Haematologica, volume 102, issue 3

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

Editor-in-Chief Jan Cools (Leuven)

Deputy Editor Luca Malcovati (Pavia)

Managing Director Antonio Majocchi (Pavia)

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

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

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

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

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

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

Information for readers, authors and subscribers Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org).

Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org).

Manuscripts should be submitted online at http://www.haematologica.org/.

Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail online (www.haematologica.org/content/policies).

Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties.

Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google Play) is free.

For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org).

Rates of the International edition for the year 2017 are as following: Print edition

Institutional Euro 500

Personal Euro 150

Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org).

Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature.

Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955.

Printing: Tipografia PI-ME, via Vigentina 136, Pavia, Italy. Printed in February 2017.

haematologica calendar of events

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

EHA Scientific Meeting on Rare Lymphomas Chairs: M Kersten and M Dreyling March 10-12, 2017 Barcelona, Spain

EHA Hematology Tutorial on Lymphoid Malignancies Chairs: R Foà, I Hus, T Robak March 17-18, 2017 Warsaw, Poland

43rd Annual Meeting of the EBMT European Society for Blood and Marrow Transplantation (EBMT) March 26-29, 2017 Marseille, France

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

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

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

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

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

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

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

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

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

Calendar of Events updated on February 1, 2017

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

Table of Contents Volume 102, Issue 3: March 2017 Cover Figure Image accompanying the review article on page 423. (Image created by www.somersault1824.com)

Editorials 411

Tailoring of medical treatment: hemostasis and thrombosis towards precision medicine Giovanni Di Minno and Elena Tremoli

418

Research in the heart of hematology: chronic myeloid leukemia 2017 R. Hehlmann

421

Research in morphology and flow cytometry is at the heart of hematology Marie C. Béné and Gina Zini

Review Article 423

Is immunotherapy here to stay in multiple myeloma?

Paula Rodríguez-Otero et al.

Guideline Article 433

ECIL-6 guidelines for the treatment of invasive candidiasis, aspergillosis and mucormycosis in leukemia and hematopoietic stem cell transplant patients Frederic Tissot et al.

Articles Hematopoiesis

445

IFNα-mediated remodeling of endothelial cells in the bone marrow niche Áine M. Prendergast et al.

Red Cell Biology & Its Disorders

454

Iron-heme-Bach1 axis is involved in erythroblast adaptation to iron deficiency Masahiro Kobayashi et al.

466

Small-molecule Factor D inhibitors selectively block the alternative pathway of complement in paroxysmal nocturnal hemoglobinuria and atypical hemolytic uremic syndrome Xuan Yuan et al.

Blood Transfusion

476

Superior survival of ex vivo cultured human reticulocytes following transfusion into mice Sabine Kupzig et al.

Platelet Biology & Its Disorders

484

Gfi1b controls integrin signaling-dependent cytoskeleton dynamics and organization in megakaryocytes Hugues Beauchemin et al.

Myelodysplatic Syndrome

498

Progression in patients with low- and intermediate-1-risk del(5q) myelodysplastic syndromes is predicted by a limited subset of mutations Christian Scharenberg et al.

Haematologica 2017; vol. 102 no. 3 - March 2017 http://www.haematologica.org/

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

509

Aberrant let7a/HMGA2 signaling activity with unique clinical phenotype in JAK2-mutated myeloproliferative neoplasms Chih-Cheng Chen et al.

Chronic Myeloid Leukemia

519

Phase 1 dose-finding study of rebastinib (DCC-2036) in patients with relapsed chronic myeloid leukemia and acute myeloid leukemia Jorge Cortes et al.

Acute Myeloid Leukemia

529

CEBPA–double-mutated acute myeloid leukemia displays a unique phenotypic profile: a reliable screening method and insight into biological features Francesco Mannelli et al.

Acute Lymphoblastic Leukemia

541

Tumor suppressors BTG1 and IKZF1 cooperate during mouse leukemia development and increase relapse risk in B-cell precursor acute lymphoblastic leukemia patients Blanca Scheijen et al.

552

Population pharmacokinetics of intravenous Erwinia asparaginase in pediatric acute lymphoblastic leukemia patients Sebastiaan D.T. Sassen et al.

Chronic Lymphocytic Leukemia

562

T cells in chronic lymphocytic leukemia display dysregulated expression of immune checkpoints and activation markers Marzia Palma et al.

Non-Hodgkin Lymphoma

573

The small FOXP1 isoform predominantly expressed in activated B cell-like diffuse large B-cell lymphoma and full-length FOXP1 exert similar oncogenic and transcriptional activity in human B cells Martine van Keimpema et al.

584

Factors related to the relative survival of patients with diffuse large B-cell lymphoma in a population-based study in France: does socio-economic status have a role?

Sandra Le Guyader-Peyrou et al.

Plasma Cell Disorders

593

Evaluation of the Revised International Staging System in an independent cohort of unselected patients with multiple myeloma Efstathios Kastritis et al.

Cell Therapy & Immunotherapy

600

Effect of high-dose plerixafor on CD34+ cell mobilization in healthy stem cell donors: results of a randomized crossover trial Jeremy Pantin et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/102/3.toc

e73

Reduced-intensity conditioning and stem cell transplantation in infants with Diamond Blackfan anemia Roman Crazzolara et al. http://www.haematologica.org/content/102/3/e73

e76

Different clinical characteristics of paroxysmal nocturnal hemoglobinuria in pediatric and adult patients Álvaro Urbano-Ispizua et al. http://www.haematologica.org/content/102/3/e76

e80

Selective silencing of α-globin by the histone demethylase inhibitor IOX1: a potentially new pathway for treatment of b-thalassemia Sachith Mettananda et al. http://www.haematologica.org/content/102/3/e80

Haematologica 2017; vol. 102 no. 3 - March 2017 http://www.haematologica.org/

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

e85

Increased serum hepcidin contributes to the anemia of chronic kidney disease in a murine model Mark R. Hanudel et al. http://www.haematologica.org/content/102/3/e85

e89

Effects of ibrutinib treatment on murine platelet function during inflammation and in primary hemostasis Robert H. Lee et al. http://www.haematologica.org/content/102/3/e89

e93

Exome sequencing identified RPS15A as a novel causative gene for Diamond-Blackfan anemia Fumika Ikeda et al. http://www.haematologica.org/content/102/3/e93

e97

Evaluation of bone marrow morphology is essential for assessing disease status in recombinant interferon Îą-treated polycythemia vera patients Elizabeth Margolskee et al. http://www.haematologica.org/content/102/3/e97

e100

Extracellular vesicles released from chronic lymphocytic leukemia cells exhibit a disease relevant mRNA signature and transfer mRNA to bystander cells Katrin S. Reiners et al. http://www.haematologica.org/content/102/3/e100

e104

Role of serum free light chain assay in the detection of early relapse and prediction of prognosis after relapse in multiple myeloma patients treated upfront with novel agents Paola Tacchetti et al. http://www.haematologica.org/content/102/3/e104

e108

Invasive fungal infections in chronic lymphoproliferative disorders: a monocentric retrospective study Maria Chiara Tisi et al. http://www.haematologica.org/content/102/3/e108

Case Report This case report is available online only at www.haematologica.org/content/102/3.toc

e112

Deep and sustained response after venetoclax therapy in a patient with very advanced refractory myeloma with translocation t(11;14) Cyrille Touzeau et al. http://www.haematologica.org/content/102/3/e112

Comments These comments are available online only at www.haematologica.org/content/102/3.toc

e115

Hematopoietic stem cell apheresis in the context of a related allogeneic transplant for acute myeloid leukemia: an unexpected outcome, medical emergency and ethical issue Marc Bourgeois et al. http://www.haematologica.org/content/102/3/e115

e117

Historical controls?

Paul S. Gaynon http://www.haematologica.org/content/102/3/e117

e118

Inclusion and response criteria for clinical trials in relapsed/refractory acute lymphoblastic leukemia and usefulness of historical control trials Nicola GĂśkbuget et al. http://www.haematologica.org/content/102/3/e118

Haematologica 2017; vol. 102 no. 3 - March 2017 http://www.haematologica.org/

EDITORIALS Tailoring of medical treatment: hemostasis and thrombosis towards precision medicine Giovanni Di Minno1 and Elena Tremoli2 1

Clinica Medica, Dipartimento di Medicina Clinica e Chirurgia, Università degli Studi di Napoli “Federico II”, Naples and 2Centro Cardiologico Monzino, IRCCS, Milano, Italy.

E-mail: diminno@unina.it

doi:10.3324/haematol.2016.156000

B

y integrating genetic, biomarker, phenotypic, and psychosocial characteristics that distinguish one patient from others with similar clinical presentations, the aim of precision medicine is to target treatments to individual needs. Presently, simplified individual pharmacokinetic analyses spare hemophilia patients from unnecessary exposure to replacement treatments and ultimately reduce costs. Likewise, successful clopidogrel use in vascular medicine is based on the integration of genomics, lifestyle and environmental data.

Beyond the development of medical devices that are unique to a patient, or treatments tailored to specific molecular and nonmolecular targets, through the use of big biobanks and electronic medical records that integrate biological information with clinical data, it is likely that algorithms will be developed to classify individuals into subpopulations differing by their susceptibility to bleeding and/or thrombosis, by the severity of their diseases, and/or by their response to specific treatments. An inherent risk of such strategies is differential access to treatments for individual patients, families, and communities, since costs for therapies depend on the size of the target population. Fostering standardization of care in the era of precision medicine implies a public health perspective and a supportive institutional environment to harmonize the interests shared by healthcare providers, patients and communities, to acknowledge the individual roles and responsibilities in decision-making, and to balance the generation of long-term knowledge and short-term health gains. Integrating efficacy, safety, and cost-effectiveness is a challenge for precision medicine and the opportunity for it to generate early measurable health benefits and to live up to its promise.

A clinical case A 19-year old patient was referred to our Hemophilia and Thrombosis Center because of an ischemic stroke (confirmed by magnetic resonance imaging) that occurred after 3 months of oral contraceptive use. The girl was the daughter of a patient already attending the Center for type I von Willebrand disease, but her personal and family history had been uneventful. The reason for the referral was to decide whether she should be given long-term treatment with a low dose of aspirin. The laboratory work-up revealed that, in addition to type I von Willebrand disease, she was homozygous for the prothrombin G20210A mutation, and the same thrombophilic mutation was found in other members of the family. Heterozygous factor V Leiden or the G20210A prothrombin mutation may compensate for low factor VIII or IX levels in hemophilia, resulting in more efficient thrombin generation and ensuing attenuation of clinical symptoms1 and the risk of thrombotic complications.2 This information was interpreted to account for the poor bleeding tendency of the patient. She was informed that: (i) despite recommendations concerning drugs to avoid in patients with von Willebrand disease, chronic daily treatment with low-dose aspirin (100 mg/day) was conceivably helpful in her case, and (ii) prophylaxis with low-molecular weight haematologica | 2017; 102(3)

heparin/warfarin would be possible in specific, at-risk situations. Over the last 10 years in which she took low-dose aspirin daily, she had no inappropriate bleeding events, no stroke recurrence, and had two successful pregnancies.

