centers since January 2005. Data on HIV-infection and lymphoma characteristics, treatment and outcome are recorded. From the time of lymphoma diagnosis, patients are followed every six months. Ethics approval was obtained from the ethics committees of the University of Cologne (IRC Cologne: 05-174), Germany, and written informed consent was given by each participating patient.

The present analysis includes only patients with aggressive Bcell lymphoma in first CR. Lymphomas were grouped in DLBCL, BL, plasmablastic lymphoma (PBL) and ARL, not further classifiable, the latter group representing aggressive B-cell non-Hodgkin lymphomas (B-NHL) that could not be classified into any subtype.

To study the impact of chemotherapy dose intensity on the risk of relapse in patients treated with either R-CHOP-based regimens or the short intensive GMALL protocol,10 we performed an analysis of dose reductions and delays in chemotherapy cycles.

(Information about the GMALL and R-CHOP protocol can be found in the Online Supplementary Tables S1 and S2, respectively).

A full-intensity treatment was considered to consist of six cycles of chemotherapy according to the R-CHOP or GMALL protocol administered at 3-week intervals without dose reductions and within a period of 120 days (5x21 days for 6 cycles plus a maximum of 3 days delay per cycle). Less than 6 cycles of chemother858

apy were classified as “cycle reduction” and the treatment duration was calculated according to the number of cycles given. If the dose of any chemotherapy drug was reduced by 20% or more, treatment intensity was considered to be reduced.

As positron emission tomography (PET) scans were not routinely performed, the 1999 standardized response criteria for nonHodgkin’s lymphomas13 were used rather than the 2007 criteria.14 CR was defined as the disappearance of all disease manifestations for at least three months. This definition also includes uncertain complete remission (CRu) that implied a residual mass of 1.5 cm or smaller that remained unchanged over at least three months.

Statistical analysis Statistical analyses were performed using IBM SPSS Statistics software (IBM, Armonk, NY, USA), v.24.0. Univariate statistics were performed using Pearson’s χ², Fisher’s exact one-way Analysis of Variance (ANOVA) with Bonferroni-corrected posthoc test, or Kruskal-Wallis test depending on data. For the multivariate Cox regression analysis, continuous clinically meaningful breakpoints that showed P-values below 0.1 in the univariate analysis were considered. Kaplan-Meier curves were used to illustrate the relapse-free survival (RFS) and overall survival.

Differences between subgroups were assessed with the log-rank test. RFS was defined as the period between first diagnosis and any lymphoma relapse according to the STEEP criteria.15 OS was defined as the period between first diagnosis and death from any cause. All-cause deaths as well as “lost to follow up” were censored. All P-values were two-sided. P<0.05 was considered statistically significant.

Results Patients' characteristics and outcome Numbers and characteristics of patients included in the present analysis are depicted in Figure 1. In total, 254 of 399 (63.7%) patients with high-grade NHL of B-cell origin (classified as ARL) reached a CR with their first-line chemotherapy. Of those, 127 had DLBCL, 91 BL, 29 PBL, and 7 ARL, not further classified. ARL was CD20-negative in 24 of 254 cases (9.5%), among them 22 PBL and 2 DLBCL cases. Among 22 PBL cases with information on Epstein-Barr-Virus (EBV) status, EBV was present in 15 (68%). Overall, 86.2% of patients with CD20+ lymphomas received rituximab. Notably, patients diagnosed before 2010 were less frequently treated with rituximab than those diagnosed from 2010 onwards (79% vs. 96%, P<0.001). Further, 73% of patients with CD4 cell counts less than 50/µl received rituximab compared to 88% with CD4 counts 50/mL or over (P=0.096). Patients' characteristics with respect to treatment outcomes are listed in Table 1. OS of patients who achieved CR with first-line therapy was significantly better than that of patients in other response groups (Figure 2). After a median follow up of 4.6 years, 5-year OS of the entire group of all 262 patients in first CR was 87.1% [standard error (SE) 2.3%] (Figure 3A) with differences between lymphoma subtypes: 87.8% (SE 3.1%) in DLBCL, 87.6% (SE 3.7%) in BL, 79.6% (SE 11.3%) in PBL, and 83.3% (SE 15.2%) in ARL, not further classified (P=0.994) (Figure 3B).

Incidence of recurrent disease in ARL After a median follow up of 4.6 years, a relapse of the ARL had occurred in 29 of 254 patients (11.4%). Relapses were observed in 14 patients with DLBCL (11.0%), 9 with haematologica | 2018; 103(5)

Relapses in AIDS-related lymphomas

BL (9.9 %), 3 with PBL (11.5%) and 3 with ARL, not further classified (42.9%), after a median follow up of 5.0, 4.6, 3.5 and 6.0 years, respectively. Isolated central nervous system (CNS) relapses were observed in 3 of 29 patients (DLBCL: n=2; BL: n=1). RFS depicted by Kaplan Meier curves is shown in Figure 3C and D. Five-year RFS (5yRFS) was 88.4% (SE 2.9%) in DLBCL, 88.9% (SE 3.5%) in BL, and 88.6% (SE 6.2%) in PBL. By contrast, 5yRFS was lower in patients with ARL, not further classified [57.1% (SE 18.7%); P=0.057) (Figure 3D).

Among patients who achieved CR with first-line RCHOP-based protocols, 5yRFS was 87.8% (SE 3.1%) and 84.4% (SE 8.3%) in DLBCL and PBL, respectively, as compared to 65.5% (SE 12.6) in BL and 40.0% in ARL, not further classified (SE 21.9%; P=0.005) (Figure 3E). No significant differences in 5yRFS between ARL subtypes were observed in patients treated with the GMALL protocol

(P=0.884) (Figure 3F), although the number of patients with subtypes other than BL was very small in this analysis. Of note, patients with BL who received the GMALL protocol had a significantly better 5yRFS than those receiving R-CHOP-based protocols [94.2% (SE 2.8%) vs.

65.5% (SE 12.6%); P=0.001].

Risk factors for recurrent disease in ARL Univariate analysis identified several factors associated with a lower risk for ARL relapse such as a low IPI, stage I or II according to the Ann Arbor Staging System, cART given during chemotherapy, CD4 T-cell counts >200x109/L, pathology other than ARL, not further classified, and chemotherapy according to the GMALL-protocol (Table 2). These factors were analyzed in a multivariate Cox proportional hazards model, with backward stepwise elimination based on a Wald statistic with Pâ&#x2030;¤0.1.

Figure 1. Flow chart of patients included in the present analysis. NHL: nonHodgkin lymphoma; T-NHL: T-cell non-Hodgkin lymphoma; CR: complete remission; BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; PBL: plasmablastic lymphoma; ARL: AIDS-related lymphoma.

haematologica | 2018; 103(5)

859

P. Schommers et al.

After two elimination steps, histology [BL: Hazard ratio (HR) 2.60 95% Confidence Interval (95%CI): 0.92 – 7.4, PBL: HR 1.28 95% CI 0.36-4.57, ARL, not further classified: HR 5.08 95% CI 1.13 – 22.90, indicator = DLBCL), stage III or IV according to the Ann Arbor Staging System (HR 4.85 95%CI 1.44 – 16.34), no concomitant cART (HR 4.28 95%CI 1.19 – 15.39) and use of R-CHOP (HR: 7.59 95%CI 1.87 – 30.81) remained in the model (Table 2). A higher IPI was no longer predictive anymore in the multivariate model.

Dose intensity of chemotherapy Since chemotherapy regimen (R-CHOP or GMALL) seems to be critical for RFS, we investigated how many patients of all aggressive B-NHL had any kind of reduction (either in the number of chemotherapy cycles or in the treatment intensity) or a delay during their treatment.

Results of dose intensity analysis are shown in Online Supplementary Table S3. Overall, 32.7% of the patients had a treatment delay, 13.8% had dose reductions, and 16.5% had reduced numbers of chemotherapy cycles

Table 1. Patients’ characteristics based on their treatment outcome.

CR after CR after Progressive Partial On Treatment No first-line further lines disease remission chemotherapy related chemotherapy chemotherapy chemotherapy (n=45) (n=22) (n=15) deaths given (n=254) (n=21) (n=23) (n=7) Median age (years) Male Median viral load (copies/mL) HIV-RNA below limit of detection Median CD4+ T cells (x109/L) CD20+ lymphoma BM involvement CNS involvement IPI score Low Intermediate High Lymphoma subtype DLBCL BL PBL ARL, not further classifiable Median follow up (years)

Total (N=387)

P

44 230 (91%) 19031

48 20 (95%) 26790

45 44 (98%) 40208

44 21 (96%) 7159

50 15 (100%) 11387

48 21 (91%) 2390

46 6 (86%) 105000

45 357 (92%) 18557

0.104b 0.514a 0.784b

74 (30%) 248

5 (24%) 190

11 (25%) 111

8 ((36%) 186

4 (29%) 153

8 (35%) 157

2 (29%) 58

112 (30%) 212

0.903a 0.006b

213 (90%) 47 (20%) 17 (8%)

20 (95%) 7 (33%) 2 (10%)

34 (81%) 15 (39%) 8 (21%)

18 (90%) 6 (29%) 2 (11%)

10 (83%) 3 (21%) 1 (9%)

13 (68%) 6 (21%) 5 (33%)

4 (67%) 1(17%) 0 (0%)

312 (87%) 85 (23%) 36 (11%)

0.044a 0.190a 0.023

100 (42%) 104 (44%) 34 (14%)

5 (24%) 11 (52%) 5 (24%)

8 (18%) 20 (46%) 16 (36%)

2 (11%) 13 (68%) 4 (21%)

5 (39%) 6 (42%) 2 (15%)

2 (9%) 11 (50%) 9 (41%)

0 (0%) 3 (43%) 4 (57%)

122 (36%) 168 (46%) 74 (20%)

<0.001a

127 (50%) 91 (36%) 29 (11%) 7 (3%) 4.64

7 (33%) 11 (52%) 3 (14%) 0 (0%) 5.4

24 (53%) 12 (27%) 7 (16%) 2 (4%) 0.5

13 (59%) 5 (23%) 3 (14%) 1 (5%) 0.7

7 (47%) 4 (27%) 3 (20%) 1 (7%) 0.1

7 (30%) 9 (39%) 7 (30%) 0 (0%) 0.2

4 (57%) 1 (14%) 2 (29%) 0 (0%) 0

189 (49%) 133 (34%) 54 (14%) 11 (3%) 2.4

0.420a <0.001b

BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; PBL: plasmablastic lymphoma; IPI: International Prognostic Index; BM: bone marrow; CNS: central nervous system. aTwo-sided Pearson’s χ2. bKruskal-Wallis test.

Figure 2. Kaplan-Meier estimates for overall survival (OS) of the different observed treatment outcomes of aggressive non-Hodgkin lymphomas. CR: complete remission; Prog. Disease: progressive disease; Part. Remission: partial remission. Dotted line indicates 3 months.

860

haematologica | 2018; 103(5)

Relapses in AIDS-related lymphomas

given. Only 37.0% of patients received their full planned course of therapy.

Patients who experienced treatment delays and/or dose reductions or received a reduced number of chemotherapy cycles were significantly older (42 vs. 45 years; P=0.042) and had received the GMALL-protocol significantly more often than patients who completed the full planned course of therapy (P=<0.001) (Online Supplementary Table S3).

Overall, 86.2% of patients with CD20+ lymphomas

received rituximab. There was no difference in the relapse rate between patients with or without administration of rituximab (P=0.75), and our results remained consistent when patients without rituximab were excluded (data not shown).

Factors influencing the RFS To investigate the influence of different factors, including treatment reduction or delay on 5yRFS, we investigat-

Table 2. Risk factors for 5-year relapse-free survival (including all aggressive non-Hodgkin lymphoma in first complete remission; n=254).

Aggressive NHL (N=254) 5-year relapse-free survival % Sex Age CNS involvement BM involvement Bulky Disease CD4+ T cells <50x109/l Prior AIDS-defining illness IPI score

Ann Arbor stage Extranodal involvement ECOG score Elevated LDH Antiretroviral Treatment

cART during Chemotherapy CD20 positive lymphoma Lymphoma subtype

Chemotherapy

Male (n=230) Female (n=24) >60Y (n=24) <60Y (n=228) Yes (n=17) No (n=204) Yes (n=47) No (n=193) Yes (n=44) No (n=137) Yes (n=37) No (n=202) Yes (n=56) No (n=193) Low (n=100) Intermediate (n=104) High (n=34) I/II (n=93) III/IV (n=156) Yes (n=71) No (n=181) 0-1 (n=159) 2-5 (n=78) Yes (n=146) No (n=96) Viral load b.d. (n=74) Naive (n=134) Therapy failure (n=39) Yes (n=234) No (n=9) Positive (n=213) Negative (n=24) DLBCL (n=127) BL (n=91) PBL (n=29) ARL, not further classifiable (n=7) CHOP (n=163) GMALL (n=87)

87 96 77 89 82 90 81 90 86 88 86 88 87 88 95 84 82 95 83 87 88 89 86 86 92 88 88 87 89 67 88 87 88 89 89 57 84 95

P (univariate)

P (multivariate)

0.229 0.178 0.305 0.101 0.637 0.573 0797

0.039

Indicator 0.760 0.853

0.005

0.011

0.936 0.635 0.143

0.986 0.033

0.026

0.888

0.064

Indicator 0.072 0.703 0.034

0.013

0.005

Univariate statistics: Log rank test. Multivariate statistics: Cox regression. Viral load b.d.: Viral load below limit of detection; cART: combination antiretroviral therapy; BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; PBL: plasmablastic lymphoma; IPI: International Prognostic Index; BM: bone marrow; CNS: central nervous system; ECOG: Eastern Cooperative Oncology Group scale.

haematologica | 2018; 103(5)

861

P. Schommers et al.

ed two selected groups: patients with DLBCL receiving RCHOP-based regimens and patients with BL receiving the GMALL-regimen (see also Figure 1). In patients with DLBCL who received a R-CHOP-based treatment, a high IPI, an elevated LDH, stage III or IV according to the Ann Arbor Staging System, and bone marrow involvement at first diagnosis were associated with a significantly increased risk of relapse (Online Supplementary Table S4).

However, none of these parameters turned out to be an independent risk factor in the Cox proportional hazards model. Of note, chemotherapy dose reductions, treatment delays or a reduced number of R-CHOP-cycles given did not adversely affect 5yRFS (Online Supplementary Table S4).

By univariate analysis, patients with BL who underwent GMALL-chemotherapy had a significantly increased risk of relapse if the following factors were present: diagnosis of another AIDS-defining disease prior to BL diagnosis, CNS-involvement, failure of cART defined as measurable viral loads despite concomitant cART, concomitant cART during chemotherapy, and reduced numbers of chemotherapy cycles administered (Online Supplementary Table S4). However, none of these factors remained significant in the Cox proportional hazards model.

Discussion In this large prospective cohort study of 254 ARL patients who had achieved a CR with first-line chemotherapy (64% of all cases), the total relapse rate was 11% after a median follow up of 4.6 years. Patients with DLBCL who were mainly treated with R-CHOP-based regimens had a relapse rate of 11%. These rates are in line with the 6-10% relapse rates reported in HIV-negative DLBCL in first CR.6,8 Notably, the 10% relapse rate of patients with BL who were mainly treated with the GMALL protocol compares favorably with the 12% relapse rate reported in the HIV-negative setting.10 Outcome of patients with ARL in which histology was not further classifiable was poor with a 5yRFS of only 57%. There was no difference in type and intensity of chemotherapy to that used in DLBCL (data not shown), therefore these cases may represent a subgroup of highly aggressive lymphomas that may benefit from intensive chemotherapy regimens.

Even though the overall survival of patients with PBL was shown to be significantly worse than that in DLBCL and BL,16,17 there was no difference in relapse rates

A

B

C

D

E

F

Figure 3. Kaplan-Meier estimates for aggressive non-Hodgkin lymphoma (NHL) that achieved complete remission (CR) after first-line chemotherapy. (A) Overall survival of all AIDS-related lymphomas (ARL) and of (B) different subtypes (Log rank test: P=0.982). (C) Relapse-free survival of all ARL and of (D) different subtypes (P=0.064). (E) Relapse-free survival of different subtypes treated with R-CHOP-based regimens (rituximab, cyclophosphamide, adriamycin, vincristine, and prednisone) and (F) GMALL-based chemotherapeutic regimens (P=0.006 and P=0.79, respectively). DLBCL: diffuse-large B-cell lymphoma; BL: Burkitt-lymphoma; PBL: plasmablastic lymphoma. Dotted line indicates 3 months.

862

haematologica | 2018; 103(5)

Relapses in AIDS-related lymphomas

between patients who have achieved a CR (12%) to those observed in DLBCL and BL. The inferior OS of the small patient group of PBL may at least in part be explained by 3 late deaths 4-7 years after first diagnosis that were unrelated to lymphoma: one case of sepsis due to pneumonia and 2 cases of secondary malignancies (lung cancer and oral cavity cancer).

AIDS-related lymphoma patients with an intermediate and high IPI had higher relapse rates and a lower 5yRFS than those with a low IPI in univariate analysis. The lack of significance in the multivariate analysis was somewhat surprising as previous studies have demonstrated strong prognostic relevance of the IPI in ARL.17,18 Whether patients with HIV-related DLBCL and intermediate or high IPI may benefit from more intensive treatments such as the CHOEP regimen, as has been shown in the HIVnegative setting, remains to be seen.19,20 Previous studies have shown that concomitant cART was associated with improved CR rates and a trend toward improved OS.21 Our results also support a concurrent use of cART as it was associated with better 5yRFS. Several cART regimens with a good safety and tolerability profile and low interaction potential are now available, strongly arguing for a simultaneous cART during ARL chemotherapy.4 The use of R-CHOP-based regimens showed significantly less treatment delays and reductions, as compared to the GMALL protocol. However, the majority of patients with BL (84%) received chemotherapy according to the GMALL-protocol which resulted in significantly lower relapse rates compared to R-CHOP-based regimens (Figure 3E and F). Notably, treatment delays and a reduced chemotherapy intensity appeared to have no impact on the relapse-rate in GMALL-treated BL, while, at least in the univariate analysis, a reduced number of chemotherapy cycles was associated with lower 5yRFS. Thus, our data indicate that HIV-infected patients with BL should be

References 1. Lewden C, May T, Rosenthal E, et al.

Changes in causes of death among adults infected by HIV between 2000 and 2005: The "Mortalite 2000 and 2005" surveys (ANRS EN19 and Mortavic). J Acquir Immune Defic Syndr. 2008;48(5):590-598.

2. Barta SK, Dunleavy K, Mounier N. Diffuse large B-cell lymphoma. Hentrich M, Barta SK (eds.). HIV-associated Hematological Malignancies. Springer International Publishing; 2016.

3. Dunleavy K, Wilson WH. How I treat HIVassociated lymphoma. Blood. 2012; 119(14):3245-3255.

4. Brunnberg U, Hentrich M, Hoffmann C, Wolf T, Hubel K. HIV-Associated Malignant Lymphoma. Oncol Res Treat.

2017;40(3):82-87.

5. Re A, Michieli M, Casari S, et al. High-dose therapy and autologous peripheral blood stem cell transplantation as salvage treatment for AIDS-related lymphoma: longterm results of the Italian Cooperative Group on AIDS and Tumors (GICAT) study with analysis of prognostic factors. Blood.

2009;114(7):1306-1313.

haematologica | 2018; 103(5)

treated with the planned number of intensive chemotherapy cycles.3,8 By contrast, reduced relative dose intensity did not negatively impact 5yRFS in patients treated with R-CHOP-based regimens for DLBCL. This finding does not correspond to data reported in HIV-negative DLBCL and warrants further investigation.22,23 It is important to note that this analysis focuses on patients in first CR, and that factors that predict RFS were not necessarily associated with initial treatment response.

Our study has several limitations. First, given the uncontrolled design selection biases cannot be ruled out. Second, the analysis of risk factors associated with outcome is more exploratory in nature. Given the relatively low number of patients in some of the selected subgroups, the statistical power of the analysis is limited and does not allow any firm conclusions to be drawn. Notably, data on potential risk factors for lymphoma relapse such as adherence to ART or cumulative viremia between CR and relapse are not available.24 Nevertheless, if a CR has been reached, the relapse rate was low regardless of whether the CR was achieved with or without dose reduction and whether rituximab was used or not. Fourth, CRs were not generally confirmed by negative positron emission tomography (PET) scans as recommended by current guidelines for HIV-negative lymphomas.14,25 However, the role of PETscanning in HIV-lymphoma remains controversial as the rate of false positive results appears to be higher than in the HIV-negative setting.26,27 Finally, the number of patients with ARL, not further classified, may have been lowered by reference pathology services which, in turn, may have slightly altered our findings.

In conclusion, both CR rates and relapse rates observed in the German HIV-related Lymphoma Cohort Study are similar to those reported in HIV-negative NHL. These data add to the growing body of evidence showing that treatment outcomes compare favorably with those in patients with NHL and no HIV infection.

6. Pfreundschuh M, Kuhnt E, Trumper L, et al.

CHOP-like chemotherapy with or without rituximab in young patients with goodprognosis diffuse large-B-cell lymphoma: 6year results of an open-label randomised study of the MabThera International Trial (MInT) Group. Lancet Oncol. 2011; 12(11):1013-1022.

7. Feugier P, Van Hoof A, Sebban C, et al.

Long-term results of the R-CHOP study in the treatment of elderly patients with diffuse large B-cell lymphoma: a study by the Groupe d'Etude des Lymphomes de l'Adulte. J Clin Oncol. 2005;23(18):41174126.

8. Rovira J, Valera A, Colomo L, et al.

Prognosis of patients with diffuse large B cell lymphoma not reaching complete response or relapsing after frontline chemotherapy or immunochemotherapy.

Ann Hematol. 2015;94(5):803-812.

9. Montoto S, Noy A, Ribera JM. Burkitt lymphoma. Hentrich M, Barta SK (eds): HIVassociated Hematological Malignancies.

Springer International Publishing; 2016.

10. Hoelzer D, Walewski J, Dohner H, et al.

Improved outcome of adult Burkitt lymphoma/leukemia with rituximab and chemotherapy: report of a large prospective

11.

12.

13.

14.

15.

16.

multicenter trial. Blood. 2014;124(26):38703879.

Dunleavy K, Pittaluga S, Shovlin M, et al.

Low-intensity therapy in adults with Burkitt's lymphoma. N Engl J Med. 2013; 369(20):1915-1925.

Ribrag V, Koscielny S, Bosq J, et al.

Rituximab and dose-dense chemotherapy for adults with Burkitt's lymphoma: a randomised, controlled, open-label, phase 3 trial. Lancet. 2016;387(10036):2402-2411.

Cheson BD, Horning SJ, Coiffier B, et al.

Report of an international workshop to standardize response criteria for nonHodgkin's lymphomas. NCI Sponsored International Working Group. J Clin Oncol.

1999;17(4):1244.

Cheson BD, Pfistner B, Juweid ME, et al.

Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579586.

Hudis CA, Barlow WE, Costantino JP, et al.

Proposal for standardized definitions for efficacy end points in adjuvant breast cancer trials: the STEEP system. J Clin Oncol.

2007;25(15):2127-2132.

Schommers P, Wyen C, Hentrich M, et al.

Poor outcome of HIV-infected patients with plasmablastic lymphoma: results from

863

P. Schommers et al.

17.

18.

19.

20.

864

the German AIDS-related lymphoma cohort study. Aids. 2013;27(5):842-845.

Schommers P, Hentrich M, Hoffmann C, et al. Survival of AIDS-related diffuse large Bcell lymphoma, Burkitt lymphoma, and plasmablastic lymphoma in the German HIV Lymphoma Cohort. Br J Haematol.

2015;168(6):806-810.

Barta SK, Xue X, Wang D, et al. A new prognostic score for AIDS-related lymphomas in the rituximab-era.

Haematologica. 2014;99(11):1731-1737.

Hentrich M, Hoffmann C, Mosthaf F, et al.

Therapy of HIV-associated lymphoma-recommendations of the oncology working group of the German Study Group of Physicians in Private Practice Treating HIVInfected Patients (DAGNA), in cooperation with the German AIDS Society (DAIG).

Ann Hematol. 2014;93(6):913-921.

Schmitz N, Nickelsen M, Ziepert M, et al.

Conventional chemotherapy (CHOEP-14)

with rituximab or high-dose chemotherapy (MegaCHOEP) with rituximab for young, high-risk patients with aggressive B-cell lymphoma: an open-label, randomised, phase 3 trial (DSHNHL 2002-1). Lancet Oncol. 2012;13(12):1250-1259.

21. Barta SK, Xue X, Wang D, et al. Treatment factors affecting outcomes in HIV-associated non-Hodgkin lymphomas: a pooled analysis of 1546 patients. Blood. 2013; 122(19):3251-3262.

22. Bosly A, Bron D, Van Hoof A, et al.

Achievement of optimal average relative dose intensity and correlation with survival in diffuse large B-cell lymphoma patients treated with CHOP. Ann Hematol. 2008; 87(4):277-283.

23. Hirakawa T, Yamaguchi H, Yokose N, Gomi S, Inokuchi K, Dan K. Importance of maintaining the relative dose intensity of CHOPlike regimens combined with rituximab in patients with diffuse large B-cell lymphoma.

Ann Hematol. 2010;89(9):897-904.

24. Zoufaly A, Stellbrink HJ, Heiden MA, et al.

Cumulative HIV viremia during highly active antiretroviral therapy is a strong predictor of AIDS-related lymphoma. J Infect Dis. 2009;200(1):79-87.

25. Younes A, Hilden P, Coiffier B, et al.

International Working Group consensus response evaluation criteria in lymphoma (RECIL 2017). Ann Oncol. 2017;28(7):14361447 26. Mhlanga JC, Durand D, Tsai HL, et al.

Differentiation of HIV-associated lymphoma from HIV-associated reactive adenopathy using quantitative FDG PET and symmetry. Eur J Nucl Med Mol Imaging. 2014;41(4):596-604.

27. Sathekge M. Differentiation of HIV-associated lymphoma from HIV-reactive adenopathy using quantitative FDG-PET and symmetry. Eur J Nucl Med Mol Imaging. 2014;41(4):593-595.

haematologica | 2018; 103(5)

ARTICLE

Chronic Lymphocytic Leukemia

Highly similar genomic landscapes in monoclonal B-cell lymphocytosis and ultra-stable chronic lymphocytic leukemia with low frequency of driver mutations Andreas Agathangelidis,1* Viktor Ljungström,2* Lydia Scarfò,1 Claudia Fazi,1 Maria Gounari,1,3 Tatjana Pandzic,2 Lesley-Ann Sutton,2,4 Kostas Stamatopoulos,3 Giovanni Tonon,5 Richard Rosenquist2,4** and Paolo Ghia1**

Strategic Research Program on CLL and B-cell Neoplasia Unit, Division of Experimental Oncology, Università Vita-Salute San Raffaele and IRCCS Istituto Scientifico San Raffaele, Milan, Italy; 2Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala University, Sweden; 3Institute of Applied Biosciences, Center for Research and Technology Hellas, Thessaloniki, Greece; 4Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden and 5Functional Genomics of Cancer Unit, Division of Experimental Oncology, IRCCS Istituto Scientifico San Raffaele, Milan, Italy 1

*

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):865-873

AA and VL contributed equally as first authors. **RR and PG contributed equally as last authors

ABSTRACT

D

espite the recent discovery of recurrent driver mutations in chronic lymphocytic leukemia, the genetic factors involved in disease onset remain largely unknown. To address this issue, we performed whole-genome sequencing in 11 individuals with monoclonal Bcell lymphocytosis, both of the low-count and high-count subtypes, and 5 patients with ultra-stable chronic lymphocytic leukemia (>10 years without progression from initial diagnosis). All three entities were indistinguishable at the genomic level exhibiting low genomic complexity and similar types of somatic mutations. Exonic mutations were not frequently identified in putative chronic lymphocytic leukemia driver genes in all settings, including low-count monoclonal B-cell lymphocytosis. To corroborate these findings, we also performed deep sequencing in 11 known frequently mutated genes in an extended cohort of 28 monoclonal B-cell lymphocytosis/chronic lymphocytic leukemia cases.

Interestingly, shared mutations were detected between clonal B cells and paired polymorphonuclear cells, strengthening the notion that at least a fraction of somatic mutations may occur before disease onset, likely at the hematopoietic stem cell level. Finally, we identified previously unreported non-coding variants targeting pathways relevant to B-cell and chronic lymphocytic leukemia development, likely associated with the acquisition of the characteristic neoplastic phenotype typical of both monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia.

Introduction Chronic lymphocytic leukemia (CLL), the most common adult leukemia in the West, is a clinically heterogeneous disease.1 At one end of the spectrum, CLL patients present with an indolent disease that does not require therapy for decades. At the other end of the spectrum, patients experience a rapidly progressive disease, need early treatment, and frequently relapse.2,3 High-throughput studies14,15 have established that, though displaying a markedly lower mutational burden compared to solid tumors,16 CLL is characterized by a diverse genetic landscape with driver gene mutations in pathways considered central for disease pathogenesis, e.g. NOTCH and NF-κB signaling.7,9,17 The frequency of most driver gene mutations in CLL tends to increase in aggressive/refractory cases supporting their involvement mainly in disease progression.18-20 Chronic lymphocytic leukemia is preceded by a condition termed monoclonal B-cell lymphocytosis (MBL) that is characterized by the presence of circulating monoclonal B cells with a CLL phenotype, however, at a lower concentration than Haematologica | 2018; 103(5)

Correspondence: ghia.paolo@hsr.it

Received: August 8, 2017.

Accepted: February 7, 2018.

Pre-published: February 15, 2018. doi:10.3324/haematol.2017.177212 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/865 ©2018 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.

865

A. Agathangelidis et al. required for a clinical diagnosis of CLL (≥5x109/L).21-24 MBL, found in otherwise healthy individuals, is divided into 2 subtypes based on the number of circulating cells: ‘high-count MBL’ (HC-MBL: 0.5-5x109/L) that evolves into CLL requiring therapy at a rate of 1%/year,25 and ‘lowcount MBL’ (LC-MBL: <0.5x109/L) that has not been observed to progress into a clinical disease,26 yet persists over time.26,27 Several typical CLL driver gene mutations have been reported in HC-MBL9,28,29 even years before the transition to CLL,30 and these correlate with adverse disease course.31 Such mutations have been reported in multipotent hematopoietic progenitor CD34+ cells from patients with CLL,32 suggesting that such aberrations may also be implicated in CLL onset.

Here, we aimed to gain insight into the genetic lesions that may be involved in the transformation from MBL to CLL, analyzing LC-MBL cases for the first time. To this end, we used whole-genome sequencing (WGS) and targeted re-sequencing to profile LC-MBL, HC-MBL and a particularly indolent subset of CLL, i.e. patients with ultrastable disease for more than ten years, thus, clinically analogous to MBL. Moreover, in order to explore the possible origin of genetic lesions at the hematopoietic progenitor cell level, we analyzed polymorphonuclear (PMN) cells from the study participants.

We report that the genomic profiles of ultra-stable CLL patients are very similar to individuals with LC-MBL and HC-MBL, characterized by infrequent CLL driver gene mutations that, however, were not associated with disease progression. Furthermore, we observed non-coding variants (NCVs) that target key pathways/cellular processes relevant to normal and neoplastic B-cell development, thus, potentially contributing to the leukemic transformation. We also found shared somatic mutations between MBL/CLL and PMN cells, strengthening the notion that at least a proportion of somatic mutations may occur before the onset of CLL.

DNA Blood Mini kit (Qiagen, Germany) was used for samples with more than 1x106 cells as well as for the buccal samples. DNA quantity and quality were assayed using the Qubit dsDNA HS Assay Kit (Life Technologies, USA).

WGS: library preparation The Nextera technology was utilized for the library construction (Nextera™ DNA Sample Prep Kit, Illumina, USA) as it requires low input material whilst maintaining library complexity.

Fifty ng of genomic DNA were used for the construction of libraries that were sequenced in paired-end mode 2x100bp on a HiSeq 2000 (Illumina, USA).

A variant allele frequency (VAF) of 10% was used as threshold for variant calling. More detailed information regarding the bioinformatics analysis is given in the Online Supplementary Appendix.

Targeted re-sequencing: library preparation Probes targeting all coding exons or hotspot regions of 11 known or postulated CLL driver genes (ATM, BIRC3, MYD88, NOTCH1, SF3B1, TP53, EGR2, POT1, NFKBIE, XPO1, FBXW7) (Online Supplementary Table S1) were designed using Agilent’s SureDesign service (https://earray.chem.agilent.com/suredesign/home. htm). The target regions were captured using the HaloPlex HS targeting enrichment kit (Agilent Technologies, USA). Paired-end sequencing (150 bp reads) was performed on the NextSeq instrument with the use of the 500/550 High Output Kit (Illumina, USA).

Gene enrichment analysis The identification of genes/gene pathways (gene enrichment analysis, GEA) enriched within the targets of NCVs and motifbreaking events caused by NCVs was performed with Enrichr31 using the KEGG 2016 gene database.

Results WGS reveals highly similar mutational profiles in MBL and ultra-stable CLL

Methods The research protocol was approved by the Institutional Ethics Committee and all participants gave written informed consent in accordance with the Declaration of Helsinki.

Study population The study cohort comprised 9 subjects with LC-MBL, 13 subjects with HC-MBL, and 7 patients with Rai stage 0 CLL, herein called ‘ultra-stable’ CLL. Detailed information about the study cohort is provided in the Online Supplementary Appendix.

Cell samples Chronic lymphocytic leukemia cells were stained with antiCD19, anti-CD5 and anti-CD20 antibodies. CD19+CD5+CD20dim cells were sorted using a High Speed FACS Sorter MoFLo (Beckman Coulter) according to previously published methods.26 PMN cells were sorted based on physical parameters. Buccal cells were collected with the use of appropriate buccal swabs (Epicentre, Madison, USA).