When guidelines cannot be relied on To give advice on an appropriate treatment, there must be a high level of evidence available, based on multiple randomized controlled clinical trials, which the guidelines can draw on to justify their recommendations. Thus, the strength of guidelines is when they are applied to areas in which large trials have provided convincing evidence of the benefit of certain interventions.3 However, in spite of the fact that each patient is treated with the treatment that everyone else with that condition receives, certain medical interventions are more effective or cause fewer side effects in some patients than in others.4 This is usually accounted for by inherent limitations of guidelines. The problem-solving approach of clinical trials employs methods (inclusion/exclusion criteria, randomization, etc.) theoretically free from bias, and is finalized at answering a single question at a time. In real-world practice, however, there is a context rather than a single question: patients simultaneously raise multiple clinical problems; there are no inclusion or exclusion criteria, and it is uncommon that the individual patient in front of us fits into the inclusion criteria for the trials used to formulate the guidelines while not having any of the exclusion criteria.5 Thus, in everyday practice, the evidence from guidelines applies to a middle segment of a patient population, but not to the two extremes [patients not entirely meeting the eligibility and/or exclusion criteria of the trial(s) on which the guidelines are based]. If the prevalence of a disorder is low, large prospective studies, and, in turn, evidence-based recommendations for clinical management, remain improbable. In such settings, registries are the only manner to collect enough data on optimal therapeutic approaches, and to attempt risk– and cost– benefit evaluations for different treatment options. However registries have a variety of limitations, the first and foremost being the lack of randomization. All in all, there are areas in medical practice in which there is uncertainty concerning the real state of a patient,6 and for which definitive conclusions cannot be drawn. In addition to uncertainty concerning the real state of a patient,6 a lack of compliance of healthcare professionals with guidelines may also be related to doctors’ preferences with respect to the outcome of the decision.7 “Less algorithmic” individualized guidelines are now available to help doctors define the best strategy for a given patient.8 Since an algorithm is a problem-solving system in which there is a single answer to a given question, the advent of individualized guidelines acknowledges (but does little to attenuate) the uncertainty associated with any medical decision. Being made in a context of uncertainty and risk, since statistics cannot take into account the individual context of each medical decision, medical decisions cannot be free from 411

Editorials

Figure 1. Implementing clinical care: precision medicine. The purpose of a comprehensive system of precision medicine is to establish the information base and infrastructure to provide more precise individual information, to make (new) clinical treatment more efficient. This system is aimed at integrating huge amounts of data with the goal of improving health, and implies targeting treatments to the needs of individual patients on the basis of genetic, phenotypic and psychosocial characteristics that distinguish a given patient from other patients with similar clinical presentations.11

biases.5 Especially (but not exclusively) in areas of uncertainty, the safety and quality of care we provide often relies on additional information that may be relevant for the patient in front of us and that the individual physician has gathered. For example, in a program for patients undergoing percutaneous coronary interventions, which was implemented in nine large US centers, informing physicians of a patient’s bleeding risk led to a reduction in the occurrence of bleeding (from 1.7% to 1%, -44%).9 Regardless of whether the reduction was truly brought about by the precision of the prediction, or by raising the general awareness of the risk, implementation of personalized bleeding risks helped doctors to identify subjects truly at risk of bleeding and to use techniques to avoid hemorrhage appropriately.

The promise of precision medicine How precision medicine will enter clinical care and affect the vision of medicine is now being delineated, and the manner in which it will allow for more efficient clinical research and facilitate scientific discoveries is being clarified.10 Precision medicine (Figure 1) implies targeting treatments to the needs of individual patients on the basis of genetic, phenotypic and psychosocial characteristics that distinguish a given patient from other patients with similar clinical presentations.11 Inherent to this definition is the concept of integrating (in electronic health records) individual-level information (e.g. genomics, biomarkers, physiological, lifestyle and other environmental factors) with the ultimate aim of providing better clinical care for each patient.12 In addition to a dramatic improvement and price reduction in genome sequencing,13 the prospect of applying precision medicine broadly is supported by largescale biological databases, powerful methods for characterizing patients (e.g. proteomics, metabolomics, genomics, diverse cellular assays), mobile healthcare technology, and computational tools for analyzing large sets of data.14 Although imaging techniques15 and individual differences in terms of the unique circumstances of the person (personality, resources, culture, individual behavior) 412

and of his/her environment (family, friends, communities, religion) are keys to improve healthcare,16 genomics is the leading driver of an early identification strategy, once individual profiles (e.g. polymorphisms) are available.17 Following the original observation of a common association between ankylosing spondylitis (or insulin-dependent diabetes) and the alleles of genes of the HLA system,18 the concept of employing genetics to make major contributions to clinical practice has been progressively extended to the entire human genome. In particular, since cancer is a disease of the genome, oncology has been the obvious target of such a strategy.19 For example, targeting HER2 overexpression with the monoclonal antibody, trastuzumab, improved outcome in metastatic breast cancer;20 the tyrosine kinase inhibitor, imatinib, transformed the care of patients with chronic myeloid leukemia to a manageable chronic disease,21 and the identification of somatic mutations in the BRAF gene in the majority of malignant melanomas22 enabled the development of vemurafenib which specifically targets the underlying molecular lesion.23 Large�scale cancer whole genome sequencing projects are now expected to provide a complete catalogue of genomic alterations in primary cancers, to elucidate the mutational patterns and influences across the natural history of cancers, and to provide targeted therapies and newer approaches to cancer prevention.24,25 In keeping with the genetically-based improved care of patients with cancer, the concept that treatments should be tailored to the individual patient - taking into account relevant personal data – is spreading into all areas of medical practice. Recombinant clotting factor concentrates have revolutionized the care of hemophilia in the western world. However, their relatively short half-lives necessitate frequent intravenous administration of concentrates (at least 2-3 times a week) associated with peaks and troughs of circulating factor levels and occasional breakthrough bleeding when levels drop below 1%. In addition to being invasive, prophylaxis is demanding and not curative, and the cost is logarithms higher in the 25%-35% of patients with severe hemophilia A who develop neutralizhaematologica | 2017; 102(3)

Editorials

Figure 2. The promise of precision medicine. “Medicine was, in its history, first of all curative, then preventive and finally predictive”.18 This change in medical attitudes was due to the advances that have entered medical practice: new imaging techniques and powerful strategies in biological investigation have dramatically improved our ability to monitor early stages of disease development. In addition, increasingly effective treatments (organ transplantation; smart drugs; targeted strategies, vaccinations) have progressively reduced the rates of failures and side effects and in turn improved the cure of (chronic) diseases.

ing inhibitors.26 Simplified pharmacokinetic studies have shown significant differences in the individual half-life of clotting factors.27 This information has major implications for the management of prophylaxis in hemophilia.

Appropriate treatments will spare patients from unnecessary exposure to replacement treatments and ultimately reduce costs. This may be especially important for novel clotting formulations with extended half-lives.28 Although antiplatelet treatment significantly reduced stroke and coronary events in secondary prevention trials, 10%-20% of patients have recurrent events during longterm follow-up.29,30 Residual platelet reactivity (i.e. an incomplete response to the antiplatelet treatment) in subjects on treatment with clopidogrel or aspirin predicts recurrent events.31,32 Similar to haplotypes of cyclooxygenase-1 that modulate platelet response to aspirin,33 polymorphisms in the CYP2C19 gene (most often CYP2C19*2) associated with a 20%-25% production of inactive metabolite, diminish the response to clopidogrel.34 Compared to wild-type subjects, carriers of polymorphic alleles of the ABCB1 gene, which modulates clopidogrel absorption, had a higher rate of cardiovascular events at 1year follow-up.35 Among subjects under treatment with clopidogrel, carriers of these polymorphisms had a 50% higher risk of cardiovascular death, acute myocardial infarction, and stroke.35 In addition to genetically determined cases, an incomplete response to clopidogrel may be the result of poor compliance by the patient, clinical conditions leading to an abnormally high platelet turnhaematologica | 2017; 102(3)

over (e.g. high pretreatment platelet reactivity, high glucose levels, inflammation, hypercoagulable states, low fibrinolytic potential), or the simultaneous administration of interfering drugs.36 Appropriate information should be collected prior to the day of a potential vascular intervention so that, at the time of prescribing, relevant data are available to the practitioner. Thus, the integration of genomics, lifestyle and environmental data is key to successful clopidogrel treatment. Prospective randomized trials did not demonstrate a clinical benefit of using platelet function testing to adjust antiplatelet treatment.37-39 However, low event rates in current antiplatelet practice would require very large numbers of enrolled patients to provide reliable conclusions. Whether, however, electronic health records should be pre-populated with other data, in addition to genetic and pharmacogenomics data, providing clinicians with new critical information about the risk of an “incomplete response” to clopidogrel is so far unknown.40,41

Toward precision medicine: challenges and opportunities It is conceivable that precision medicine-based interventions will improve clinical outcomes for individual patients and minimize the risk of failure or adverse health outcomes in those less likely to have a response to a particular treatment. (Figure 2) It is also obvious that our ability to handle vast amounts of new knowledge and treatment options within the framework of everyday practice is critical to define the impact of an initiative that is 413

Editorials

expected to transform morbidity and mortality patterns.

Hurdles to overcome and directions to be followed in the early phases of precision medicine in order for it to gather strength and to live up to its promise have been identified (Table 1). In particular, it is imperative that the investment in precision medicine is oriented to a public health perspective to help ensure accessibility and generalizability, to assess methods of implementation, and to provide an appropriate balance between generation of long-term knowledge and short-term health gains.42 Nevertheless, how precision medicine, by identifying the needs and improving the outcomes of an individual patient, might be a means of providing the best available health care at a population level is matter of debate.11,43 Points for debate, stemming from real-life practice in hemostasis and thrombosis, are summarized in the following paragraphs.

• Costs should be affordable. Considerable planning of daily activities that would be taken for granted by most people living without hemophilia is still required, particularly in children and adolescents. However, compared to the on-demand strategy, the prophylactic use of clotting products to maintain circulating clotting factor levels ≥1% of normal has resulted in a dramatic reduction in bleeding frequency and associated complications e.g. hemophilic arthropathy. By paying a cost for factor concentrates (≈€ 150,000/adult/year) that is higher than that of on-demand treatments, the life expectancy of European and America patients with severe hemophilia on prophylactic treatment has normalized to that of age-matched healthy males in their community.44 In terms of calculation of human capital, the healthier a population, the more productive it is.45 The quality of life of patients with severe

Table 1. Providing the appropriate health care infrastructure to deliver precision medicine in routine clinical settings: health plans and issues for debate.

Issues

Outline(s)

General

By means of preemptive pharmacogenomic data and clinical decision support integrated into an electronic medical record, prescribers can deliver genome-guided therapy at the point of care.

Coping costs with changes Avoiding ethnicity/ race-driven discrimination

Suggested readings (14),(17), (40), (68) * (69)

The investment in precision medicine should not worsen existing health disparities and should generate early measurable health benefits.

Using genomic, clinical, personal and environmental data collected from very large numbers of individuals (70),(71) from various populations, and connecting their health records, “non-responders“ to a treatment might be ** identified as largely belonging to definite minority, racial, ethnic groups or underserved populations.

Appropriate protections needed against discrimination in the access to treatments.

The Detection of susceptibility genes in the absence of the simultaneous development of a preventive or (71),(72) "unpatients' issue” therapeutic ad hoc strategy will increase physician visits, laboratory tests, and patient anxiety. The poor *** information about the pathogenicity of most genetic variants is a barrier to the translation of experimental findings to clinical care. A systematic approach to determining genetic causality is mandatory.

The impact Similar to “omics”, the unique attributes of people have a major impact on an individual’s susceptibility (16),(42) of “Personomics” to disease: how that disease will reveal itself phenotypically, and how the individual with the disease will respond to treatment.

Re-classifying (i)Identify the true penetrance of certain inherited conditions as ascertained via a population-based approach; (74) diseases (ii)Develop new diagnostic tests to allow for newer prognostic implications; (iii)Handle and interpret massive **** amounts of genomic/non-genomic data, far beyond the expertise of medical professionals not trained to deal with complex data sets.

Precision prevention Family history was the most important genetic risk factor in the Framingham Heart Study, and also accounts (63),(75),(76) for gene-environment interactions. Health data collection in families is an inexpensive tool for identifying ***** individuals/families that require earlier and more intensive screening for major diseases.

The availability of molecular profiling tests, e.g. individual germline DNA sequencing, calls for educating clinicians to the knowledge bases needed to assist them in taking actions based on genetic test results.

Monitoring Creating ad hoc regulatory agencies; (77) the implementation Safeguard against the marketing and distribution of fraudulent products.

Identifying new Handling the information gathered: descriptive statistical associations will not necessarily advance (10),(11),(42), areas of research the information of how molecules interact mechanistically to produce diseases or lead to rational therapies. (78),(79),(80) N-of-1 trials: the inclusion of N-of-1 trial data into randomized controlled trial meta-analyses improves the precision of yielded treatment effects. Improving the possibility that N-of-1 trial data allow for individual information to be shared with individuals who do not share a specific genome.

Targeted therapies: (i) identifying patients who will best respond to already proven interventions; (ii) searching for reliable biomarkers to target the patients likely to present the best benefit-risk balance for a given active compound. (iii) dose adjustment, methods to optimize the benefit-risk ratio of the drugs chosen; biomarkers of efficacy; toxicity, and treatment withdrawal.

Revised clinical guidelines: when clinical trials will assess the efficacy, safety, and cost-effectiveness of targeted therapies, new guidelines are mandatory. *Enabling clinicians and patients to acquire and process molecular testing, to interpret results, and identify specific pathways affected. **Example: low levels of 25-hydroxyvitamin D as an independent risk factor for coronary artery disease or fatal stroke in white people, but not blacks. ***Asymptomatic subjects carrying a mutation: emotional and ethical issues. ****Taxonomy of disease based on molecular and clinical parameters. *****Targeting preventive strategies to the specific subsets of a population that will derive maximal benefit.

414

haematologica | 2017; 102(3)

Editorials

hemophilia on prophylactic treatment has significantly improved, and these patients are increasingly involved in working activities. However, the cost for prophylaxis is not affordable for the large majority of countries. Thus, most hemophilia patients worldwide receive no treatment or only sporadic on-demand therapy. Such patients are condemned to shortened lives of pain and disability.

• Costs for therapies depend on the size of the target population: the smaller the population, the more expensive the drug. At least initially, gene therapy in hemophilia is likely to command a high price to recoup research and development costs. Such economic considerations may have major implications for differential access to treatment for hemophilia families, communities and society. One of the claims is expected to be that successful gene therapy offers the advantage of continuous endogenous expression of clotting factor. While improving quality of life, this would eliminate breakthrough bleeding and micro-hemorrhages and comorbidities, thereby reducing the cost of care for the healthcare system, and the need for frequent medical interventions.

Point-of-care ultrasound detection of affected joints – reliably correlating with magnetic resonance imaging detection of cartilage damage, effusion, and synovial hypertrophy carried out inside Hemophilia Centers at each patient’s visit, could have major implications for the management of hemophilia patients.46,47 Repeated evaluations of patients’ joints will provide hints as to whether this strategy is worth the cost required.

• The results of testing should be actionable, thus informing prognosis and/or supporting rational prescribing.48 In patients with severe hemophilia B, a single intravenous administration of an adeno-associated virus vector (AAV8) encoding an optimized F9 gene resulted in long-term (>4 years), dose-dependent increases in circulating factor IX to levels between 1% to 6% of the normal value without persistent or late toxicity.49 By providing stable, long-term therapeutic levels of coagulation factor IX, gene therapy has the potential to change the treatment paradigm for hemophilia.50 With the availability of ad hoc genomic data, actionable tests should help to iden-

tify: (i) which patients will have loss or reduction of transgene expression and/or persistence of high titers of antiAAV8 IgG - with subsequent successful gene transfer with vector of the same serotype (in the event that transgene expression falls below therapeutic levels); (ii) spread of vector particles to non-hepatic tissues, including the gonads; (iii) insertional mutagenesis (deep sequencing studies have shown that integration of the AAV genome can occur in the liver).51 • Stroke, the fourth leading cause of mortality and the leading cause of neurological disability in western countries, involves an intricate interplay of environmental and genetic factors. Genes not only influence susceptibility to stroke but also affect the response to pharmacological agents and in turn the outcome of the disease. A significant number of patients experience drug-induced adverse reactions; a poor clinical outcome, and recurrent stroke events. Several single nucleotide polymorphisms in genes encoding for metabolizers, transporters and target receptors influence the pharmacokinetics and pharmacodynamics of drugs used in the treatment of stroke.52 Clinical trials have related candidate gene variants with abnormal drug response in stroke treatment.53 However, these results need to be replicated in genome-wide association studies.

A broad research program prospectively testing targeted therapeutic strategies based on pharmacogenetics, and ultimately encouraging approaches to build the evidence base needed to guide clinical practice, is likely to be costsaving in this setting.54 • Defining effective strategies to improve communication and outcomes is mandatory in precision medicinebased approaches. Because of the limited time spent on direct care of patients, these days physicians and medical students know very little about their patients as people.16 Trainees spend less time interviewing and examining their patients, and duty-hour regulations progressively erode the time residents devote to history taking and physical examination skills.55 On the other hand, residents spend a good amount of time at the computer,56 getting to know an electronic facsimile of a patient – the “iPatient”57 – well

Figure 3. Perspectives in precision medicine. Tests of increased susceptibility to prevent the development of diseases are steadily increasing in different areas of clinical medicine. In a precision prevention-based health system, curative medicine is expected to be needed only in a very limited number of cases "only in desperation".18

haematologica | 2017; 102(3)

415

Editorials

before they have met the actual person in a hospital bed or outpatient clinic. Genomics of diseased tissues document mutations that confer sensitivity to drugs not approved for that specific indication.58 Under such newer scenarios, empathy and humanity are required to let doctors enter into a process aimed at clarifying the context of the treatment. Negotiating an effective and newer equitable partnership with practitioners was the obvious direction to be followed after hepatitis and human immunodeficiency virus transmitted by blood products called for a newer doctor-patient relationships in the bleeding disorder community.59,60 • Forging an effective and newer patient-physician alliance61 in the era of precision medicine should be aimed at precision medicine-oriented information. Major directions to be pursued are how to handle patients’ expectations, educate them regarding new medical concepts, and deliver results to individuals. In this respect, if information available on the Internet is increasingly generating more aggressive patients, a surreptitious and coercive interpretation of guidelines will further and definitely mark the advent of the age of medical defensiveness, and healthcare systems will continue to contribute poorly to the wellbeing and life expectancy of the least-advantaged people and the potential for a renewed patient-physician alliance will fade to become merely virtual. If, instead, through shared reasoning, strong public health-healthcare partnerships ensure that all people have access to the intended benefits of technology and track efficacy, safety, and effectiveness outcomes in the real world; if dying is accepted as part of the natural history of each one of us; if patients consider their doctor as a travel companion whose life and commitment have been a constant, prolonged challenge to the power of death, disease, solitude, and pain, then there is room for hope.

Attempting to remedy a not optimal genome: perspectives.

Faced with the spread of information from guidelines, the concept that, when appropriately formulated, every clinical question can be solved by an approach that employs the theory of probability has gathered strength and relevance.5 Thus, medical students and patients now envisage medical practice as relatively simplified problemsolving. However, although the “one-size fits all” approach is broadly used in prevention and management across the vast majority of clinical settings, the need for the right drug at the right dose in the right patient is being increasingly recognized in real-life practice,19 given that there are areas in which guidelines cannot be relied on for how to deal with a specific issue.62 In these areas, which are often very important for the clinician, it remains imperative to identify in advance which patients are less likely to benefit from a given intervention. Although it is still too often a theoretical concept - because of the lack of convenient diagnostic methods or treatments, and of drugs corresponding to each subtype of pathology - precision medicine argues for improved risk predictions, behavioral changes; lower costs, and gains in public health, and the community should be aware and engaged in its progress. The promise of precision medicine involves every aspect of medical care, and calls for active collabora416

tion between researchers, doctors, patients and other stakeholders. It demands newer levels of medical education and up-to-date diagnostics, informatics and algorithms to assist healthcare providers with information management and decision making. Integrating efficacy, safety, and cost-effectiveness is a challenge for precision medicine and the opportunity for it to generate early measurable health benefits and to live up to its promise.

Indeed, the investment in precision medicine-based treatment decisions19 should be harmonized with information emerging from evidence-based medicine to help the standardization of care. Within the framework of open discussions and a supportive institutional environment - to acknowledge and fulfil the individual roles and responsibilities in decision-making - interests shared by individual patients, families, and communities, the diagnostics and pharmaceutical industries, and healthcare providers (governments, industry, payers, and other stakeholders) should be aligned.63 Major advances in medical practice lend credence to the possibility that we are heading towards hospitals that ban relatives and visitors (mobile telephones and webcams will keep patients in touch with their relatives and friends), where hospital files are e-files, where remote consultations will be the rule, and where nursing will be based on the least possible contact – a virtually total “asepsis”. While some might suppose that doctors will no longer have to stumble to find the right words for the unavoidable, it is highly likely that in the near future medical decisions will have to be taken by integrating patients’ “omics” with characteristics and preferences and other individual-level data. The number of genes that presently confer susceptibility to (or protection from) diseases is steadily increasing. The availability of molecular profiling tests calls for new school curricula to educate clinicians of the 21st century to the knowledge needed to assist them in taking actions based on genetic test results.