DNA extraction The NucleoSpin® Tissue XS kit (Macherey-Nagel, Germany) was used for DNA extraction in samples with less than 5x104 cells and the QIAamp DNA Micro kit (Qiagen, Germany) in samples with cell numbers ranging between 5x104 and 1x106. The QIAamp 866

Whole-genome sequencing was performed on 6 individuals with LC-MBL, 5 individuals with HC-MBL, and 5 patients with ultra-stable CLL. For each individual/patient, samples from MBL/CLL cells and PMN cells were evaluated against buccal (control) cells resulting in a total of 48 samples sequenced with an average autosomal coverage of 32X (Online Supplementary Table S2). Basic demographic and biological characteristics of the MBL/CLL cases included in the WGS analysis are provided in Online Supplementary Table S3. Overall, 37,033 somatic variants were detected in MBL/CLL samples with an average of 2040 somatic variants in LC-MBL (range: 298-2871), 2558 in HC-MBL (range: 1428-3483), and 2400 in CLL (range: 1650-3176), respectively. Notably, 2792 variants were identified in the 15 PMN control samples compared with buccal DNA, with an average of 186 variants/sample (the PMN sample from case CLL_3 was excluded from the analysis due to tumor cell contamination) (Figure 1A).

Highly analogous mutation rates were observed in HCMBL and CLL (0.79 and 0.74 mutations per Mb, respectively), while a slightly lower rate was seen in LC-MBL (0.63 mutations per Mb); this latter finding was due to sample LC-MBL_1 (excluding this sample, the average mutation rate for LC-MBL would have been 0.74 mutations per Mb) (Figure 1B). The ratio of single-nucleotide variants (SNVs)/small indels was again almost identical in haematologica | 2018; 103(5)

WGS in MBL and ultra-stable CLL

the three entities (ranging from 11.4 to 12.6), whereas it was significantly lower in the PMN samples (3.9 for LCMBL and CLL and 4.1 for HC-MBL, respectively) (P<0.005 for all cases) (Figure 1C).

The transition to transversion (Ti/Tv) ratio ranged from 0.99 to 1.13 in the MBL/CLL samples, while it was slightly lower in the PMN samples (0.94) (Figure 2A). No clear differences were observed between MBL/CLL and PMN samples when the distribution of mutations among the six types was examined (Figure 2B). We then evaluated the sequence context of each mutation by incorporating information on the bases immediately upstream and downstream of the mutated base, hence leading to 96 possible (Online mutation types in this classification16 Supplementary Table S4). Almost all major differences were identified between MBL/CLL samples and PMN samples with the former group exhibiting mainly C>T mutations at NpCpG trinucleotides (P<0.05 for ultra-stable CLL and HC-MBL).

The MutationalPatterns package33 was used to delineate the mutational signatures in our cohort. Mutational patterns identified in the MBL/CLL samples resembled those reported by Puente et al.;9 this finding was corroborated by

calculating the pairwise similarity with the 30 previously published signatures, where signature 9 and, to some extent, signature 1 where the main contributors (Figure 2C). The same analysis in the PMN samples gave different results, with a strong impact of mutational signatures 3 and 5 (Figure 2D). Signature 3 had previously been identified in solid tumors and is associated with failure of DNA double-strand break-repair by homologous recombination.16 Signature 5 exhibits transcriptional strand bias for C>T and T>C mutations at ApTpN context and displays a correlation between smoking history and mutation contribution.16

MBL and ultra-stable CLL display a paucity of mutations in putative CLL driver genes Whole-genome sequencing identified 186 non-synonymous exonic variants amongst MBL/CLL samples and 15 amongst PMN samples. The average number was 8.9 for LC-MBL (range: 1-16), 14.8 for HC-MBL (range: 9-27), 11.6 for ultra-stable CLL (range: 7-19), and 0.9 for the PMN samples (range: 0-6), respectively (Figure 3A). In MBL/CLL samples, the vast majority of non-synonymous mutations were missense [LC-MBL: 47 of 53 (88.7%); HC-

A

B

C

Figure 1. Somatic mutational analysis of ultra-stable chronic lymphocytic leukemia (CLL), high-count monoclonal B-cell lymphocytosis (HC-MBL), low-count monoclonal B-cell lymphocytosis (LC-MBL) and control polymorphonuclear (PMN) cell samples. (A) Total number of somatic mutations identified by whole-genome sequencing (WGS) in CLL cell samples from MBL, CLL and the respective PMN samples. All samples carried similar mutational loads with the exception of a single LC-MBL sample (LC-MBL_1) that displayed a very low number of mutation events; as can be seen, the corresponding PMN sample had a mutation load similar to the other PMN samples, where comparison of mutation profiles between the MBL and PMN sample showed few common hits, thus excluding the likelihood of contamination. Concerning PMN control samples, they were also characterized by high homogeneity regarding the mutational load. There was a single sample with a very high mutational load; detailed comparison against its respective CLL sample showed a high overlap of mutations indicating potential tumor cell contamination, hence this sample was removed from downstream analysis. (B) Average mutation rates Âą Standard Deviation (SD) for LC-MBL, HC-MBL and CLL. Highly analogous mutation rates were observed in the HC-MBL (0.79 mutations per Mb) and CLL (0.74 mutations per Mb) samples, while LC-MBL samples had a slightly lower ratio (0.63 mutations per Mb). (C) Average SNV to small indels ratio Âą SD for all sample groups. All 3 entities displayed similar ratios in clear contrast to the PMN samples where the ratio was much lower.

haematologica | 2018; 103(5)

867

A. Agathangelidis et al.

MBL: 61 of 73 (83.7%); ultra-stable CLL: 50 of 59 (84.7%)], whereas the remainder concerned either nonsense mutations or frameshift indels (Figure 3B and Online Supplementary Table S5). Forty-nine of the 186 mutations (26.3%) had a VAF more than 50%. Concerning the 7 PMN samples harboring mutations, only a single mutation (6.7%) had a VAF more than 50% (Figure 3C) (Online Supplementary Table S6). The most commonly mutated gene was IGLL5, in accordance with a recently reported study,6 carrying mutations in 5 different samples (2 LCMBL, 2 HC-MBL and 1 CLL samples), likely introduced by the somatic hypermutation (SHM) process. Only 6 of 186 mutations (1.6%) detected in the MBL/CLL samples concerned putative CLL driver genes, according to 2 recently reported lists.7,9 In detail, 3 were identified in individuals with HC-MBL: i) a single NOTCH1 p.P2514Rfs*4 deletion (VAF 20%), a known hotspot mutation in CLL10,28,34-36 in HC-MBL_4; ii) a single FBXW7 p.W307S mutation (VAF 26%) in HC-MBL_2; and iii) a single KIAA0947 p.L2093X (VAF 43%) in HC-MBL_5. Two mutations concerned individuals with LC-MBL: i) a KLHL6 p.A91D mutation (VAF 45%) in LC-MBL_5; and ii) a single CD79A p.E200G mutation (VAF 53%) in LC-MBL_6. Finally, a CD79B p.N68S mutation (VAF 41%) was identified in a single CLL sample (CLL_5). Although most of these exact mutations have not previously been reported in CLL, functional

prediction using Polyphen-2 classified all but the CD79B mutation as probably damaging. No CLL driver gene mutations were found in the PMN samples.

To assess whether the non-synonymous mutations identified here might be potentially relevant to CLL, we compared our findings to the variants reported by Puente et al.9 and the International Cancer Genome Consortium (ICGC) database.37 Overall, the vast majority of genes carrying mutations in our series were also reported as mutated in either or both datasets: 94% in LC-MBL, 89% in HC-MBL, and 97% in CLL. We extended our analysis by performing targeted resequencing of 11 putative CLL driver genes in 8 LC-MBL, 13 HC-MBL and 7 ultra-stable CLL samples as well as 24 corresponding PMN samples. All but one LC-MBL case (LC-MBL_4) subjected to WGS were included in this analysis (Online Supplementary Table S7). In total, 5 variants were detected in 3 different HC-MBL samples, including 4 missense variants and 1 frameshift deletion. Two variants (targeting the NOTCH1 and FBXW7 genes) had been already identified by WGS, whereas the remaining 3 concerned the POT1 (n=2) and SF3B1 (n=1) genes. In detail, a single HC-MBL case (HC-MBL_5) harbored an SF3B1 mutation (p.K700E; VAF 1.1%), a known hotspot mutation in CLL,18,29,38,39 and a POT1 mutation (p.M1V; 4.3%), while the other POT1 mutation (p.S38R; 6.7%) was found

A

B

C

D

Figure 2. Detailed analysis of mutation types. (A) Transition to transversion (Ti/Tv) ratios were comparable in all monoclonal B-cell lymphocytosis (MBL)/chronic lymphocytic leukemia (CLL) samples and somewhat lower in the polymorphonuclear (PMN) cell samples. (B) Similar distribution of mutations among the 6 mutation classes for each MBL/CLL entity and PMN samples (average values Âą Standard Deviation). Similar profiles were evident for all entities with the G>A mutation predominating in all cases. (C) Mutational signatures that contribute to the somatic mutations observed in the MBL/CLL samples. (D) Mutational signatures that dominate in the PMN samples.

868

haematologica | 2018; 103(5)

WGS in MBL and ultra-stable CLL

in HC-MBL_2, which also harbored the FBXW7 mutation.

Two somatic non-synonymous variants were identified in 2 different PMN samples: an ATM variant (p.R337C; VAF 20%) detected in an LC-MBL case, frequently reported and probably with a low functional impact, and a TP53 mutation (p.G245A, VAF 3%) found in a HC-MBL case, previously reported in several human cancers and lymphomas40 (Online Supplementary Table S8).

Non-coding variants in MBL and ultra-stable CLL target genes in pathways relevant to CLL pathogenesis Coding and non-coding regions enriched for mutations were detected using Fishhook29 in 10 kilobase windows across the genome and with compensation for replication timing. In line with previous findings,9 the analysis revealed highest mutational enrichment in the IG loci and within sites known to be recurrently affected by off-target somatic hypermutation (e.g. BTG2, BCL6 and TCL1A) (Online Supplementary Figure S1).

Funseq2,41 a bioinformatics tool investigating the linkage between NCVs and target genes using integrated bisulfite sequencing, ChIP-Sequencing, and RNA-sequencing data from the Roadmap Epigenomics Project, was used for the examination of the NCVs. This analysis revealed a total of 1517 variants in the MBL/CLL samples and 39 in the PMN samples. After stringent filtering, 106 NCVs of potential relevance to MBL and CLL were identified (Online Supplementary Table S9): 29 in LC-MBL (average 4.8), 45 in HC-MBL (average 9), and 32 in CLL samples (average 6.4), respectively; only 4 NCVs were found in 2 PMN samples. Since we intentionally selected for NCVs in transcription factor (TF) highly occupied regions (see Online Supplementary Appendix), not unexpectedly most variants were located in gene promoter sites (Figure 4A).

Twenty-nine variants (26.4%) concerned 16 cancer-associated genes and were evenly distributed amongst the 3 entities: 9 in LC-MBL samples, 11 in HC-MBL, and 9 in ultra-stable CLL samples. Three of these cancer-associated genes were recurrently targeted: 9 variants concerned the BTG2 gene in 4 samples (2 CLL, 1 HC-MBL and 1 LCMBL), 5 variants involved the BCL6 gene in 2 samples (1 HC-MBL and 1 LC-MBL), and 2 variants targeted the BIRC3 gene in 2 samples (1 CLL and 1 HC-MBL). We also identified 6 variants concerning the ST6GAL1 gene in 3 CLL samples and the same NKIRAS1-related variant in 2 CLL samples (Figure 4B). Moreover, pathway analysis with Enrichr31 showed that 30 of 110 (27.3%) of the variants targeted genes were implicated in key CLL pathways and cellular processes, such as the PI3K-AKT pathway (TCL1A, CCND1, BCL2, PKN1, DDIT4 and SGK3) (P<0.05), the NF-κB pathway (BIRC3, BCL2 and PLAU) (P<0.05) and the spliceosome machinery (DDX46 and HSPA2). In most of these cases (22 of 30, 73.3%), the variants were located at promoter sites. Comparison to the series by Puente et al.9 identified 4 common gene targets: BTG2, BCL6, BACH2 and TCL1A; none of our samples carried variants affecting either the 3’ UTR of the NOTCH1 gene or the PAX5 gene enhancer.

We also analyzed the predicted impact of the NCVs on TF binding and found that 72 of 110 (65.5%) of the variants could result in a motif-breaking event (LC-MBL: n=21, HC-MBL: n=33, ultra-stable CLL: n=18) (Online Supplementary Table S10). We subsequently investigated genes and gene pathways that may be affected by such TF motif breaks. In 55 of 72 (76.4%) cases, variants disrupted haematologica | 2018; 103(5)

a DNase I hypersensitive site, while enrichment analysis of the implicated target genes using Enrichr31 led to the identification of genes participating in pathways relevant to CLL pathogenesis, such as the MAPK, WNT and AP-1 pathways (P<0.0005) (Online Supplementary Table S11).

Moreover, we examined the potential relation of the NCVs that affect TF binding to AID activity by checking if they occurred in the known hotspots (WRCY, RGYW, WA, TW). According to our findings, 21 of 72 (29.2%) NCVs were located at AID hotspots. Gene enrichment analysis of the remaining 51 target genes revealed similar pathways as in the original analysis (namely AP-1 and DNA damage response pathways).

Shared mutations between CLL and polymorphonuclear cells indicate that somatic variants can arise before CLL onset Shared mutations between MBL/CLL samples and their respective PMN samples were identified in all samples irrespective of origin. Regarding exonic mutations, the same synonymous GSE1 mutation was found in an HCMBL case and its paired PMN sample with comparable VAF (28% vs. 26%). In addition, a LC-MBL sample and its paired PMN sample carried an identical mutation within the ncRNA gene LOC339874, though with different VAF (16% vs. 31%). In the case of non-exonic mutations, 179 shared NCVs were identified between MBL/CLL and PMN samples (Online Supplementary Table S12); the average number per sample was 15.8 for LC-MBL, 8.2 for HCMBL, and 9 for ultra-stable CLL (range: 2-34), respectively.

Most of these mutations were intergenic (128 of 179, 71.5%) (Figure 4C). Interestingly, 6 NCVs were recurrently found in more than one MBL/CLL-PMN sample pair: 3 were intergenic and the other 3 were intronic (Online Supplementary Table S13). Finally, we also examined the mutational signatures for shared mutations between MBL/CLL and PMN samples but did not observe clear correlations with any signature (data not shown).

In order to exclude the possibility of contamination of the PMN cell fraction by MBL/CLL DNA, we designed allele-specific primers (Online Supplementary Table S14) and performed PCR amplification of the clonotypic IGH gene rearrangement in both the MBL/CLL and the respective PMN samples in 11 of 16 cases with available material.

We identified the clonotypic rearrangement in all 11 MBL/CLL samples but in none of the corresponding PMN samples examined, effectively ruling out the possibility that the observed results were due to contamination.

Somatic copy-number analysis sCNA analysis was performed in 3 samples from each entity and their respective PMN control samples, as well as in 4 additional PMN samples from 1 HC-MBL and 3 LC-MBL cases. In total, 16 sCNAs were identified in the MBL/CLL samples (average: 1.8, range: 1-6): 7 in LC-MBL, 4 in HC-MBL, and 5 in the ultra-stable CLL samples, all but one concerning deletion events (Online Supplementary Table S15). Of the recurrent cytogenetic aberrations included in the Döhner hierarchical model,42 del(17p), del(11q) and trisomy 12 were not identified in any of the samples, whereas del(13q) was detected in 7 of 9 MBL/CLL cases (2 LC-MBL, 3 HC-MBL and 2 CLL cases).

FISH analysis gave concordant results in 5 of 7 cases where data from both techniques were available; in the remaining 2 cases del(13q) was detected with a single 869

A. Agathangelidis et al.

A

B

C

Figure 3. Exonic mutations in our monoclonal Bcell lymphocytosis (MBL)/chronic lymphocytic leukemia (CLL) cohort and polymorphonuclear (PMN) cell samples. (A) Average numbers of exonic non-synonymous mutations in MBL/CLL entities and PMN samples. (B) The vast majority of non-synonymous mutations were missense in all 3 MBL/CLL entities. PMN samples failed to show such predominance. (C) Average numbers of mutations with VAFâ&#x2030;Ľ50% for all 3 MBL/CLL entities were comparable. A single PMN sample carried a clonal mutation.

technique each (Online Supplementary Table S16). All other sCNAs represented unique events. In terms of distribution across the chromosomes, 7 of the 32 (21.9%) sCNAs were found in the vicinity of centromeres, whereas 11 of 32 (34.4%) were located close to a telomere (distance <10x107 bp). None of the PMN samples demonstrated sCNAs typical of CLL. Only one MBL case showed a shared del(8)(p11.22) between the LC-MBL sample and its paired PMN sample.

Discussion Limited information is available concerning the genomic landscape at the very early or indolent phases of CLL. To this end, we compared the genomes of ultra-stable CLL cases, defined as those cases stable for more than ten years after diagnosis, to genomes from individuals with: i) LCMBL, a condition that does not progress into a clinically relevant leukemia;26 and, ii) HC-MBL, a clinically identifiable pre-leukemic state.25 Both types of MBL and ultra-stable CLL exhibited the same low level of genomic complexity, similar genomewide mutation rates, and average number of exonic mutations, which were distinct from those of the control samples. Reflecting this similarity, analysis relating to published mutational signatures revealed similar patterns in samples from all 3 entities. In more detail, signature 9 that predominated in the MBL/CLL cohort has been previously identified in CLL and B-cell lymphomas and is attributed to polymerase Ρ that is involved in AID-induced 870

somatic hypermutation.16 The second ranking signature 1 is an age-related signature stemming from spontaneous deamination of 5-methylcytosine that has been detected in many cancer types.16 Analogies between MBL and ultrastable CLL extended also to sCNAs in that all samples, irrespective of origin, carried very few sCNA. Del(13q) predominated in all three entities, as shown in previous studies.26 Most of the sCNAs were located in close proximity to either centromeres or telomeres, in keeping with previous findings reporting significant over-representations in these regions due to duplication rates.43 Thus, most of the sCNAs identified here may not be directly related to the MBL/CLL phenotype.

Interestingly, PMN cells harbored a significantly higher load of mutations compared to buccal cells. Mutations detected in the PMN samples were characterized by the dominance of distinct mutational signatures compared to the MBL/CLL cohort. However, these samples carried shared somatic mutations with the respective MBL/CLL cell samples in all analyzed cases. Most shared mutations concerned intergenic regions, yet we also identified a single shared exonic mutation. This finding supports the notion that some mutations present in the CLL clone could be acquired prior to disease onset, as previously suggested.32 Almost all genes that were found mutated in HC-MBL and/or LC-MBL had been previously described as recurrently mutated in CLL.9,37 In contrast to our recent WES study on relapsing CLL,8 where the great majority of cases carried at least one CLL driver mutation, such mutations were relatively scarce in our cohort. Most importantly, the haematologica | 2018; 103(5)

WGS in MBL and ultra-stable CLL

A

B

C

Figure 4. Analysis of non-coding variants and shared mutations between monoclonal B-cell lymphocytosis (MBL)/chronic lymphocytic leukemia (CLL) and polymorphonuclear (PMN) cell samples in the present cohort. (A) Topology of the 106 relevant non-coding variants identified in the present study. The majority concerned gene promoter sites. (B) Recurrent non-coding variants in genes relevant to CLL. The BIRC3, BCL6 and BTG2 genes are known to be associated with various types of cancer. (C) Topology of shared mutations between MBL/CLL and PMN samples. Intergenic mutations predominated, followed by intronic mutations.

lack of any obvious impact of the identified mutations on disease progression after a prolonged follow up highlights the fact that the mere presence of a given driver mutation does not axiomatically equate with disease progression, as previously reported.31,44 Additional studies are required in order to clarify this phenomenon.

Chronic lymphocytic leukemia cases have been shown to harbor detrimental gene mutations in subclones not visible at diagnosis that are progressively selected, e.g. following the use of chemotherapy.45,46 Such mutations were identified by targeted re-sequencing, yet only in a minor fraction of the present cohort. In this context, it has been recently proposed that sequence depths greater than 4000X will be essential in order to robustly identify all subclonal mutations and predict aggressive cases.12 Arguably, the absence of such driver mutations when applying highly sensitive methods may potentially help to identify individuals with very indolent disease for whom less frequent, if any follow up will be warranted.

Whole-genome sequencing revealed that the non-codhaematologica | 2018; 103(5)

ing mutome commonly targets gene pathways and cellular processes involved in CLL pathobiology. A sizeable proportion of variants affected the promoter sites of genes previously associated with cancer, e.g. BTG2, BCL6, BACH2 and TCL1A. Interestingly, all 4 genes have been recently reported as non-IG targets of the SHM process in patients with lymphomas,47 implicating this otherwise normal process in the emergence of CLL-like clones.

Additional studies need to be performed to address the relevance of such mutations; however, since AID-related mutations are common in CLL,16 they may indeed be relevant to disease pathogenesis. We did not identify any “poor-prognostic” 3’ UTR NOTCH1 mutations;9 however, we did discover two NCVs targeting indirectly the BIRC3 gene. We also found a number of variants within the promoter sites of genes implicated in pathways relevant to CLL biology, including the PI3K-AKT and NF-κB pathways as well as the spliceosome machinery.

Furthermore, we identified TF binding motif breaking events that may arise due to NCVs; most concerned the 871

A. Agathangelidis et al.

MAPK, WNT and AP-1 signaling pathways. In this context, preliminary results (data not shown) from our ongoing high-throughput study on aggressive CLL cases showed a great degree of consistency in the targeting of NCVs: the same “CLL-relevant” gene pathways were again among the most common targets of NCVs, further corroborating our present findings. Having said that, some of these variants could represent bystander SHM targets of unknown significance or minor contributors to disease pathogenesis, therefore requiring further studies before definitive conclusions can be drawn regarding their actual significance.

It is important to note that a recent study on the epigenetic profile of CLL48 reported a novel pathogenic role of TF dysregulation in CLL, with increased activity of EGR and NFAT as well as loss of EBF and AP-1, causing imbalances in the normal B-cell epigenetic program.

Interestingly, certain members of these networks (e.g.

EBF1, JUN and FOS) were among the most commonly affected TFs across all sample types tested. Collectively, our findings support the notion that gene pathways could be indirectly targeted by NCVs with the targets being either the genes themselves or other interacting genes, e.g.

TFs. Limitations of the present work involve the relatively small size of the cohort, mainly due to the rarity of samples meeting the selection criteria. In particular, CLL patients had to have stable disease after a prolonged follow up, whereas all individuals with MBL had to have a persistent monoclonal B-cell population. Concerning LCMBL, low CLL cell number was an additional challenging factor. Furthermore, although our targeted re-sequencing approach covered almost 50% of reported mutations in putative driver genes (as reported by Puente et al.),9 by definition this approach is not exhaustive.

References 1. Chiorazzi N, Rai KR, Ferrarini M. Chronic lymphocytic leukemia. N Engl J Med. 2005;352(8):804-815.

2. Parikh SA, Shanafelt TD. Prognostic factors and risk stratification in chronic lymphocytic leukemia. Semin Oncol. 2016; 43(2):233-240.

3. Scarfo L, Ferreri AJ, Ghia P. Chronic lymphocytic leukaemia. Crit Rev Oncol Hematol. 2016;104:169-182.

4. Fabbri G, Khiabanian H, Holmes AB, et al.

Genetic lesions associated with chronic lymphocytic leukemia transformation to Richter syndrome.

J Exp Med. 2013;210(11):2273-2288.

5. Fabbri G, Rasi S, Rossi D, et al. Analysis of the chronic lymphocytic leukemia coding genome: role of NOTCH1 mutational activation. J Exp Med. 2011;208(7):1389-1401.

6. Kasar S, Kim J, Improgo R, et al. Wholegenome sequencing reveals activationinduced cytidine deaminase signatures during indolent chronic lymphocytic leukaemia evolution. Nat Commun. 2016; 6(8866).

7. Landau DA, Tausch E, Taylor-Weiner AN, et al. Mutations driving CLL and their evolution in progression and relapse. Nature.

2015;526(7574):525-530.

872

In summary, we report that MBL and ultra-stable CLL are virtually indistinguishable at the genomic level. While this may be reflective of a passive and slow accumulation of mutations, we identified both exonic and NCV-targeted pathways central for B-cell biology and CLL development, likely linked to the acquisition of the MBL/CLL phenotype. Importantly, ultra-stable CLL cases carried few known driver gene mutations, even after ten years of follow up, perhaps reflecting the central role of microenvironmental signals rather than cell-intrinsic defects in shaping clonal behavior. In other words, cell-extrinsic triggering, specifically mediated through the B-cell receptor, might represent the major driving force in the early stages of CLL, whereas disease progression will require acquisition of genetic driver mutations.

Funding This research project was supported by the Associazione Italiana per la Ricerca sul Cancro, AIRC (Investigator Grant #15189 to PG and Special Program Molecular Clinical Oncology – 5 per mille #9965), Milano, Italy, Ricerca Finalizzata 2010 (#2318823 to PG); Swedish Cancer Society, the Swedish Research Council, Uppsala University, Uppsala University Hospital, Lion’s Cancer Research Foundation, and Selander’s Foundation, Uppsala; and, H2020 “AEGLE, An analytics framework for integrated and personalized healthcare services in Europe”, by the European Union; “MEDGENET, Medical Genomics and Epigenomics Network” (No.692298) by the European Union; “GCH-CLL” funded by the General Secretariat for Research and Technology (GSRT) of Greece and the Italian Ministry of Health (MoH); and IMI2 “HARMONY”, funded by the European Union. AA is a fellow of Associazione Italiana per la Ricerca sul Cancro AIRC (Triennial fellowship “Guglielmina Lucatello é Gino Mazzega”).

8. Ljungstrom V, Cortese D, Young E, et al.

Whole-exome sequencing in relapsing chronic lymphocytic leukemia: clinical impact of recurrent RPS15 mutations.

Blood. 2016;127(8):1007-1016.

9. Puente XS, Bea S, Valdes-Mas R, et al. Noncoding recurrent mutations in chronic lymphocytic leukaemia. Nature. 2015; 526(7574):519-524.

10. Puente XS, Pinyol M, Quesada V, et al.

Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature. 2011;475(7354):101105.

11. Quesada V, Conde L, Villamor N, et al.

Exome sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in chronic lymphocytic leukemia. Nat Genet. 2012;44(1):47-52.

12. Rose-Zerilli MJ, Gibson J, Wang J, et al.

Longitudinal copy number, whole exome and targeted deep sequencing of 'good risk' IGHV-mutated CLL patients with progressive disease. Leukemia. 2016;30(6):13011310.

13. Schuh A, Becq J, Humphray S, et al.

Monitoring chronic lymphocytic leukemia progression by whole genome sequencing reveals heterogeneous clonal evolution patterns. Blood. 2012;120(20):4191-4196.

14. Wang L, Lawrence MS, Wan Y, et al. SF3B1 and other novel cancer genes in chronic

15.

16.

17.

18.

19.

20.

lymphocytic leukemia. N Engl J Med. 2011;365(26):2497-2506.

Burns A, Alsolami R, Becq J, et al. Wholegenome sequencing of chronic lymphocytic leukaemia reveals distinct differences in the mutational landscape between IgHV(mut) and IgHV(unmut) subgroups.

Leukemia. 2018;32(2):332-342.

Alexandrov LB, Nik-Zainal S, Wedge DC, et al. Signatures of mutational processes in human cancer.

Nature.

2013; 500(7463):415-421.

Landau DA, Carter SL, Stojanov P, et al.

Evolution and impact of subclonal mutations in chronic lymphocytic leukemia.

Cell. 2013;152(4):714-726.

Rossi D, Bruscaggin A, Spina V, et al.

Mutations of the SF3B1 splicing factor in chronic lymphocytic leukemia: association with progression and fludarabine-refractoriness. Blood. 2011;118(26):6904-6908.

Rossi D, Fangazio M, Rasi S, et al.

Disruption of BIRC3 associates with fludarabine chemorefractoriness in TP53 wild-type chronic lymphocytic leukemia.

Blood. 2012;119(12):2854-2862.

Schnaiter A, Paschka P, Rossi M, et al.

NOTCH1, SF3B1, and TP53 mutations in fludarabine-refractory CLL patients treated with alemtuzumab: results from the CLL2H trial of the GCLLSG. Blood.

2013;122(7):1266-1270.

haematologica | 2018; 103(5)

WGS in MBL and ultra-stable CLL

21. Landgren O, Albitar M, Ma W, et al. B-cell clones as early markers for chronic lymphocytic leukemia. N Engl J Med. 2009; 360(7):659-667.

22. Marti GE, Rawstron AC, Ghia P, et al.

Diagnostic criteria for monoclonal B-cell lymphocytosis. Br J Haematol. 2005; 130(3):325-332.

23. Dagklis A, Fazi C, Scarfo L, et al.

Monoclonal B lymphocytosis in the general population. Leuk Lymphoma. 2009; 50(3):490-492.

24. Ghia P, Prato G, Scielzo C, et al.

Monoclonal CD5+ and CD5- B-lymphocyte expansions are frequent in the peripheral blood of the elderly. Blood. 2004; 103(6):2337-2342.

25. Rawstron AC, Bennett FL, O'Connor SJ, et al. Monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia. N Engl J Med. 2008;359(6):575-583.

26. Fazi C, Scarfo L, Pecciarini L, et al. General population low-count CLL-like MBL persists over time without clinical progression, although carrying the same cytogenetic abnormalities of CLL. Blood. 2011; 118(25):6618-6625.

27. Matos DM, Furtado FM, Falcao RP. Monoclonal B-cell lymphocytosis in individuals from sporadic (non-familial) chronic lymphocytic leukemia families persists over time, but does not progress to chronic B-cell lymphoproliferative diseases. Rev Bras Hematol Hemoter. 2015;37(5):292295.

28. Rasi S, Monti S, Spina V, et al. Analysis of NOTCH1 mutations in monoclonal B-cell lymphocytosis. Haematologica. 2012; 97(1):153-154.

29. Imielinski M, Guo G, Meyerson M. Insertions and Deletions Target LineageDefining Genes in Human Cancers. Cell.

2017;168(3):460-472 e414.

30. Kreutzer DA, Essigmann JM. Oxidized,

haematologica | 2018; 103(5)

31.

32.

33.

34.

35.

36.

37.

38.

39.

deaminated cytosines are a source of C --> T transitions in vivo. Proc Natl Acad Sci USA. 1998;95(7):3578-3582.

Kuleshov MV, Jones MR, Rouillard AD, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 2016; 44(W1):W90-97.

Damm F, Mylonas E, Cosson A, et al.

Acquired initiating mutations in early hematopoietic cells of CLL patients. Cancer Discov. 2014;4(9):1088-1101.

Blokzijl FJ, R.; van Boxtel, R.; Cuppen, E. MutationalPatterns: comprehensive genome-wide analysis of mutational processes. bioRxiv 071761; doi: https:// doi.org/10.1101/071761.

Baliakas P, Hadzidimitriou A, Sutton LA, et al. Recurrent mutations refine prognosis in chronic lymphocytic leukemia. Leukemia.

2015;29(2):329-336.

Sutton LA, Young E, Baliakas P, et al.

Different spectra of recurrent gene mutations in subsets of chronic lymphocytic leukemia harboring stereotyped B-cell receptors. Haematologica. 2016;101(8):959967.

Lionetti M, Fabris S, Cutrona G, et al. Highthroughput sequencing for the identification of NOTCH1 mutations in early stage chronic lymphocytic leukaemia: biological and clinical implications. Br J Haematol.

2014;165(5):629-639.

Zhang J, Baran J, Cros A, et al. International Cancer Genome Consortium Data Portal--a one-stop shop for cancer genomics data.

Database (Oxford). 2011;2011:bar026.

Strefford JC, Sutton LA, Baliakas P, et al.

Distinct patterns of novel gene mutations in poor-prognostic stereotyped subsets of chronic lymphocytic leukemia: the case of SF3B1 and subset #2. Leukemia. 2013; 27(11):2196-2199.

Jeromin S, Weissmann S, Haferlach C, et al.

40.

41.

42.

43.

44.

45.

46.

47.

48.

SF3B1 mutations correlated to cytogenetics and mutations in NOTCH1, FBXW7, MYD88, XPO1 and TP53 in 1160 untreated CLL patients. Leukemia. 2014;28(1):108117.

Wilda M, Bruch J, Harder L, et al.

Inactivation of the ARF-MDM-2-p53 pathway in sporadic Burkitt's lymphoma in children. Leukemia. 2004;18(3):584-588.

Fu Y, Liu Z, Lou S, et al. FunSeq2: a framework for prioritizing noncoding regulatory variants in cancer. Genome Biol. 2014; 15(10):480.

Dohner H, Stilgenbauer S, Benner A, et al.

Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343(26):1910-1916.

Nguyen DQ, Webber C, Ponting CP. Bias of selection on human copy-number variants.

PLoS Genet. 2006;2(2):e20.

Hurtado AM, Chen-Liang TH, Przychodzen B, et al. Prognostic signature and clonality pattern of recurrently mutated genes in inactive chronic lymphocytic leukemia. Blood Cancer J. 2015;5:e342.

Ojha J, Ayres J, Secreto C, et al. Deep sequencing identifies genetic heterogeneity and recurrent convergent evolution in chronic lymphocytic leukemia. Blood.

2015;125(3):492-498.

Rossi D, Khiabanian H, Spina V, et al.

Clinical impact of small TP53 mutated subclones in chronic lymphocytic leukemia.

Blood. 2014;123(14):2139-2147.

Khodabakhshi AH, Morin RD, Fejes AP, et al. Recurrent targets of aberrant somatic hypermutation in lymphoma. Oncotarget.

2012;3(11):1308-1319.

Oakes CC, Seifert M, Assenov Y, et al.

DNA methylation dynamics during B cell maturation underlie a continuum of disease phenotypes in chronic lymphocytic leukemia. Nature Genetics. 2016;48(3):253264.

873

ARTICLE

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):874-879

Toxicities and outcomes of 616 ibrutinib-treated patients in the United States: a real-world analysis

Anthony R. Mato,1 Chadhi Nabhan,2 Meghan C. Thompson,1 Nicole Lamanna,3 Danielle M. Brander,4 Brian Hill,5 Christina Howlett,6,7 Alan Skarbnik,7 Bruce D. Cheson,8 Clive Zent,9 Jeffrey Pu,10 Pavel Kiselev,11 Andre Goy,7 David Claxton,10 Krista Isaac,12 Kaitlin H. Kennard,1 Colleen Timlin,1 Daniel Landsburg,1 Allison Winter,5 Sunita D. Nasta,1 Spencer H. Bachow,3 Stephen J. Schuster,1 Colleen Dorsey,1 Jakub Svoboda,1 Paul Barr13* and Chaitra S. Ujjani8*

Hematology and Oncology, University of Pennsylvania, Philadelphia, PA; 2Cardinal Health, Dublin, OH 3Hematology/Oncology, Presbyterian/Columbia University Medical Center, New York, NY; 4Hematology/Oncology, Duke University, Durham, NC; 5Hematology and Medical Oncology, Cleveland Clinic, OH; 6Pharmacy, Ernest Mario School of Pharmacy, New Brunswick, NY; 7Hematology/Oncology, John Theurer Cancer Center, Hackensack, NY; 8 Hematology/Oncology, Georgetown University Hospital, Washington DC; 9 Hematology/Oncology, University of Rochester Medical Center, NY; 10Hematology/Oncology, Penn State Milton S Hershey Medical Center, PA; 11Lymphoma, Celgene Corp, Summit, NY; 12 Hematology/Oncology, Lankenau Hospital, Wynnewood, PA and 13Division of Hematology and Oncology, University of Rochester, NY, USA 1

*Indicates shared last authorship

ABSTRACT

C

Correspondence: amato@mskcc.org

Received: October 20, 2017.

Accepted: January 26, 2018.

Pre-published: February 1, 2018.

doi:10.3324/haematol.2017.182907 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/874 ©2018 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.

874

linical trials that led to ibrutinib’s approval for the treatment of chronic lymphocytic leukemia showed that its side effects differ from those of traditional chemotherapy. Reasons for discontinuation in clinical practice have not been adequately studied. We conducted a retrospective analysis of chronic lymphocytic leukemia patients treated with ibrutinib either commercially or on clinical trials. We aimed to compare the type and frequency of toxicities reported in either setting, assess discontinuation rates, and evaluate outcomes. This multicenter, retrospective analysis included ibrutinib-treated chronic lymphocytic leukemia patients at nine United States cancer centers or from the Connect® Chronic Lymphocytic Leukemia Registry. We examined demographics, dosing, discontinuation rates and reasons, toxicities, and outcomes. The primary endpoint was progression-free survival. Six hundred sixteen ibrutinib-treated patients were identified. A total of 546 (88%) patients were treated with the commercial drug. Clinical trial patients were younger (mean age 58 versus 61 years, P=0.01) and had a similar time from diagnosis to treatment with ibrutinib (mean 85 versus 87 months, P=0.8). With a median follow-up of 17 months, an estimated 41% of patients discontinued ibrutinib (median time to ibrutinib discontinuation was 7 months).

Notably, ibrutinib toxicity was the most common reason for discontinuation in all settings. The median progression-free survival and overall survival for the entire cohort were 35 months and not reached (median follow-up 17 months), respectively. In the largest reported series on ibrutinibtreated chronic lymphocytic leukemia patients, we show that 41% of patients discontinued ibrutinib. Intolerance as opposed to chronic lymphocytic leukemia progression was the most common reason for discontinuation. Outcomes remain excellent and were not affected by line of therapy or whether patients were treated on clinical studies or commercially. These data strongly argue in favor of finding strategies to minimize ibrutinib intolerance so that efficacy can be further maximized. Future clinical trials should consider time-limited therapy approaches, particularly in patients achieving a complete response, in order to minimize ibrutinib exposure.

haematologica | 2018; 103(5)

Ibrutinib: treatment toxicities and outcomes

Introduction Ibrutinib is an orally bioavailable, irreversible inhibitor of Bruton tyrosine kinase (BTK). It is a standard of care in the treatment of chronic lymphocytic leukemia (CLL) in the relapsed/refractory1-3 as well as front-line4 settings.

The toxicities of ibrutinib and reasons for its discontinuation were initially defined through several landmark studies comparing ibrutinib to chlorambucil (front-line, RESONATE 2),4 ofatumumab (RESONATE)2 and bendamustine and rituximab +/- ibrutinib versus placebo (HELIOS).5 Higher proportions of patients discontinuing therapy (ranging from 25% to 51%) have been found with long-term follow-up (median follow-ups ranging between 20 months and approximately 5 years) of these and other ibrutinib clinical trials. Although the percentage of discontinuations due to progression of disease (21% to 45%) remained similar, there were high proportions of discontinuations due to adverse events, ranging from 12% to 32%.6-10 Woyach et al. suggested that discontinuation due to CLL progression (distinguished from disease transformation) occurred later in the disease course (cumulative incidence of 7.3% at 2 years, 19.1% at 4 years) while other reasons for discontinuation, such as Richter transformation, peaked early and reached a plateau by 3 years (18.7% at 2 years, 23.9% at 3 years).8 In clinical practice, adverse events were found to be the most common cause of ibrutinib discontinuation among 143 patients, approaching 50%.11 Atrial fibrillation, infectious complications, and cytopenias were the most commonly described adverse events.11 High percentages of patients discontinuing therapy, ranging from 19-41%, were encountered in additional studies conducted in the USA and outside of the USA; however, these studies were limited by relatively small sample sizes and short follow-up periods.12-15 We, therefore, aimed to characterize patterns of care among ibrutinib-treated CLL patients in clinical practice in the USA focusing on rates and reasons for discontinuation, and how these affect outcomes. To our knowledge, this series is the largest report on ibrutinib-treated CLL patients.

Methods We conducted a multicenter, retrospective cohort study of CLL patients at nine USA cancer centers and from the Connect® CLL Registry (199 USA centers, 80% community sites) who were treated with ibrutinib either as part of a clinical trial or with commercially available drug.16 The institutional review board of each participating institution approved this study. Investigators at each institution were asked to utilize chart review, institutional electronic medical records and clinical/pathological databases to obtain required information for all CLL patients treated with ibrutinib. Data collected included: patients’ demographics, genetic characteristics, number of prior therapies, dosing and dose adjustments, discontinuation rates and reasons, toxicities and outcomes.

The period of enrollment was January 2014 to August 2016.

The primary study endpoint was progression-free survival, which was defined as time from ibrutinib treatment to progression or death from any cause as per the Kaplan Meier method.17 Patients were otherwise censored, regardless of progression status, at the time of last follow-up and at the time of next therapy. When interpreting medical records, investigators were advised to use the haematologica | 2018; 103(5)

International Workshop on Chronic Lymphocytic Leukemia (IWCLL) criteria to define response and progression of disease.18,19 Patients were stratified by line of therapy (front-line versus relapsed/refractory), reason for discontinuation (intolerance versus progressive disease), clinical trial participation versus commercial use, and depth of response (complete response versus partial response and partial response with lymphocytosis). Secondary endpoints included overall survival and reasons for ibrutinib discontinuation. Overall survival was defined as the time in months from initiation of ibrutinib to death.

Reasons for ibrutinib discontinuation were categorized as follows: toxicity, progressive disease, Richter transformation to either diffuse large B-cell lymphoma or Hodgkin lymphoma, planned cellular therapy (allogeneic hematopoietic stem cell transplantation or chimeric antigen receptor genetically modified T-cell therapy), secondary malignancies, physician’s or patient’s preference, financial concerns, and other/unrelated death. Toxicities leading to discontinuation were categorized as: hematologic toxicity, infection, atrial fibrillation, congestive heart failure, druginduced pneumonitis, drug-induced colitis, transaminitis, bleeding, arthralgia/myalgia, dermatological, neurotoxicity, other, or unknown.

Survival data were compared using the log-rank test.20 Univariate Cox regression analyses were used to estimate hazard ratios.21 All other comparison analyses were descriptive. All tests were two-sided at the 5% level. Statistical analyses were performed using STATA 10.1 (Stata Statistical Software: Release 10.

2007; StataCorp LP, College Station, TX, USA).

Results Patients We identified 616 patients who received ibrutinib, including 536 relapsed-refractory and 80 previously untreated patients. Data from the nine contributing academic centers were collected retrospectively and included information for 399 patients treated with ibrutinib; data from the Connect CLL registry were collected prospectively and included information on 217 patients largely collected from community sites (80% community). The patients’ baseline characteristics stratified by line of therapy are available in Table 1. A total of 546 (88%) patients were treated with commercially available drug/off study.

Clinical trial patients were younger (mean age 58 versus 61 years, P=0.01), had a similar time from diagnosis to treatment with ibrutinib (mean 85 versus 87 months, P=0.8) and were more consistently initiated at a dose of 420 mg daily (100% versus 89%).

Reasons for discontinuation, toxicities and timing of events At a median follow-up of 17 months (range, 1-60 months), 41% of patients discontinued ibrutinib. The median time to ibrutinib discontinuation was 7 months (range, 0.1–41), with a median time of 6 months for patients who discontinued due to intolerance and 10 months for those who discontinued due to progression of disease (Online Supplementary Figure S1). Among patients on ibrutinib monotherapy, 41% discontinued therapy while the percentage of discontinuation among patients receiving ibrutinib-based combination therapy was 43.9%. Ibrutinib starting dose (420 mg daily versus <420 mg daily) did not correlate with the proportion of patients who discontinued ibrutinib due to toxicity (51% versus 875

A.R.â&#x20AC;&#x2C6;Mato et al.

50%) or disease progression (19.6% versus 21.4%).

Reasons for ibrutinib discontinuation are listed in Table 2. Percentages listed indicate the proportion of discontinuations due to each category. Toxicity was the most common reason for discontinuation in all settings, accounting for 63.1% of discontinuations in front-line use (n=12/80 front-line patients) and 50.2% of discontinuations in relapsed/refractory use (n=116/536 relapsed/refractory patients). Toxicity was the most common reason for discontinuation in several settings including: commercial use and clinical trial use (50% of discontinuations in front-line commercial use, 77.7% of discontinuations in front-line clinical trial use, 52.5% of discontinuations in relapsed/refractory commercial use, and 39.7% of discontinuations in relapsed/refractory trial use). Notably, the proportion of discontinuations due to progressive disease was lower: 15.8% in the front-line setting and 20.9% in relapsed/refractory use. Richter transformation to diffuse large B-cell lymphoma or Hodgkin lymphoma accounted

for 5.3% of the discontinuations in the front-line setting and 5.0% in the relapsed/refractory setting.

Among the patients treated front-line with ibrutinib, the three most common toxicities leading to discontinuation were arthralgia (41.6%), atrial fibrillation (25%), and rash (16.7%). In the relapsed/refractory population, the most common toxicities leading to discontinuation were atrial fibrillation (12.3%), infection (10.7%), pneumonitis (9.9%), bleeding (9%) and diarrhea (6.6%).

The median time to ibrutinib discontinuation varied by toxicity: bleeding (8 months), diarrhea (7.5 months), atrial fibrillation (7 months), infection (6 months), arthralgia (5 months), pneumonitis (4.5 months) and rash (3.5 months).

Outcomes At a median follow-up of 17 months, the median progression-free and overall survival for the entire cohort were 35 months and not reached, respectively (Figure 1A,B). Overall survival from the start of ibrutinib therapy,

Table 1. Baseline characteristics.

Baseline characteristics Total number Median age at diagnosis, years (range) Median time from diagnosis to ibrutinib start, months (range) del17p(+) del11q(+) Tp53mut(+) Complex karyotype (+) (>3) Median time diagnosis to ibrutinib, months (range) Median ibrutinib starting dose Ibrutinib administered as monotherapy Ibrutinib suspension required Ibrutinib dose adjusted Median follow-up

Ibrutinib in front line

Ibrutinib in relapse

80 62 (37-88), n=78 25.5 (1-232), n=80 37%, n=76 19%, n=75 12%, n=42 40%, n=60 26 (1-232) 420 mg 68%, n=80 30%, n=79 15%, n=79 17 months

536 60 (22-95), n=532 78.3 (1-366), n=536 26%, n=368 35%, n=367 13%, n=142 34%, n=214 78 (1-660) 420 mg 89%, n=536 37%, n=310 20%, n=309

del17p(+) (deletion 17p mutation positive); del11q(+) (deletion 11q mutation positive); Tp53mut(+) (Tp53 mutation positive); Complex karyotype (+)(>3) (complex karyotype positive defined as three or more abnormalities).

A

B

Figure 1. Outcomes for the entire cohort. Kaplan Meier curves at a median follow-up of 17 months showing (A) progression-free survival (PFS) for the entire cohort and (B) overall survival (OS) for the entire cohort.

876

haematologica | 2018; 103(5)

Ibrutinib: treatment toxicities and outcomes

stratified by whether the drug was being used in the front-line versus relapsed/refractory setting, is shown in Online Supplementary Figure S2A,B. Notably, there was no significant difference in progression-free survival by front-line versus relapsed/refractory use (P=0.27, log-rank test) (Figure 2A) or at first, second or third relapse (P=0.45) (Online Supplementary Figure S3). Progression-free survival was similar when stratified by ibrutinib use in the clinical practice setting as compared to the clinical trial setting (P=0.14, log-rank test) (Figure 2B). Patients who discontinued due to toxicity had significantly longer progression-free survival and overall survival than those who discontinued due to disease progression (P=0.01 and P=0.02, respectively, log-rank test) (Figure 2C,D).

Investigator-assessed depth of response (complete response versus partial response versus partial response with lymphocytosis versus stable disease versus progressive disease) appeared to correlate with a longer progression-free survival (Figure 2E). We also stratified progression-free survival by deletion 17p status and complex karyotype status (â&#x2030;Ľ3 abnormalities) in CLL patients treated in the relapsed/refractory setting. Progression-free survival was not significantly different in patients with deletion 17p (P=0.70), but was significantly shorter in patients with a complex karyotype (hazard ratio=1.8, 95% confidence interval: 1.1-3.0, P=0.01). The Kaplan Meier curves for these analyses are shown in Online Supplementary Figure S4A-C.

A

B

C

D

E

Figure 2. Outcomes stratified by line of therapy, clinical trial participation, reason for discontinuation and depth of response. Kaplan Meier curves showing outcomes stratified by (A) line of therapy (progression-free survival), (B) clinical trial participation (progression-free survival), (C) reason for discontinuation (progression-free survival), (D) reason for discontinuation (overall survival), and (E) depth of response. CR: complete response; PR: partial response; PR-L: partial response with lymphocytosis; SD: stable disease; PD: progressive disease.

haematologica | 2018; 103(5)

877

A.R. Mato et al.

Discussion In the largest reported series of ibrutinib-treated CLL patients so far, we found that the median progression-free survival was 35 months. Interestingly, this outcome was comparable between previously untreated and relapsed/refractory patients. Our observed progressionfree survival of 35 months is shorter than that previously described in clinical trials in which the median progression-free survival was 52 months in relapsed/refractory disease.9 These observations suggest that outcomes vary when comparing clinical trial patients and those treated in clinical practice, underscoring the need to better understand outcomes and toxicities in a real-world setting.

Patient-specific factors, including molecular prognostic markers, performance status and prior therapies, may partially account for these discrepancies. Our front-line cohort of patients had a greater number of molecular abnormalities than typically seen in the treatment-naïve population; 37% of patients had del17p and 40% had a complex karyotype. Both features are associated with an inferior progression-free survival.22 For the relapsed cohort, patients may have received prior idelalisib which could have affected their subsequent response to ibrutinib; notably, this was prohibited in the earliest clinical trials of ibrutinib.11 At a median follow-up of 17 months, overall discontinuation in our study was high at 41%, suggesting that ibrutinib discontinuation is an emerging issue in clinical practice. This mirrors the high overall discontinuation patterns seen in longer follow-up studies of clinical trial patients,6-9 particularly the 51% estimated discontinuations seen at a median follow-up of 3.4 years by Woyach et al.8 In addition, 15% of front-line and 20% of relapsed ibrutinibtreated patients required a dose reduction, which is significantly higher than the 4% of patients requiring dose reduction due to adverse events in the RESONATE trial due to adverse events (4%).2 Early ibrutinib trials suggested that progressive disease was the cause of the majority of cases of discontinuation,2 a pattern that has persisted in many,6,8,9 but not all,5,7,10 studies including longer follow-up periods. Surprisingly, our results demonstrate that intolerance (50.2% in the relapsed/refractory setting, 63% in the front-line setting), rather than progressive disease (21% in the relapsed/refractory setting, 16% in the front-line setting), accounts for the majority of cases of discontinuation.

Similar findings were previously noted in a smaller series reported by our group.11 With respect to adverse events leading to discontinuation, these values are greater than those reported in a pooled analysis from the phase III relapsed RESONATE and front-line RESONATE-2 studies in which the discontinuation rate was 10%.23 Because at least 50% of discontinuations are due to intolerance rather than progression of disease, it is unlikely that these patients harbor ibrutinib resistance mutations and, therefore, should likely be sensitive to alternate kinase inhibitors with different side effect profiles. Two clinical trials are being conducted at this time studying umbralisib (NCT02742090) and acalabrutinib (NCT02717611) in kinase inhibitor-intolerant patients.

For the first time we have established a timeline associated with time to discontinuation due to specific ibrutinibrelated toxicities. Similar to the experience with idelalisib where specific toxicities appear to occur after different 878

Table 2. Reasons for Ibrutinib discontinuation.

Reason for ibrutinib discontinuation

Ibrutinib in Ibrutinib in relapse front-line (n=19) (n=231)

Toxicity CLL progression Other/unrelated death Physician’s or patient’s preference RT DLBCL Stem cell transplantation/CAR T-cell Financial concerns Secondary malignancy RT Hodgkin lymphoma

63.1% (n=12) 15.8% (n=3) 5.3% (n=1) 10.5% (n=2) 5.3% (n=1) 0 0 0 0

50.2% (n=116) 20.9% (n=49) 12.1% (n=28) 6.7% (n=15) 4.6% (n=10) 3.3% (n=8) 0.8% (n=2) 0.8% (n=2) 0.4% (n=1)

CLL: chronic lymphocytic leukemia; RT DLBCL: Richter transformation to diffuse large B-cell lymphoma; CAR T-cell: chimeric antigen receptor T-cell); RT: Richter transformation.

periods of time on treatment,24 these data may be of value in developing monitoring strategies for specific toxicities and educational material related to ibrutinib toxicities.

Reasons for discontinuation also varied by line of therapy. Our analysis revealed a strikingly higher number of discontinuations due to toxicity in the front-line setting; 63% compared to the previously reported 9% in RESONATE-2.4 Similar findings, although not as marked, were noted in the relapsed setting; 50% of patients in our analysis discontinued therapy due to intolerance compared to 12% discontinuing due to adverse events in the 4-year follow-up of RESONATE data.6 In line with prior reports, we found that atrial fibrillation and pulmonary complications were common reasons for discontinuation.

Arthralgias and rash were frequently noted as well, particularly in treatment-naïve patients. The higher discontinuation rate for toxicity may reflect lack of physicians’ comfort in toxicity management, a higher incidence of toxicity in clinical practice, differences in patients’ comorbidities and age, or a lower threshold for discontinuation given an increasing number of available alternative treatment choices. This may be in contrast to the limited number of therapies available to ibrutinib-treated patients in early clinical trials. In a recent series by Lampson et al. of treatment-naïve patients treated with idelalisib plus ofatumumab adverse events (particularly liver toxicity) were the most common reasons for discontinuation.25 This study suggested that younger patients’ age and intact immunity may lead to autoimmune treatment-related toxicities in treatment-naïve patients.25 Progression of disease was the second most common indication for ibrutinib discontinuation. Sixteen percent of previously untreated patients experienced progression compared to 10% progressions/deaths at an 18-month follow-up in the front-line RESONATE-2 study.4 This difference may be related to the exclusion of patients with del17p from the RESONATE-2 study.4 Rates of discontinuation due to progression in the relapsed population in our cohort were more comparable with the 27% found in the 4-year follow-up of the RESONATE study.6 Outcomes of patients in our series who discontinued therapy due to toxicity were superior to those of patients who discontinued due to progressive disease.

We also considered limitations in the study design.

Conducted by physicians and research coordinators from several institutions in a retrospective manner, data collechaematologica | 2018; 103(5)

Ibrutinib: treatment toxicities and outcomes

tion may have been affected by inconsistencies in chart interpretation, as well as clinical experience and practice style. There were a disproportionate number of relapsed/refractory patients compared to front-line patients.

It is vital to develop strategies to mitigate ibrutinib intolerance so that efficacy can be further maximized.

Examples include the creation of guidelines for the evaluation and management of problematic side effects such as atrial fibrillation, rash, and arthralgias. An educational forum focused on oncologists, physician educators, and nurses should be implemented. In addition, the design of future clinical trials should allow for cessation of therapy in order to minimize ibrutinib exposure, particularly in the

References 1. Byrd J, Furman R, Coutre S, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-34.

2. Byrd J, Brown J, O'Brien S, et al. Ibrutinib versus ofatumumab in previously treated chronic lymphoid leukemia. N Engl J Med. 2014;371(3):213-223.

3. O'Brien S, Jones J, Coutre S, et al. Ibrutinib for patients with relapsed or refractory chronic lymphocytic leukaemia with 17p deletion (RESONATE-17): a phase 2, openlabel, multicentre study. Lancet Oncol.

2016;17(10):1409-1418.

4. Burger J, Tedeschi A, Barr P, et al. Ibrutinib as initial therapy for patients with chronic lymphocytic leukemia N Engl J Med. 2015;373(25):2425-2437.

5. Chanan-Khan A, Cramer P, Demirkan F, et al. Ibrutinib combined with bendamustine and rituximab compared with placebo, bendamustine, and rituximab for previously treated chronic lymphocytic leukaemia or small lymphocytic lymphoma (HELIOS): a randomised, double-blind, phase 3 study.

Lancet Oncol. 2016;17(2):200-211.

6. Byrd J, Hillmen P, O’Brien S, et al. Long-term efficacy and safety with ibrutinib (ibr) in previously treated chronic lymphocytic leukemia (CLL): up to four years follow-up of the RESONATE study. J Clin Oncol.

2017;35(suppl; abstr 7510).

7. Maddocks K, Ruppert A, Lozanski G, et al.

Etiology of ibrutinib therapy discontinuation and outcomes in patients with chronic lymphocytic leukemia. JAMA Oncol. 2015; 1(1):80-87.

8. Woyach J, Ruppert A, Guinn D, et al.

BTKC481S-mediated resistance to ibrutinib in chronic lymphocytic leukemia. J Clin Oncol. 2017;35(13):1437-1443.

9. O’Brien S, Furman R, Coutre S, et al. Fiveyear experience with single-agent ibrutinib in patients with previously untreated and relapsed/refractory chronic lymphocytic leukemia/small Lymphocytic Leukemia.

Blood. 2016; 128:233.

haematologica | 2018; 103(5)

small subset of patients who achieve complete remission.

This strategy has been successfully demonstrated in patients receiving venetoclax who were able to achieve minimal residual disease negativity.26 For example, the incorporation of BCL-2 inhibitors and/or anti-CD20 monoclonal antibody therapies in combination with ibrutinib may enable patients to experience minimal residual disease-negative responses that may translate into shorter durations of treatment.27,28 Acknowledgments The authors thank Joseph and Cindy Riggs for their ongoing support of this work. They also thank the Center for CLL, University of Pensylvania.

10. Jain P, Thompson P, Keating M, et al. Longterm outcomes for patients with chronic lymphocytic leukemia who discontinue ibrutinib. Cancer. 2017;123(12): 2268-2273.

11. Mato A, Nabhan C, Barr P, et al. Outcomes of CLL patients treated with sequential kinase inhibitor therapy: a real world experience. Blood. 2016;128(18):2199-2205.

12. Parikh S, Chaffee K, Call T, et al. Ibrutinib therapy for chronic lymphocytic leukemia (CLL): an analysis of a large cohort of patients treated in routine clinical practice.

Blood. 2015;126(23):2935.

13. Sandoval-Sus J, Chavez J, Dalia S, et al.

Outcomes of patients with relapsed/refractory chronic lymphocytic leukemia after ibrutinib discontinuation outside clinical trials: a single institution experience. Blood.

2015;126:2945.

14. Winqvist M, Asklid A, Andersson P, et al.

Real-world results of ibrutinib in patients with relapsed or refractory chronic lymphocytic leukemia: data from 95 consecutive patients treated in a compassionate use program. Haematologica. 2016;101(12): 15731580.

15. UK CLL Forum.

Ibrutinib for relapsed/refractory chronic lymphocytic leukemia: a UK and Ireland analysis of outcomes in 315 patients. Haematologica.

2016;101(12):1563-1572.

16. Mato A, Nabhan C, Kay N, et al. Real-world clinical experience in the Connect chronic lymphocytic leukaemia registry: a prospective cohort study of 1494 patients across 199 US centres. Br J Haematol. 2016;175(5):892903.

17. Bland J, Altman D. Survival probabilities (the Kaplan-Meier method).

BMJ. 1998;317(7172):1572.

18. Hallek M, Cheson B, 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. 2008;111(12):54455456.

19. Cheson B, Byrd J, Rai K, et al. Novel targeted agents and the need to refine clinical end

20.

21.

22.

23.

24.

25.

26.

27.

28.

points in chronic lymphocytic leukemia. J Clin Oncol. 2012;30(23):2820-2822.

Matthews D, Farewell V. 7 The log-rank or Mantel-Haenszel test for the comparison of survival curves in: Basel S, Karger A, editors.

Using and Understanding Medical Statistics 2007:67-75.

Anderson P, Gill R. Cox’s regression model for counting processes: a large sample study.

Ann Statist. 1982;10(4):1100-1120.

Thompson PA, O'Brien SM, Wierda WG, et al. Complex karyotype is a stronger predictor than del(17p) for an inferior outcome in relapsed or refractory chronic lymphocytic leukemia patients treated with ibrutinibbased regimens. Cancer. 2015;121(20): 36123621.

O'Brien SM, Byrd JC, Hillmen P, et al.

Outcomes with ibrutinib by line of therapy in patients with CLL: analyses from phase III data. J Clin Oncol. 2016;34(15_suppl): 7520.

Coutré SE, Barrientos JC, Brown JR, et al.

Management of adverse events associated with idelalisib treatment: expert panel opinion. Leuk Lymphoma. 2015;56(10): 27792786.

Lampson B, Kasar S, Matos T, et al. Idelalisib given front-line for treatment of chronic lymphocytic leukemia causes frequent immune-mediated hepatotoxicity. Blood 2016;128(12):195-203.

Seymour JF, Ma S, Brander DM, et al.

Venetoclax plus rituximab in relapsed or refractory chronic lymphocytic leukaemia: a phase 1b study. Lancet Oncol. 1017;18(2): 230-240.

Sharman J, Brander D, Mato A, et al.

Ublituximab and ibrutinib for previously treated genetically high-risk chronic lymphocytic leukemia: results of the GENUINE phase 3 study.

J Clin Oncol.

2017;35(Supplemental, ASCO abstract 7504).

Hillmen P, Rawstron A, Munir T, et al. The initial report of bloodwise tap clarity study combining ibrutinib and venetoclax in relapsed, refractory CLL shows acceptable safety and promising early indications EHA Oral Presentation June 22-25, Madrid Spain.

2017.

879

ARTICLE

Plasma Cell Disorders

Ferrata Storti Foundation

A novel nano-immunoassay method for quantification of proteins from CD138-purified myeloma cells: biological and clinical utility

Irena Misiewicz-Krzeminska,1,2,3 Luis Antonio Corchete,1,2 Elizabeta A. Rojas,1,2 Joaquín Martínez-López,4 Ramón García-Sanz,1,2,5 Albert Oriol,7 Joan Bladé,8 Juan-José Lahuerta,6 Jesús San Miguel,9 María-Victoria Mateos1,2,5 and Norma C. Gutiérrez1,2,5

Haematologica 2018 Volume 103(5):880-889

Cancer Research Center-IBMCC (USAL-CSIC), Salamanca, Spain, 2Institute of Biomedical Research of Salamanca (IBSAL), Spain; 3National Medicines Institute, Warsaw, Poland; 4 Hematology Department, Hospital 12 de Octubre, CNIO, Complutense University, CIBERONC, Madrid, Spain; 5Hospital Universitario de Salamanca, CIBERONC, Spain; 6 Hospital 12 de Octubre, Madrid, Spain; 7Hospital Germans Trias i Pujol, Barcelona, Spain; 8 Hospital Clinic, Barcelona, Spain and 9Clínica Universidad de Navarra, CIMA, IDISNA, CIBERONC, Pamplona, Spain 1

ABSTRACT

P

doi:10.3324/haematol.2017.181628

rotein analysis in bone marrow samples from patients with multiple myeloma has been limited by the low concentration of proteins obtained after CD138+ cell selection. A novel approach based on capillary nano-immunoassay could make it possible to quantify dozens of proteins from each myeloma sample in an automated manner. Here we present a method for the accurate and robust quantification of the expression of multiple proteins extracted from CD138-purified multiple myeloma samples frozen in RLT Plus buffer, which is commonly used for nucleic acid preservation and isolation. Additionally, the biological and clinical value of this analysis for a panel of 12 proteins essential to the pathogenesis of multiple myeloma was evaluated in 63 patients with newly diagnosed multiple myeloma. The analysis of the prognostic impact of CRBN/Cereblon and IKZF1/Ikaros mRNA/protein showed that only the protein levels were able to predict progression-free survival of patients; mRNA levels were not associated with prognosis.

Interestingly, high levels of Cereblon and Ikaros proteins were associated with longer progression-free survival only in patients who received immunomodulatory drugs and not in those treated with other drugs. In conclusion, the capillary nano-immunoassay platform provides a novel opportunity for automated quantification of the expression of more than 20 proteins in CD138+ primary multiple myeloma samples.

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/880

Introduction

Correspondence: normagu@usal.es

Received: September 27, 2017.

Accepted: January 31, 2018.

Pre-published: March 15, 2018.

©2018 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.

880

Genomics has come to dominate biomedical research in recent years. For example, high-throughput genomic technologies have been used for the comprehensive analysis of multiple myeloma (MM). In particular, gene expression profiling has enabled the molecular classification of MM, which is widely used in biological research.1 However, a knowledge of protein expression is essential for identifying therapeutic targets, since proteins are the molecules through which most new drugs achieve their efficacy. The limited amount of sample remaining after plasma cell purification means that messenger RNA (mRNA) quantification is still used as an indirect measure of protein expression in most cases. However, several studies have shown that protein levels cannot be predicted from mRNA measurements.2 Immunohistochemistry and flow cytometry have been used to analyze expression at the protein level, although to a limited extent. These methods are of great value and are of proven clinical utility, but they have some limitations that make them less useful for studying intracellular protein levels. They mostly use directly haematologica | 2018; 103(5)

Nano-scale protein quantification in multiple myeloma

marked antibodies that reduce the sensitivity of detection, and fewer antibodies are available for these techniques, even when used in an indirect assay.3 Immunohistochemistry allows only semiquantitative analysis of protein expression, and requires a well-trained pathologist to interpret the results. Moreover, neither technique is able to identify non-specific antibody binding to other proteins.3 Western blotting (WB) remains the “gold standard” technique for protein characterization in most laboratories.

However, WB consumes large quantities of reagents, has a low throughput, and requires a great deal of time and effort involving many laborious manual processing steps.

Moreover, WB only yields semiquantitative data of poor repeatability, making it a challenge to go beyond using the assay in discovery research to apply it reliably in the clinical setting.4–6 A further drawback is that it is not always possible to obtain the quantity of protein extract required for WB from primary cancer samples.

MM is a clear prototype of a bone marrow-infiltrating tumor for which a relatively small quantity of sample is available after the diagnostic procedure, which involves morphological evaluation, immunophenotypic characterization by flow cytometry, and CD138+ plasma cell separation for routine fluorescence in situ hybridization analysis.

The recent development of a method based on the combination of capillary nano-electrophoresis with immunoassay (CNIA), also known as ‘simple western’, requires only very small amounts of sample to be able to measure protein expression.3,7 This technical advance makes it possible to analyze the expression of 50-100 proteins in a single MM sample. Here we present the results of a pilot study using this platform in MM patients. The main goal was to quantify accurately and robustly the proteins extracted from CD138-purified MM samples frozen in RLT Plus buffer, which is commonly used as a method for RNA and DNA preservation. Additionally, we attempted to establish the clinical value of this analysis using a panel of proteins essential to MM pathogenesis, comparing it with that of the corresponding RNA expression.

extracted separately. After overnight incubation at -20ºC, the proteins were centrifuged at 13,000 x g for 30 min at 4ºC, and washed twice with 70% ice-cold ethanol followed by centrifugation for a further 10 min. The protein precipitate was dried at 39ºC and dissolved with 50 μL 0.2 M NaOH for 10 min at room temperature and 4x WB sample buffer for at least 15 min at room temperature.

Samples were then denatured at 95ºC for 5 min, cooled to room temperature and stored in aliquots at -80ºC. Before any assay, samples were heated to room temperature, then kept at 37ºC for 30 min in order to re-dissolve any protein that had precipitated during freezing.

Capillary electrophoresis immunoassay Capillary electrophoresis immunoassay or simple western analysis was performed using the WESTM machine (ProteinSimple, San Jose, CA, USA) in accordance with the manufacturer’s protocols. The Total Protein Assay (ProteinSimple) was used to quantify the protein concentration. In brief, 5 μL of proteins were loaded on the plate, separated by size, labeled with a biotin reagent and detected by chemiluminescence using streptavidin-horseradish peroxidase. At the end of the run, the proportion of the protein of interest in the total protein in the sample was measured, in comparison to a standard curve previously generated using JJN3 cell line extracts of known protein concentrations.

Primary antibodies used in the study and the optimized conditions for each antibody are presented in Table 1. Data were analyzed using CompassTM software. Each protein peak was measured automatically and normalized with respect to the GAPDH median area under the peak. Expression of each protein is presented as its abundance relative to GAPDH.