Precision prevention (Figure 3) implies that doctors will become advisers for individuals who are not ill. With appropriate protections (see the "unpatients’ issue” in the Table 1), each individual will have full knowledge of his/her health capital and will be able to manage it as his/her banking account. Longevity will, it is to be hoped, continue to increase each year, not only in industrialized countries. Whether, in addition to definitely changing the ethical dimensions of the relation among patients, their doctors and other healthcare providers, precision prevention-based health systems will also lower costs of future medical practice is so far unknown.

In addition to the search for drugs and medical devices that are unique to a patient or to a limited group of patients (e.g. the use of mobile technology to serve as an effective reminder to monitor medication adherence such as the international normalized ratio), the future in hemostasis and thrombosis is expected to help physicians to classify patients into sub-populations that differ by their susceptibility to bleeding and/or thrombosis, by the biology/prognosis of those diseases they may develop, and/or by their response to a treatment. According to one metaanalysis, genotype-guided treatment schemes have so far been less informative than clinical dosing for warfarin and its analogues.64 However, a prolonged period of observation would be needed to address this issue properly. haematologica | 2017; 102(3)

Editorials

Follow-ups ranged from 4 weeks to 6 months (median, 12 weeks) for both the genotype-guided and clinical-dosing algorithm studies evaluated in that meta-analysis.64 These are the early days of precision medicine. It has long been known that much of cancer biology is based on the central tenet that it is a genetic disease.65 However, the concept that our treatments should be more precisely tailored to specific molecular targets contributed little to cancer treatment until the 21st century. Major achievements sometimes take more time than anticipated by the original hype.66 To obtain sufficient funding, researchers need to create hype, and this may lead to unrealistic expectations on the part of patients and clinicians.67

References 1. Kurnik K, Kreuz W, Horneff S, et al. Effects of the factor V G1691A mutation and the factor II G20210A variant on the clinical expression of severe hemophilia A in children--results of a multicenter study. Haematologica.

2007;92(7):982-985.

2. Franchini M, Lippi G. Factor V Leiden and hemophilia. Thromb Res. 2010;125(2):119-123.

3. Shaw JA, Cooper ME. Guidelines and their use in clinical practice. Diab Vasc Dis Res. 2014;11(1):3-4.

4. Illing PT, Mifsud NA, Purcell AW. Allotype specific interactions of drugs and HLA molecules in hypersensitivity reactions. Curr Opin Immunol.

2016;42:31-40.

5. Reach G. Clinical inertia, uncertainty and individualized guidelines.

Diabetes Metab. 2014;40(4):241-245.

6. Phillips LS, Branch WT, Cook CB, et al. Clinical inertia. Ann Intern Med. 2001;135(9):825-834.

7. Kerr EA, Zikmund-Fisher BJ, Klamerus ML, Subramanian U, Hogan MM, Hofer TP. The role of clinical uncertainty in treatment decisions for diabetic patients with uncontrolled blood pressure. Ann Intern Med. 2008;148(10):717-727.

8. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycaemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia.

2012;55(6):1577-1596.

9. Spertus JA, Decker C, Gialde E, et al. Precision medicine to improve use of bleeding avoidance strategies and reduce bleeding in patients undergoing percutaneous coronary intervention: prospective cohort study before and after implementation of personalized bleeding risks. BMJ. 2015;350:h1302.

10. Antman EM, Loscalzo J. Precision medicine in cardiology. Nat Rev Cardiol. 2016;13(10):591-602.

11. Jameson JL, Longo DL. Precision medicine--personalized, problematic, and promising. N Engl J Med. 2015;372(23):2229-2234.

12. Shah SH, Arnett D, Houser SR, et al. Opportunities for the cardiovascular community in the precision medicine initiative. circulation.

2016;133(2):226-231.

13. Hayden EC. Technology: the $1,000 genome. Nature. 2014;507 (7492):294-295.

14. Baker M. Big biology: the 'omes puzzle. Nature. 2013;494(7438):416419.

15. Evens AM, Kostakoglu L. The role of FDG-PET in defining prognosis of Hodgkin lymphoma for early-stage disease. Blood. 2014;124(23):33563364.

16. Ziegelstein RC. Personomics. JAMA Intern Med. 2015;175(6):888-889.

17. Ashley EA. The precision medicine initiative: a new national effort.

JAMA. 2015;313(21):2119-2120.

18. Dausset J. Journal of Biomedicine and Biotechnology. J Biomed Biotechnol. 2001;1(1):1-2.

19. Collins FS, Varmus H. A new initiative on precision medicine. N Engl J Med. 2015;372(9):793-795.

20. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med. 2001;344(11):783-792.

21. Hughes TP, Kaeda J, Branford S, et al. Frequency of major molecular responses to imatinib or interferon alfa plus cytarabine in newly diagnosed chronic myeloid leukemia. N Engl J Med. 2003;349(15):1423-1432.

22. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in human cancer. Nature. 2002;417(6892):949-954.

23. Flaherty KT, Puzanov I, Kim KB, et al. Inhibition of mutated, activated

haematologica | 2017; 102(3)

BRAF in metastatic melanoma. N Engl J Med. 2010;363(9):809-819.

24. Garraway LA, Lander ES. Lessons from the cancer genome. Cell.

2013;153(1):17-37.

25. Stratton MR. Journeys into the genome of cancer cells. EMBO Mol Med. 2013;5(2):169-172.

26. Di Minno MN, Di Minno G, Di Capua M, Cerbone AM, Coppola A. Cost of care of haemophilia with inhibitors. Haemophilia.

2010;16(1):e190-201.

27. Valentino LA, Mamonov V, Hellmann A, et al. A randomized comparison of two prophylaxis regimens and a paired comparison of on-demand and prophylaxis treatments in hemophilia A management. J Thromb Haemost. 2012;10(3):359-367.

28. Pipe SW. The hope and reality of long-acting hemophilia products. Am J Hematol. 2012;87 (Suppl 1):S33-39.

29. Angiolillo DJ. Variability in responsiveness to oral antiplatelet therapy.

Am J Cardiol. 2009;103(3 Suppl):27A-34A. 30. Heras M, Bueno H, Bardaji A, Fernandez-Ortiz A, Marti H, Marrugat J. Magnitude and consequences of undertreatment of high-risk patients with non-ST segment elevation acute coronary syndromes: insights from the DESCARTES Registry. Heart. 2006;92(11):1571-1576.

31. Krasopoulos G, Brister SJ, Beattie WS, Buchanan MR. Aspirin "resistance" and risk of cardiovascular morbidity: systematic review and metaanalysis. BMJ. 2008;336(7637):195-198.

32. Sofi F, Marcucci R, Gori AM, Giusti B, Abbate R, Gensini GF. Clopidogrel non-responsiveness and risk of cardiovascular morbidity. An updated meta-analysis. Thromb Haemost. 2010;103(4):841-848.

33. Maree AO, Curtin RJ, Chubb A, et al. Cyclooxygenase-1 haplotype modulates platelet response to aspirin. J Thromb Haemost.

2005;3(10):2340-2345.

34. Wang L, McLeod HL, Weinshilboum RM. Genomics and drug response.

N Engl J Med. 2011;364(12):1144-1153.

35. Simon T, Verstuyft C, Mary-Krause M, et al. Genetic determinants of response to clopidogrel and cardiovascular events. N Engl J Med. 2009;360(4):363-375.

36. Di Minno MN, Guida A, Camera M, Colli S, Di Minno G, Tremoli E. Overcoming limitations of current antiplatelet drugs: a concerted effort for more profitable strategies of intervention. Ann Med. 2011;43(7):531544.

37. Price MJ, Berger PB, Teirstein PS, et al. Standard- vs high-dose clopidogrel based on platelet function testing after percutaneous coronary intervention: the GRAVITAS randomized trial. JAMA. 2011;305(11):1097-1105.

38. Trenk D, Stone GW, Gawaz M, et al. A randomized trial of prasugrel versus clopidogrel in patients with high platelet reactivity on clopidogrel after elective percutaneous coronary intervention with implantation of drug-eluting stents: results of the TRIGGER-PCI (Testing Platelet Reactivity In Patients Undergoing Elective Stent Placement on Clopidogrel to Guide Alternative Therapy With Prasugrel) study. J Am Coll Cardiol. 2012;59(24):2159-2164.

39. Collet JP, Cuisset T, Range G, et al. Bedside monitoring to adjust antiplatelet therapy for coronary stenting. N Engl J Med. 2012;367 (22):2100-2109.

40. Altman RB. PharmGKB: a logical home for knowledge relating genotype to drug response phenotype. Nat Genet. 2007;39(4):426.

41. Hiatt WR, Fowkes FG, Heizer G, et al; EUCLID Trial Steering Committee and Investigators. Ticagrelor versus clopidogrel in symptomatic peripheral artery disease. N Engl J Med. 2016 Nov 13. [Epub ahead of print] 42. Khoury MJ, Evans JP. A public health perspective on a national precision medicine cohort: balancing long-term knowledge generation with early health benefit. JAMA. 2015;313(21):2117-2118.

43. Joyner MJ, Paneth N. Seven questions for personalized medicine. Jama.

2015;314(10):999-1000.

44. Mannucci PM. Treatment of haemophilia: building on strength in the third millennium. Haemophilia. 2011;17 (Suppl 3):1-24.

45. Onarheim KH, Iversen JH, Bloom DE. Economic benefits of investing in women's health: a systematic review. PloS One. 2016;11(3):e0150120.

46. Martinoli C, Della Casa Alberighi O, Di Minno G, et al. Development and definition of a simplified scanning procedure and scoring method for haemophilia early arthropathy detection with ultrasound (HEAD-US).

Thromb Haemost. 2013;109(6):1170-1179.

47. Di Minno MN, Iervolino S, Soscia E, et al. Magnetic resonance imaging and ultrasound evaluation of "healthy" joints in young subjects with severe haemophilia A. Haemophilia. 2013;19(3):e167-173.

48. Lander ES. Cutting the Gordian helix--regulating genomic testing in the era of precision medicine. N Engl J Med. 2015;372(13):1185-1186.

49. Nathwani AC, Reiss UM, Tuddenham EG, et al. Long-term safety and efficacy of factor IX gene therapy in hemophilia B. N Engl J Med. 2014;371(21):1994-2004.

50. Lheriteau E, Davidoff AM, Nathwani AC. Haemophilia gene therapy: progress and challenges. Blood Rev. 2015;29(5):321-328.

51. Li H, Malani N, Hamilton SR, et al. Assessing the potential for AAV vec-

417

Editorials tor genotoxicity in a murine model. Blood. 2011;117(12):3311-3319.

52. Mega JL, Close SL, Wiviott SD, et al. PON1 Q192R genetic variant and response to clopidogrel and prasugrel: pharmacokinetics, pharmacodynamics, and a meta-analysis of clinical outcomes. J Thromb Thrombolysis. 2016;41(3):374-383.

53. Bhatt DL, Pare G, Eikelboom JW, et al. The relationship between CYP2C19 polymorphisms and ischaemic and bleeding outcomes in stable outpatients: the CHARISMA genetics study. Eur Heart J. 2012;33(17):2143-2150.

54. Munshi A, Sharma V. Genetic signatures in the treatment of stroke. Curr Pharm Des. 2015;21(3):343-354.

55. Block L, Habicht R, Wu AW, et al. In the wake of the 2003 and 2011 duty hours regulations, how do internal medicine interns spend their time? J Gen Intern Med. 2013;28(8):1042-1047.

56. Oxentenko AS, Manohar CU, McCoy CP, et al. Internal medicine residents' computer use in the inpatient setting. J Grad Med Educ.

2012;4(4):529-532.

57. Verghese A. Culture shock--patient as icon, icon as patient. N Engl J Med. 2008;359(26):2748-2751.

58. Rubin R. Precision medicine: the future or simply politics? JAMA. 2015;313(11):1089-1091.

59. Steinhart B. Patient autonomy: evolution of the doctor-patient relationship. Haemophilia. 2002;8(3):441-446.

60. Aledort LM. The evolution of comprehensive haemophilia care in the United States: perspectives from the frontline. Haemophilia.

2016;22(5):676-683.

61. Balint J, Shelton W. Regaining the initiative. Forging a new model of the patient-physician relationship. JAMA. 1996;275(11):887-891.

62. Paneni F. 2013 ESC/EASD guidelines on the management of diabetes and cardiovascular disease: established knowledge and evidence gaps. Diab Vasc Dis Res. 2014;11(1):5-10.

63. Dzau VJ, Ginsburg GS, Van Nuys K, Agus D, Goldman D. Aligning incentives to fulfil the promise of personalised medicine. Lancet.

2015;385(9982):2118-2119.

64. Stergiopoulos K, Brown DL. Genotype-guided vs clinical dosing of warfarin and its analogues: meta-analysis of randomized clinical trials. JAMA Intern Med. 2014;174(8):1330-1338.

65. Ciardiello F, Arnold D, Casali PG, et al. Delivering precision medicine in oncology today and in future-the promise and challenges of personalised cancer medicine: a position paper by the European Society for Medical Oncology (ESMO). Ann Oncol. 2014;25(9):1673-1678.

66. Pitt GS. Cardiovascular precision medicine: hope or hype? Eur Heart J. 2015;36(29):1842-1843.

67. Di Minno G, de Gaetano G. Human genome, environment and medical practice. Intern Emerg Med. 2013;8(8):645-649.

68. Huser V, Sincan M, Cimino JJ. Developing genomic knowledge bases and databases to support clinical management: current perspectives.

Pharmgenomics Pers Med. 2014;7:275-283.

69. Bayer R, Galea S. Public health in the precision-medicine era. N Engl J Med. 2015;373(6):499-501.

70. Robinson-Cohen C, Hoofnagle AN, Ix JH, et al. Racial differences in the association of serum 25-hydroxyvitamin D concentration with coronary heart disease events. JAMA. 2013;310(2):179-188.

71. Michos ED, Reis JP, Post WS, et al. 25-Hydroxyvitamin D deficiency is associated with fatal stroke among whites but not blacks: the NHANESIII linked mortality files. Nutrition. 2012;28(4):367-371.

72. Jonsen AR, Durfy SJ, Burke W, Motulsky AG. The advent of the "unpatients'. Nat Med. 1996;2(6):622-624.

73. Bloss CS, Schork NJ, Topol EJ. Direct-to-consumer pharmacogenomic testing is associated with increased physician utilisation. J Med Genet.

2014;51(2):83-89.

74. National Research Council (US) Committee on A Framework for Developing a New Taxonomy of Disease. Toward Precision Medicine: Building a Knowledge Network for Biomedical Research and a New Taxonomy of Disease. Washington (DC), 2011.

75. Ashley EA, Butte AJ, Wheeler MT, et al. Clinical assessment incorporating a personal genome. Lancet. 2010;375(9725):1525-1535.

76. Nherera L, Marks D, Minhas R, Thorogood M, Humphries SE. Probabilistic cost-effectiveness analysis of cascade screening for familial hypercholesterolaemia using alternative diagnostic and identification strategies. Heart. 2011;97(14):1175-1181.

77. Ahima RS. Rethinking the definition of diabetes for precision medicine.

Mol Endocrinol. 2015;29(3):335-337.

78. Marquet P, Longeray PH, Barlesi F, et al. Translational research: precision medicine, personalized medicine, targeted therapies: marketing or science? Therapie. 2015;70(1):1-19.

79. Punja S, Xu D, Schmid CH, et al. N-of-1 trials can be aggregated to generate group mean treatment effects: a systematic review and meta-analysis. J Clin Epidemiol. 2016;76:65-75.

80. Vohra S. N-of-1 trials to enhance patient outcomes: identifying effective therapies and reducing harms, one patient at a time. J Clin Epidemiol.

2016;76:6-8.

Research in the heart of hematology: chronic myeloid leukemia 2017 Rüdiger Hehlmann Medizinische Fakultät Mannheim, Universität Heidelberg, Germany E-mail: r.hehlmann@urz.uni-heidelberg.de

O

doi:10.3324/haematol.2016.159848

ne of the great success stories of modern hematology is reaching its next and possibly final phase: the achievement of treatment-free remissions in stable deep molecular responders with chronic myeloid leukemia (CML) which may well be equivalent to cure.

Although only the minority of patients achieve treatmentfree remissions, the absolute numbers of patients currently in discontinuation studies (Table 1) and in durable treatment-free remissions (40- 60%) are impressive and argue for a change in the treatment strategy for CML. The progress since last year cannot be overlooked.1 The goal is to define patients in whom treatment can be stopped safely and to establish a strategy for treatment discontinuation.2 This is not the first amazing success in some 50 years of basic and clinical research underlying the success story of CML: the detection of oncogenes and of kinase activity in many of them was fortuitous, since it was a byproduct of the search for human leukemia viruses. In realization that most animal leukemias could be induced by viruses, this 418

was a high priority research field in the late 1960s and early 1970s. Large national programs funded with billions of dollars, such as the Special Virus Cancer Program and the National Cancer Act for the “conquest of cancer”, had been started in the USA. With modern molecular biology methods, so-called footsteps of viruses were looked for.

The detection of reverse transcriptase in human leukemic cells3 and of virus-related RNA and DNA in human cells and in the human genome4,5 were at the time interpreted as breakthroughs on the path to detection of human leukemia viruses. Whereas ultimately no such viruses were found associated with common human leukemias, oncogenes proved central in human carcinogenesis. An example is the role in CML of the ABL oncogene which, in 1980, was detected in the acutely transforming defective Abelson leukemia virus in which parts of the virus genome had been replaced by cellular sequences.6 It was shown that most retroviral oncogenes were present as so called protooncogenes in the human genome pointing early to ubiquity and important functions of these genes in haematologica | 2017; 102(3)

Editorials

the biology of normal and malignant cells. The discovery that the ABL oncogene was located on chromosome 9 at the breakpoint of the t(9;22) translocation,7 and of fusion transcripts of ABL with the BCR region on chromosome 228 paved the way to the stunning observation that BCRABL sequences could induce leukemia in mice.9,10 Since ABL, like many other oncogenes, had tyrosine kinase activity, legions of tyrosine kinase inhibitors (TKI) were produced.11 It was the logical next step to define an inhibitor specific for BCR-ABL and suitable for therapeutic use in humans.12 The current global strategies are aimed at recognizing patient- and treatment-related factors indicating that treatment discontinuation would be successful and safe.

Optimization of current treatments has priority over the development and characterization of new drugs. New and better drugs may become more important again when strategies for TKI discontinuation have been optimized and the specific needs for drug treatment of patients who do not qualify for discontinuation are better known. More than 20 studies on treatment discontinuation have been published and even more were submitted for presentation at the 2016 American Society of Hematology (ASH) annual congress (Table 1). The total number of patients in published and ongoing studies on this subject is well above 3000. The largest of the studies, the EURO-SKI study,13 reports on 750 TKI (mostly imatinib) pre-treated patients with a follow-up after discontinuation of up to 36 months.

Evolving factors that have been identified to predict successful discontinuation and stable treatment-free remissions are duration of TKI treatment (≥5.8 years better) and duration of deep molecular response (each additional year

increases the probability of staying in major molecular remission by 16%). The impact of high Sokal risk score, younger age, gender, prior suboptimal response, TKI resistance, line of therapy, depth of remission (molecular response versus greater than molecular response) and other factors require confirmation or longer follow-up in larger cohorts. Treatment discontinuation can even be successful at a second attempt: based on a study of 60 patients, Legros et al. reported that the chances of success are not much different from those after first attempts.14 It came as a surprise that treatment discontinuation may induce adverse effects and that quality of life before and after stopping TKI treatment may not be much different.

The condition, which is termed TKI-discontinuation syndrome,15 with joint and muscle pain resembles polymyalgia rheumatica and occurs in about 30% of patients. In the majority of cases it seems to subside after some time and rarely requires reinstitution of TKI treatment.

An interesting observation is that no type or dose of TKI has thus far been shown to produce a clear survival advantage. An explanation could be that current TKI treatment is so efficient and survival so close to that of the general population that further improvement becomes difficult to prove, particularly in view of the fact that currently more patients die of comorbidities than of CML. Proof for this needs to be obtained by long-term observation of sufficiently large cohorts with survival as an endpoint.

An apparent limitation of progress is the concern that the same quality of CML management is not provided everywhere, not even in Europe or North America. The European LeukemiaNet (ELN) management recommendations for CML try to provide uniform definitions and rec-

Table 1.