Methods For more specific information see the Online Supplementary File.

Patients and multiple myeloma cell lines Sixty-three samples from patients diagnosed with MM between October 2013 and November 2015 were included in the study (Online Supplementary Table S1). Forty-three had been enrolled in two Spanish Myeloma Group clinical trials: GEM2010 [bortezomib/melphalan/prednisone and lenalidomide/dexamethasone in a sequential or alternating manner; (n=24)] and BenVelPres [bendamustine/bortezomib/prednisone; (n=19)]. The other 20 patients were not treated as part of a clinical trial.

The impact of RNA and protein expression on patients’ survival was evaluated only in the group of patients that took part in the clinical trials (Online Supplementary Table S1). The scheme of the study is presented in Figure 1.

Protein extraction from RLT Plus buffer Proteins were extracted by ice-cold acetone precipitation from RNA-column flow-through liquid. To increase the rate of protein precipitation 10 mM NaCl was added to the acetone at 80% (v/v).

For technical reasons, each sample was divided into two tubes and haematologica | 2018; 103(5)

Figure 1. Scheme of the study.

881

I. Misiewicz-Krzeminska et al.

DNA/RNA extraction and quantitative real-time polymerase chain reaction analysis mRNA expression was evaluated by Taqman assay quantitative real-time polymerase chain reaction (qRT-PCR) analysis using the respective GAPDH Taqman assay as a control, by the 2-ΔCt method.

Statistical analysis Spearman correlations were calculated. Progression-free survival (PFS) was calculated for each gene and protein. Survival curves were plotted using the Kaplan–Meier method and statistical significance was evaluated with the log-rank test.

Results Protein extraction from RLT Plus buffer results in optimal quality and quantity Firstly, we evaluated the amount and quality of the protein extracted with our protocol. The data generated by the WESTM system were visualized as virtual blots (Figure 2A) or peaks (Figure 2B) that were quantified as the area under the curve using the inbuilt algorithm of the CompassTM software. Using JJN3 myeloma cell line lysates of known concentration, we generated a protein standard curve that proved to be linear over the evaluated range of concentrations (Figure 2C). The amount of protein obtained from each sample ranged between 0.00 and 0.36 mg protein, with a median quantity of 0.06 mg per sample. Three of the 63 samples had insufficient material to analyze protein expression (Figure 2D). We compared the expression of the various proteins extracted from MM cells stored in RLT Plus buffer with that obtained using the standard RIPA protocol and found the signals to be similar for the two protocols (Figure 2E,F).

Optimization of protein quantification by capillary nano-electrophoresis with immunoassay For each analyzed protein, we first searched in the ProteinSimple antibody database for the optimized conditions (http://www.proteinsimple.com/antibody/antibodies.html). If the antibody was present, we re-evaluated it in our system, using the antibody at the indicated concentra-

tion and at double and half the indicated concentration. In the event that the protein evaluated was not present in the database, we performed a full optimization, which consisted of running the assays in the cell line samples at two concentrations (0.1 mg/mL and 0.2 mg/mL) with at least five antibody dilutions in order to determine the optimal concentration at which the antigen-antibody binding was saturated and no change in antibody concentration influenced the result. The optimized concentrations for each antibody, the molecular weights at which the peaks were observed, and the coefficients of variation arising from the validation of each protein are shown in Table 1.

Standard curves were produced for each protein to evaluate the range of linearity over which the expression of each protein could be quantified. Briefly, each capillary contained the sample at a different dilution, and the protein detection was visualized as virtual blots, as exemplified by the use of Aiolos in Figure 3A. The peaks obtained for each dilution, which were obtained automatically by the program, have a distinct height and width, depending on the sample dilution (Figure 3B). Once they had been quantified the standard curve was generated (Figure 3C).

After protein quantification, we compared the value obtained for each sample and each protein with the respective standard curve to ensure correct measurement.

The limit of quantitation was set as signal-to-noise ratio of 10:1 in accordance with the guidelines from the European Directorate for the Quality of Medicine set out in the European Union Pharmacopoeia.8 The results of Aiolos quantification in six samples are shown in Figure 3D, in which virtual blots for both Aiolos and GAPDH are presented.

Analysis of mRNA and protein expression We analyzed the expression of 12 genes and their encoded proteins, together with GAPDH as a control (Figure 1). We decided to select proteins involved in MM or cancer pathogenesis: Cyclin D1 and Cyclin D2, whose overexpression is a unifying event for most MM;9–11 c-myc, which is consistently found to be involved in the transformation of monoclonal gammopathy of undetermined significance into MM;12,13 HSP90, which is upregulated in

Table 1. Summary of proteins and antibodies used in the study.

1 2 3 4 5 6 7 8 9 10 11 12 13

Target protein

Company

Cat number

Species

Clonality

Dilution Ab

MW peak

Intra-assay CV (%)

Aiolos Calnexin Cereblon c-myc Cyclin D1 Cyclin D2 DDX21 HSP90 Ikaros PSME1 RIPK1 XAF1 GAPDH

Cell Signaling Enzolifesciences NovusBio Cell Signalling Abcam Cell Signaling Abcam Cell Signaling Santa Cruz NovusBio Cell Signaling Cell Signaling Cell Signaling

12720 ADI-SPA-860 NBP1-91810 5605 Ab134175 3741 Ab182156 4877 Sc-13039 NBP1-83121 3493 13805 2118

Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit

Polyclonal Polyclonal Polyclonal Monoclonal Monoclonal Monoclonal Monoclonal Monoclonal Polyclonal Polyclonal Monoclonal Monoclonal Monoclonal

1:100 1:250 1:80 1:50 1:50 1:100 1:100 1:50 1:50 1:50 1:50 1:25 1:50

85 119 59 75 40 38 110 95 70 37 79 46, 110 42

10.9 4.9 9.7 8.3 9.9 7.8 11.4 7.0 10.8 10.2 5.8 15.5 8.6

Ab: antibody; MW: molecular weight; CV: coefficient of variation.

882

haematologica | 2018; 103(5)

Nano-scale protein quantification in multiple myeloma

many solid and hematologic malignancies, including MM;14 Calnexin, which forms endoplasmic reticulum and is upregulated in MM relative to normal plasma cells in genetically identical twin samples;15 and DDX21 or RIPK1, with known involvement in several tumors.16â&#x20AC;&#x201C;18 In addition, proteins involved in the mechanism of action of antimyeloma drugs were included: Cereblon, Ikaros, Aiolos for immunomodulatory drugs; XAF1 for melphalan; and PSME1 for bortezomib.19â&#x20AC;&#x201C;22 At the protein level, PSME1 and Calnexin showed the highest median level of expression, while HSP90 was the most strongly expressed mRNA (Figure 4A,B). Conversely, proteins Cyclin D1, Cyclin D2 and c-myc had the lowest

A

B

median level of expression, and CRBN, RIPK1 and XAF1 were the least expressed mRNA. In general, the expression of mRNA was more homogenous than that of proteins, as indicated by the higher coefficients of variation for the latter (Figure 4C). In fact, the coefficients of variation were significantly lower than those for c-myc, DDX21, HSP90, IKZF1 and PSME1 mRNA than for their respective encoded proteins. The highest variability in expression, both at the mRNA and protein levels, was observed for CCND2/Cyclin D2 and CCND1/Cyclin D1, as well as for c-myc protein.

Next, we analyzed the correlation between the two levels of gene expression, mRNA and protein. Interestingly,

D

C

E

F

Figure 2. Optimization of protein extraction from RLT Plus samples. Due to the various additives in the sample buffer there is marked incompatibility with most of the normally used protein quantification methods. The Total Protein assay was therefore used, as it is insensitive to high SDS concentrations. The standard curve was generated using JJN3 cell line extracts at 0.25 mg/mL concentration, and serial dilutions thereof. Each capillary contained one sample of a known concentration.

Results were visualized as virtual gels (A) and the numbers correspond to the areas under the curves of the peaks (B). A standard curve was generated, plotting the result for each capillary quantification (C). Amount of protein obtained from each sample (D). Comparison of results from Calnexin, Cyclin D2, GAPDH and Ikaros quantification in the U266 cell line, from which protein extracts were obtained by RLT Plus and RIPA extraction, and visualized as virtual blots or two distinct dilutions of the sample (E), and one dilution extracted with both protocols visualized as peaks (F).

haematologica | 2018; 103(5)

883

I. Misiewicz-Krzeminska et al.

only Cyclin D1 and Cyclin D2 protein levels were strongly correlated with the respective CCND1 and CCND2 mRNA levels (Figure 4D). We observed a modest correlation for Aiolos, Calnexin and DDX21 proteins with their respective mRNA. Although the number of proteins analyzed was limited, we examined the correlation between the levels of the different proteins. A positive correlation was observed between most of the protein pairs (Online Supplementary Figure S2). We confirmed the previously described relationship between Ikaros/Aiolos and c-myc.19 Additionally, cmyc protein expression was positively correlated with Cereblon, Calnexin, and RIPK1, and negatively correlated with DDX21 (Online Supplementary Figure S2). We found that protein levels of Cereblon, Ikaros and Aiolos, all of which are required for the activity of immunomodulatory drugs, were correlated with each other (Online Supplementary Figure S2).

The potential association between the expression of proteins and mRNA tested in the study and chromosomal abnormalities was also explored. We confirmed the wellestablished pattern of CCND1/Cyclin D1 and CCND2/Cyclin D2 expression in t(11;14) and t(4;14) (Online Supplementary Figure S3). A lower level of expression of PSME1 and RIPK1 proteins in MM with 1q gains, and a higher level of IKZF1 mRNA expression in MM with t(11;14) were also observed.

A

B

Influence of mRNA and protein levels on survival of myeloma patients Since clinical data were available for 43 MM patients, 24 enrolled in GEM 2010 and 19 in BenVelPres clinical trials, we also performed survival analysis for proteins and mRNA using PFS as the endpoint (Table 2). Cereblon and Ikaros were the only proteins able to predict PFS. Interestingly, mRNA levels of CRBN and IKZF1 were not associated with prognosis (Figure 5). Accordingly, patients with a high level of Cereblon protein had a longer PFS than those with a low level (50.4 versus 16.3 months, P<0.001). Similarly, high levels of Ikaros protein were associated with longer PFS (45.1 versus 17.8 months, P<0.01). The levels of two mRNA were associated with longer PFS: a high level of PSME1 (50.4 versus 23.5 months, P<0.05) and a low level of XAF1 (20.3 versus 45.1 months, P<0.05).

Since Cereblon, Ikaros and Aiolos are involved in the mechanism of action of immunomodulatory drugs, and only GEM2010 patients were treated with lenalidomide, we examined whether the prognostic value of these proteins was influenced by the type of treatment. Indeed, high levels of Cereblon and Ikaros were both associated with longer PFS only in patients who received immunomodulatory drugs and not in those treated with other drugs (Figure 6).

D

C

Figure 3. Optimization of protein expression quantification. For each protein the standard curve was generated using the sample with the strongest signal to prepare the serial dilutions. Each dilution was run in a separate capillary. The sample result for Aiolos is visualized as a virtual blot (A) or as peaks (B). The standard curve established the linear range for each protein (C). The sample result for Aiolos together with the respective GAPDH for each sample was run in separate capillaries (D). Each peak area was quantified and Aiolos was normalized with respective to GAPDH, as shown in the example.

884

haematologica | 2018; 103(5)

Nano-scale protein quantification in multiple myeloma

Discussion MM has been comprehensively studied at the DNA and RNA levels using high-throughput technologies such as microarrays for detecting copy number abnormalities, and gene expression profiling, and, more recently, next-generation sequencing for DNA mutation analysis. MM sam-

A

ples after CD138+ separation are usually stored in buffers such as TRIZOL or RLT Plus, which preserve nucleic acids for subsequent use in genomic studies. While it is technically possible to extract proteins from these buffers,23â&#x20AC;&#x201C;25 the quantity of protein would not be sufficient for multiple WB to be carried out. In the classic WB, the amount of the purified plasma cells, even if all of it were available for the

C

B

D

Figure 4. Two levels of analysis of each gene RNA and protein. mRNA expression of each gene was assessed by qRT-PCR and normalized relative to GAPDH and expressed as 2-Î&#x201D;Ct (A). Abundance of each protein was assessed by CNIA and normalized relative to GAPDH abundance in each case (B). The Y axis of graphs (A) and (B) are expressed on a log scale. The variability of each mRNA and protein measurement in the analyzed population of patients with MM, measured as percentage coeffcient of variation (CV%). The threshold of statistical significance (*P<0.05) was determined as described in the Methods section (C). Spearman correlation coefficient for each mRNA/protein pair ranked by increasing P value (D).

haematologica | 2018; 103(5)

885

I. Misiewicz-Krzeminska et al.

Table 2. Univariate analysis of progression-free survival.

mRNA

HR (CI)

Group

n

IKZF3

1.84 (0.74-4.6) 0.44 (0.18-1.11) 0.49 (0.14-1.67) 2.55 (0.91-7.14) 0.38 (0.13-1.05) 1.54 (0.51-4.64) 0.49 (0.2-1.17) 0.52 (0.22-1.25) 1.93 (0.77-4.83) 0.33 (0.12-0.86) 0.52 (0.22-1.24) 2.39 (0.98-5.8)

H L H L H L H L H L H L H L H L H L H L H L H L

10 33 21 22 10 33 27 16 17 26 33 10 29 14 24 19 25 18 18 25 28 15 11 32

CANX1 CRBN CCND1 CCND2 MYC DDX21 HSP90 IKZF1 PSME1 RIPK1 XAF1

Median PFS (months) P-value 21.7 45.1 50.4 23.5 50.4 29.8 29.8 NR NR 29.8 30 45.1 43.8 23.5 50.4 23.5 29.8 43.8 50.4 23.5 50.4 23.5 20.3 45.1

Protein

HR

Group (CI)

0.19

Aiolos

0.074

Calnexin

0.24

Cereblon

0.066

CyclinD1

0.053

CyclinD2

0.44

c-myc

0.1

Ddx21

0.14

Hsp90

0.15

Ikaros

0.018

Psme1

0.13

Ripk1

0.048

Xaf1

0.51 (0.19-1.34) 1.82 (0.72-4.6) 0.23 (0.09-0.55) 1.55 (0.64-3.78) 0.24 (0.05-1.03) 0.63 (0.26-1.56) 0.47 (0.15-1.46) 1.5 (0.53-4.27) 0.31 (0.12-0.77) 0.48 (0.19-1.26) 1.41 (0.53-3.73) 1.56 (0.63-3.85)

H L H L H L H L H L H L H L H L H L H L H L H L

n Median PFS (months) P-value 33 10 22 18 31 12 23 20 11 32 27 13 14 23 12 25 32 11 30 10 11 29 15 25

45.1 23.5 30 43.8 50.4 16.3 30 NR NR 29.8 45.1 29.8 50.4 30 29.8 50.4 45.1 17.8 45.1 16.3 30 45.1 30 45.1

0.16 0.2 0.00034 0.33 0.051 0.31 0.18 0.44 0.0075 0.13 0.49 0.34

HR: hazard ratio; CI: confidence interval; PFS: progression-free survival; H: high level; L: low level.

protein studies, would have allowed at most six proteins to be evaluated, the median amount of sample being sufficient to analyze two proteins.

Here we present for the first time a method for quantifying the expression of multiple proteins from myeloma cells stored in the buffer commonly used for nucleic acid preservation and isolation. We report a protocol for protein extraction from MM cells stored in RLT Plus buffer based on a well-known acetone precipitation procedure.26,27 We decided to add NaCl to the sample before precipitation, since a greater inorganic salt content is known to improve protein yield.28 After testing several types of salts and concentrations, we chose the one with the best performance. We also assessed several methods of protein pellet dissolution, finally settling for a 0.2 M NaOH and 4x WB sample buffer, since slight changes in the pH of the environment change protein solubility.26,29 In contrast to the classic WB, which provides only semiquantitative (blot-based) results, CNIA quantifies the area under the curve of the signal in each capillary, enabling expression relative to the control protein to be calculated.3,30 To determine whether the CNIA method is suitable for evaluating the expression of multiple proteins, using various antibodies, we tested the performance of each protein in the WESTM system. We first optimized the concentration of the antibody to be employed using RIPAextracted proteins from MM cell lines, as suggested by the system provider. The antibody dilution used has to be the one that saturates the epitope-antibody binding, so that the additional increase in antibody concentration would not have caused the increase in the signal. Although the 886

concentration of antibodies used by the CNIA platform is higher than that used in WB, lower amounts of antibody are required because of the small volume of antibody.

Comparing the signal detected by each antibody when the sample was extracted by the RIPA method with that obtained using our protocol revealed no significant differences for any of the proteins evaluated, which supports the suitability of the present protocol for extracting proteins from RLT Plus buffer. We observed differences between the predicted molecular weight and that detected by the CNIA system for some proteins, regardless of the extraction method. The most probable explanation for this phenomenon is that migration depends on the mobility in the matrix.31 In fact, each system provides a unique molecular weight value that depends on the particular interactions between the matrix and protein, and the true molecular weight can only be determined by mass spectrometry or sequencing.32 In our assay, we analyzed only the data obtained from the antibodies that detected peaks whose signal intensity was linear in the serial dilutions, enabling the frequent biases in WB to be eliminated.

Although such standard curves are not usually calculated as part of the WB method, this is highly recommendable.33 Even though, at first glance, the optimization step required for each antibody seems laborious, one CNIA run allows 24 samples to be analyzed, so it would be possible to optimize, for example, four antibodies in 3 hours.

Therefore, bearing in mind the subsequent possibility of quantifying protein abundances, it is not such a time-consuming process.

After demonstrating that our approach accurately quanhaematologica | 2018; 103(5)

Nano-scale protein quantification in multiple myeloma

B

A

Figure 5. Progression-free survival according to levels of MRNA and protein expression.

Progression-free survival in patients with low and high levels of mRNA (A) and protein (B) expression. The log-rank test was performed for each gene and protein and Kaplan-Meier curves represent the PFS of MM patients depending on mRNA and protein status.

Cutoff Finder software (http://molpath.charite.de/cut off) was used to obtain the optimal cutoff, which was defined as that producing the most significant split that discriminates between good and poor survival by examining all the possible cutoffs using the log-rank test.

tified the proteins extracted at the same time as the DNA and RNA from the RLT Plus buffer, we investigated the applicability of the method to the analysis of the expression of key proteins in MM biology, such as D Cyclins, cmyc, Cereblon, Ikaros, and Aiolos, among others. We also wanted to compare protein expression with the corresponding mRNA level, since many basic studies have revealed that only 30-40% of protein abundance can be explained by the mRNA level.34 Our results showed a moderate or low correlation between mRNA and protein levels of expression, and are consistent with the general observation that ~60% of the variation in protein concentration cannot be explained by measuring mRNA alone.34 We also observed that the mRNA level was less variable than protein expression among MM patients for all the mRNA and protein pairs analyzed. Indeed, protein abundance is regulated by a variety of complex mechanisms, such as post-transcriptional and post-translational modifications, and protein degradation regulation.34,35 By measuring mRNA abundance, only the early steps in a long chain of regulatory events are considered.36 However, the mRNA level is still often employed as a proxy for protein abundance, mostly because of the lack of appropriate technology to quantify proteins quickly and efficiently.

Our results reproduce the well-known pattern of CCND1/Cyclin D1 and CCND2/Cyclin D2 expression in MM with t(11;14) and t(4;14).9 We also found a correlation between c-myc and Ikaros and Aiolos levels, analyzing either mRNA or protein expression, consistent with the previously demonstrated regulation of c-myc by Ikaros and Aiolos in MM.19 Interestingly, the correlation between Ikaros and Aiolos levels was stronger for the protein than for the mRNA. To our knowledge, this is the first time that the protein levels of c-myc, Ikaros and haematologica | 2018; 103(5)

Aiolos have been quantified and the relationship between their expressions analyzed in MM. In T-cell acute lymphoblastic leukemia, for example, the levels of mRNA encoding Ikaros and Aiolos were weakly, but significantly correlated.37 Among the proteins included in our study, we observed a significant association between protein level and PFS for Cereblon and Ikaros, while this association was not observed for the respective mRNA levels. Cereblon forms an E3 ubiquitin ligase complex together with the damaged DNA binding protein 1 (DDB1), Cullin4A (CUL4) and Roc1. Immunomodulatory drugs, such as lenalidomide or pomalidomide, bind to Cereblon in a region located at the C-terminus of this protein.38,39 Our results did not demonstrate a correlation between Cereblon protein and mRNA level, and showed that only high levels of Cereblon protein were associated with a good prognosis in MM. These findings are concordant with those of previous studies and support the usage of protein expression to evaluate Cereblon levels.40 Several independent groups have identified Ikaros and Aiolos as the downstream targets of Cereblon after immunomodulatory drug activation.41â&#x20AC;&#x201C;43 The role of the level of Ikaros in MM survival is controversial. When Ikaros expression was investigated at the RNA level, a low level of mRNA IKZF1 expression was associated with better prognosis in newly diagnosed patients treated with immunomodulatory drugs.44 On the other hand, low IKZF1 levels were found to predict a lack of responsiveness to immunomodulatory drugs and a shorter overall survival in refractory MM patients.45 We observed that a high level of Ikaros protein was associated with longer PFS, while no significant impact on prognosis was observed when PFS was estimated from mRNA levels.

887

I. Misiewicz-Krzeminska et al.

Figure 6. Progression-free survival in patients with low and high Cereblon, Aiolos and Ikaros protein levels, depending on the treatment scheme (only patients treated according to GEM2010 trial received lenalidomide). The log-rank test was performed for each protein and Kaplan-Meier curves represent progressionfree survival of MM patients depending on protein status.

Cutoff Finder software (http://molpath.charite.de/cut off) was used to obtain the optimal cutoff, which was defined as that producing the most significant split that discriminates between good and poor survival by examining all the possible cutoffs using the log-rank test.

These results are consistent with the longer survival displayed by relapsed/refractory MM patients treated with lenalidomide who expressed high levels of IKZF1/3 protein, as evaluated by immunohistochemical staining.46 Although the number of patients analyzed in this study is relatively small, the survival analysis carried out dividing patients according to drug therapy showed that high levels of Cereblon and Ikaros proteins are associated with a longer PFS only in patients who receive immunomodulatory drugs and not in those who are treated with other drugs. Other studies reached the same conclusion that the level of Cereblon can predict the outcome of patients with MM mainly in those treated with immunomodulatory drugs.47–50 By contrast, in the present series of MM patients, the level of Aiolos did not influence the outcome of the patients treated with immunomodulatory drugs.

In summary, we present the implementation of a novel technique based on capillary nano-immunoassay for quantifying protein expression in MM samples in the clinical setting. The requirement for only a relatively small 888

amount of material means that, for the first time, more than 20 proteins can be analyzed using the same sample frozen for DNA and RNA analysis. This makes the CNIA platform a fast, effective and accurate tool for exploring the impact of different proteins on the survival of patients with MM and for investigating new protein biomarkers that could help to predict the response to new drugs that directly target specific proteins. These encouraging results require further validation in a larger cohort of patients with MM or other hematologic malignancies.

Acknowledgments The authors thank Isabel Isidro, Teresa Prieto and Vanesa Gutierrez for their technical assistance.

Funding This work was funded by a grant from the International Myeloma Foundation's Black Swan Research Initiative® and “Gerencia Regional de Salud, Junta de Castilla y León” (BIO/SA35/14). WES™ platform was acquired thanks to INNOCAMPUS Program (CEI10-1-0010). haematologica | 2018; 103(5)

Nano-scale protein quantification in multiple myeloma

References 1. Broyl A, Hose D, Lokhorst H, et al. Gene expression profiling for molecular classification of multiple myeloma in newly diagnosed patients. Blood. 2010;116(14):2543– 2553.

2. Tian Q, Stepaniants SB, Mao M, et al.

Integrated genomic and proteomic analyses of gene expression in mammalian cells. Mol Cell Proteomics. 2004;3(10):960–969.

3. Chen J-Q, Wakefield LM, Goldstein DJ. Capillary nano-immunoassays: advancing quantitative proteomics analysis, biomarker assessment, and molecular diagnostics. J Transl Med. 2015;13:182.

4. Chen J-Q, Heldman MR, Herrmann MA, et al. Absolute quantitation of endogenous proteins with precision and accuracy using a capillary western system. Anal Biochem.

2013;442(1):97–103.

5. Ghosh R, Gilda JE, Gomes AV. The necessity of and strategies for improving confidence in the accuracy of western blots. Expert Rev Proteomics. 2014;11(5):549–560.

6. Gassmann M, Grenacher B, Rohde B, Vogel J. Quantifying western blots: pitfalls of densitometry. Electrophoresis. 2009;30(11): 1845–1855.

7. Fan AC, Deb-Basu D, Orban MW, et al.

Nanofluidic proteomic assay for serial analysis of oncoprotein activation in clinical specimens. Nat Med. 2009;15(5):566–571.

8. Technical guide for the eleboration of monographs.

9. Bergsagel PL, Kuehl WM, Zhan F, Sawyer J, Barlogie B, Shaughnessy J. Cyclin D dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood.

2005;106(1):296–303.

10. Chesi M, Bergsagel PL. Molecular pathogenesis of multiple myeloma: basic and clinical updates. Int J Hematol. 2013;97(3):313–323.

11. Misiewicz-Krzeminska I, Sarasquete ME, Vicente-Dueñas C, et al. Post-transcriptional modifications contribute to the upregulation of Cyclin D2 in multiple myeloma. Clin Cancer Res. 2016;22(1):207–217.

12. Chiecchio L, Dagrada GP, Protheroe RKM, et al. Loss of 1p and rearrangement of MYC are associated with progression of smouldering myeloma to myeloma: sequential analysis of a single case. Haematologica.

2009;94(7):1024–1028.

13. Chng W-J, Huang GF, Chung TH, et al.

Clinical and biological implications of MYC activation: a common difference between MGUS and newly diagnosed multiple myeloma. Leukemia 2011;25(6):1026–1035.

14. Zhang L, Fok JHL, Davies FE. Heat shock proteins in multiple myeloma. Oncotarget.

2014;5(5):1132–1148.

15. Munshi NC, Hideshima T, Carrasco D, et al.

Identification of genes modulated in multiple myeloma using genetically identical twin samples. Blood. 2004;103(5):1799–1806.

16. Cimino D, Fuso L, Sfiligoi C, et al.

Identification of new genes associated with breast cancer progression by gene expression analysis of predefined sets of neoplastic tissues. Int J Cancer. 2008;123(6):1327–1338.

17. Schneider AT, Gautheron J, Feoktistova M, et al. RIPK1 suppresses a TRAF2-dependent pathway to liver cancer. Cancer Cell.

2017;31(1):94–109.

18. Yoo JY, Jaime-Ramirez AC, Bolyard C, et al.

haematologica | 2018; 103(5)

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Bortezomib treatment sensitizes oncolytic HSV-1-treated tumors to NK cell immunotherapy. Clin Cancer Res. 2016;22 (21):5265–5276.

Bjorklund CC, Lu L, Kang J, et al. Rate of CRL4(CRBN) substrate Ikaros and Aiolos degradation underlies differential activity of lenalidomide and pomalidomide in multiple myeloma cells by regulation of c-Myc and IRF4. Blood Cancer J. 2015;5e354.

Kortüm KM, Zhu YX, Shi CX, Jedlowski P, Stewart AK. Cereblon binding molecules in multiple myeloma.

Blood Rev. 2015;29(5):329–334.

Park GB, Kim Y-S, Kim D, et al. Melphalaninduced apoptosis of EBV-transformed B cells through upregulation of TAp73 and XAF1 and nuclear import of XPA. J Immunol. 2013;191(12):6281–6291.

Dytfeld D, Luczak M, Wrobel T, et al.

Comparative proteomic profiling of refractory/relapsed multiple myeloma reveals biomarkers involved in resistance to bortezomib-based therapy.

Oncotarget.

2016;7(35):56726-56736.

Chomczynski P. A reagent for the singlestep simultaneous isolation of RNA, DNA and proteins from cell and tissue samples.

BioTechniques. 1993;15(3):532–534, 536– 537.

Vorreiter F, Richter S, Peter M, Baumann S, von Bergen M, Tomm JM. Comparison and optimization of methods for the simultaneous extraction of DNA, RNA, proteins, and metabolites. Anal Biochem. 2016;50825–33.

Mathieson W, Thomas GA. Simultaneously extracting DNA, RNA, and protein using kits: is sample quantity or quality prejudiced? Anal Biochem. 2013;433(1):10–18.

Pace CN, Treviño S, Prabhakaran E, Scholtz JM. Protein structure, stability and solubility in water and other solvents. Philos Trans R Soc Lond B Biol Sci. 2004;359(1448):12251234; discussion 1234-1235.

Race J. The determination of blood-proteins by acid-acetone.

Biochem J. 1932;26(5):1571–1584.

Crowell AMJ, Wall MJ, Doucette AA. Maximizing recovery of water-soluble proteins through acetone precipitation. Anal Chim Acta. 2013;79648–54.

Cutler P, editor. Protein Purification Protocols. 2nd. Totowa, NJ: Humana Press; 2004. 484 p.

Rustandi RR, Loughney JW, Hamm M, et al.

Qualitative and quantitative evaluation of SimonTM, a new CE-based automated Western blot system as applied to vaccine development. Electrophoresis. 2012;33(17): 2790–2797.

Copeland RA. Methods for Protein Analysis.

Boston, MA: Springer US. Strupat K. Molecular weight determination of peptides and proteins by ESI and MALDI. Meth Enzymol. 2005;405:1–36.

McDonough AA, Veiras LC, Minas JN, Ralph DL. Considerations when quantitating protein abundance by immunoblot. Am J Physiol Cell Physiol. 2015;308(6):C426-433.

Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet. 2012;13(4):227–232.

Liu Y, Beyer A, Aebersold R. On the Dependency of cellular protein levels on mRNA abundance. Cell. 2016;165(3):535– 550.

36. Payne SH. The utility of protein and mRNA correlation. Trends Biochem Sci. 2015;40 (1):1–3.

37. Mitchell JL, Yankee TM. Variations in mRNA and protein levels of Ikaros family members in pediatric T cell acute lymphoblastic leukemia. Ann Transl Med. 2016;4(19):363–363.

38. Ito T, Ando H, Suzuki T, et al. Identification of a primary target of thalidomide teratogenicity. Science. 2010;327(5971):1345– 1350.

39. Ito T, Ando H, Handa H. Discovery of the target for immunomodulatory drugs (IMiDs). Rinsho Ketsueki. 2016;57(5):556– 562.

40. Gandhi AK, Mendy D, Waldman M, et al.

Measuring cereblon as a biomarker of response or resistance to lenalidomide and pomalidomide requires use of standardized reagents and understanding of gene complexity. Br J Haematol. 2014;164(2):233–244.

41. Lu G, Middleton RE, Sun H, et al. The myeloma drug lenalidomide promotes the cereblon-dependent destruction of Ikaros proteins. Science. 2014;343(6168):305–309.

42. Krönke J, Udeshi ND, Narla A, et al.

Lenalidomide causes selective degradation of IKZF1 and IKZF3 in multiple myeloma cells. Science. 2014;343(6168):301–305.

43. Zhu YX, Braggio E, Shi C-X, et al. Cereblon expression is required for the antimyeloma activity of lenalidomide and pomalidomide.

Blood. 2011;118(18):4771–4779.

44. Krönke J, Kuchenbauer F, Kull M, et al.

IKZF1 expression is a prognostic marker in newly diagnosed standard-risk multiple myeloma treated with lenalidomide and intensive chemotherapy: a study of the German Myeloma Study Group (DSMM).

Leukemia. 2017;31(6):1363-1367.

45. Zhu YX, Braggio E, Shi C-X, et al.

Identification of cereblon-binding proteins and relationship with response and survival after IMiDs in multiple myeloma. Blood.

2014;124(4):536–545.

46. Pourabdollah M, Bahmanyar M, Atenafu EG, Reece D, Hou J, Chang H. High IKZF1/3 protein expression is a favorable prognostic factor for survival of relapsed/refractory multiple myeloma patients treated with lenalidomide. J Hematol Oncol. 2016;9 (1):123.

47. Bila J, Sretenovic A, Jelicic J, et al. Prognostic significance of cereblon expression in patients with multiple myeloma. Clin Lymphoma Myeloma Leuk. 2016;16(11): 610–615.

48. Jung S-H, Choi H-J, Shin M-G, et al.

Thalidomide-based induction regimens are as effective as bortezomib-based regimens in elderly patients with multiple myeloma with cereblon expression. Ann Hematol.

2016;95(10):1645–1651.

49. Huang S-Y, Lin C-W, Lin H-H, et al.

Expression of cereblon protein assessed by immunohistochemicalstaining in myeloma cells is associated with superior response of thalidomide- and lenalidomide-based treatment, but not bortezomib-based treatment, in patients with multiple myeloma. Ann Hematol. 2014;93(8):1371–1380.

50. Broyl A, Kuiper R, van Duin M, et al. High cereblon expression is associated with better survival in patients with newly diagnosed multiple myeloma treated with thalidomide maintenance. Blood. 2013;121(4):624–627.

889

ARTICLE

Plasma Cell Disorders

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):890-897

Impact of extramedullary disease in patients with newly diagnosed multiple myeloma undergoing autologous stem cell transplantation: a study from the Chronic Malignancies Working Party of the EBMT

Nico Gagelmann,1 Diderik-Jan Eikema,2 Simona Iacobelli,3 Linda Koster,2 Hareth Nahi,4 Anne-Marie Stoppa,5 Tamás Masszi,6 Denis Caillot,7 Stig Lenhoff,8 Miklos Udvardy,9 Charles Crawley,10 William Arcese,11 Clara Mariette,12 Ann Hunter,13 Xavier Leleu,14 Martin Schipperus,15 Michel Delforge,16 Pietro Pioltelli,17 John A. Snowden,18 Maija Itälä-Remes,19 Maurizio Musso,20 Anja van Biezen,2 Laurent Garderet21 and Nicolaus Kröger1

Department of Stem Cell Transplantation, University Medical Center HamburgEppendorf, Hamburg, Germany; 2EBMT Data Office, Leiden, the Netherlands; 3 Dipartimento di Biologia, Università degli Study di Roma “Tor Vergata”, Italy; 4Center for Hematology and Regenerative Medicine, Karolinska Institutet, Stockholm, Sweden; 5 Institut Paoli Calmettes, Marseille, France; 6St. István and St. László Hospital, Budapest, Hungary; 7Hématologie Clinique, Dijon University Hospital, Dijon, France; 8 Department of Hematology, Skane University Hospital Lund, Sweden; 9Department of Hematology, Bone Marrow Transplant Unit, Debrecen Medical University, Hungary; 10 Department of Haematology, Cambridge University Hospitals, UK; 11University Tor Vergata, Roma, Italy; 12Department of Hematology, Grenoble University Hospital, France; 13 Leicester Royal Infirmary, Leicester; 14Hematology, Hôpital La Mileterie, Poitiers, France; 15Haga Teaching Hospital, the Hague, the Netherlands; 16Department of Hematology, UZ Leuven, Belgium; 17Hematology, Ospedale San Gerardo ASST MonzaUniversità degli Studi di Milano Bicocca, Monza, Italy; 18Department of Haematology, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK; 19Turku University Hospital, Finland; 20Division of Hematology, Casa di Cura "La Maddalena", Palermo, Italy and 21Hopital St Antoine, Paris, France 1

ABSTRACT

Correspondence: nkroeger@uke.uni-hamburg.de

Received: August 15, 2017.

Accepted: January 26, 2018.

Pre-published: February 1, 2018. doi:10.3324/haematol.2017.178434 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5890 ©2018 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.

890

W

e investigated extramedullary disease in newly diagnosed multiple myeloma patients and its impact on outcome following first-line autologous stem cell transplantation. We identified 3744 adult myeloma patients who received up-front single (n=3391) or tandem transplantation (n=353) between 2005 and 2014 with available data on extramedullary involvement at diagnosis. The overall incidence of extramedullary disease was 18.2% (n=682) and increased per year from 6.5% (2005) to 23.7% (2014). Paraskeletal involvement was found in 543 (14.5%) and extramedullary organ involvement in 139 (3.7%).

More patients with extramedullary organ involvement had multiple involved sites (≥2; P<0.001). In a comparison of patients with single sites with patients without the disease, up-front transplantation resulted in at least similar 3-year progression-free survival (paraskeletal: P=0.86, and extramedullary organ: P=0.88). In single paraskeletal involvement, this translated less clearly into worse 3-year overall survival (P=0.07) while single organ involvement was significantly worse (P=0.001). Multiple organ sites were associated with worse outcome (P<0.001 and P=0.01).

First-line treatment with tandem compared with single transplantation resulted in similar survival in patients with extramedullary disease at diagnosis (P=0.13 for both).

Introduction Multiple myeloma (MM) accounts for approximately 2% of all new cancer cases and 13% of hematologic cancers with an age-adjusted incidence of 6 per 100,000 per year in the USA and Europe.1 Autologous stem cell transplantation (ASCT) and the development of new agents have considerably increased the median survival of MM patients.2 The disease is characterized by a clonal proliferation of malignant plasma cells with a strong dependence on the bone marrow (BM) microenvironment.3 haematologica | 2018; 103(5)

Autologous SCT for extramedullary myeloma

However, in some MM patients, myeloma cells escape the BM, resulting in extramedullary disease (EMD), which can be further characterized by two different types of involvement: 1) paraskeletal (PS), consisting of masses that arose from bone lesions; and 2) extramedullary organ involvement (EM), resulting from hematogenous spread into different organs, skin and lymph nodes.4,5 At the time of MM diagnosis, the incidence of EM involvement in observational studies ranges from 1.7% to 4.5 using a baseline staging that includes whole-body magnetic resonance imaging (MRI) or positron emission tomographycomputed tomography (PET-CT).6 Paraskeletal involvement is more frequent and varies from 7% to 34.2% due to different definitions and access of sensitive imaging techniques.7-10 Rates are also considered to be higher at relapse or after surgery.11,12 Several studies reported that EMD was associated with shorter survival rates, and thus considered EMD as a high-risk feature. However, the evidence of the effect of EMD at diagnosis is limited due to small populations, heterogenous patient or intervention selection, and relapse settings.13-16 Therefore, very limited data are available to assess the role of EMD at diagnosis of MM patients after up-front ASCT. This lack of evidence is striking, since ASCT is standard therapy in firstline therapy in eligible patients.17,18 Therefore, the objective of this study was to determine the demographic and clinical characteristics of EMD in MM patients at diagnosis and to evaluate its impact on outcome after up-front ASCT as first-line therapy. For this purpose, we analyzed 3744 patients with or without EMD at diagnosis after up-front single or tandem ASCT who had been reported to the European Society for Blood and Marrow Transplantation (EBMT) registry between 2005 and 2014.

Methods Study design and data collection We included adult patients with MM who had available data on extramedullary involvement at time of diagnosis who received an up-front single ASCT within 12 months of diagnosis or a tandem ASCT within six months from first ASCT as firstline therapy and who had been reported to the EBMT registry between January 2005 and December 2014. Patients were considered eligible for analysis if there were full data available on extramedullary involvement (yes or no) at time of diagnosis, its location, and the number of sites. This study was performed in accordance with the principles of the Declaration of Helsinki and was approved by the Chronic Malignancies Working Party of the EBMT. The EBMT is a non-profit, scientific society representing more than 600 transplant centers, mainly in Europe.

Data are entered, managed, and maintained in a central database with internet access. Audits are routinely performed to determine the accuracy of the data. Data on extramedullary involvement were extracted from the database using Med-B forms.

Patients whose transplant data are reported provided informed consent to use the information for research purposes and data are anonymized.

Definitions and statistical analysis The primary end point was 3-year progression-free survival (PFS), which was defined as the time from ASCT to disease progression or death from any cause. The secondary end points were 3-year overall survival (OS), non-relapse mortality (NRM) haematologica | 2018; 103(5)

and response. Overall survival was defined as the time from ASCT to death from any cause or last follow up. Non-relapse mortality was defined as death without evidence of relapse or progression, with relapse or progression as competing events.

Remission, progression and relapse were defined according to standard EBMT criteria.19 On the basis of type of extramedullary involvement, we defined three groups of myeloma patients: 1) without EMD (MM group); 2) with paraskeletal (PS group); and 3) extramedullary organ involvement (EM group). In addition, we determined and analyzed the impact of the number of involved sites as one or multiple (≥ 2) sites. Disease stage at diagnosis was determined according to the International Staging System (ISS; IIII),20 Salmon and Durie stages I, II or III, and also according to renal function A or B.21 Performance status at ASCT was assessed with the Karnofsky score (≤80 indicating poor and >80 good status).22 Categorical variables were compared with the use of the Fisher’s exact test or the χ² test. Continuous variables were analyzed using the Kruskal-Wallis test for independent samples.

Survival probabilities were estimated by the Kaplan-Meier method,23 and the Log-Rank test was used for univariate comparison. Median follow up was calculated according to the reverse Kaplan-Meier method.24 Outcomes were artificially censored at three years. We used cumulative incidence analysis to assess NRM, and labeled death from relapse as a competing event.25,26 The proportional hazards assumption was verified using graphical methods. Scaled Schoenfeld27 residuals and graphical checks proposed by Klein and Moeschberger28 were performed to find evidence of violations. To minimize the effect of selection bias, we used a landmark analysis at six months whenever single and tandem ASCT were compared.

To assess the multivariate effect of factors on each end point, we used the Cox proportional hazards model to estimate hazard ratios (HR).29 Only complete cases were included in the analysis.

All tests were two-sided, with the type I error rate fixed at a=0.05. All analyses were performed using the statistical software R, v.3.1.0 (R Foundation for Statistical Computing, Vienna, Austria) and SPSS Statistics 23 (SPSS, IBM Corp, Armonk, NY, USA).

Results Incidence and sites Among the 3744 patients identified in the registry, 14.5% (n=543) had paraskeletal involvement (PS group) and 3.7% (n=139) extramedullary organ involvement (EM group), while 81.8% (n=3062) had no EMD (MM group).

Between 2005 and 2014, the EMD incidence per year increased from 6.5% to 23.7%.

Within the EM group, the involved sites were: kidney (27.3%, n=38), skin (23.0%, n=32), lymph nodes (17.3%, n=24), central nervous system (CNS; 10.1%, n=14), lung and respiratory tract (6.5%, n=9), gastrointestinal tract (GI) and liver (5.8%, n=8), pleura and heart (5.0%, n=7), and spleen, ovaries and testes (5.0%, n=7). Most patients with EMD (93.5%, n=639) presented with one involved site (PS1 and EM1), 5.7% (n=36) had two sites, 0.7% (n=5) had three sites, while four and five sites were present in 0.1% (n=1) of patients, respectively. Notably, within the PS group, all 19 patients with multiple (≥2) sites had only additional paraskeletal involvement (PS2), while further involvement in all 24 EM patients was also restricted to other organs (EM2).

891

N. Gagelmann et al.

Table 1. Patientsâ&#x20AC;&#x2122;, disease and transplantation characteristics.

Characteristic N. of patients (%) Sex, n. (%) Female Male Age at diagnosis in years Median Range ISS, n. (%) I II III Unknown Renal function, n. (%) A B Unknown Karnofsky score, n. (%) Good Poor Unknown Status at ASCT, n. (%) CR PR < PR Unknown Type of myeloma, n. (%) Light chain only Non-secretory Heavy and light chain Unknown Ig-type, n. (%) G A D/E/M Unknown Light chain type, n. (%) Kappa Lambda Unknown Years of ASCT, n. (%) 0 1 â&#x2030;Ľ2 Years of ASCT, n. (%) < 2009 2009-11 > 2011 Time to 1st ASCT in months Median Range Type of ASCT, n.

Tandem Single

Patients without EMD MM group PS group

Patients with EMD EM group Total

3062 (81.8)

543 (14.5)

139 (3.7)

3744

1279 (41.8) 1783 (58.2)

240 (44.2) 303 (55.8)

57 (41.0) 82 (59.0)

1576 (42.1) 2168 (57.9)

59.8 27.4 to 77.7

59.8 26.8 to 76.8

59.0 31.8 to 72.8

781 (36.9) 759 (35.8) 578 (27.3) 944

158 (38.6) 148 (36.2) 103 (25.2) 134

29 (30.5) 29 (30.5) 37 (38.9) 44

968 (36.9) 936 (35.7) 781 (27.4) 1122

2188 (82.7) 458 (17.3) 416

410 (83.2) 83 (16.8) 50

85 (65.9) 44 (34.1) 10

2683 (82.1) 585 (17.9) 476

1877 (67.3) 914 (32.7) 271

344 (67.7) 164 (32.3) 35

81 (62.8) 48 (37.2) 10

2302 (67.2) 1126 (32.8) 316

580 (19.1) 2262 (74.7) 187 (6.2) 33

115 (21.5) 389 (72.6) 32 (6.0) 7

16 (11.7) 109 (79.6) 12 (8.8) 2

711 (19.2) 2760 (74.6) 231 (6.2) 42

672 (22.1) 74 (2.4) 2292 (75.4) 24

122 (22.5) 29 (5.4) 389 (72.0) 3

39 (28.3) 3 (2.2) 96 (69.6) 1

833 (22.4) 106 (2.9) 2777 (74.7) 28

1648 (70.8) 634 (27.2) 45 (1.9) 735

282 (72.1) 101 (25.8) 8 (2.0) 152

72 (73.5) 22 (22.4) 4 (4.1) 41

2002 (71.1) 757 (26.9) 57 (2.0) 928

1853 (63.7) 1054 (36.3) 155

327 (65.1) 175 (34.9) 41

73 (55.7) 58 (44.3) 8

2253 (63.6) 1287 (36.4) 204

0.13

115 (82.7) 24 (17.3)

3062 (81.8) 639 (17.1) 43 (1.1)

<0.001

524 (96.5) 19 (3.5) 518 (16.9) 1400 (45.7) 1144 (37.4)

82 (15.1) 204 (37.6) 257 (47.3)

25 (18.0) 62 (44.6) 52 (37.4)

625 (16.7) 1666 (44.5) 1453 (38.8)

0.001

6.2 1.1 to 11.9

6.2 2.1 to 11.9

6.1 3.9 to 11.9

249 2813

89 454

15 124

3062 (100)

P

0.55

0.59

0.11

<0.001

0.55

0.10

0.002

0.51

0.81

353 3391

EMD: extramedullary disease; MM: patients without extramedullary disease; PS: paraskeletal involvement; EM: extramedullary organ involvement; N.: number; ISS: International Staging System; ASCT: autologous stem cell transplantation; CR: complete remission; PR: partial remission.

892

haematologica | 2018; 103(5)

Autologous SCT for extramedullary myeloma

Figure 1. Progression-free survival (A) and overall survival (B) with numbers at risk of myeloma patients following up-front autologous stem cell transplantation according to presence of involvement. MM: no extramedullary disease; PS: paraskeletal involvement; EM: extramedullary organ involvement; N: number.

Patientsâ&#x20AC;&#x2122; and disease characteristics Median age at diagnosis was 59.8 years in both MM and PS, and 59.0 years in EM patients (P=0.59). In all groups there were more males (57.9%) than females (42.1%).

More EM patients (34.1%) had worse renal function (stage B) in comparison to PS (16.8%) and MM patients (17.3%; P<0.001). Patients with EM involvement (28.3%) were more likely to have light chain disease compared to PS (22.5%) and MM patients (22.1%; P=0.002). More detailed patientsâ&#x20AC;&#x2122; characteristics are listed in Table 1.

Transplantation characteristics and responses The source of stem cells for all patients was peripheral blood. A total of 3391 patients underwent up-front single ASCT and 353 patients up-front tandem ASCT; there was no difference in time to first ASCT between the two groups (P=0.81). Complete remission (CR) before the first ASCT was reported in 21.5% PS, 11.7% EM and 19.1% MM patients; partial remission (PR) was achieved by 72.6% PS, 79.6% EM and 74.7% MM patients (P=0.1) (Table 1). After ASCT, complete response was achieved by 41.6% PS, 36.1% EM and 43.9% MM patients, while 54.0% PS, 51.9% EM and 49.8% MM patients showed partial response (P=0.001).

EMD and survival Median follow up was 36.3 months (range: 1-118.9 months) after the date of ASCT. In the univariate analysis, the MM and PS groups showed similar 3-year PFS of 47.9% (95%CI: 45.8-50.1) versus 50.0% (95%CI: 44.6-55.3; P=0.78) and similar 3-year OS of 80.1% (95%CI: 78.4-81.8) versus 77.7% (95%CI: 73.3-82.1; P=0.09) (Figure 1A and B).

In contrast, EM patients had a significantly worse 3-year PFS of 39.9% (95%CI: 30.3-49.5) in comparison to MM (P=0.001) and PS patients (P=0.007), and a significantly worse 3-year OS of 58.0% (95%CI: 48.1-67.9) compared to MM and PS patients (P<0.001, respectively). Within the EM group, 3-year PFS differed according to involved organs: kidney (59.5%), skin (20.1%), lymph nodes (37.6%), CNS (47.9%), lung/respiratory tract (44.4%), haematologica | 2018; 103(5)

GI/liver (22.5%), and spleen, ovaries and testes (60.0%) (Table 2).

Comparing the MM group without EMD to those with EMD, one involved site resulted in a similar 3-year PFS of 49.4% (95%CI: 44.6-54.3; P=0.36) while multiple involved sites showed a worse PFS of 22.7% (95%CI: 5.2-40.2; P=0.001) (Figure 2A). Both one and multiple involved sites showed worse 3-year OS rates of 73.5% (95%CI: 69.277.7; P<0.001) and 71.4% (95%CI: 55.1-87.7; P=0.05) in comparison to patients without EMD (80.1%) (Figure 2B).

After stratification of EMD groups according to one versus multiple involved sites (PS1 vs. PS2, and EM1 vs.

EM2), PS patients showed no significant difference in 3year PFS of 50.5% (95%CI: 45.0-55.9%) versus 36.0% (95%CI: 5.2-66.8%; P=0.71), and OS of 77.2% (95%CI: 72.7-81.7%) versus 91.7% (95%CI: 76.0-100; P=0.27). In EM patients, this comparison resulted in a significantly worse 3-year PFS of multiple sites in the univariate analysis: 44.7% (95%CI: 34.1-55.3%) versus 13.9% (95%CI: 035.5%; P=0.03) (Figure 3A). In contrast, 3-year OS was 58.7% (95%CI: 47.9-69.5%) for EM1 versus 57.5% (95%CI: 34.2-80.8%; P=0.51) (Figure 3B).

Tandem transplantation and survival A landmark analysis was used to compare tandem and single ASCT, considering a total of 3139 patients who were alive at six months. In patients without EMD, the comparison of tandem versus single ASCT resulted in similar 3year PFS, 53.8% (95%CI: 46.7-60.9) versus 51.3% (95%CI: 48.9-53.7; P=0.37), and similar 3-year OS: 84.7% (95%CI: 79.6-89.8) versus 81.6% (95%CI: 79.8-83.4; P=0.26).

Patients with EMD showed a 3-year PFS of 59.0% (95%CI: 46.3-71.8) after tandem versus 53.0% (95%CI: 47.5-58.6) after single (P=0.43) ASCT, while 3-year OS was 77.0% (95%CI: 66.1-87.9) versus 76.9% (95%CI: 72.4-81.4; P=0.91).

Within each EMD group, PS patients showed a similar 3year PFS of 59.4% (95%CI: 45.3-73.6) after tandem versus 54.3% (95%CI: 48.0-60.5; P=0.44) after single ASCT and similar 3-year OS of 82.6% (95%CI: 72.3-92.8) versus 893

N. Gagelmann et al.

80.3% (95%CI: 75.6-85.1; P=0.88). Patients with EM involvement showed no significant difference in both 3year PFS and OS after tandem versus single transplantation: 56.2% (95%CI: 27.2-85.3) versus 48.3% (95%CI: 36.6-60.1; P=0.98), and 52.0% (95%CI: 20.0-84.0) versus 64.9% (95%CI: 54.2-75.7; P=0.39).

and Salmon and Durie (P=0.02), and lower remission status at transplantation (P<0.001).

Non-relapse mortality at three years occurred in 3.0% (95%CI: 2.0-4.0) of MM, 3.0% (95%CI: 2.0-5.0) of PS patients, and 7.0% (95%CI: 2.0-12.0) of EM patients (P=0.05). Main causes of death were relapse or progression (86.3%), infection (7.1%), secondary malignancy or posttransplant lymphoproliferative disorder (3.6%), organ damage or failure (1.8%), toxicity (0.4%), and unknown in 83 patients.

Role of other factors on survival and causes of death All patients in CR before first ASCT showed a significantly better 3-year PFS of 59.8% (95%CI: 55.3-64.3) compared to 30.7% (95%CI: 28.2-33.2) in PR and 24.7% (95%CI: 17.6-31.8; P<0.001) in less than PR. There was also a significant difference in 3-year OS, with patients in CR showing 83.6% (95%CI: 80.2-87.0) compared to 78.8% (95%CI: 76.9-80.6) in patients with PR and 27.8% (95%CI: 20.8-34.9) in patients with less than PR (P<0.001).

Other factors associated with worse PFS in patients with EMD were: older age (P=0.04), transplantation before 2011 (P=0.01), higher disease stage according to ISS (P=0.01) and Salmon and Durie (P=0.02), and lower remission status at transplantation (P<0.001). Factors associated with worse OS in EMD patients were: transplantation before 2011 (P=0.02), higher disease stage according to ISS (P=0.002)

Multivariate analyses A multivariable model was constructed to examine the effect of EMD on 3-year PFS and OS after adjusting for possible prognostic factors. All factors and covariates including corresponding references are listed in Table 3. To avoid linearly dependent covariates, we merged the disease group and the new variable of the number of involved sites into a 5-level variable consisting of patients without EMD (MM group) and patients with EMD according to number of involved sites (PS1, PS2, EM1 and EM2). Cox proportional hazards regression considering independent factors for worse PFS yielded significant results for EM2

Table 2. Involved sites in extramedullary organ involvement (EM) group and survival after autologous stem cell transplantation (ASCT).

Site Kidney CNS Lung / respiratory tract GI tract / liver Pleura / heart Spleen / ovaries / testes Skin Lymph nodes

N. of patients (%)

N. of deaths

3-year PFS in % (95% CI)

3-year OS in % (95% CI)

38 (27.3) 14 (10.1) 9 (6.5) 8 (5.8) 7 (5.0) 7 (5.0) 32 (23.0) 24 (17.3)

7 4 3 3 5 2 10 10

59.5 (41.1 to 77.9) 47.9 (18.3 to 77.4) 44.4 (7.4 to 81.5) 22.5 (0 to 58.8) NE 60.0 (17.1 to 100) 20.1 (3.4 to 36.7) 37.6 (16.4 to 58.7)

75.3 (59.0 to 91.7) 64.3 (35.5 to 93.1) 41.7 (0 to 85.1) 58.3 (22.0 to 94.7) NE 60.0 (17.1 to 100) 53.3 (30.5 to 76.0) 48.2 (25.1 to 71.3)

PFS: progression-free survival; OS: overall survival; N.: number; CI: Confidence Interval; CNS: central nervous system; GI: gastrointestinal; NE: not estimable.

Figure 2. Progression-free survival (A) and overall survival (B) with numbers at risk of myeloma patients following up-front autologous stem cell transplantation according to number of involvements: 0, 1 and â&#x2030;Ľ 2. N: number.

894

haematologica | 2018; 103(5)

Autologous SCT for extramedullary myeloma

with an HR of 3.40 (95%CI: 1.74-6.61; P<0.001).

Interestingly, there was no difference in PFS in EM1 compared to MM with an HR of 1.03 (95%CI: 0.66-1.62; P=0.88). Comparison of PFS between PS and MM showed no difference for PS1, with an HR of 1.02 (95%CI: 0.831.27; P=0.86), and a less clear HR of 2.46 (95%CI: 0.926.62; P=0.07) for PS2.

In the OS analysis, EM1 and EM2 were associated with worse outcome, showing HRs of 2.30 (95%CI: 1.43-3.70; P=0.001) and 3.64 (95%CI: 1.48-8.94; P=0.01). The difference between patients with one site of PS involvement and those without EMD was less clear, with an HR of 1.33 (95%CI: 0.98-1.83; P=0.07), while PS2 resulted in a similar outcome with an HR of 0.74 (95%CI: 0.10-5.32; P=0.77).

Tandem ASCT showed similar 3-year PFS and OS compared to single ASCT, with HRs of 0.83 (95%CI: 0.66-1.06; P=0.13) and 0.74 (95%CI: 0.51-1.09; P=0.13). However, other factors made a significant contribution to an increased risk of worse outcome (Table 3). For PFS, these factors were: ISS stage II and III, PR and less than PR at ASCT. OS was significantly influenced by ISS stage II and III, male sex, PR and less than PR at ASCT, and the presence of heavy and light chains.

Discussion Extramedullary disease in patients with MM is considered a poor prognostic factor. This EBMT registry study including 682 EMD patients identified an increase per year of EMD incidence at diagnosis from 2005 to 2014. We demonstrated that first-line ASCT resulted in at least similar 3-year PFS in patients with single sites of EMD compared to patients without EMD. Another finding, even though this was less clear, was that this translated into worse 3-year OS in single PS involvement while single sites of EM were significantly associated with worse outcome, which worsened still further when multiple sites of organs were involved. As far as treatment options for EMD at diagnosis are concerned, we found both first-line tandem and single ASCT resulted in similar 3-year PFS and OS. Evidence on the role of EMD at diagnosis after first-line ASCT is still limited. A retrospective single center study30 of 27 patients concluded that ASCT might overcome poor prognosis at onset compared to patients without EMD, while another study showed extramedullary organ involvement in only 4 patients and that its impact on outcome could be under-estimated.5 A prospective study31 of

Table 3. Multivariate analysis.

Factors – reference Group – MM without EMD PS1 PS2 EM1 EM2 Sex – male Female Age in years, > 60 < 50 50 to 60 ISS – I II III Renal function – A B Status at ASCT – CR PR < PR Type of myeloma - light chain Non-secretory Heavy and light Year of ASCT - > 2011 < 2009 2009 to 2011 Type of ASCT - single Tandem

3-year PFS Hazard ratio (95% CI)

P

3-year OS Hazard ratio (95% CI)

P

1.02 (0.82 to 1.27) 2.46 (0.92 to 6.62) 1.03 (0.66 to 1.62) 3.40 (1.74 to 6.61)

< 0.001 0.86 0.07 0.88 < 0.001

1.33 (0.98 to 1.83) 0.74 (0.10 to 5.32) 2.30 (1.43 to 3.70) 3.64 (1.48 to 8.94)

< 0.001 0.07 0.77 0.001 0.01

0.86 (0.74 to 1.01) 0.81 (0.74 to 1.03) 0.95 (0.80 to 1.11) 1.48 (1.23 to 1.77) 1.81 (1.46 to 2.24) 0.99 (0.79 to 1.24)

0.06 0.22 0.08 0.50 < 0.001 < 0.001 < 0.001

1.22 (0.97 to 1.53) 1.05 (0.88 to 1.26)

0.93 < 0.001 < 0.001 < 0.001 0.10 0.41 0.08 0.21 0.09 0.61

0.83 (0.66 to 1.06)

0.13

1.58 (1.26 to 1.97) 2.18 (1.54 to 3.10) 0.77 (0.42 to 1.43) 1.19 (0.98 to 1.46)

0.71 (0.56 to 0.91) 1.04 (0.73 to 1.48) 1.17 (0.91 to 1.50) 1.75 (1.29 to 2.37) 2.68 (1.92 to 3.74) 1.25 (0.92 to 1.69)

0.01 0.45 0.85 0.21 < 0.001 < 0.001 < 0.001

1.07 (0.75 to 1.53) 1.00 (0.76 to 1.33)

0.16 0.02 0.03 0.01 0.09 0.17 0.04 0.91 0.71 0.98

0.74 (0.51 to 1.09)

0.13

1.48 (1.05 to 2.08) 2.08 (1.22 to 3.54) 1.70 (0.80 to 3.61) 1.38 (1.01 to 1.88)

PFS: progression-free survival; OS: overall survival; CI: Confidence Interval; MM: patients without extramedullary disease; PS: patients with paraskeletal involvement arising from bone lesions; PS1: patients with paraskeletal involvement having one involved site; PS2: patients with paraskeletal involvement and multiple involved sites; EM1: patients with extramedullary organ involvement having one involved site; EM2: patients with extramedullary organ involvement and multiple involved sites; ISS: International Staging System; CR: complete remission; PR: partial remission; ASCT: autologous stem cell transplantation.

haematologica | 2018; 103(5)

895

N. Gagelmann et al.

Figure 3. Progression-free survival (A) and overall survival (B) with numbers at risk of myeloma patients with extramedullary organ involvement following up-front autologous stem cell transplantation according to number of involvements: 1 and â&#x2030;Ľ 2. EM: patients with extramedullary organ involvement; EM1: patients with one site of EM; EM2: patients with two or more sites of EM; N: number.

patients in relapse with either soft-tissue or bone-related involvement at a single institution found that bone-related relapses were associated with better OS. However, treatments before diagnosis of extramedullary relapse significantly differed between groups. Since different types of involvement were reported, this variable was examined closely.

In our study, especially EM involvement in 139 MM patients was associated with lower rate of CR before and after ASCT, a higher frequency of ISS stage III, and worse renal function. Importantly, the impact of the number of involved sites on outcome in EMD at diagnosis had not previously been described. We found 20% of all EMD patients having multiple sites of involvement, which is in line with previous reports (16%).13 Notably, the location of further involvement was only paraskeletal in the PS group and was also restricted to other organs in the EM group.32 The use of radiation therapy might contribute to the difference in PFS and OS of patients with single sites of EMD compared to patients without EMD, because it is considered effective in reducing progression in patients with solitary osseous and extraosseous involvement,33,34 in particular because reports about the efficacy of novel agents in these cohorts at diagnosis are very limited. Some results would suggest an induction bortezomib-based regimen followed by high-dose melphalan/ASCT for patients with paraskeletal rather than extramedullary involvement.14,35-37 In a retrospective study38 investigating carfilzomib alone or in combination as salvage therapy in relapse, presence of extramedullary involvement resulted in shorter duration of response compared to absent EMD, suggesting limited treatment effect. Smaller reports on the possible impact of immunomodulatory drugs showed partial efficacy regarding response rates in EMD patients.10,39,40 Retrospective studies highlighted an extremely poor prognosis for CNS involvement with a median OS of less than six months.41,42 However, in addition to systemic antiMM therapy, CNS irradiation and the use of novel combination therapies have been shown to improve the duration 896

of response.42 With regard to these analyses outside transplantation settings, we investigated survival according to involved sites in EM patients, finding most of the patients had kidney, skin or lymph node involvement. After upfront ASCT, best outcomes were found in kidney and CNS involvement while skin and lymph node involvement showed worse outcome. Interestingly, our CNS cohort showed higher rates of OS compared to previous reports, which might be due to the selection of patients with CNS involvement at diagnosis, while most reports evaluated patients at later phases of the disease.41,42 A pooled analysis of prospective studies regarding transplantation strategies suggested the superiority of tandem ASCT in patients with poor prognostic features at diagnosis.4,43 Our landmark analyses of EMD patients who received either tandem or single ASCT as first-line therapy found no difference in PFS and OS. However, this analysis was conducted with the use of retrospective data and is, therefore, subject to the attendant limitations. Regression modeling and landmark analyses were performed as a means of controlling for differences between the patients, but such adjustment cannot account for all discrepancies in clinical and diagnostic characteristics between groups. The increasing incidence of EMD might be caused by a more frequent use of whole-body MRI or PET-CT in recent years. However, although recent evidence promotes the use of more sensitive imaging techniques,44 data are not routinely documented, and they are still not part of routine diagnostics and were thus not available in our study.45,46 A randomized trial is the only way to overcome these challenges and to assess the definite impact of EMD in newly diagnosed MM patients after ASCT. In conclusion, this EBMT study identified an increase in incidence per year of EMD in newly diagnosed MM patients from 2005 to 2014. We revealed that first-line ASCT in patients with single sites of EMD (PS or EM) resulted in at least similar 3-year PFS compared to patients without EMD. Nevertheless, single EM involvement was associated with worse 3-year OS, which worsened still further when multiple sites of organs were involved. haematologica | 2018; 103(5)

Autologous SCT for extramedullary myeloma

References 1. National Cancer Institute. Surveillance epidemiology and end results program. SEER stat fact sheets: Myeloma. 2013; http://seer.cancer.gov/statfacts/html/mulmy .html [Last accessed 14 April 2017].

2. 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.

3. Palumbo A, Anderson K. Multiple myeloma.

N Engl J Med. 2011; 364(11):1046-1060.

4. Touzeau C, Moreau P. How I treat extramedullary myeloma. Blood. 2016; 127(8):971-976.

5. Bladé J, Fernández de Larrea C, Rosiñol L, Cibeira MT, Jiménez R, Powles R. Soft-tissue plasmacytomas in multiple myeloma: incidence, mechanisms of extramedullary spread, and treatment approach. J Clin Oncol. 2011;29(28):3805-3812.

6. Weinstock M, Aljawai Y, Morgan EA, et al.

Incidence and clinical features of extramedullary multiple myeloma in patients who underwent stem cell transplantation. Br J Haematol. 2015;169(6):851-858.

7. Cavo M, Terpos E, Nanni C, et al. Role of 18F-FDG PET/CT in the diagnosis and management of multiple myeloma and other plasma cell disorders: a consensus statement by the International Myeloma Working Group. Lancet Oncol. 2017;18(4):e206-e217.

8. Bartel TB, Haessler J, Brown TL, et al. F18fluorodeoxyglucose positron emission tomography in the context of other imaging techniques and prognostic factors in multiple myeloma. Blood. 2009; 114(10):20682076.

9. Usmani SZ, Heuck C, Mitchell A, et al.

Extramedullary disease portends poor prognosis in multiple myeloma and is overrepresented in high-risk disease even in the era of novel agents.

Haematologica.

2012;97(11):4761-4767.

10. Short KD1, Rajkumar SV, Larson D, et al.

Incidence of extramedullary disease in patients with multiple myeloma in the era of novel therapy and the activity of pomalidomide in extramedullary myeloma.

Leukemia. 2011;25(6):906-908.

11. Rosiñol L, Fernández de Larrea C, Bladé J. Extramedullary myeloma spread triggered by surgical procedures: an emerging entity?

Acta Haematol. 2014;132(1):36-38.

12. Pérez-Simón JA, Sureda A, Fernández-Aviles F, et al. Reduced-intensity conditioning allogeneic transplantation is associated with a high incidence of extramedullary relapses in multiple myeloma patients. Leukemia.

2006;20(3):542-545.

13. Varettoni M1, Corso A, Pica G, Mangiacavalli S, Pascutto C, Lazzarino M. Incidence, presenting features and outcome of extramedullary disease in multiple myeloma: a longitudinal study on 1003 consecutive patients. Ann Oncol. 2010;21(2):325330.

14. Rosiñol L, Oriol A, Teruel AI, et al.

Superiority of bortezomib, thalidomide, and dexamethasone (VTD) as induction pretransplantation therapy in multiple myeloma: a randomized phase 3 PETHEMA/GEMstudy.

Blood.

2012;120(8):1589-1596.

15. Usmani SZ, Rodriguez-Otero P, Bhutani M, Mateos MV, Miguel JS. Defining and treating high-risk myeloma. Leukemia.

2015;29(11):2119-2125.

16. Sonneveld P, Avet-Loiseau H, Lonial S, et al.

Treatment of multiple myeloma with high-

haematologica | 2018; 103(5)

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

risk cytogenetics: a consensus of the International Myeloma Working Group.

Blood. 2016;127(24):2955-2962.

Cavo M, Rajkumar SV, Palumbo A, et al.

International Myeloma Working Group consensus approach to the treatment of multiple myeloma patients who are candidates for autologous stem cell transplantation.