Study

TKI Min. treatment duration (years) 1

Euro-SKI STIM2 TWISTER3 A-STIM4 KIDS5 HOVON6 STIM27 ISAV8 STOP 2G-TKI9 DADI10 NILST11 TRAD12 Dasfree13 ENESTop14 STAT215 ENEST freedom16 D-STOP17 RE-STIM18 Total: 18

IM IM IM IM IM IM IM IM DAS / NIL DAS 2nd line NIL IM / DAS DAS IM / NIL IM / NIL NIL IM / DAS (2nd stop)

3 2 3 3 3 20 months 2 2 2 ND* 2 3 2 3 2 3 ND* 35 months

N. 750 100 40 80 90 18 200 108 60 63 87 75 130 126 96 190 54 67 2334

Depth of MR 4

Min. duration of MR (years) RFS with at least MMR

MR Not determined individually MR4.5 UMRD* MR4.5 MR4.5 MR4.5 UMRD* MR4.5 MR4 MR4.5 MR4.5 MR4.5 MR4.5 MR4.5 MR4.5 MR4 mostly UMRD*

1 2 2 2 2 2 2 1.5 2 1 2 2 1 1 2 1 2 31 mos

52% at 2 years 38% at 7 years 45% at 42 months 64% at 23 months 58% at 2 years 33% at 3 years 46% at 2 years 52% at 22 months ≈ 55% at 4 years 49% at 6 months 59% at 1 year 58% at 6 months 63% at 1 year 58% at 4 years 68% at 1 year 52% at 4 years 63% at 1 year 44% at 22 months 33% – 68% after 0.5 – 7 years

TKI: tyrosine kinase inhibitor; IM: imatinib; DAS: dasatinib; NIL: nilotinib;*ND: not defined, MR: molecular response, UMRD: undetectable minimal residual disease; RFS: relapse-free survival; MMR: major molecular response. 1Mahon et al., ASH-Abstract 2016; 2Mahon et al., Lancet Oncology 2010;11:1029; 3Ross et al., Blood 2013;122:515; 4Rousselot et al., JCO 2014;32:424; 5Lee et al., Haematologica 2016;104:717; 6Thielen et al., EJC 2013;49:3242; 7Nicolini et al., ASH-Abstract 2013 # 654; 8Mori et al., AJH 2015;90:910; 9Rea et al. Blood 2016, epub ahead of print; 10 Imagawa et al., Lancet Haematol. 2015; 11Kadowaki et al., ASH-Abstract 2016; 12Kim et al., ASH-Abstract 2016; 13Shah et al. ASH-Abstract 2010; 14Hughes et al., ASH-Abstract 2016; 15Takahashi et al. ASH-Abstract 2016; 16Hochhaus et al., ASCO-Abstract 2016; 17Keenajai et al., ASH-Abstract 2016; 18Legros et al., ASH-Abstract 2016.

haematologica | 2017; 102(3)

419

Editorials

ommendations globally. The population-based registry of the European Treatment and Outcome Study (EUTOS) for CML project, a public-private partnership between the ELN and Novartis, now reports that in 20 European countries most patients are managed according to ELN recommendations.16 The prospects are therefore excellent that in Europe the new treatment discontinuation strategy in CML will be available to most CML patients even in routine care. The reliable availability of high quality, standardized, molecular monitoring is of the utmost importance.

Another registry study (Simplicity) involving 1,494 TKItreated CML patients from North America and Europe found lower rates of molecular monitoring in the USA than in Europe. In an analysis of switching therapies within the first 12 months, intolerance of first-line TKI was, at >70%, the most frequent reason for switching treatment.17 Suboptimal tolerability of TKI and adverse effects, particularly of second- and third-generation TKI, may be obstacles to achieving the best possible outcome. The ELN has, therefore, appointed an international panel of experts, including experts from North America and Asia, to provide evidence-guided recommendations for the management of TKI-related adverse events.18 The recommendations will help to provide uniform and high quality management of CML patients globally.

In spite of the excellent long-term tolerability of imatinib, careful monitoring for late toxicities remains important. Some reports indicate that kidney function should be more carefully monitored since a decrease of glomerular filtration rate has been observed after long-term imatinib treatment. Older age and lower estimated glomerular filtration rate at the initiation of imatinib were found to be associated with later development of chronic kidney disease.19 This confirms an earlier report by Marcolino et al.20 Since nilotinib has been reported beneficial for renal function, a switch to nilotinib may be considered in patients with these characteristics.

Pricing has been a topic of serious concern in CML treatment for some time. The now general availability of generic imatinib alleviates this concern, but the question remains whether the quality of generic imatinib preparations will be equal to that of branded imatinib. At least five contributions to the ASH 2016 conference have addressed this question and agree that generic imatinib is of similar quality. The largest cohort was from the Polish Imatinib Generics Registry: Sacha et al. prospectively observed 726 patients treated with various generic imatinib preparations (mostly Nibix and Meaxin) for 1 year and concluded that the clinical efficacy and tolerability of the tested generics are not inferior to those of branded imatinib.21 A long-neglected, but central determinant of the natural course of CML is additional chromosomal aberrations as a consequence of BCR-ABL-induced genetic instability.

After the frequency of such aberrations in blast crisis had been observed22 and the relevance of clonal evolution for progress of CML recognized, more detailed information was provided recently on which additional chromosomal aberrations may be drivers and which are merely bystanders.23 Fabarius et al. analyzed the prognostic impact of unbalanced additional chromosomal aberrations at 420

Figure 1. Translocation t(9;22) with iso-chromosome i(17)(q10) which has been identified as a poor prognosticator at diagnosis and in the course of CML. The long arm of chromosome 17 is doubled and replaces the short arm in an inverse direction. Multi-color fluorescence in situ hybridization. Courtesy of C. Haferlach and A. Fabarius.

diagnosis and found that only major route aberrations (+8, +Ph, i(17)(q10) and +19) had a negative impact on survival.24 Wang et al. determined the impact of additional chromosomal aberrations arising de novo in the course of CML in more than 2,000 patients and categorized the aberrations according to their impact on survival. The most unfavorable were chromosome 17, 7 and 3 aberrations [i(17)(q10), 3q26, -7] (Figure 1). Trisomy 8 was found to be less unfavorable unless it was combined with other additional chromosomal aberrations.25 Chen et al. examined the differential impact of additional chromosomal aberrations on lymphoid and myeloid blast crisis and found that such aberrations confer an inferior prognosis in myeloid, but not in lymphoid blast crisis.

The challenge for the coming years will be how best to prevent additional chromosomal aberrations in the remaining patients who cannot be treated satisfactorily with TKI.

References 1. Hehlmann R. Innovation in hematology. Perspectives: CML 2016.

Haematologica. 2016;101(6):657-659.

2. Hughes TP, Ross DM. Moving treatment-free remission into mainstream clinical practice in CML. Blood. 2016;128(1):17-23.

3. Gallo RC, Yang SS, Ting RC. RNA-dependent DNA polymerase of human acute leukemic cells. Nature. 1970;228(5275):927-929.

4. Hehlmann R, Kufe D, Spiegelman S. RNA in human leukemic cells related to the RNA of a mouse leukemia virus. Proc Natl Acad Sci USA. 1972;69(2):435-439.

5. Leib-Mรถsch C, Brack R, Werner T, Erfle V, Hehlmann R. Isolation of an SSAV-related endogenous sequence from human DNA. Virology.

1986;155(2):666-677.

6. Witte ON, Dasgupta A, Baltimore D. Abelson murine leukaemia virus protein is phosphorylated in vitro to form phosphotyrosine. Nature.

1980;283(5750):826-831.

7. Heisterkamp N, Stephenson JR, Groffen J, et al. Localization of the c-ab1 oncogene adjacent to a translocation break point in chronic myelocytic leukaemia. Nature. 1983;306(5940):239-242.

8. Shtivelman E, Lifshitz B, Gale RP, Canaani E. Fused transcript of abl and bcr genes in chronic myelogenous leukaemia. Nature. 1985;315 (6020):550-554.

9. Daley GQ, van Etten RA, Baltimore D. Induction of chronic myeloge-

haematologica | 2017; 102(3)

Editorials

10.

11.

12.

13.

14.

15.

16.

17.

nous leukemia in mice by the P210bcr/abl gene of the Philadelphia chromosome. Science. 1990;247(4944):824-830.

Heisterkamp N, Jenster G, ten Hoeve J, Zovich D, Pattengale PK, Groffen J. Acute leukaemia in bcr/abl transgenic mice. Nature. 1990;344 (6263):251-253.

Levitzki A, Gazit A. Tyrosine kinase inhibition: an approach to drug development. Science. 1995;267(5205):1782-1788.

Druker BJ, Tamura S, Buchdunger E, et al. Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells. Nat Med. 1996;2(5):561-566.

Mahon FX. Cessation of tyrosine kinase inhibitors treatment in chronic myeloid leukemia patients with deep molecular response: results of the Euro-Ski trial. ASH-Abstract # 787 2016.

Legros L. Second TKI siscontinuation in CML patients that failed first discontinuation and subsequently regained deep molecular response after TKI re-challenge. ASH-Abstract # 788 2016.

Richter J, Soderlund S, Lubking A, et al. Musculoskeletal pain in patients with chronic myeloid leukemia after discontinuation of imatinib: a tyrosine kinase inhibitor withdrawal syndrome? J Clin Oncol.

2014;32(25):2821-2823.

Hoffmann VS, Baccarani M, Hasford J, et al. Treatment and outcome of 2904 CML patients from the EUTOS population-based registry.

Leukemia. 2016 Sep 23. [Epub ahead of print] Goldberg S. Tyrosine kinase inhibitor (TKI) switching patterns during the first 12 months in simplicity, an observational study of chronic-phase chronic myeloid leukemia (CP-CML) patients (Pts) in routine clinical practice. ASH-Abstract # 937 2016.

18. Steegmann JL, Baccarani M, Breccia M, et al. European LeukemiaNet recommendations for the management and avoidance of adverse events of treatment in chronic myeloid leukaemia. Leukemia. 2016;30(8):16481671.

19. Sakurai M. Long-term treatment with imatinib is associated with decreased estimated glomerular filtration rate and hemoglobin level in patients with chronic myelogenous leukemia. ASH-Abstract # 1888 2016.

20. Marcolino MS, Boersma E, Clementino NC, et al. Imatinib treatment duration is related to decreased estimated glomerular filtration rate in chronic myeloid leukemia patients. Ann Oncol. 2011;22(9):2073-2079.

21. Sacha T. Imatinib generics in treatment of chronic myeloid leukemia; a prospective observation in large cohort of patients from polish (PALG) Imatinib Generics Registry. ASH-Abstract # 629 2016.

22. Johansson B, Fioretos T, Mitelman F. Cytogenetic and molecular genetic evolution of chronic myeloid leukemia. Acta Haematol. 2002;107(2):7694.

23. Fabarius A, Kalmanti L, Dietz CT, et al. Impact of unbalanced minor route versus major route karyotypes at diagnosis on prognosis of CML. Ann Hematol. 2015;94(12):2015-2024.

24. Wang W, Cortes JE, Tang G, et al. Risk stratification of chromosomal abnormalities in chronic myelogenous leukemia in the era of tyrosine kinase inhibitor therapy. Blood. 2016;127(22):2742-2750.