Blood. 2011;117(23):6063-6073.

Bianchi G, Richardson PG, and Anderson KC. Best Treatment Strategies in High-Risk Multiple Myeloma: Navigating a Gray Area.

J Clin Oncol. 2014;32(20):2125-2132.

Bladé J, Samson D, Reece D, et al. Criteria for evaluating disease response and progression in patients with multiple myeloma treated by high-dose therapy and haemopoietic stem cell transplantation.

Myeloma Subcommittee of the EBMT. European Group for Blood and Marrow Transplant. Br J Haematol. 1998; 102(5):1115-1123.

Greipp PR, San Miguel J, Durie BG, et al.

International staging system for multiple myeloma. J Clin Oncol. 2005;23(15):34123420.

Durie BG, Salmon SE. A clinical staging system for multiple myeloma. Correlation of measured myeloma cell mass with presenting clinical features, response to treatment, and survival. Cancer. 1975;36(3):842-854.

Karnofsky DA, Abelmann WH, Craver LF, Burchenal JH. The Use of the Nitrogen Mustards in the Palliative Treatment of Carcinoma – with Particular Reference to Bronchogenic Carcinoma.

Cancer.

1948;1(4):634-56.

Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457-481.

Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time.

Control Clin Trials. 1996;17(4):343-346.

Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94(446):496509.

Gray RJ. A class of k-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16(3):1141-1154.

Iacobelli S, EBMT Statistical Committee.

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.

Klein JP, Moeschberger ML (eds): Survival Analysis: Techniques for Censored and Truncated Data, 2nd edn. Springer-Verlag: New York, USA, 2003.

Cox DR: Regression models and life tables. J R Stat Soc B. 1972;34(2):187-220.

Lee SE, Kim JH, Jeon YW, et al. Impact of extramedullary plasmacytomas on outcomes according to treatment approach in newly diagnosed symptomatic multiple myeloma. Ann Hematol. 2015; 94(3):445452.

Pour L, Sevcikova S, Greslikova H, et al.

Soft-tissue extramedullary multiple myeloma prognosis is significantly worse in comparison to bone-related extramedullary relapse. Haematologica. 2014;99(2):360364.

Ghobrial IM. Myeloma as a model for the process of metastasis: implications for therapy. Blood. 2012;120(1):20-30.

Hu K, Yahalom J. Radiotherapy in the management of plasma cell tumors. Oncology (Williston Park). 2000;14(1):101-108.

Terpos E, Morgan G, Dimopoulos MA, et al. International Myeloma Working Group

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

Recommendations for the Treatment of Multiple Myeloma–Related Bone Disease. J Clin Oncol. 2013;31(18):2347-2357.

Paubelle E, Coppo P, Garderet L, et al.

Complete remission with bortezomib on plasmocytomas in an end-stage patient with refractory multiple myeloma who failed all other therapies including hematopoietic stem cell transplantation: possible enhancement of graft-vs-tumor effect. Leukemia. 2005;19(9):1702-1704.

Sonneveld P, Goldschmidt H, Rosiñol L, et al. Bortezomib-based versus nonbortezomib-based induction treatment before autologous stem-cell transplantation in patients with previously untreated multiple myeloma: a meta-analysis of phase III randomized, controlled trials. J Clin Oncol.

2013;31(26):3279-3287.

Mateos MV, Oriol A, Matínez-López J, et al.

Bortezomib, melphalan, and prednisone versus bortezomib, thalidomide, and prednisone as induction therapy followed by maintenance treatment with bortezomib and thalidomide versus bortezomib and prednisone in elderly patients with untreated multiple myeloma: a randomised trial.

Lancet Oncol. 2010;11(10):934-941.

Muchtar E, Gatt ME, Rouvio O, et al.

Efficacy and safety of salvage therapy using Carfilzomib for relapsed or refractory multiple myeloma patients: a multicentre retrospective observational study. Br J Haematol.

2016;172(1):89-96.

Calvo-Villas JM, Alegre A, Calle C, et al.

Lenalidomide is effective for extramedullary disease in relapsed or refractory multiple myeloma. Eur J Haematol.

2011;87(3):281-284.

Rosiñol L, Cibeira MT, Bladé J, et al.

Extramedullary multiple myeloma escapes the effect of thalidomide. Haematologica.

2004;89(7):832-836.

Nieuwenhuizen L, Biesma DH. Central nervous system myelomatosis: review of the literature. Eur J Haematol. 2008;80(1):19.

Chen CI, Masih-Khan E, Jiang H, et al.

Central nervous system involvement with multiple myeloma: long term survival can be achieved with radiation, intrathecal chemotherapy, and immunomodulatory agents. Br J Haematol. 2013;162(4):483-488.

Cavo M, Salwender H, Rosiñol L, et al.

Double vs single autologous stem cell transplantation after bortezomib-based induction regimens for multiple myeloma: an integrated analysis of patient-level data from phase European III studies. Blood.

2013;122:767.

Moreau P, Attal M, Caillot D, et al.

Prospective evaluation of magnetic resonance imaging and [18F] Fluorodeoxyglucose positron emission tomography-computed tomography at diagnosis and before Maintenance Therapy in Symptomatic Patients With Multiple Myeloma Included in the IFM/DFCI 2009 Trial: Results of the IMAJEM Study. J Clin Oncol. 2017; 35(25):2911-2918.

Kumar S, Paiva B, Anderson KC, et al.

International Myeloma Working Group consensus criteria for response and minimal residual disease assessment in multiple myeloma. Lancet Oncol. 2016;17(8):e328346.

Pawlyn C, Fowkes L, Otero S, et al. Wholebody diffusion-weighted MRI: a new gold standard for assessing disease burden in patients with multiple myeloma?

Leukemia. 2016;30(6):1446-1448.

897

ARTICLE

Hemostasis

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):898-907

Immobilized fibrinogen activates human platelets through glycoprotein VI

Pierre H Mangin,1 Marie-Blanche Onselaer,2 Nicolas Receveur,1 Nicolas Le Lay,3 Alexander T Hardy,2 Clare Wilson,4 Ximena Sanchez,5 Stéphane Loyau,3 Arnaud Dupuis,1 Amir K Babar,6 Jeanette LC Miller,6 Helen Philippou,4 Craig E Hughes,2,8 Andrew B Herr,6 Robert AS Ariëns,4 Diego Mezzano,5 Martine Jandrot-Perrus,3,7 Christian Gachet1 and Steve P. Watson2,9

Université de Strasbourg, INSERM, EFS Grand-Est, BPPS UMR-S 1255, FMTS, France; Institute of Cardiovascular Sciences, IBR Building, College of Medical and Dental Sciences, University of Birmingham, UK; 3Université de Paris Diderot, INSERM UMR_S1148, Hôpital Bichat, Paris, France ; 4Thrombosis and Tissue Repair Group, Institute of Cardiovascular and Metabolic Medicine, University of Leeds, UK; 5Laboratorio de Hemostasia, Pontificia Universidad Catolica de Chile, Santiago, Chile; 6Division of Immunobiology, Center for Systems Immunology & Division of Infectious Diseases, Cincinnati, OH, USA; 7Acticor Biotech, Hôpital Bichat, INSERM, UMR-S 1148, Paris, France; 8 Institute for Cardiovascular and Metabolic Research, Harborne Building, University of Reading, UK and 9Centre of Membrane Proteins and Receptors (COMPARE), Universities of Birmingham and Nottingham, Midlands, UK 1 2

ABSTRACT

G

doi:10.3324/haematol.2017.182972

lycoprotein VI, a major platelet activation receptor for collagen and fibrin, is considered a particularly promising, safe antithrombotic target. In this study, we show that human glycoprotein VI signals upon platelet adhesion to fibrinogen. Full spreading of human platelets on fibrinogen was abolished in platelets from glycoprotein VIdeficient patients suggesting that fibrinogen activates platelets through glycoprotein VI. While mouse platelets failed to spread on fibrinogen, human-glycoprotein VI-transgenic mouse platelets showed full spreading and increased Ca2+ signaling through the tyrosine kinase Syk. Direct binding of fibrinogen to human glycoprotein VI was shown by surface plasmon resonance and by increased adhesion to fibrinogen of human glycoprotein VI-transfected RBL-2H3 cells relative to mock-transfected cells.

Blockade of human glycoprotein VI with the Fab of the monoclonal antibody 9O12 impaired platelet aggregation on preformed platelet aggregates in flowing blood independent of collagen and fibrin exposure.

These results demonstrate that human glycoprotein VI binds to immobilized fibrinogen and show that this contributes to platelet spreading and platelet aggregation under flow.

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/898

Introduction

Correspondence: pierre.mangin@efs.sante.fr or s.p.watson@bham.ac.uk Received: October 20, 2017.

Accepted: February 13, 2018.

Pre-published: February 22, 2018.

©2018 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.

898

The immunoglobulin receptor glycoprotein (GP) VI is expressed on megakaryocytes and platelets. GPVI associates with the Fc receptor (FcR) γ-chain in the membrane, and with the Src family kinases (SFK) Lyn and Fyn through its cytosolic tail.1 Ligand binding clusters GPVI at the platelet surface promoting phosphorylation of the immunoreceptor tyrosine-based motif (ITAM) of the FcR γ-chain by SFK.2–4 This results in the recruitment of Syk and formation of a LAT-based signalosome that activates PLCγ2 leading to an increase in Ca2+, activation of integrin and secretion of granules.5 GPVI is widely known as a platelet activation receptor for fibrillar collagen.5 However, in recent years, GPVI has been shown to bind to additional ligands, including subendothelial and plasma adhesive proteins such as laminins and fibrin,6–8 the hormone adiponectin and the transmembrane protein emmprin.9,10 Several of these interactions are relatively weak and of unclear significance. For example, GPVI supports adhesion and efficient activation of platelets to collagen haematologica | 2018; 103(5)

Fibrinogen, GPVI and platelet activation

and fibrin but is only involved in post-adhesive events on laminin.6,11 Determining the importance of the interaction of GPVI with each ligand in mediating platelet activation in vivo will require development of selective inhibitors.

GPVI is involved in arterial thrombosis and in several of the more recently discovered roles of platelets, including maintenance of vascular integrity at sites of inflammatory challenge.12 We and others have reported that the absence of GPVI reduces experimental thrombosis in mouse models of atherosclerotic plaque rupture13,14 and abolishes occlusive thrombus formation following FeCl3 injury.15 In contrast, the absence or blockade of GPVI has a relatively minor impact on hemostasis in mice16 and patients deficient in GPVI have a relatively mild bleeding diathesis.17–19 These results highlight that GPVI is a promising antithrombotic target with inhibitors predicted to have a minor effect on hemostasis.20 Following ligand binding, GPVI stimulates signals that convert integrin aIIbβ3 from a low to a high affinity state for fibrinogen and other physiological ligands.21 Ligand engagement of integrin aIIbβ3 has been reported to generate outside-in signals that are similar to those of GPVI, including activation of Src and Syk kinases, PLCγ2 and Ca2+ mobilization.22–24 Paradoxically, however, human but not mouse platelets generate extensive lamellipodial sheets and stress fibers on fibrinogen whereas on collagen, which stimulates similar signals, full spreading of platelets is seen in both species.25 One explanation for this difference is the presence of the low affinity immune receptor, FcγRIIA, in the human but not the rodent genome, as FcγRIIA-transfected transgenic mouse platelets exhibit increased spreading and Syk activation upon adhesion to fibrinogen, although the increase in spreading is only partial.26–28 Outside-in signaling by aIIbβ3 is also mediated by two conserved tyrosines present in a NxxY motif in the integrin β3 intracytoplasmic domain independent of Src and Syk activation. Mutation of these two tyrosine residues to phenylalanine leads to a re-bleeding diathesis that has been attributed to a defect in clot retraction.29 This shows that engagement of integrin aIIbβ3 leads to activation of multiple signaling pathways.

In the present study, we showed that full spreading of human platelets on fibrinogen is abolished in patients deficient in GPVI and that transgenic mouse platelets expressing human GPVI, in contrast to wild-type platelets, spread fully on fibrinogen. Direct binding of fibrinogen to human GPVI was demonstrated using human GPVI-transfected cell lines and recombinant GPVI. Inhibiting the binding of fibrinogen to GPVI limited platelet aggregation under conditions that excluded involvement of collagen and fibrin.

Mice Wild-type mice were generated from breeding of heterozygotes or purchased from Harlan Laboratories (Hillcrest, UK) or Charles River (Lyon, France). GPVI-/- mice were provided by Dr Jerry Ware.31 Mouse platelets expressing human GPVI have been described and characterized previously.32 Syk chimera mice have been described previously.27 Ethical approval for animal experimentation was obtained from the French Ministry of Research and UK Home Office in accordance with the European Union guidelines, the Guide for the Care and Use of Laboratory Animals.

Reagents PRT-060318 was obtained from Caltag Medsystems (Buckingham, UK), REOPRO from E. Lilly (Indianapolis, IN, USA). Recombinant GPVI was made as described elsewhere.33 Fibrinogen was from Kabi (Bad Homburg, Germany) or from ERL (South Bend, IN, USA). RAM.1 (anti-GPIbβ) was generated in U949.34 The blocking Fab fragment of monoclonal antibody directed against human GPVI, 9O12.2, and its humanized version are referred to as 9O12 in this manuscript.35,36 All other reagents were from previously described sources.11,37 The antifibrin antibody 59D8 was obtained from CT Esmon (Oklahoma Medical Research Foundation, OK, USA).38

Generation and characterization of RBL-2H3 cells The cDNA of WT human GPVI33 was inserted in pSRαNeo between the 5’-XhoI and 3’-BamHI restriction sites. Rat basophilic leukemia cells, RBL-2H3, were cultured in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum albumin and stably transfected with 1 μg of DNA corresponding to the empty vector or WThGPVI vector mixed with FuGENE6 (Roche, Boulogne-Billancourt, France) and selected in growth medium containing G418 0.7%; 1 mg/mL geneticin (GibcoBRL, Invitrogen, Cergy Pontoise). Cell surface expression of recombinant GPVI and constitutively expressed integrin aIIbβ3 was confirmed by flow cytometry and immunoblot (data not shown).

Cell adhesion to fibrinogen

LAB TEK 4 wells were coated with 400 μL/well of collagen (50 µg/mL) or fibrinogen (100 μg/mL) overnight at 37°C. Wells were saturated with human serum albumin (10 mg/mL) for 1 h at 37°C. Trypsinised RBL cells (3x105 cells/mL) were incubated with phosphate-buffered saline or 9O12 (50 μg/mL) and/or REOPRO (40 μg/mL) for 15 min at 37°C. Subsequently, 100,000 cells were added to the wells (300 μL) for 1 h at 37°C. After three washing steps, cells were fixed with 400 μL paraformaldehyde 4% for 20 min. Pictures were taken with an EVOS optic microscope (x10). Actin was stained with Alexa-488-phalloidin and the nucleus with DAPI.

Washed platelets Methods Patients Family 1 and family 2 are two families who are heterozygous or homozygous for an adenine insertion in exon 6 of GP6 that generates a premature stop codon in position 242 of the protein.17 They have been described previously.30 Patient 3 is a 10-year old boy suffering from an autoimmune disease with anti-GPVI antibodies. The platelets of this patient do not aggregate to collagen and GPVI is not detected at the platelet surface using flow cytometry and western blot (data not shown). haematologica | 2018; 103(5)

Human blood was taken from patients or from healthy donors using 3.8% (v/v) sodium citrate (1:9) as the anticoagulant. Human and mouse washed platelets were obtained by centrifugation using prostacyclin (2.8 μmol/L) and resuspended in modified Tyrode-HEPES buffer as described elsewhere.37,39

Platelet spreading Platelet adhesion to immobilized fibrinogen was achieved as described previously.37 Images of the platelets were obtained with a Zeiss Axiovert 200 mol/L microscope or with a Leica DMI400 microscope. Platelet surface area was analyzed using ImageJ software (NIH, Bethesda, USA).

899

P.H. Mangin et al.

Western blotting For stimulation on fibrinogen, washed platelets were pre-treated with 10 μmol/L indomethacin and 2 U/mL apyrase. Platelets (1.5 mL containing 5x108/mL) were allowed to adhere to 10 cm dishes coated with 100 μg/mL fibrinogen or heat-inactivated bovine serum albumin for 45 min at 37°C. Non-adherent platelets were removed and lysed by addition of 2X lysis buffer (150 mmol/L NaCl, 10 mmol/L Tris, 1 mmol/L EGTA, 1 mmol/L EDTA, 1% NP-40; pH 7.4, plus 1.25 mmol/L Na3VO4, 50 μg/mL AEBSF, 2.5 μg/mL leupeptin, 2.5 μg/mL aprotinin and 0.25 μg/mL pepstatin). Adherent platelets were washed twice with Tyrode buffer then lysed with 1X lysis buffer on ice for 15 min before scraping.

Proteins were immunoprecipitated with a-Syk antibody and protein A-sepharose beads for 2 h. The beads were washed, proteins eluted in sodium dodecyl sulfate (SDS) sample buffer, separated

by SDS-polyacrylamide gel electrophoresis (PAGE), electro-transferred, and western blotted with the stated antibodies. For whole platelet lysates, washed platelets (5x108/mL) were lysed directly with an equal volume of 2xSDS sample buffer, separated by SDSPAGE, electro-transferred, and western blotted with the stated antibodies.

Ca2+ assay and in vitro perfusion assay Intraplatelet Ca2+ concentrations following platelet adhesion to fibrinogen were measured using a dual-dye ratiometric method and hirudinated blood perfusion was performed as previously described.40 Three-dimensionsal reconstructed images were obtained using the 3D module of Leica LAS X software.

Solid-based binding assay Binding studies were performed with the recombinant proteins, GPVI-Fc fusion (dimer) and GPVI-His tagged (monomer). Cover slips were coated with collagen or fibrinogen overnight at 4°C. The plates were blocked with 3% bovine serum albumin – phosphate-buffered saline for 1 h and washed prior to addition of monomeric or dimeric GPVI at a concentration of 100 nmol/L for 1 h. After washing, 4 μg/mL of secondary antibodies, horseradish peroxidase (HRP)-conjugated goat anti-human IgG Fc or HRP-conjugated anti-His Tag, were added for 1 h. GPVI binding was detected using 3,3′,5,5′-tetramethylbenzidine. The reaction was stopped with H2SO4 (2 mol/L) and absorbance was measured at 450 nm with a spectrofluorometer.

Surface plasmon resonance Surface plasmon resonance was performed on a Pioneer platform from PALL® FortéBio® (Portsmouth, UK). IF-1 purified fibrinogen was diluted to 100 μg/mL using 10 mmol/L NaAc pH 5.0.

IF-1 fibrinogen was adsorbed to the chip surface via amine coupling to a level of 3825 resonance units (RU) at flow-cell 1 and 3423 RU on flow-cell 3. Flow-cell 2 was activated using amine coupling and blocked using 1 mol/L ethanolamine and was the designated reference channel. GPVI analytes were dialysed and diluted to 1 µmol/L using the same batch of running buffer as used for the blanks. Analytes were injected using the OneStep® titration function at a flow rate of 30 μL/min with a 100% loop-inject and 400 s dissociation. The chip surface was regenerated by flushing with 1 mol/L NaCl at 60 μL/min for 10 s, followed by a further 400 s dissociation. Qdat data analysis software (PALL® FortéBio®, UK) was used to analyze the data. Binding data were fitted using a one site KA/KD model and analyte aggregation parameters adjusted per binding curve according to goodness of fit and curve type.

Statistics The statistical analyses were performed using the GraphPad Prism program, version 5.0 (Prism, GraphPad, LaJolla, CA, USA).

900

The values are indicated as mean ± standard error of the mean (SEM). The statistical analysis is described in the Figure legends.

Results Abolition of spreading on fibrinogen in glycoprotein VI-deficient human platelets Human platelets undergo robust spreading on immobilized fibrinogen, generating lamellipodial sheets and stress fibers.41 This is illustrated in Figure 1A with over 90% of platelets from a control donor undergoing full spreading on fibrinogen over 30 min; the small number of partiallyspread platelets most likely represent newly adhered cells.

In 2013, Matus et al. described four unrelated families with index cases who are homozygous for an adenine insertion in exon 6 of human GP6, which leads to a premature ‘stop codon’ in position 242 prior to the transmembrane domain.17 All four homozygous patients lack expression of GPVI on their platelets and heterozygous relatives express approximately 50% of the receptor. Since then, two further unrelated families with the same mutation have been identified by the same group and also been shown to lack surface expression of GPVI with absent platelet aggregation to collagen.30 Unexpectedly, in studying platelets from two unrelated index cases in these families, we observed reduced adhesion on immobilized fibrinogen and a failure to form lamellipodial sheets and stress fibers (Figure 1A).

The absence of GPVI was confirmed by flow cytometry and by abolition of aggregation to collagen but not to other agonists in both cases30 (data not shown). In contrast, spreading and adhesion of platelets from heterozygote carriers from each family and platelets from a control were similar (Figure 1A). The same result was also seen in a patient with an auto-immune thrombocytopenia associated with the absence of GPVI expression (Figure 1Bi).

Adhesion of platelets was blocked by the aIIbβ3 receptor antagonist, REOPRO (Figure 1B), as previously shown in controls. These results demonstrate that adhesion of human platelets on fibrinogen is critically dependent on integrin aIIbβ3 with a minor contribution from GPVI, but that full spreading requires GPVI.

Mouse platelets expressing human glycoprotein VI undergo full spreading on fibrinogen Mouse platelets adhere and undergo limited spreading on human or mouse fibrinogen, forming filopodia and limited lamellipodia but not stress fibers (Figure 2A). A similar response is seen in platelets deficient in GPVI (Figure 2A), whereas adhesion is abolished in the absence of the integrin β3-subunit (Figure 2B). A similar level of adhesion is seen in human GPVI transgenic mouse platelets but is associated with the formation of lamellipodial sheets and stress fibers (Figure 2C). These results demonstrate that full spreading but not adhesion of mouse platelets is dependent on human GPVI and not mouse GPVI. One potential explanation for these results is that human but not mouse GPVI is able to bind to fibrinogen and mediate platelet activation.

Fibrinogen binds to monomeric human glycoprotein VI To test whether fibrinogen is able to bind to GPVI, increasing concentrations of recombinant soluble GPVI extracellular domain, expressed either as a monomer (GPVI-ex) or dimer (GPVI-Fc), was flowed over immobihaematologica | 2018; 103(5)

Fibrinogen, GPVI and platelet activation

lized fibrinogen and binding monitored by surface plasmon resonance. As shown in Figure 3A, clear binding of monomeric GPVI (ka = 1.17 ± 0.01 x 104 M-1s-1) was observed with a kd of 3.94 ± 0.01 x 10-3 s-1. Binding was fitted to a single site with an equilibrium dissociation constant (KD) of 336 ± 1 nmol/L. In contrast, binding of dimeric GPVI to fibrinogen was not detected at concentrations up to 1 μmol/L (Figure 3Ai). In a second approach, fibrinogen was immobilized on a plastic surface and a solid phase binding assay was performed. There was increased binding of monomeric GPVI, but not dimeric GPVI, which was inhibited by D-dimer (Figure 3Aii) where the binding motif in fibrin resides.30 To further investigate the ability of GPVI to bind to fibrinogen, we transfected rat RBL-2H3 basophilic cells, which constitutively express integrin aIIbβ3 at low levels, with human GPVI and studied adhesion to immobilized fibrinogen. We observed a 3-fold increase in adhesion of GPVI-transfected cells to fibrinogen and to collagen relative to the adhesion of mock-transfected control cells (Figure 3Bi, ii). RBL-2H3 cells expressing human GPVI also formed stress fibers upon adhesion to fibrinogen. The human GPVI-blocking monoclonal

antibody 9012 Fab blocked the increase in adhesion.

Blocking the integrin aIIbβ3 with REOPRO reduced cell adhesion to immobilized fibrinogen to the same level as 9O12 Fab, with no further inhibition in the presence of both inhibitors (data not shown), indicating the presence of additional binding proteins for fibrinogen in the adherent cell line although binding to these was not sufficient to induce spreading (Figure 3Biii and not shown). These results demonstrate that GPVI binds to immobilized fibrinogen and is able to contribute to cell adhesion.

Spreading of human platelets but not mouse platelets is dependent on Syk The formation of lamellipodial sheets and stress fibers in human platelets on fibrinogen and collagen is blocked by the inhibitors of Src and Syk tyrosine kinases, PP2 and PRT060318, respectively (Figure 4Ai & ii). Adhesion of human platelets to fibrinogen induces phosphorylation of Syk which co-precipitates with the phosphorylated FcR γchain (Figure 4Aiii). These results provide further evidence of GPVI activation in human platelets by immobilized fibrinogen. In contrast, the morphological modifications of

A

Figure 1. Glycoprotein VI supports platelet adhesion and spreading on immobilized fibrinogen. (A, B) Washed human platelets were allowed to adhere to immobilized fibrinogen (10 μg/mL) for 30 min.

(A)(i).

Representative epifluorescence images of fibrinogen-adherent platelets from healthy donors (Control) or members of a family with a mutation in the GP6 gene (Heterozygotes: Family 1 +/-; homozygotes: Family 1 -/-). Scale bars represent 10 μm. (A)(ii). Bar graph representing the percentage of platelets spreading on fibrinogen.

Spreading is expressed as the mean±SEM in five random fields, in two separate experiments (one-way ANOVA, Kruskal-Wallis post-hoc test, ***P<0.0002; ****P<0.0001).

(A)(iii). Bar graph representing the number of platelets adhering to immobilized fibrinogen per mm² of a control (Control) and two families with a mutation in the GP6 gene. Adhesion is expressed as mean±SEM in five random fields, in two separate experiments (one-way ANOVA, Kruskal-Wallis post-hoc test, **P<0.002; ****P<0.0001). (B). Washed platelets from a control or a patient with an immune thrombocytopenic purpura presenting with undetectable levels of GPVI on platelets (Patient 3) were allowed to adhere to fibrinogen (100 μg/mL). (B)(i). Representative epifluorescence images of washed platelets from patient 3 adhering to immobilized fibrinogen for 30 min, in the presence or absence of REOPRO (40 μg/mL).

Scale bars represent 10 μm. (B)(ii). Bar graph represents the number of platelets adhering to fibrinogen per mm2. Adhesion is expressed as mean±SEM in eight random fields, in two separate experiments.

B

haematologica | 2018; 103(5)

901

P.H. Mangin et al.

mouse platelets on fibrinogen is blocked by the Src kinase inhibitor PP2 but not by the Syk kinase inhibitor PRT060318 (Figure 4B). Morphological changes of mouse platelets on fibrinogen are also not altered in platelets from irradiated mice transplanted with Syk-deficient fetal liver (Figure 4B) or from PF4.Cre-Sykfl/fl mice (Online Supplementary Figure S1). Western blotting for Syk confirmed lack of expression of the tyrosine kinase in the two transgenic models (Figure 4B and not shown). Thrombin stimulated full spreading of wild-type and Syk-deficient platelets (Figure 4B). Formation of lamellipodia and stress

A

B

C

902

fibers in human GPVI transgenic mouse platelets was blocked by PRT060318 (Figure 4C). The ability of Src and Syk inhibitors to block spreading of human platelets and human GPVI transgenic mouse platelets on fibrinogen is consistent with platelet activation by GPVI. This is supported by demonstration of phosphorylation of the FcR γchain. The limited spreading of mouse platelets on fibrinogen is mediated through a Src-dependent but Syk-independent pathway. Together, these results support a model in which immobilized fibrinogen activates human but not mouse platelets through GPVI.

Figure 2. Human but not mouse glycoprotein VI supports platelet adhesion and spreading on immobilized fibrinogen. (A). Washed platelets from wildtype mice (WT mice), GPVI-deficient mice (GPVI-/- mice) or from healthy donors (Human) were allowed to adhere to human or mouse fibrinogen (FGN) for 30 or 45 min, respectively, and fixed with PFA and stained with Alex-488phalloidin (4 μg/mL).

(A)(i).

Representative epifluorescence images of washed platelets adhering to fibrinogen. Scale bars represent 10 µm. (A)(ii).

Bar graph representing the number of platelets adhering to immobilized fibrinogen per mm². Adhesion is expressed as mean±SEM in five random fields, in three separate experiments (two-way ANOVA, Bonferroni post-hoc test: P>0.05). (A)(iii). Bar graph representing the percentage of platelets spreading on immobilized fibrinogen. Spreading is expressed as the mean±SEM in five random fields, in six separate experiments.

Significance was attained using a twoway ANOVA, Bonferroni post-hoc test: ****P<0.001. (B). Washed control (WT) or β3-deficient (β3-/-) platelets were allowed to adhere to fibrinogen for 60 min, fixed with PFA and stained with TRITC-phalloidin (2 μg/mL). (B)(i).

Representative epifluorescence images of washed mouse platelets adhering to fibrinogen. Scale bars represent 10 μm.

(B)(ii). Bar graph representing the number of platelets adhering to immobilized fibrinogen per mm². Adhesion is expressed as the mean±SEM in eight random fields, in four separate experiments (Mann-Whitney test, **P<0.001). (C). Washed platelets from wild-type mice (WT mice) or mice expressing human GPVI (hGPVI mice) were allowed to adhere to fibrinogen for 60 min, fixed with PFA and stained with TRITC-phalloidin (2 μg/mL). (C)(i).

Representative epifluorescence images of washed platelets adhering to fibrinogen. Scale bars represent 10 μm. (C)(ii).

Bar graph (left) representing the number of platelets adhering to immobilized fibrinogen per mm². Adhesion is expressed as the mean±SEM in eight random fields, in four separate experiments (Mann-Whitney test, P>0.05).

Bar graph (right) representing the percentage of platelets spreading on immobilized fibrinogen. Spreading is expressed as the mean±SEM in eight random fields, in four separate experiments (Mann-Whitney test, **P<0.001).

haematologica | 2018; 103(5)

Fibrinogen, GPVI and platelet activation

Fibrinogen stimulates an increase of Ca2+ in human glycoprotein VI-transgenic mouse platelets The observation that platelets expressing human but not mouse GPVI undergo full spreading suggests that signals from human GPVI are of significance. To investigate this, a dual-dye Ca2+ assay was used to monitor cytoplasmic Ca2+ levels as a marker of PLCγ2 activation. Analysis of single platelet Ca2+ profiles by confocal microscopy highlighted that signals generated on fibrinogen are composed of Ca2+ spikes (Figure 5A). The number of Ca2+ spikes in mouse platelets expressing human GPVI was significantly increased relative to the number in wild-type

platelets and was blocked in the presence of PRT-060318 (Figure 5Ai-ii) highlighting the critical role of Syk in Ca2+ mobilization. These results demonstrate that human GPVI stimulates Ca2+ signaling in fibrinogen-adherent mouse platelets.

Fab 9O12 blocks aggregate growth of humanized glycoprotein VI mouse platelets Fibrinogen plays a critical role in hemostasis and arterial thrombosis through crosslinking of platelets in the growing thrombus. In addition, we now show that fibrinogen induces platlet activation by GPVI. To establish whether

A

B

Figure 3. Monomeric but not dimeric GPVI binds to immobilized fibrinogen and supports cellular adhesion. (A)(i). IF-1 purified fibrinogen was immobilised to a COOH-V chip surface using amine coupling covalent linkage, to a level of 3,825 RU. Monomeric or dimeric GPVI was titrated over three to four orders of magnitude across the chip surface to a maximum of 1 µmol/L, generating a single binding curve (1 RU represents the binding of approximately 1 pg/mm2 protein). The graphs shown are representative of three repeats and the KD is expressed as mean ± SEM. (A)(ii). Solid-phase binding assays were performed in Nunc maxisorb 96-well plates coated overnight with BSA and fibrinogen (FGN).

Monomeric or dimeric GPVI (100 nmol/L) was incubated as described. Bound GPVI was detected using HRP-coupled to an anti-6×His monoclonal antibody for monomeric GPVI or an anti-human IgG for dimeric GPVI. The histogram (mean ± SD) shows the results from five independent experiments.

Significance was determined using a one-way ANOVA, Dunnett post-hoc test : *P<0.05. (B). RBL2H3 cells (3x105 cells/mL; 300 μL) transduced with an empty vector (RBL-2H3 Ctrl) or with human fulllength GP6 cDNA (RBL-2H3 huGPVI) were pre-incubated with PBS, REOPRO (20 μg/mL), and 9O12 (50 μg/mL) and allowed to adhere to immobilized fibrinogen for 1 h at 37°C, in 5% CO2. After three gentle washing steps, cells were permeabilised with TRITON x100 0.2% and stained with Alexa-568 phalloidin (1.5 U/mL). (B)(i). Representative epifluorescence images of RBL cells adhering to immobilized fibrinogen. (B)(ii). and (B)(iii). Cells were manually counted in six to eight different experiments (*P<0.05 Mann Whitney t-test in (ii) and one-way ANOVA was followed by the Bonferroni multiple comparison test).

haematologica | 2018; 103(5)

903

P.H. Mangin et al.

activation of GPVI by platelet-bound fibrinogen participates in platelet aggregation we performed an in vitro flow adhesion assay under conditions that prevent activation of GPVI by collagen and by fibrin. To achieve this, we generated a platelet aggregate over type I fibrillar collagen using hirudin-treated blood to prevent formation of fibrin. We

then perfused additional blood from the same donor over the aggregate at a wall shear rate of 300 s-1 in the presence or absence of the Fab fragment of the GPVI blocking monoclonal antibody 9O12. As expected, we were unable to detect the presence of fibrin in the aggregate using a specific antibody (data not shown). The aggregate continued to

A

B

C

904

Figure 4. Syk promotes platelet spreading on fibrinogen downstream of glycoprotein VI but not in integrin aIIbβ3 outside-in signaling.

Washed human platelets, or washed platelets from wild-type mice (WT mice), Syk chimera mice (Syk chimera mice) and mice expressing human GPVI (hGPVI mice) were allowed to adhere to fibrinogen or to collagen in the presence of either the Src inhibitor PP2 (20 μmol/L), the Syk inhibitor (5 μmol/L), thrombin (0.1 U/mL), or vehicle control, for 30 min (human platelets) or 45 min (mouse platelets) at 37°C followed by fixation with PFA. (A)(i).

Representative DIC images of human platelets adhering to fibrinogen or collagen.

Scale bars represent 5 μm. (A)(ii). Bar graph representing the surface area of platelets spreading on fibrinogen. Spreading is expressed as the mean±SEM in five or more random fields, in three separate experiments. Significance was determined by oneway ANOVA and Bonferroni multiple comparison test: *P<0.01. (A)(iii) Representative western blot from human platelets following adhesion to fibrinogen. Following immunoprecipitation of Syk, proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and western blot for phosphotyrosine. Membranes were then stripped and reprobed for Syk to confirm equal protein loading. Blots are representative of three separate experiments. (B)(i).

Representative DIC images of mouse platelets adhering to fibrinogen. Scale bars represent 5 μm. (B)(ii). Bar graph representing the surface area of platelets spreading on fibrinogen. Spreading is expressed as the mean±SEM in five or more random fields, in three separate experiments.

Significance was determined using one-way ANOVA, with the Bonferroni post-hoc test: *P<0.05, **P<0.001. (B)(iii). Expression of Syk was measured by western blotting platelet whole-cell lysates with an a-Syk antibody. Membranes were then stripped and reprobed with an anti-a-tubulin antibody to confirm equal protein loading. Blots are representative of three separate experiments. (C)(i). Representative epifluorescence images of mouse platelets adhering to fibrinogen. Scale bars represent 10 µm.

(C)(ii). Bar graph representing the surface area of platelets spreading on fibrinogen.

Spreading is expressed as the mean±SEM in eight random fields, in five separate experiments (Mann-Whitney test, **P<0.01).

haematologica | 2018; 103(5)

Fibrinogen, GPVI and platelet activation

grow in the presence of a Fab control but was dramatically inhibited in the presence of Fab 9O12 (Figure 6A and Online Supplementary Figure S2). In contrast, and as previously shown, blockade of GPVI did not impair aggregation measured by light transmission aggregometry in response to ADP, U46619 and thrombin.35 These results demonstrate a critical role for GPVI in aggregate growth under flow when the roles of collagen and fibrin are negated. In contrast, fibrinogen does not induce activation of platelets in suspension either because the interaction is dependent on activation of integrin aIIbβ3 or because it cannot crosslink GPVI.

Discussion In this study we show that human GPVI binds to immobilized fibrinogen and that this leads to intracellular signals, which drive the formation of lamellipodial sheets and stress fibers in human platelets and in human GPVIexpressing mouse platelets. This explains the previously paradoxical observation that only human platelets form lamellipodial sheets and stress fibers on a fibrinogen surface, despite mouse platelets being able to form both actin structures in the presence of G protein-coupled receptor agonists such as thrombin. We also show that the interaction of fibrinogen with GPVI is important for aggregate growth providing a new understanding of hemostasis and thrombosis.

We recently identified fibrin as a novel ligand for GPVI7,8,42 and have shown that binding resides in the Ddimer region.30 The observation that fibrinogen also activates GPVI should not, therefore, be a surprise.

Nevertheless, this was unexpected and came from the observation that human platelets deficient in GPVI adhere to but do not spread on fibrinogen. This raises the question as to why this has been previously overlooked. One reason is that mouse platelets do not spread on fibrinogen and thus there is no defect in the absence of GPVI. A second reason is the low level of phosphorylation of the FcR γ-chain induced by fibrinogen in human platelets relative to that by collagen and other GPVI-agonists. This may reflect the extent to which each ligand is able to cluster GPVI and, in the case, of fibrinogen, the dependency on the interaction with integrin aIIbβ3. A third reason is that fibrinogen binds selectively to monomeric GPVI whereas the original binding studies were performed with dimeric GPVI.7,8 It is worth noting that we have reported a reduced number of dimers on immobilized fibrinogen relative to collagen.42 Adhesion of human platelets to fibrinogen is dependent on integrin aIIbβ3. At present, it is not known whether binding to integrin aIIbβ3 is critical for activation of GPVI or simply to promote adhesion such that activation of GPVI can occur. As a dimer, fibrinogen should be able to bind two GPVI monomers, but alternatively the interaction with integrin aIIbβ3 may be required to support activation of monomeric GPVI. A similar role for an integrin in the activation of an ITAM receptor has been reported in other hematopoietic cells with the postulate that the integrin and the ITAM receptor would be associated via a linker protein.43 Fibrinogen is present in whole blood at a concentration of 2 - 4 mg/mL but does not induce platelet activation.

This may be explained by an inability of soluble fibrinohaematologica | 2018; 103(5)

gen to bind GPVI in suspension due to conformational differences between circulating and immobilized fibrinogen.

Alternatively it may be due to the inability of the dimeric fibrinogen to cluster GPVI on the platelet surface in suspension or because of a dependency on binding to integrin aIIbβ3. While the affinity of fibrinogen for GPVI is in the range of that for collagen for GPVI,44,45 we have shown that fibrinogen (and fibrin) bind selectively to monomeric GPVI and this would not be sufficient to induce activation because of the absence of crosslinking. The reason why human platelets, but not mouse platelets, spread on fibrinogen is unclear. Based on the fact that human and mouse GPVI share 64% homology,33 one could imagine that only human GPVI binds to fibrinogen or that both bind to this adhesive protein but only human GPVI is able to promote activation.

GPVI is primarily known as the major signaling receptor for collagen. However, in recent years, GPVI has been shown to bind to other ligands including laminin, the transmembrane protein emmprin, adiponectin, histones and fibrin.6-8,10,47 The physiological significance of many of these interactions is uncertain, in part because of their low affinity or because of whether they acutally occur in vivo.

The interaction that has received the greatest attention is that with fibrin which lies at the interface of the core and shell of the growing platelet aggregate.7,8 This interaction

A

Figure 5. Human glycoprotein VI, but not integrin aIIbβ3 plays a major role in the regulation of the calcium signaling after platelet adhesion to fibrinogen.

Washed platelets from wild-type (WT) or mice expressing human GPVI (hGPVI) were loaded with Oregon-green Bapta-1-AM and Calcein red orange and deposited on immobilized fibrinogen (100 μg/mL). Modifications in fluorescence of individual adherent platelets were monitored for 7 min by confocal microscopy and the Ca2+ concentrations were determined as detailed in the Methods section. (A)(i). Typical time-course Ca2+ profile of one representative platelet adhering to fibrinogen. (A)(ii). Dot plot representing the number of calcium spikes over a period of 5 min. Each point represents an individual platelet. The results are presented as the mean±SEM of five independent experiments (one-way ANOVA, Bonferroni post-hoc test, ***P<0.001).

905

P.H. Mangin et al.

A

Figure 6. Blockade of human glycoprotein VI limits platelet accumulation to a growing aggregate. (A)(i). Hirudinated human whole blood labeled with DIOC6 (1 μmol/L) was perfused over immobilized collagen (200 μg/mL) to preform aggregates for 2 min 30 s, before perfusing hirudinated blood from the same donor in the presence of the Alexa Fluor 647-conjugated monoclonal antibody against GPIbβ (5 μg/mL) and with a Fab control (Control) or the blocking anti-GPVI antibody 9O12 (50 μg/mL). (A)(i). Representative 3D reconstructions from confocal images of aggregates obtained after 7 min 30 s of blood perfusion at 300 s-1. Preformed aggregates are represented in gray, aggregates formed in the presence of a Fab control are represented in red and aggregates formed in the presence of the Fab 9O12 are depicted in orange. (A)(ii). Bar graph representing the volume of the platelet aggregates (mean±SEM) in eight random fields, in six separate experiments performed with different blood donors (Mann-Whitney test, ***P<0.001). The gray, red and orange colors represent the volume of the preformed aggregates, the aggregates formed in the presence of a Fab control and the aggregates formed in the presence of the Fab 9O12, respectively.

takes place at a critical checkpoint in aggregate consolidation and aggregate growth. Thus, GPVI has the potential to both initiate and propagate thrombus formation through interactions with collagen and fibrin and, it appears now, also with fibrinogen. Moreover, while collagen and fibrin are localized at the base of the thrombus and in the core, respectively, fibrinogen is found throughout the aggregate. This suggests a model in which thrombus growth could be sustained by GPVI/fibrinogen, potentially in association with other adhesive proteins. Indeed, in addition to fibrinogen other adhesive proteins such as von Willebrand factor and fibronectin have been shown to support thrombus growth.47–50 Whether these proteins participate in GPVI-mediated platelet aggregation is unclear since von Willebrand factor is not known to be a ligand of GPVI and fibronectin does not directly promote platelet adhesion and activation through GPVI.51 Selective inhibition of the interaction of GPVI with collagen, fibrinogen and fibrin is required to establish their respective contributions to platelet activation in hemostasis and thrombosis.

The discovery that GPVI initiates and propagates platelet aggregation at sites of vessel injury suggests a major role in hemostasis and thrombosis. Paradoxically, however, mice and humans deficient in GPVI only have at most a mild bleeding diathesis, which in the case of humans may be due to additional confounders such as a low platelet count as seen in patients with immuneinduced thrombocytopenia caused by antibodies to GPVI. The relatively minor role of GPVI in hemostasis

References 1. Suzuki-Inoue K, Tulasne D, Shen Y, et al.

Association of Fyn and Lyn with the prolinerich domain of glycoprotein VI regulates intracellular signaling. J Biol Chem.

2002;277(24):21561-21566.

906

can be explained by redundancy in pathways of platelet activation, with the GPIb-von Willebrand factor axis initiating hemostasis, and ADP, thromboxane and thrombin inducing powerful activation. Additionally, the reactive fibrillar type I and III collagen present in deeper layers of the vessels would limit the role of GPVI in the hemostasic response following superficial injury. On the other hand, the discovery that fibrin and immobilized fibrinogen activate GPVI may be of significance at sites of fibrinogen deposition or fibrin formation in diseased vessels following inflammation or loss of vascular integrity. The ability of fibrin and immobilized fibrinogen to activate GPVI may also reflect yet-to-be-discovered new roles for GPVI. In conclusion, in the present study, we have identified immobilized fibrinogen as a novel activator of human but not mouse GPVI and have shown that this interaction supports platelet aggregation under flow. This further emphasizes the contribution of GPVI to platelet activation in thrombosis.

Acknowledgments This work was supported by the British Heart Foundation (RG/13/18/30563); SPW holds a BHF Chair (CH03/003) and ATH holds a BHF Studentship (FS/15/71/31677). NLL’s contract was funded by the Agence Nationale pour la Recherche ANR-14-CE35-0022-02. The authors would like to thank Victor Tybulewicz and Edina Schweighoffer for providing critical reagents.

2. Ezumi Y, Shindoh K, Tsuji M, Takayama H. Physical and functional association of the Src family kinases Fyn and Lyn with the collagen receptor glycoprotein VI-Fc receptor γ chain complex on human platelets. J Exp Med. 1998;188(2):267-276.

3. Quek LS, Pasquet JM, Hers I, et al. Fyn and Lyn phosphorylate the Fc receptor gamma

chain downstream of glycoprotein VI in murine platelets, and Lyn regulates a novel feedback pathway. Blood. 2000;96(13):42464253.

4. Severin S, Nash CA, Mori J, et al. Distinct and overlapping functional roles of Src family kinases in mouse platelets. J Thromb Haemost. 2012;10(8):1631-1645.

haematologica | 2018; 103(5)

Fibrinogen, GPVI and platelet activation 5. Nieswandt B, Watson SP. Platelet-collagen interaction: is GPVI the central receptor?

Blood. 2003;102(2):449-461.

6. Inoue O, Suzuki-Inoue K, McCarty OJ, et al.

Laminin stimulates spreading of platelets through integrin alpha6beta1-dependent activation of GPVI. Blood. 2006;107(4):14051412.

7. Mammadova-bach E, Ollivier V, Loyau S, et al. Platelet glycoprotein VI binds to polymerized fibrin and promotes thrombin generation. Blood. 2016;126(5):683-692.

8. Alshehri OM, Hughes CE, Montague S, et al. Fibrin activates GPVI in human and mouse platelets. Blood. 2015;126(13):16011608.

9. Riba R, Hughes CE, Graham A, Watson SP, Naseem KM. Globular adiponectin induces platelet activation through the collagen receptor GPVI-Fc receptor chain complex. J Thromb Haemost. 2008;6(6):1012-1020.

10. Seizer P, Borst O, Langer HF, et al. EMMPRIN (CD147) is a novel receptor for platelet GPVI and mediates platelet rolling via GPVIEMMPRIN interaction. Thromb Haemost.

2009;101(4):682-686.

11. Schaff M, Tang CJ, Maurer E, et al. Integrin 6 1 is the main receptor for vascular laminins and plays a role in platelet adhesion, activation, and arterial thrombosis.

Circulation. 2013;128(5):541-552.

12. Boulaftali Y, Hess PR, Getz TM, et al.

Platelet ITAM signaling is critical for vascular integrity in inflammation. J Clin Invest.

2013;123(2):908-916.

13. Kuijpers MJE, Gilio K, Reitsma S, et al.

Complementary roles of platelets and coagulation in thrombus formation on plaques acutely ruptured by targeted ultrasound treatment: a novel intravital model. J Thromb Haemost. 2009;7(1):152-161.

14. Hechler B, Gachet C. Comparison of two murine models of thrombosis induced by atherosclerotic plaque injury. Thromb Haemost. 2011;105 (Suppl):S3-12.

15. Bender M, Hagedorn I, Nieswandt B. Genetic and antibody-induced glycoprotein VI deficiency equally protects mice from mechanically and FeCl3-induced thrombosis. J Thromb Haemost. 2011;9(7):14231426.

16. Nieswandt B, Schulte V, Bergmeier W, et al.

Long-term antithrombotic protection by in vivo depletion of platelet glycoprotein VI in mice. J Exp Med. 2001;193(4):459-469.

17. Matus V, Valenzuela G, Sáez CG, et al. An adenine insertion in exon 6 of human GP6 generates a truncated protein associated with a bleeding disorder in four Chilean families.

J Thromb Haemost.

2013;11(9):1751-1759.

18. Dumont B, Lasne D, Rothschild C, et al.

Absence of collagen-induced platelet activation caused by compound heterozygous GPVI mutations. Blood. 2009;114(9):19001903.

19. Hermans C, Wittevrongel C, Thys C, Smethurst PA, Van Geet C, Freson K. A compound heterozygous mutation in glycoprotein VI in a patient with a bleeding disorder.

J Thromb Haemost. 2009;7(8):1356-1363.

20. Zahid M, Mangin P, Loyau S, et al. The future of glycoprotein VI as an antithrombotic target. J Thromb Haemost.

2012;10(12):2418-2427.

21. Bennett JS, Berger BW, Billings PC. The structure and function of platelet integrins. J Thromb Haemost. 2009;7(Suppl 1):200-205.

haematologica | 2018; 103(5)

22. Obergfell A, Judd BA, Del Pozo MA, Schwartz MA, Koretzky GA, Shattil SJ. The molecular adapter SLP-76 relays signals from platelet integrin aIIbβ3 to the actin cytoskeleton. J Biol Chem. 2001;276(8): 5916-5923.

23. Goncalves I, Hughan SC, Schoenwaelder SM, Yap CL, Yuan Y, Jackson SP. Integrin aIIbβ3-dependent calcium signals regulate platelet-fibrinogen interactions under flow: Involvement of phospholipase Cγ2. J Biol Chem. 2003;278(37):34812-34822.

24. Wonerow P, Pearce AC, Vaux DJ, Watson SP. A critical role for phospholipase Cγ2 in aIIbβ3-mediated platelet spreading. J Biol Chem. 2003;278(39):37520-37529.

25. Watson SP, Auger JM, McCarty OJ, Pearce AC. GPVI and integrin aIIbβ3 signaling in platelets. J Thromb Haemost. 2005;3(8): 1752-1762.

26. Boylan B, Gao C, Rathore V, Gill JC, Newman DK, Newman PJ. Identification of FcgammaRIIa as the ITAM-bearing receptor mediating alphaIIbbeta3 outside-in integrin signaling in human platelets. Blood.

2008;112(7):2780-2786.

27. Hughes CE, Finney BA, Koentgen F, Lowe KL, Watson SP. The N-terminal SH2 domain of Syk is required for (hem) ITAM, but not integrin, signaling in mouse platelets. Blood.

2015;125(1):144-155.

28. Zhi H, Rauova L, Hayes V, et al. Cooperative integrin/ITAM signaling in platelets enhances thrombus formation in vitro and in vivo. Thromb Haemost. 2013;121(10): 1858-1867.

29. Law DA, DeGuzman FR, Heiser P, Ministri-Madrid K, Killeen N, Phillips DR. Integrin cytoplasmic tyrosine motif is required for outside-in aIIbβ3 signalling and platelet function. Nature. 1999;401 (6755):808-811.

30. Onselaer M-B, Hardy AT, Wilson C, et al.

Fibrin and D-dimer bind to monomeric GPVI. Blood Adv. 2017; 1(19):1495-1504.

31. Kato K, Kanaji T, Russell S, et al. The contribution of glycoprotein VI to stable platelet adhesion and thrombus formation illustrated by targeted gene deletion. Blood.

2003;102(5):1701-1707.

32. Mangin PH, Tang C, Bourdon C, et al. A humanized glycoprotein VI (GPVI) mouse model to assess the antithrombotic efficacies of anti-GPVI agents. J Pharmacol Exp Ther. 2012;341(1):156-163.

33. Jandrot-Perrus M, Busfield S, Lagrue AH, et al. Cloning, characterization, and functional studies of human and mouse glycoprotein VI: a platelet-specific collagen receptor from the immunoglobulin superfamily. Blood.

2000;96(5):1798-1807.

34. Perrault C, Moog S, Rubinstein E, et al. A novel monoclonal antibody against the extracellular domain of GPIbbeta modulates vWF mediated platelet adhesion. Thromb Haemost. 2001;86(5):1238-1248.

35. Lecut C, Feeney LA, Kingsbury G, et al.

Human platelet glycoprotein VI function is antagonized by monoclonal antibodyderived Fab fragments. J Thromb Haemost.

2003;1(12):2653-2662.

36. Lebozec K, Jandrot-Perrus M, Avenard G, Favre-Bulle O, Billiald P. Design, development and characterization of ACT017, a humanized Fab that blocks platelet’s glycoprotein VI function without causing bleeding risks. MAbs. 2017:9(6):945-958.

37. Hughes CE, Sinha U, Pandey A, Eble JA,

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

O’Callaghan CA, Watson SP. Critical role for an acidic amino acid region in platelet signaling by the HemITAM (hemi-immunoreceptor tyrosine-based activation motif) containing receptor CLEC-2 (C-type lectin receptor2). J Biol Chem. 2013;288(7):5127-5135.

Hui KY, Haber E, Matsueda GP,Monoclonal antibodies to a synthetic fibrin-like peptide bind to human fibrin but not fibrinogen.

Science 1983,222(4628):1129-1132.

Cazenave J-P, Ohlmann P, Cassel D, Eckly A, Hechler B, Gachet C. Preparation of washed platelet suspensions from human and rodent blood. Methods Mol Biol. 2004;272:13-28.

Schaff M, Receveur N, Bourdon C, et al.

Novel function of tenascin-C, a matrix protein relevant to atherosclerosis, in platelet recruitment and activation under flow.

Arterioscler Thromb Vasc Biol. 2011;31(1): 117-124.

Haimovich B, Lipfert L, Brugge JS, Shattil SJ. Tyrosine phosphorylation and cytoskeletal reorganization in platelets are triggered by interaction of integrin receptors with their immobilized ligands. J Biol Chem.

1993;268(21):15868-15877.

Poulter NS, Pollitt AY, Owen DM, et al.

Clustering of glycoprotein VI (GPVI) dimers upon adhesion to collagen as a mechanism to regulate GPVI signaling in platelets. J Thromb Haemost. 2017;15(3):549-564.

Evans R, Patzak I, Svensson L, et al. Integrins in immunity. J Cell Sci. 2009;122(2):215-225.

Miura Y, Takahashi T, Jung SM, Moroi M. Analysis of the interaction of platelet collagen receptor glycoprotein VI (GPVI) with collagen: a dimeric form of GPVI, but not the monomeric form, shows affinity to fibrous collagen. J Biol Chem. 2002;277 (48):46197-46204.

Loyau S, Dumont B, Ollivier V, et al. Platelet glycoprotein VI dimerization, an active process inducing receptor competence, is an indicator of platelet reactivity. Arterioscler Thromb Vasc Biol. 2012;32(3):778-785.

Aslan JE, Phillips KG, Healy LD, et al.

Histone deacetylase 6-mediated deacetylation of a-tubulin coordinates cytoskeletal and signaling events during platelet activation. AJP Cell Physiol. 2013;305(12):C1211C1213.

Kulkarni S, Dopheide SM, Yap CL, et al. A revised model of platelet aggregation. J Clin Invest. 2000;105(6):783-791.

Yang H, Reheman A, Chen P, Zhu G. Fibrinogen and von Willebrand factor-independent platelet aggregation in vitro and in vivo. J Thromb Haemost. 2006;4(10):22302237.

Reheman A, Yang H, Zhu G, et al. Plasma fibronectin depletion enhances platelet aggregation and thrombus formation in mice lacking fibrinogen and von Willebrand factor plasma fibronectin depletion enhances platelet aggregation and thrombus formation in mice lacking fibrinogen and von Willebrand factor. 2012;113(8):18091817.

Cho J, Mosher DF. Impact of fibronectin assembly on platelet thrombus formation in response to type I collagen and von Willebrand factor. Blood. 2006;108(7):22292236.

Maurer E, Schaff M, Receveur N, Bourdon C, Mercier L, Nieswandt B. Fibrillar cellular fibronectin supports efficient platelet function and procoagulant activity. Thromb Haemost. 2015;114(6):1175-1188.

907

ARTICLE

Coagulation & its Disorders

Ferrata Storti Foundation

Haematologica 2018 Volume 103(5):908-918

Circulating microRNAs as biomarkers of disease and typification of the atherothrombotic status in antiphospholipid syndrome

Carlos Pérez-Sánchez,1* Iván Arias-de la Rosa,1* María Ángeles Aguirre,1,2 María Luque-Tévar,1 Patricia Ruiz-Limón,1 Nuria Barbarroja,1 Yolanda Jiménez-Gómez,1 María Carmen Ábalos-Aguilera,1 Eduardo Collantes-Estévez,1,2,3 Pedro Segui,1,4 Francisco Velasco,5 María Teresa Herranz,6 Jesús Lozano-Herrero,6 María Julia Hernandez-Vidal,6 Constantino Martínez,7 Rocío González-Conejero,7 Massimo Radin,8 Savino Sciascia,8 Irene Cecchi,8 María José Cuadrado9** and Chary LópezPedrera1,2**

1 Instituto Maimónides de Investigación Biomédica de Córdoba (IMIBIC), Spain; 2Unidad de Gestión Clínica Reumatología, Hospital Universitario Reina Sofía, Córdoba, Spain; 3 Departamento de Medicina (Medicina, Dermatología y Otorrinolaringología), Universidad de Córdoba, Spain; 4Unidad de Gestión Clínica Radiología, Hospital Universitario Reina Sofía, Córdoba, Spain; 5Unidad de Gestión Clínica Hematología, Hospital Universitario Reina Sofía, Córdoba, Spain; 6Servicio de Medicina Interna, Hospital Morales Meseguer, Murcia, Spain; 7Centro Regional de Hemodonación, Universidad de Murcia, IMIB-Arrixaca, Spain; 8Department of Clinical and Biological Sciences, Center of Research of Immunopathology and Rare Diseases, Torino, Italy and 9 Lupus Research Unit and St. Thomas’ Hospital, London, UK

*CP-S and IA-R shared first authorship and contributed equally to this work. **MJC and CL-P shared last authorship and contributed equally to this work.

ABSTRACT

Correspondence: rosario.lopez.exts@juntadeandalucia.es

Received: November 10, 2017.

Accepted: February 22, 2018.

Pre-published: March 15, 2018. doi:10.3324/haematol.2017.184416 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/5/908 ©2018 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.

908

W

e aimed to identify the plasma miRNA profile of antiphospholipid syndrome (APS) patients and to investigate the potential role of specific circulating miRNAs as non-invasive disease biomarkers. Ninety APS patients and 42 healthy donors were recruited.

Profiling of miRNAs by PCR-array in plasma of APS patients identified a set of miRNAs differentially expressed and collectively involved in clinical features. Logistic regression and ROC analysis identified a signature of 10 miRNA ratios as biomarkers of disease. In addition, miRNA signature was related to fetal loss, atherosclerosis, and type of thrombosis, and correlated with parameters linked to inflammation, thrombosis, and autoimmunity. Hard clustering analysis differentiated 3 clusters representing different thrombotic risk profile groups. Significant differences between groups for several miRNA ratios were found. Moreover, miRNA signature remained stable over time, demonstrated by their analysis three months after the first sample collection. Parallel analysis in two additional cohorts of patients, including thrombosis without autoimmune disease, and systemic lupus erythematosus without antiphospholipid antibodies, each displayed specific miRNA profiles that were distinct from those of APS patients. In vitro, antiphospholipid antibodies of IgG isotype promoted deregulation in selected miRNAs and their potential atherothrombotic protein targets in monocytes and endothelial cells. Taken together, differentially expressed circulating miRNAs in APS patients, modulated at least partially by antiphospholipid antibodies of IgG isotype, might have the potential to serve as novel biomarkers of disease features and to typify patients’ atherothrombotic status, thus constituting a useful tool in the management of the disease.

Introduction Antiphospholipid syndrome (APS) is a clinical disorder characterized by the occurrence of thrombosis and/or pregnancy morbidity associated with the persistent presence of antiphospholipid antibodies (aPL), including anti-cardiolipin antihaematologica | 2018; 103(5)

MicroRNAs as antiphospholipid syndrome biomarkers

Table 1. Clinical and laboratory parameters of the antiphospholipid syndrome (APS) patients and the healthy donors (HDs).

Females/males Age, years Arterial thrombosis Venous thrombosis Recurrences Pregnancy morbidity Pathological CIMT ABI – left* ABI – right* LA positivity aCL-IgG# GPL aCL-IgM# MPL Anti-β2GPI IgG# SGU Anti-β2GPI IgM# SMU Antiplatelet agents† Anticoagulant agents‡ Total cholesterol level,* mg/dL Cholesterol HDL level,* mg/dL Cholesterol LDL level,* mg/dL Triglycerides level,* mg/dL ESR,* mm/h

APS (total n. 90)

HDs (total n. 42)

48/42 51.2 ± 13.1 35/90 55/90 37/90 23/90 24/90 1.3 ± 0.11 1.27 ± 0.11 85/90 23.4 (0.5-448) 21.8 (0-354) 23.9 (0-361) 14.8 (0-289) 30/90 62/90 191.7 ± 32.06 51.09 ± 12.4 112.1 ± 33.2 155.1 ± 163.2 13.5 ± 13.6

22/20 46.2 ± 13.4 0/42 0/42 0/42 0/42 6/42 1.2 ± 0.09 1.2 ± 0.09 0/42 1.3 (0.5-5) 4.9 (0.8-17) 1 (1-2) 1.2 (1-2.6) 0/42 0/42 190 ± 41.8 55.4 ± 13.1 118.4 ± 26.9 88.3 ± 50.07 6.6 ± 4

P

NS

0.00 0.02 0.02 0.00 0.00 0.00 0.02 0.02

NS NS NS NS 0.05

n: number; NS: not significant; CIMT: carotid intima-media thickness; ABI: ankle brachial index; LA: lupus anticoagulant; aCL: anti-cardiolipin antibodies; GPL: IgG phospholipid units; MPL: IgM phospholipid units; anti-β2GPI: anti-β2 glycoprotein 1 antibodies; SGU: stantard IgG units; SMU: standard IgM units; HDL: high-density lipoprotein; LDL: low-density lipoprotein; ESR: erythrocyte sedimentation rate. *Results expressed as mean±Standard Deviation. #Results expressed in mean and values range. †Antiplatelet agents include acetylsalicylic acid and clopidogrel. ‡Anticoagulant agents indicate vitamin K antagonists, including warfarin and acenocumarol.

bodies (aCL), anti-β2-glycoprotein 1 antibodies (antiβ2GPI) and/or lupus anticoagulant (LA). Cardiac, cerebral and vascular strokes in these patients are responsible for a significant reduction in life expectancy.1 The course of cardiovascular disease (CVD) in APS patients may rapidly change from asymptomatic to severe life-threatening manifestations that are difficult to deal with. Timely diagnosis and accurate monitoring of the course of APS are essential to improve the quality of therapy and avoid approaches based on medical empirical protocols. In the same way, like many other autoimmune diseases, APS is a heterogeneous entity, and this has a dramatic impact on diagnosis and treatment.2 Understanding of the pathophysiological mechanisms explaining how atherosclerosis and CVD are associated to APS has been greatly broadened with the application of genomic technologies.3 One emerging and important mechanism controlling gene expression is epigenetics, which regulates gene packaging and independent expression of alterations in the DNA sequence. Epigenetics, which comprises DNA methylation, histone modifications, and microRNAs (miRNAs) activity, is providing new directions linking genomics and environmental factors.4 miRNAs are small, non-coding RNAs that, depending upon base pairing to messenger RNA (mRNA), mediate mRNA cleavage, translational repression or mRNA destabilization. miRNAs are known to be involved in crucial cellular processes and their dysregulation has been described in many cell types and fluids in a broad range of diseases.5-7 haematologica | 2018; 103(5)

In the setting of APS, a previous study by our group recognized that aPL modulate the expression of 2 miRNAs in monocytes (miR-19b and miR-20a) that control the expression of key proteins involved in the pathology of the disease, such as tissue factor (TF).8 Moreover, we recently demonstrated that both aPL and the anti-double stranded DNA antibodies (anti-dsDNA) promote specific changes in the expression of proteins related to the biogenesis of miRNAs in leukocytes of APS and systemic lupus erythematosus (SLE) patients, which are translated in the altered expression of the miRNAs profile and that of their protein targets (related to CVD) in these disorders.9 Extensive analyses have shown that miRNAs are released into the circulation where they are present in concentration levels that differ between healthy subjects and patients. Although little is known about the origin and function of such circulating miRNAs, these molecules are increasingly recognized as non-invasive and readily accessible biomarkers for risk stratification, diagnosis and prognosis of multiple forms of CVD.10 Specific profiles of circulating miRNAs are also associated to the pathophysiology of different systemic autoimmune diseases, including SLE, systemic sclerosis, and rheumatoid arthritis (RA), and some of them appear to be of diagnostic and, possibly, of prognostic value.11 To date, in the context of APS, no study has analyzed the potential role of the circulating miRNAs as biomarkers of the disease. Therefore, the present study was designed to determine the plasma miRNA specific profile of APS patients, 909

C. Perez-Sánchez et al.

A

B

C

Figure 1. Antiphospholipid syndrome (APS) patients showed a specific circulating miRNAs profile related to clinical features of this autoimmune disorder. (A) To identify the changes that occurred in the expression levels of microRNAs (miR) in plasma from antiphospholipid syndrome versus controls, Human Serum & Plasma miRNA PCR-array (Qiagen) was performed in the study cohort. Expression levels of 19 miRNAs were found up-regulated in antiphospholipid syndrome, while 20 miRNAs were down-regulated. (B) Ingenuity Pathway Analysis (IPA) uncovered the main enriched biological functions and pathways in which these microRNAs are involved. The analysis included only the functions and pathways with average IPA score >2 [indicated as -log (P value)]. (C) Validation of selected miRNAs by RT-PCR in the whole cohort of APS patients and healthy donors. *P<0.05.

their modulation by autoantibodies, and their potential role as non-invasive biomarkers of disease features.

ical parameters, B-Mode Ultrasound IMT and Ankle Brachial Index measurements see the Online Supplementary Appendix.

Methods

Isolation of miRNAs and analysis of miRNAs expression profiling

Patients Ninety patients with primary APS and 42 healthy donors (HDs) were included in this study over a period of 24 months (see Online Supplementary Appendix). All experimental protocols were approved by the ethics committee of the Reina Sofia Hospital, Cordoba, Spain, and written informed consent was obtained. The characteristics of patients and HDs are shown in Table 1.

The adjusted global anti-phospholipid syndrome score (aGAPSS) was calculated for each APS patient, as previously described.12 Briefly, aGAPSS was calculated by adding the points corresponding to both the cardiovascular and thrombotic risk factors, based on a linear transformation derived from the β regression coefficient as follows: 3 for hyperlipidemia, 1 for arterial hypertension, 5 for aCL IgG/IgM, 4 for anti-β2GPI IgG/IgM, and 4 for LA. Two additional cohorts of patients were further analyzed as disease control, including 23 patients with thrombosis in the absence of an associated autoimmune disease [12 non-pregnant women and 11 men; mean age 44 (range: 21-73 years), including patients with objectively verified thrombotic events: 14 deep venous thrombosis and 9 thrombosis in intra-cerebral vessels], and 25 SLE patients without aPLs (Online Supplementary Table S1).

For details of blood sample collection and assessment of biolog910

Total RNA, including the miRNA fraction, was extracted from both plasma and supernatants obtained from in vitro studies by using the QIAzol miRNeasy kit (Qiagen, Valencia, CA, USA) following the manufacturer's instructions13 (Online Supplementary Appendix). To identify the changes that occurred in the expression levels of miRNAs in plasma from APS patients and HDs, a Human Serum & Plasma miRNA PCR-array (Qiagen) was performed (Online Supplementary Appendix) on an exploratory cohort (Online Supplementary Table S2).