25. Chen Z, Cortes JE, Jorgensen JL, et al. Differential impact of additional chromosomal abnormalities in myeloid vs lymphoid blast phase of chronic myelogenous leukemia in the era of tyrosine kinase inhibitor therapy. Leukemia. 2016;30(7):1606-1609.

Research in morphology and flow cytometry is at the heart of hematology Marie C. BĂŠnĂŠ1 and Gina Zini,2 EHA Scientific Working Group "Diagnosis", European LeukemiaNet WP10.

1

Hematology Biology, Nantes University Hospital, France and 2Blood Bank and UNICATT Cord Blood Bank, Fondazione Policlinico Gemelli, UniversitĂ Cattolica del S. Cuore, Roma, Italy E-mail: mariecbene@gmail.com

T

doi:10.3324/haematol.2016.163147

hrough its inherent implication in such major physiological systems as oxygenation, coagulation, protection against infections and tumor proliferation, hematology is a nearly boundless field for research and discovery. Indeed, hematology has been at the heart of such fundamental work as the understanding of iron regulation in red blood cell mediated tissue oxygenation,1 or the deciphering of the intricate molecular interactions between endothelial cells, platelets and plasmatic proteins in the early stages of hemostasis.2 Deeper insights into cell biology have also been attained through the use of the easily available blood or bone marrow cells and cell lines derived from hematological malignancies. Moreover, the notion of the importance of the microenvironment of many cells has certainly found fertile ground in the study of the bone marrow or diseased lymph nodes involved in leukemia and myeloproliferative or lymphoproliferative disorders.3-5 Hematology is thus entertaining strong transdisciplinary interactions with pathology, immunology, biochemistry, cytogenetics, molecular biology and also, more recently, advanced imaging techniques.6 These various aspects deal with the physiological maintenance of homeostasis, a precise and tightly regulated phenomenon in an extremely active system producing and eliminating trillions of cells every day. They are also at the heart of our increasing understanding of the mechanisms of disease and targeted therapeutic approaches.

In the field of hematology, two sub-disciplines are at the basis of diagnosis for malignant and non-malignant haematologica | 2017; 102(3)

processes, i.e., the morphology of hematopoietic cells and flow cytometry. Both are considered indispensable, but it is sometimes forgotten that both represent true specialties, requiring thorough training and experience. It could be argued that the progresses of automation could alleviate these prerequisites, but the reality seems more subtle.

Indeed, blood cell counters performing a complete blood count (CBC) are becoming more and more sophisticated, as previously underscored in the 2016 editorial from our scientific working group on innovation.7 With different technological approaches, latest generation instruments perform accurate and reproducible quantitative measurements of peripheral blood cell composition, and have good sensitivity and specificity for flagging the presence of abnormal cells.8 Understanding the flags or messages generated by these intelligent machines, in particular through morphological cell identification on a stained blood smear, still requires the knowledge initially placed in the design of automated interpretation. In addition, as with most expert systems, when the machine is at a loss, only the brain of the biologist/hematologist can come to the rescue, a task made increasingly difficult by the lack of experience in normal or benign situations taken care of by the instrument. Furthermore, cytomorphological analysis of bone marrow aspirates remains a cornerstone in the most recent WHO classification, and requires additional skill and knowledge.

The same is true for flow cytometry, with new instruments managing the whole process of handling samples 421

Editorials

from HIV patients for instance.9 In most cases, the robot is performing and provides straightforward and accurate results. But all biologists who have ever confronted this specific population know that some patients' samples refuse to behave as predicted, thus requiring the skills of the scientist to solve the riddle of, for instance, red blood cells not lysed efficiently. More is to be expected along these lines with the development of antibody cocktails adapted to specific disorders, where patients, in spite of all hope, cannot be standardized. What would happen, for example, in the case of double clones of lymphoproliferative disorders with an entirely automated system? Indeed, research is required in these fields to find the best compromise between time- and cost-saving robust methodology and hematological expertise in interpretation as well as “plan B” technology for unconventional cases.

Thus, research belongs at the heart of the hematology platform, dealing with the most widely prescribed blood test, CBC, and its flow cytometric corollaries.

But such a limited view would be far too reductionist.

Indeed, both morphology and flow cytometry are also at the heart of clinical research. How would any clinical trial in malignant hematology be possible without an accurate diagnosis which initially relies on these tests? How would the hematological toxicity of sometimes daring/drastic therapies be interpreted without repeated CBC and blood and bone marrow cell examinations? How would the large amount of recently publicized10 searches for fast chemosensitivity and high sensitivity minimal residual disease assessment be possible without flow cytometry?

Indeed, although often not financed, these tests are at the core of evaluations in malignant, and also non-malignant, hematological diseases. The progress of clinical research in the field of myelodysplastic syndromes (MDSs) also rely on both morphology and flow cytometry. Indeed, a Perls prussian blue staining and microscopic evaluation is far easier and more broadly performed than the search for SF3B1 mutations, even if the latter can be a surrogate classifier in cases where there are less than 5% ring sideroblasts.11 The growing literature underscores the value of searching for immunophenotypic anomalies in cytopenias with suspected MDS.12 In a broader sense of appreciation, it can also be pointed out that these “basic” methods are also extremely useful and necessary in more fundamental research. Examples are numerous, but some can be drawn from the latest meeting of the American Society of Hematology. In an exciting zebra fish model allowing for the manipulation of hematopoietic stem cells, Leonard Zon showed how the addition of mutations could indeed lead to the development of leukemia in these animals, as demonstrated morphologically by the clonal proliferation of leukocytes.13 Similarly, in the Ernest Beutler lecture, Hugues de Thé and Zhu Chen reviewed how all-trans retinoic acid (ATRA) and arsenic can drive the leukemic cells of promyelocytic leukemia to proceed to morphological and immunophenotypic differentiation.14 Even in the blossoming field of chimeric antigen receptor (CAR) T cells, the characterization of the most efficient cell products relies on flow cytometric assessment of their immunophenotype and pluripotency in chemokine secretion.15

422

Therefore, each year brings more knowledge and understanding in the complex world of hematology. Research progresses fast, using sophisticated new technology to unravel molecular mechanisms from bioinformatics analyses of huge sequences databases. Yet, in parallel, other methods are developed to get a better grasp on the “physical” structure/morphology of organelles such as ribosomes, i.e., with cryo-electron microscopy.16 Meanwhile, classical methods retain their intrinsic value to check on the morphology of manipulated cells or to use antibodies to assess their immunophenotypic or secretory profile.

In the expanding field of hematological research, which apparently is still far from having unveiled all its idiosyncrasies, the importance of maintaining the competence of experts in morphology and flow cytometry appears to be as crucial as ever in the daily process of diagnosis.

References 1. Camaschella C, Pagani A, Nai A, Silvestri L. The mutual control of iron and erythropoiesis. Int J Lab Hematol. 2016;38 (Suppl 1):20-26.

2. Ruggeri ZM, Mendolicchio GL. Interaction of von Willebrand factor with platelets and the vessel wall. Hamostaseologie. 2015;35(3):211-24.

3. Dhami SP, Kappala SS, Thompson A, Szegezdi E. Three-dimensional ex vivo co-culture models of the leukaemic bone marrow niche for functional drug testing. Drug Discov Today. 2016;21(9):1464-1471.

4. Chiarini F, Lonetti A, Evangelisti C, Buontempo F, Orsini E, Evangelisti C, et al. Advances in understanding the acute lymphoblastic leukemia bone marrow microenvironment: From biology to therapeutic targeting.

Biochim Biophys Acta. 2016;1863(3):449-463.

5. Nicholas NS, Apollonio B, Ramsay AG. Tumor microenvironment (TME)-driven immune suppression in B cell malignancy. Biochim Biophys Acta. 2016;1863(3):471-482.

6. Lin S, Ouyang T, Kanekar S. Imaging of Bone Marrow. Hematol Oncol Clin North Am. 2016;30(4):945-971.

7. Béné MC, Zini G; EHA Scientific Working Group "Diagnosis", European LeukemiaNet WP10. Innovation in hematology: morphology and flow cytometry at the crossroads. Haematologica. 2016;101(4):394-395.

8. Verbrugge SE, Huisman A. Verification and standardization of blood cell counters for routine clinical laboratory tests. Clin Lab Med. 2015;35(1):183-196.

9. Gossez M, Malcus C, Demaret J, Frater J, Poitevin-Later F, Monneret G. Evaluation of a novel automated volumetric flow cytometer for absolute CD4+ T lymphocyte quantitation. Cytometry B Clin Cytom. 2016 Jan 25. [E-pub ahead of print].

10. Grimwade D, Freeman SD. Defining minimal residual disease in acute myeloid leukemia: which platforms are ready for "prime time"?

Hematology Am Soc Hematol Educ Program. 2014;2014(1):222-233.

11. Arber DA, Orazi A, Hasserjian R, Thiele J, Borowitz MJ, Le Beau MM, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):23912405.

12. Porwit A, van de Loosdrecht AA, Bettelheim P, Brodersen LE, Burbury K, Cremers E, et al.. Revisiting guidelines for integration of flow cytometry results in the WHO classification of myelodysplastic syndromes-proposal from the International/European LeukemiaNet Working Group for Flow Cytometry in MDS. Leukemia. 2014;28(9):1793-1798.

13. Liu W, Wu M, Huang Z, Lian J, Chen J, Wang T, et al. c-myb hyperactivity leads to myeloid and lymphoid malignancies in zebrafish. Leukemia.

2016 Jul 26. [Epub ahead of print] 14. Giannì M, Koken MH, Chelbi-Alix MK, Benoit G, Lanotte M, et al.

Combined arsenic and retinoic acid treatment enhances differentiation and apoptosis in arsenic-resistant NB4 cells. Blood. 1998;91(11):43004310.

15. Watanabe K, Terakura S, Martens AC, van Meerten T, Uchiyama S, Imai M, et al. Target antigen density governs the efficacy of anti-CD20-CD28CD3 ζ chimeric antigen receptor-modified effector CD8+ T cells. J Immunol. 2015;194(3):911-920.

16. Fu Z, Kaledhonkar S, Borg A, Sun M, Chen B, Grassucci RA, et al. Key intermediates in ribosome recycling visualized by time-resolved cryoelectron microscopy. Structure. 2016;24(12):2092-2101.

haematologica | 2017; 102(3)

REVIEW ARTICLE

Is immunotherapy here to stay in multiple myeloma?