Quantitative real-time PCR

A fixed volume of 3 μl of RNA solution from the 14 μl-eluate from RNA isolation of 200 μl plasma sample was used as input into the reverse transcription. Input RNA was reverse transcribed using the TaqMan miRNA Reverse Transcription kit and miRNAspecific stem-loop primers (Life Technologies, Madrid, Spain) (Online Supplementary Appendix). The expression levels of miRNAs were calculated by using 2-Ct and reciprocal ratios were performed [Ratio miR-A/miR-B = log2 (2-Ct miR-A/2-Ct miR-B)], as previously described.14-19 Reciprocal ratios analysis is an approach that bypasses the controversial issue of data normalization of miRNAs in plasma (self-normalization). Furthermore, miRNAs whose concentrations are changed because of a pathology in opposite directions can be effective in differentiating investigated populations. haematologica | 2018; 103(5)

MicroRNAs as antiphospholipid syndrome biomarkers

Figure 2. Interaction network of microRNAs identified potential mRNA targets involved in clinical features of antiphospholipid syndrome. Using microRNA Target Filter of QIAGENâ&#x20AC;&#x2122;s Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, CA, USA, www.qiagen.com/ingenuity), the software generated a network including the selected microRNAs (miRNAs or miR) and their mRNA targets, filtered by coronary artery disease, thrombosis, abortion and cerebrovascular dysfunction. Only targets experimentally observed and predicted with high confidence are shown and related by direct interactions to their specific miRNA regulators.

Target gene prediction and integrated analysis by Ingenuity Pathway Analysis The altered miRNAs were further analyzed to obtain information about biological functions, pathways and networks by using the web-based bioinformatics tool QIAGENâ&#x20AC;&#x2122;s Ingenuity Pathway Analysis (IPA; Ingenuity Systems, http://www.INGENUITY.com).

For this purpose, all differentially regulated miRNAs and fold changes were imported into IPA20 (Online Supplementary Appendix).

Details of purification of IgG, in vitro exposure of monocytes and endothelial cells to aPL antibodies, and the statistical analysis are available in the Online Supplementary Appendix.

Results Differentially expressed miRNAs in the plasma of APS patients and HDs In the discovery phase (exploratory cohort), we identified 39 miRNAs that were differentially expressed between APS patients and HDs (cut off: 1.7-fold change), including 19 up-regulated and 20 down-regulated (Figure 1A). Using the IPA software, the functional analysis of the altered miRNAs in APS patients showed that a large number of them had validated and putative target mRNAs mainly involved in connective tissue disorders, inflammatory response, reproductive system disease, CVD or skeletal and muscular disorders (Figure 1C). haematologica | 2018; 103(5)

Bioinformatic identification and analysis of deregulated miRNAs related to the pathophysiology of APS and analysis of potential protein targets In silico studies were performed to identify the altered miRNAs that might have as potential targets a number of genes/proteins involved in the development of clinical manifestations related to APS, such as coronary artery disease, thrombosis, abortion, and cerebrovascular dysfunction. IPA identified 11 altered miRNAs as the main regulators of proteins involved in the pathology of APS, including miRNA 34a-5p, 15a-5p, 145a-5p, 133b-3p, 124-3p, 206, 20a-5p, 19b-3p, 210-3p, 296-5p and 374a-5p. This set of 11 miRNAs included, among others, the top 5 up-regulated miRNAs and 3 out of the top 5 down-regulated miRNAs in the PCR-array. The expression levels of the 11 selected miRNAs were analyzed in all study subjects by RT-PCR (Figure 1B). MiR-124 and miR-34a were found increased in APS patients in relation to healthy donors, while miR-20a, miR-19b and miR145a were found reduced. The remaining microRNAs were also found to be altered, showing a trend to either increase or reduction as observed in the discovery phase, thus validating the data obtained by PCR-array.

We further developed a network that defined the interaction between miRNA-mRNA targets (Figure 2). Key proteins involved in the pathophysiology of APS, and identi911

C. Perez-Sánchez et al.

A

B

Figure 3. A circulating miRNA signature in antiphospholipid syndrome (APS) might have potential value as biomarkers of disease. (A) Selected microRNAs (miRNAs or miR) were analyzed in the whole cohort, including 90 APS patients and 42 healthy donors, and reciprocal ratios were performed. Beeswarm plot of each differentially expressed miR ratio is shown, along with mean, Standard Deviation, and P-value. For statistical analysis, after normality and equality of variance tests, comparisons were made by paired Student t-test or a non-parametric test (Mann-Whitney rank sum test). (B) A combination of the 10 miRNA ratios as a panel was carried out by using logistic regression on the data set. ROC curve of miRNA panel and cut off were generated based on the predicted probability (P) for each subject as a single score. The equation used in our model was: “Combined miRNA-ratio panel [Logit(p)] = - 0.64 + 0.034x(miR-19b/miR-34a) + 1.061x(miR-19b/miR-15a) + 0.248x(miR-19b/miR-124) – 1.704x(miR-19b/miR-145) + 2.34x(miR-20a/miR-145) – 0.729x(miR-20a/miR-374a) – 0.624x(miR-20a/miR-210) + 0.088x(miR20a/miR-133b) + 0.166x(miR-206/miR-34a) + 0.056x(mir-124/miR-296)”. The area under the curve (AUC), sensitivity and specificity are displayed, and a cut-off value with higher specificity was selected.

fied as potential mRNA targets of those miRNAs, were quantified in the plasma of APS patients and HDs. As previously reported,20-23 APS patients showed significantly increased plasma levels of TF, PAI-1, MCP-1, VEGF-A and VEGFR-1 (Online Supplementary Figure S1).

Circulating miRNA signature as potential biomarkers of disease in APS It has been shown that the combination of miRNAs improves their predictive potential to differentiate two pathological conditions.14-19 Thus, to assess the potential of specific circulating miRNAs in APS patients as biomarkers of disease features, reciprocal ratios of the miRNAs analyzed were performed by using statistical tools. By this approach, we identified 10 miRNA ratios, integrated by the 11 selected miRNAs, and differentially expressed in plasma of APS patients in comparison with HDs, including miR-19b/miR-34a, miR-19b/miR-15a, miR19b/miR124, miR-19b/miR-145, miR-20a/miR-145, miR-20a/miR374a, miR-20a/miR-210, miR-20a/miR-133b, miR206/miR-34a and miR-124/miR-296 (Figure 3A). To further explore the efficiency of these biomarkers to identify 912

APS patients, a combination of the 10 miRNA ratios as a panel was carried out by using a logistic regression on the data set, as previously described.24 Thus, all miRNA-ratios were integrated into a single model or equation, which provided a single ‘score’ that allowed us to perform the ROC analysis and establish the cut off for prediction. The ROC curve for the 10 miRNA ratios signature revealed a marked accuracy, evidenced by an AUC of 0.81. At the optimal cut-off value of 0.6, the sensitivity and specificity of the combined miRNA panel for APS identification were 78% and 80%, respectively (Figure 3B).

Stability of miRNA expression profile over time in APS Plasma from 21 APS patients included in the study was evaluated again three months after the first blood sample collection to analyze the stability of the circulating miRNA profile. Results demonstrated that miRNA expression in the second sample collection did not change in relation to the first analysis (Online Supplementary Figure S2A). Moreover, the levels of miRNA ratios at baseline correlated significantly with the levels of these ratios three months later (Online haematologica | 2018; 103(5)

MicroRNAs as antiphospholipid syndrome biomarkers

Figure 4. Antiphospholipid syndrome (APS) patients show a specific miRNA profile distinct from both non-autoimmune patients with previous thrombotic events and aPL-negative systemic lupus erythematosus (SLE) patients. Twenty-three thrombotic non-antiphospholipid syndrome patients (non-APS) and 25 aPL-negative SLE patients were included, and the circulating microRNA (miRNA or miR) signatures of APS were compared. One-way ANOVA was used for statistical comparisons.

A Bonferroni correction was applied for multiple testing. P<0.05 was considered statistically significant. Beeswarm plot of each differentially expressed miRNA ratio is shown along with mean, Standard Deviation and P-value. n.s.: no significant statistical differences.

Supplementary Figure S2B). Thus, our data support the theory that there is a specific circulating miRNA signature in APS which remains stable over time.

APS patients show a specific miRNA profile different from both non-autoimmune patients with previous thrombotic events and aPL-negative SLE patients To assess the specificity of the miRNA signature found in APS patients, and in order to analyze whether the altered expression of the circulating miRNAs evaluated was linked to their thrombophilic status, an additional disease group including 23 patients with thrombosis in the absence of an associated autoimmune disease was evaluated. In these patients, the ratios formed by the expression levels of the 11 selected miRNAs were significantly different from those described in APS patients, except for the ratios miR-19b/miR-15a and miR-19b/miR-145 which exhibited non-significant differences (Figure 4). To evaluate if the altered expression of the miRNA signature was a sign of an autoimmune status, an additional disease group, including 25 SLE patients negative for aPL, was analyzed. In this SLE cohort, the ratios produced by the selected circulating miRNAs were significantly different from those found in APS patients, except for the ratios miR-19b/miR-34a, miR-20a/miR-374a and miR-124/miR296 which exhibited non-significant differences (Figure 4).

ences in miRNA signature, except for the ratio miR20a/374 (Online Supplementary Figure S3).

Circulating miRNAs are associated with clinical features of APS and show potential as biomarkers for the development of atherosclerosis The levels of some circulating miRNA ratios that integrate the signature in APS were associated with the ocurrence of fetal losses in these patients, including elevated levels of miR-19b/miR-124 and miR-20a/miR-374, and reduced levels of miR-124/miR-296 (Figure 5A).

Associations between miRNA ratios and the type of thrombosis suffered by APS patients were also identified.

Thus, elevated levels of ratios miR-20a/miR-145 and miR20a/miR-374 were significantly associated with the ocurrence of arterial thrombosis in APS patients (Figure 5B).

Furthermore, the ratios of miR-19b/miR-124 and miR124/miR-296 were also found to be associated with the presence of a pathological CIMT in these patients (Figure 5C). To accurately evaluate their relevance as biomarkers of early atherosclerosis, we conducted combined ROC analyses of these miRNA ratios. The combination of both circulating miRNA ratios as a panel showed an evident accuracy, with an AUC of 0.76 at a sensitivity of 67% and specificity of 78% from a cut-off value of 0.41 (Figure 5D).

Cluster analysis Potential influence of standard therapy on the profile of circulating miRNAs in APS APS patients were classified into two groups based on the treatment received, including 30 primary APS patients treated with antiplatelet agents and 62 primary APS patients treated with anticoagulant drugs. The statistical comparison between patients treated with antiplatelet and anticoagulant agents showed no significant differhaematologica | 2018; 103(5)

Hard clustering analysis in the APS cohort differentiated 3 clusters representing different thrombotic risk profile groups. Clinical and laboratory parameters of each cluster are resumed (Figure 6A). Briefly, cluster 1 (50% of the clustered cohort) was characterized by lower prevalence of cardiovascular risk factors and aPL multiple positivity.

Conversely, cluster 1 shows a higher rate of venous thrombotic event when compared to the other clusters.

913

C. Perez-Sánchez et al.

B

A

C

D

Figure 5. Circulating miRNAs are related to clinical features of antiphospholipid syndrome (APS) and show potential as biomarkers for the development of atherosclerosis. Association studies of altered circulating microRNA (miRNA or miR) ratios and the occurrence of previous fetal loss (A), the type of thrombosis suffered (B) and the presence of a pathological carotid intima-media thickness (CIMT) (C). Beeswarm plot of each miR ratio is shown, along with mean, Standard Deviation, and P-value. (D) A combination as a panel of the 2 miRNA ratios associated to the pathological CIMT was carried out by using logistic regression on the data set and receiver operator characteristics (ROC) curve analyses were performed. ROC curve of miRNA panel and cut off were generated based on the predicted probability (P) for each patient as a single score. The equation used was: “Combined miRNA-ratio panel [Logit(p)] = 0.599 – 0.133x(miR-19b/miR-124) + 0.007x(miR-124/miR296)”. The area under the curve (AUC), sensitivity and specificity are shown, and a cut-off value with higher specificity was selected.

Cluster 2 (17.6% of the clustered cohort) was characterized by a higher rate of cardiovascular risk factors, arterial thrombotic events, recurrences and a low prevalence of multiple aPL positivity. Cluster 3 (32.4% of the clustered cohort) was represented by a higher rate of multiple aPL positivity, arterial thrombotic events, and lower rate of cardiovascular risk factors. When evaluating different miRNA ratio expression among clusters, we found a statistically significant difference between groups for the following miRNA ratios: miR-19b/miR-124 (P<0.001, ANOVA), miR-20a/miR-374 (P<0.05, ANOVA), miR20a/miR-210 (P<0.001, ANOVA) and miR-124/miR-296 (P<0.05, ANOVA). miRNA ratio expression in the different clusters are summarized in Figure 6C. When comparing the aGAPSS values among the different clusters, we found a significant difference (P=0.008, t-test) between cluster 1 [mean aGAPSS 5.38; 1.628±Standard Deviation (SD)] and Cluster 2 (mean aGAPSS 8,67; 3.67±SD). Similarly, we found a significant difference (P<0.001, t-test) between cluster 1 and cluster 3 (mean aGAPSS 10.82; 2.316±SD). aGAPSS values stratifying for clusters are represented in Figure 6B.

Circulating miRNAs correlate with clinical and serological parameters in APS The miRNA ratios that integrate the signature in APS were linked with clinical parameters, such as ABI, presence of elevated titers of aPL, particularly aCL and antiβ2GPI antibodies, and erythrocyte sedimentation rate (Online Supplementary Table S3). Correlation analyses with serological markers related to atherothrombosis further 914

showed significant positive correlations with the expression levels of various miRNA ratios and with levels of TF, PAI-1, VEGF-A, VEGF-R1 and MCP-1 (Online Supplementary Table S3). Some of these correlations were also found among various miRNA ratios in plasma of APS patients.

Antiphospholipid antibodies modulate the expression of both the circulating miRNAs that integrate the signature in APS and their potential protein targets The expression of the 11 selected miRNAs was significantly altered in the supernatant of HUVECs treated with aPL-IgG in relation to those treated with a non-immuneIgG (Figure 7A), except for the miR-124 and miR-206.

Accordingly, this treatment promoted in HUVECs the secretion of atherothrombotic proteins, such as TF, PAI-1 and VEGF-R1 (Figure 7B), potential targets of the miRNAs analyzed. On the other hand, the expression levels of several miRNAs were deregulated in the supernatant of monocytes treated with aPL-IgG, including miR-19b, miR20a, miR-145, miR-210 and miR-296 (Figure 7C).

Concomitantly, aPL-IgG treatment promoted in monocytes an increase in the secretion of TF, PAI-1 and MCP-1 (Figure 7D).

Discussion The present study identifies, for the first time, a specific signature of circulating miRNAs in APS patients that might serve as potential biomarkers of clinical features of haematologica | 2018; 103(5)

MicroRNAs as antiphospholipid syndrome biomarkers

A

B

C

Figure 6. Specific miRNA signatures might identify subgroups of antiphospholipid syndrome (APS) patients showing different thrombotic risk profiles: cluster analysis. (A) Clinical and laboratory parameters of the 3 clusters. (B) Comparison of the adjusted global anti-phospholipid syndrome score (aGAPSS) values among the different clusters. (C) Evaluation of different microRNA (miR) ratios expression among clusters.

HTA: arterial hypertension; LA: lupus anticoagulant; aCL: anti-cardiolipin IgG/IgM; antiβ2GPI: anti-β2 glycoprotein 1 IgG/IgM.

this autoimmune disorder. Moreover, this signature could represent a useful tool to typify and stratify patients based on their thrombotic status and cardiovascular risk profile (Online Supplementary Figure S3).

Circulating miRNAs were firstly described in peripheral blood as promising specific biomarkers for a wide range of diseases, such as cancer and other inflammatory pathologies.25,26 Thereafter, several studies revealed the altered expression of numerous miRNAs in plasma, blood cells, and tissues of systemic autoimmune conditions, such as RA and SLE, which were directly associated to disease activity, making them potential useful biomarkers for clinical features and follow up.9,26-29 However, to date, the specific profile of circulating miRNAs in APS patients has not been evaluated. In the present study, the profiling of miRNAs by PCR-array in plasma of APS patients has helped to identify a set of miRNAs differentially expressed and collectively associated to clinical features of the disease, such as inflammatory response, reproductive system disease, and CVD, among others. Using logistic regression, we further developed a model that identified 10 miRNA ratios, differentially expressed, that showed great potential as biomarkers of disease of APS patients.

Recent studies support the evidence that an miRNAs signature has a higher diagnostic value than individual miRNAs.14-19 The use of ratios is a feasible approach that overcomes the controversial question of normalizing plasma levels of miRNAs, given the lack of a reliable normalizer for circulating miRNAs. In addition, the establishment haematologica | 2018; 103(5)

of these ratios allows the identification of a combination of expression profiles closer to reality in vivo in patients, where the interactions between miRNAs and their specific potential targets never occur in a unique or individualized way. In fact, it is likely that, in some cases, various miRNAs, whose concentrations are shifted in opposite directions in a particular pathology, contribute together and specifically to certain clinical profiles.

The signatures of circulating miRNAs identified in APS patients integrated miRNAs previously described to be altered in other autoimmune and CVD. Thus, miR-19b and miR-20a have been shown to be essential modulators of TF expression in APS and SLE patients,8 so that reduced expression of such miRNAs contributes to the overexpression of TF in monocytes, which is directly associated with the occurrence of thrombotic events in APS.21 On the other hand, miR-124, found altered in APS, SLE and RA patients at both cellular and plasma levels, modulates the overexpression of MCP-1, a key chemokine directly involved in CVD associated to these autoimmune conditions.30-33 Likewise, miR-133b and miR-145 have been identified as the most promising biomarkers of the pathogenesis of CVD. Both miRNAs participate in the differentiation of vascular smooth muscle cells. In addition, miR133b regulates angiogenesis and endothelial function, while miR-145 participates in the stabilization of atheromatous plaque.34 The miR-34a is highly expressed in endothelial cells, and elevated circulating levels of this miRNA have been associated to myocardial infarction.35 915

C. Perez-Sรกnchez et al.

A

B

C

D

Figure 7. Antiphospholipid antibodies modulate the expression of both the circulating miRNAs that integrate the signature in antiphospholipid syndrome (APS) and their putative protein targets. Human umbilical vein endothelial cells (HUVECs) were treated with antiphospholipid antibodies and secreted selected microRNAs (miRNAs) (A) and putative target protein (B) levels were determined in the supernatant. Monocytes were also treated with antiphospholipid antibodies and secreted selected miRNAs (C) and putative target proteins (D) levels were evaluated in the supernatant of culture. Differences were analyzed by Student t-test. Values are the means and Standard Error of Mean of 4 independent experiments performed in triplicate. P<0.05 was considered statistically significant. TF: tissue factor; PAI-1: plasminogen activator inhibitor-1; VEGF-A: vascular endothelial growth factor A; VEGF-R1: VEGF-Receptor-1; MCP-1: monocyte chemotactic protein.

Moreover, the main target of miR-34a is VEGF-A, a key inflammatory protein involved in numerous cardiovascular and autoimmune pathologies, including APS.23,36 In the same way, miR-374 has been described as regulator of maintenance of vascular integrity.37 The remaining miRNAs members of the signature, including miR-296, miR-210, miR-206 and miRNA-15, have been found altered in severe pre-eclampsia, one of the leading causes of maternal mortality and neonatal morbidity worldwide.38-40 Thus, all the processes regulated by these miRNAs seem to orchestrate distinct aspects of APS pathogenesis.

To assess the specificity of the circulating miRNA signature in APS we evaluated the miRNA profile in an additional cohort of patients characterized by the presence of previous thrombotic events in the absence of an associated autoimmune disease. The miRNAs analysis revealed a differential pattern of expression between these two cohorts. Those results substantiate previous studies that evidenced the presence of a distinct miRNA profile in monocytes and neutrophils of thrombotic non-autoimmune patients compared to APS patients.9 This could reflect a differential mode of regulation and activity of miRNAs in thrombotic patients compared to APS patients, on which the role of autoantibodies might be crucial.

Moreover, the analysis of a parallel autoimmune population (SLE patients) negative for aPL, also identified an miRNA signature distinct from that of APS, thus underlying the potential role of aPLs as regulators of thrombosisrelated miRNAs in APS, and pointing to the presence of a specific miRNA profile relative to the pathogenesis of each disease.

Antiphospholipid syndrome patients recruited in this 916

study were mainly treated with anticoagulant and/or antiplatelet agents. All of them have been shown to influence miRNAs expression, an epigenetic process that might help to delineate the mechanisms underlying their effects.9,41,42 Thus, we evaluated the potential effect of these treatments on the circulating miRNA expression profile. No significant differences were observed in our cohort of APS patients between those who received antiplatelet and those treated with anticoagulant agents, suggesting that the prothrombotic status induced by effects of aPLs, and the consequently deregulated miRNAs, were not differentially modulated among these drugs.

In order to understand the clinical relevance of the altered circulating miRNA signature, association and correlation studies were perfomed. Altered expression of various miRNA ratios was associated with the presence of previous fetal losses. In line with these findings, several studies have shown that the misregulation of circulating placental miRNAs in maternal blood might lead to pregnancy complications, thus acting as non-invasive diagnostic and prognostic biomarkers for pregnancy monitoring.4244 Association studies further established a significantly increased expression of 2 miRNA ratios in APS patients that had suffered arterial thrombosis in comparison with those who experienced venous thrombotic events.

Interestingly, both miRNA ratios were integrated by the miR-20a, previously reported to be the main regulator of TF, whose expression levels have been found to be related to the development of arterial thrombosis in the setting of APS.8,45 Finally, we identified 2 miRNA ratios as clinical relevant biomarkers related to early atherosclerosis development in APS patients, which were integrated by the miRhaematologica | 2018; 103(5)

MicroRNAs as antiphospholipid syndrome biomarkers

19b and miR-124, both of them critical players in the expression of proteins related to inflammation and thrombosis in APS and SLE.8,9 Correlation studies revealed that the altered circulating miRNA signature in APS is linked to parameters related to increased risk of peripheral artery disease such as ABI. Moreover, correlations between circulating miRNA levels and numerous altered parameters related to inflammation and thrombosis were also identified. These correlations support the relationship observed in the in silico study between the selected miRNAs and potential target proteins involved in various clinical features of APS. The influence of the autoimmunity in the circulating profile of miRNAs in APS was also revealed by the significant correlation between high titers of aPL-IgG and the altered expression of several miRNAs integrating the signature.

These relationships further sustain those previously identified among the altered profile of miRNAs in APS and SLE at cellular level and the autoimmune and inflammatory profile of both autoimmune conditions.9 Therefore, our data suggest that the altered plasma profile of miRNAs is an important mechanisin that might contribute to the regulation of the pro-atherothrombotic status of APS patients, on which aPL seem to play a key role.

Our in vitro studies further confirmed this hypothesis, demonstrating that aPL-IgG antibodies promoted a significant deregulation in the expression levels of both the selected miRNAs and their potential protein targets in the supernatant of cultured monocytes and HUVECs, the main drivers of the CVD in the setting of APS. These results also confirm and complement previous studies which showed the in vitro effects of aPL-IgG in the induction of prothrombotic/inflammatory mediators3,31 and the modulation of specific cellular miRNAs involved in their modulation.8,9 Nevertheless, although our data show specific effects of aPL-IgG on the secretion of several circulating microRNAs related to CVD, the contribution of other components of the vascular and immune system to the altered profile of circulating miRNAs still has to be

References 5.

1. Young SP, Kapoor SR, Viant MR, et al. The impact of inflammation on metabolomic profiles in patients with arthritis. Arthritis Rheum. 2013;65(8):2015-2023.

2. Palisi A, Grimaldi M, Sabatini P, et al. A serum nuclear magnetic resonance-based metabolomic signature of antiphospholipid syndrome. J Pharm Biomed Anal.

2017;133:90-95.

3. Perez-Sanchez C, Barbarroja N, Messineo S, et al. Gene profiling reveals specific molecular pathways in the pathogenesis of atherosclerosis and cardiovascular disease in antiphospholipid syndrome, systemic lupus erythematosus and antiphospholipid syndrome with lupus. Ann Rheum Dis. 2015;74(7):1441-1449.

4. López-Pedrera C, Pérez-Sánchez C, Ramos-Casals M, Santos-Gonzalez M, Rodriguez-Ariza A, Cuadrado MJ. Cardiovascular risk in systemic autoimmune diseases: epigenetic mechanisms of

haematologica | 2018; 103(5)

6.

7.

8.

9.

10.

defined. In addition, since we did not perform a complete plasma human microarray analysis, we cannot exclude the complementary role of other circulating miRNAs in the physiopathology of APS. Interestingly, our analysis supports a clinical role for the use of miRNA ratios when stratifying patients for their thrombotic risk. While studying the miRNA expression profile has widened the understanding of APS pathogenesis,9 its clinical utility is still a question of debate. Our data support the view that specific miRNA signatures could identify subgroups of APS patients showing different clinical profiles (in terms of site of thrombosis and risk of recurrences), potentially paving the way for their use as useful biomarkers that will increase the specificity and sensitivity of thrombotic risk assessment.

Taken together, our data suggest that differentially expressed miRNAs in the plasma of APS patients, modulated at least partially by aPL-IgG antibodies, might have the potential to serve as novel biomarkers of disease features and could help typify the atherothrombotic status of patients, thus constituting a useful tool in the management of this disease.

Acknowledgments We thank all patients for their participation in this study.

Funding This study was supported by grants from the Junta de Andalucia (CTS-7940), the Instituto de Salud Carlos III (ref. n.

PI15/01333), Cofinanciado por el Fondo Europeo de Desarrollo Regional de la Unión Europea 'Una manera de hacer Europa', Spain, and the Spanish Inflammatory and Rheumatic Diseases Network (RIER), Instituto de Salud Carlos III (RD16/0012/0015). CL-P was supported by a contract from the Spanish Junta de Andalucía. YJ-G was supported by a contract from the University of Cordoba (Co-financing of the Research Plan of the University of Cordoba and the Operating Program of the European Regional Development Funds -ERDF- for Andalusia).

immune regulatory functions. Clin Dev Immunol. 2012;2012:9746-9748.

Huang Y, Shen XJ, Zou Q, Wang SP, Tang SM, Zhang GZ. Biological functions of microRNAs: a review. J Physiol Biochem.

2011;67(1):129-139.

Xiao C, Rajewsky K. MicroRNA control in the immune system: basic principles. Cell.

2009;136(1):26-36.

Alevizos I, Illei GG. MicroRNAs as biomarkers in rheumatic diseases. Nat Rev Rheumatol. 2010;6(7):391-398.

Teruel R, Pérez-Sánchez C, Corral J, et al.

Identification of miRNAs as potential modulators of tissue factor expression in patients with systemic lupus erythematosus and antiphospholipid syndrome. J Thromb Haemost. 2011;9(10):1985-1992.

Pérez-Sánchez C, Aguirre MA, Ruiz-Limón P, et al. Atherothrombosis-associated microRNAs in Antiphospholipid syndrome and Systemic Lupus Erythematosus patients. Sci Rep. 2016;6:31375.

Creemers EE, Tijsen AJ, Pinto YM. Circulating microRNAs: novel biomarkers

11.

12.

13.

14.

15.

and extracellular communicators in cardiovascular disease? Circ Res. 2012;110(3): 483-495.

Heegaard NHH, Carlsen AL, Skovgaard K, Heegaard PMH. Circulating Extracellular microRNA in Systemic Autoimmunity.

EXS. 2015;106:171-195.

Sciascia S, Sanna G, Murru V, Roccatello D, Khamashta MA, Bertolaccini ML. GAPSS: The global anti-phospholipid syndrome score.

Rheumatology (Oxford).

2015;54(1):134-138.

Castro-Villegas C, Pérez-Sánchez C, Escudero A, et al. Circulating miRNAs as potential biomarkers of therapy effectiveness in rheumatoid arthritis patients treated with anti-TNFa. Arthritis Res Ther.

2015;17:49.

Boeri M, Verri C, Conte D, et al. MicroRNA signatures in tissues and plasma predict develoment and prognosis of computed tomography detected lung cancer. Proc Natl Acad Sci USA. 2011;108(9):3713-3718.

Hennessey PT, Sanford T, Choudhary A, et al. Serum microRNA biomarkers for detec-

917

C. Perez-Sánchez et al.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

918

tion of non-samll cell lung cancer. PLoS One. 2012;7(2):e32307.

Sheinerman KS, Tsivinsky VG, Abdullah L, Crawford F, Umansky SR. Plasma microRNA biomarkers for detection of mild cognitive impairment: biomarker validation study. Aging (Albany NY). 2013;5(12):925938.

Matthaei H, Wylie D, Lloyd MB, et al. miRNA biomarkers in cyst fluid augment the diagnosis and management of pancreatic cysts. Clin Cancer Res. 2012;18(17):47134724.

Fortunato O, Boeri M, Verri C, et al.

Assessment of circulating microRNAs in plasma of lung cancer patients. Molecules.

2014;19(3):3038-3054.

Sharova E, Grassi A, Marcer A, et al. A circulating miRNA assay as a first-line test for prostate cancer screening. Br J Cancer.

2016;114(12):1362-1366.

Mayburd AL, Martlínez A, Sackett D, et al.

Ingenuity network-assisted transcription profiling: Identification of a new pharmacologic mechanism for MK886. Clin Cancer Res. 2006;12(6):1820-1827.

Cuadrado MJ, López-Pedrera C, Khamashta MA, et al. Thrombosis in primary antiphospholipid syndrome: a pivotal role for monocyte tissue factor expression. Arthritis Rheum. 1997;40(5):834-841.

López-Pedrera C, Aguirre MA, Buendía P, et al. Differential expression of protease activated receptors in monocytes from patients with primary Antiphospholipid syndrome.

Arthritis Rheum. 2010; 62(3):869-877.

Cuadrado MJ, Buendía P, Velasco F, et al.

Vascular endothelial growth factor expression in monocytes from patients with primary antiphospholipid syndrome. J Thromb Haemost. 2006;4(11):2461-2469.

Xiao B, Wang Y, Li W, et al. Plasma microRNA signature as a noninvasive biomarker for acute graft-versus-host disease.

Blood. 2013;122(19):3365-3375.

Mitchell PS, Parkin RK, Kroh EM, et al.

Circulating microRNAs as stable bloodbased markers for cancer detection. Proc

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Natl Acad Sci USA. 2008;105(30):1051310518.

Chen X, Ba Y, Ma L, et al. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res. 2008;18(10):997-1006.

Dai Y, Huang YS, Tang M, et al. Microarray analysis of microRNA expression in peripheral blood cells of systemic lupus erythematosus patients. Lupus. 2007;16: 939-946.

Rodriguez A, Vigorito E, Clare S, et al.

Requirement of bic/microRNA-155 for normal immune function.

Science.

2007;316(5824):608-611.

Wang J, Chen J, Sen S. MicroRNA as Biomarkers and Diagnostics. J Cell Physiol.

2016;231(1):25-30.

Nakamachi Y, Kawano S, Takenokuchi M. MicroRNA-124a is a key regulator of proliferation and monocyte chemoattractant protein 1 secretion in fibroblast-like synoviocytes from patients with rheumatoid arthritis. Arthritis Rheum. 2009;60(5):1294-1304.

Pérez-Sánchez C, Ruiz-Limón P, Aguirre MA, et al. Mitochondrial dysfunction in antiphospholipid síndrome: implications in the pathogenesis of the disease and effects of coenzyme Q10 treatment. Blood. 2012; 119(24):5859-5870.

Ruiz-Limón P, Barbarroja N, Pérez-Sánchez C, et al. Atherosclerosis and cardiovascular disease in systemic lupus erythematosus: effects of in vivo statin treatment. Ann Rheum Dis. 2015;74(7):1450-1458.

Barbarroja N, Pérez-Sanchez C, Ruiz-Limon P, et al. Anticyclic citrullinated protein antibodies are implicated in the development of cardiovascular disease in rheumatoid arthritis. Arterioscler Thromb Vasc Biol.

2014;34(12):2706-2716.

Navickas R, Gal D, Laucevičius A, Taparauskaitė A, Zdanytė M, Holvoet P. Identifying circulating microRNAs as biomarkers of cardiovascular disease: a systematic review.

Cardiovasc Res. 2016;111(4):322-337.

Huang Y, Qi Y, Du JQ, Zhang DF.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

MicroRNA-34a regulates cardiac fibrosis after myocardial infarction by targeting Smad4. Expert Opin Ther Targets. 2014; 18(12):1355-1365.

Li N, Wang K, Li PF. MicroRNA-34 Family and Its Role in Cardiovascular Disease. Crit Rev Eukaryot Gene Expr. 2015;25(4):293297.

Licholai S, Blaż M, Kapelak B, Sanak M. Unbiased Profile of MicroRNA Expression in Ascending Aortic Aneurysm Tissue Appoints Molecular Pathways Contributing to the Pathology. Ann Thorac Surg.

2016;102(4):1245-1252 Choi SY, Yun J, Lee OJ, et al. MicroRNA expression profiles in placenta with severe preeclampsia using a PNA-based microarray.

Placenta. 2013;34(9):799-804.

Sheikh AM, Small HY, Currie G, Delles C. Systematic Review of Micro-RNA Expression in Pre-Eclampsia Identifies a Number of Common Pathways Associated with the Disease. PLoS One. 2016;11(8): e0160808.

Akehurst C, Small HY, Sharafetdinova L, et al. Differential expression of microRNA-206 and its target genes in preeclampsia. J Hypertens. 2015;33(10):2068-2074.

Chan ES, Cronstein BN. Methotrexate--how does it really work? Nat Rev Rheumatol.

2010;6(3):175-178.

Tsochandaridis M, Nasca L, Toga C, LevyMozziconacci A. Circulating microRNAs as clinical biomarkers in the predictions of pregnancy complications. Biomed Res Int. 2015;2015:294954.

Zhao Z, Moley KH, and Gronowski AM. Diagnostic potential for miRNAs as biomarkers for pregnancy-specific diseases.

Clinical Biochemistry. 2013;46(10-11):953960.

Wu L, Zhou H, Lin H, et al. Circulating microRNAs are elevated in plasma from severe preeclamptic pregnancies.

Reproduction. 2012;143(3):389-397.

Tatsumi K, Mackman N. Tissue Factor and Atherothrombosis. J Atheroscler Thromb.

2015;22(6):543-549

haematologica | 2018; 103(5)