Paula Rodríguez-Otero,1 Bruno Paiva,1 Monika Engelhardt,2 Felipe Prósper1 and Jesús F. San Miguel 1 Clínica Universidad de Navarra, Centro de Investigación Médica Aplicada, Instituto de Investigación Sanitaria de Navarra, Pamplona, Spain and 2Department of Medicine I, Hematology, Oncology and Stem Cell Transplantation, Medical Center - University of Freiburg, Faculty of Medicine, Germany

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

1

ABSTRACT

Haematologica 2017 Volume 102(3):423-432

I

mmune escape and impaired immune surveillance have been identified as emerging hallmarks of cancer.1 Multiple myeloma represents a genuine example of disrupted immune surveillance characterized by: impaired antibody production, deregulation of the T and natural killer cell compartment, disruption of antigen presentation machinery, upregulation of inhibitory surface ligands, and recruitment of immunosuppressive cells. Although the potential value of immunotherapeutic interventions had a clear antecedent in the graft-versus-myeloma effect induced by allogeneic stem cell transplant and donor lymphocyte infusions, it is only recently that this field has faced a real revolution. In this review we discuss the current results obtained with immune approaches in patients with multiple myeloma that have placed this disease under the scope of immuno-oncology, bringing new therapeutic opportunities for the treatment of multiple myeloma patients.

Correspondence: Introduction Multiple myeloma (MM) is a malignant disorder of clonal plasma cells (PCs) that represents approximately 1% of cancers and 10% of hematological malignancies.2 The survival of MM patients has significantly improved over the past two decades, first through the introduction of high-dose therapy followed by autologous stem cell transplantation (ASCT), and more recently due to the use of proteasome inhibitors (PIs) and immunomodulatory drugs (IMiDs) (Figure 1).3 It is expected that such improvement in patient outcomes will continue in the years to come.

Continuous drug development and understanding of the MM biology has led to a landmark achievement, with the approval in 2015 of four new drugs for the treatment of relapse and relapse/refractory MM patients. Two out of these four new drugs are monoclonal antibodies (mABs), elotuzumab and daratumumab, that represent a new passive immune-mediated therapeutic approach for MM patients, and have placed myeloma under the spotlight of the utmost promising field of immunooncology (IO). In this article we review the current knowledge of immune disturbance in MM together with the most relevant immunotherapeutic strategies in this disease.

Impaired immune surveillance in MM The generation of anti-cancer immunity is a complex, multistep process that starts with the release of cancer cell antigens after cell death. Tumor antigens are then processed and presented by antigen-presenting cells (APCs) to effector T cells, that will migrate to the tumor site once they are primed and activated; there, they may recognize the tumor antigens and launch an immune response to eradicate the cancer cells.4 Unfortunately, tumors display a wide variety of mechanisms that allow them to evade immune control, such as: (a) the production of pro-inflammatory cytokines mediating the suppression of dendritic and T-cell activation and proliferation,5,6 (b) the disruption of antigen presentation machinery through the downhaematologica | 2017; 102(3)

sanmiguel@unav.es

Received: August 2, 2016.

Accepted: October 17, 2016.

Pre-published: January 12, 2017. doi:10.3324/haematol.2016.152504 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/423 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

423

P. Rodriguez-Otero et al.

Figure 1. Evolution of the multiple myeloma treatment landscape: multiple myeloma treatment has evolved rapidly over the last years. The first active MM drug developed was melphalan in 1958. From then until 2003 the management of MM patients was mainly focused on the use of high-dose chemotherapy with stem cell rescue. In 2003 the first lMiD was approved (thalidomide), and straightaway bortezomib and lenalidomide were incorporated into the drug repertoire. For ten years these drugs were pivotal in the management of MM treatment, but in the last two years five new drugs have been approved, and immuno-oncology strategies are under development with promising activity. Th: T helper; TGF-β: transforming growth factor-β; VEGF: vascular endothelial growth factor; PGE2: prostaglandin E2; Ab: antibody; HLA: human leucocyte antigen; PD-L1: programmed death-ligand 1; Tregs: regulatory T cells; MDSCs: myeloid-derived suppressor cells; DC: dendritic cell; MDCS: myeloid dendritic cells; CCL2: CC motif chemokine ligand 2; CXCL12: CX-C motif chemokine ligand 12.

regulation of human leucocyte antigen (HLA) costimulatory molecules,7 (c) the upregulation of inhibitory surface ligands that induce T-cell anergy and exhaustion,8 or (d) the recruitment of immunosuppressive cells.9,10 Virtually all these mechanisms of tumor escape have been described in MM (Figure 2), and have been postulated to contribute to disease progression. Firstly, there is a reduction of bone marrow (BM) B-cell precursors that leads to impaired antibody production.11 Secondly, the T-cell compartment is deregulated due to reduced numbers of CD4+ T cells, altered CD4/CD8 ratio,12 abnormal T helper (Th)1/Th2 profile in favor of a Th2 immune response,13 and an increase in the number of regulatory T cells (Tregs).14,15 Furthermore, MM clonal PCs also express increased levels of inhibitory ligands, such as programmed death-ligand 1 (PD-L1), that inhibits the activation and proliferation of programmed cell death protein 1 (PD-1) 424

positive T cells.16 Thirdly, MM patients show disruption in antigen presentation, with some studies reporting defects in peripheral blood dendritic cells (DC), such as a reduced number of plasmacytoid dendritic cells (pDCs), myeloid DCs (mDCs) or peripheral blood monocytes, and also a lower expression level of both major histocompatibility complex (MHC) class II (HLA-DR) and costimulatory molecules (CD40, CD80).17 Fourth, stromal cells produce proinflammatory cytokines and other chemokines that recruit immunosuppressive cell populations, such as Tregs and myeloid-derived suppressor cells (MDSCs), thereby creating a permissive microenvironment allowing the tumor to evade immune control.18-20 Interestingly, at the same time that a growing body of evidence supported a role for immune dysfunction in the pathogenesis of MM, other examples have also emerged that clearly illustrate the importance of active immune haematologica | 2017; 102(3)

Immunotherapy strategies in Multiple Myeloma

Figure 2. Multiple myeloma is one example of disrupted immunosurveillance and immune evasion. Some evidence underscoring the disturbed immune system in MM are: (a) Impaired induction of allogeneic T-cell responses due to a decrease in the number of CD4+ T cells, and an abnormal Th1/TH2 cytokine profile; (b) reduction in the B-cell compartment with altered B-cell differentiation and antibody response; (c) decrease in the expression of tumor antigens and HLA costimulatory molecules leading to inadequate T-cell costimulation; (d) upregulation of inhibitory ligands such as PD-L1 which mediate anergy and T-cell exhaustion; (e) recruitment of immunosuppressive cell populations like MDSCs or Tregs. IMiDs: immunomodulatory drugs; inh: inhibitor.

surveillance in this disease. In the QUIREDEX trial, early treatment with lenalidomide and low-dose dexamethasone was compared to abstention in high-risk smoldering multiple myeloma (SMM) patients. It was observed that high-risk SMM patients already presented an impaired immune system at the moment of diagnosis as compared to healthy individuals of the same age, with a decreased expression of activation, Th1 and proliferation-related markers in immune cells. After nine induction cycles with lenalidomide and low-dose dexamethasone, the expression of these markers was restored and a shift in Tcell and natural killer (NK) cell phenotype was induced, with an increase in central memory T cells (CMTs), effector memory T cells (EMTs), the induction of activation markers and an increase in proliferating CD4+ and CD8+ cells.21 Another example is the graft-versus-myeloma effect of allogeneic stem cell transplant (SCT), which has been highlighted recently by Ladetto et al. upon comparing outcomes between minimal residual disease (MRD) positive and negative patients after tandem auto-allo haematologica | 2017; 102(3)

SCT using allele-specific oligonucleotide polymerase chain reaction (ASO-PCR). With a median follow-up of 12 years, 73% of the MRD-negative cases remained relapse-free; such promising results in MRD-negative patients have never been observed outside of the allogeneic setting.22 Furthermore, by using 8-color flow cytometry to simultaneously assess MRD and characterize patients’ immune profile, we were able to show that a few MRD positive cases - those showing a strong recovery of the normal B-cell lymphopoiesis - have similar outcomes to those of MRD negative patients.23 This suggests that despite MRD positivity, the intact immune surveillance was profitable in these patients. Finally, we have observed that MM patients attaining long-term survival (i.e., progression-free survival [PFS] for more than 10 years) showed a unique immune profile with a higher number of effector cells (T cells, NK cells, DCs, normal PCs), and a lower number of Tregs, underlying the importance of an active immune system to control disease evolution.24 425

P. Rodriguez-Otero et al.

Four major targets of immunotherapy in multiple myeloma The first evidence supporting a role for immunotherapy in MM comes from the graft-versus-myeloma effect induced by allogeneic SCT (allo-SCT) and donor lymphocyte infusions, that may cure some MM patients.25-28 However, the substantial toxicity of this procedure, along with the occurrence of relapse of the disease after the transplant, has hampered its extensive use. As for pharmacological approaches, interferon was the first drug used to stimulate the immune system, but its efficacy was only modest, and thus it is not currently considered as part of the myeloma treatment armamentarium; despite this, recent publications underlining its potential role in combination with other IO drugs, such as checkpoint inhibitors, may reactivate its use in the future.29,30 Subsequently, we had the opportunity to experience the anti-myeloma efficacy of a new class of compounds, the IMiDs (thalidomide, lenalidomide and pomalidomide) that represent one of the key backbones in the treatment of MM. Herein, we review four novel and most promising approaches that are currently under investigation to enhance the immune system against MM cells (Figure 3).

These are: 1. direct targeting of surface tumor antigens with monoclonal antibodies, 2. boosting immune effector using adoptive cell therapy, 3. improving immunity against tumors with vaccines, and 4. overcoming immune suppression with checkpoint blockade.

Direct targeting of surface tumor antigens using monoclonal antibodies (mAb) Monoclonal antibodies exert their cytotoxic function through different mechanisms: antibody-dependent cellular cytotoxicity (ADCC) through the engagement of immune effector cells, complement activation, antibodydependent phagocytosis, and direct effect on target cells acting through different signaling pathways. Although these mechanisms are postulated based on in vitro studies, their relative contribution to the clinical responses of mAb therapy is difficult to determine.31 While mAb therapy is already a standard of care in the treatment of some hematological malignancies, such as B-cell lymphoproliferative disorders,32 it was not until recently that this therapeutic approach was made available for MM patients. The development of effective mAb therapies in MM has probably been hindered due to both the lack of knowledge about specific PC targets (e.g., SLAMF7 or BCMA), and the concern that other highly expressed molecules on PCs were also relatively abundant in other hematopoietic cells, which would result in significant off-target effects. Nowadays, there are two mAbs, elotuzumab and daratumumab, approved for the treatment of MM. Elotuzumab is an IgG1Îş mAb with specificity against SLAMF7, an antigen expressed on both normal and malignant PCs as well as NK and T cells.33 Elotuzumab used as a single agent does not induce objective responses in MM, but in combination with lenalidomide plus dexamethasone (Rd) in a phase II trial showed high activity with an overall response rate (ORR) of up to 92%.34 These results were the basis for the randomized phase III Eloquent-2 trial comparing elotuzumab plus Rd versus Rd in relapsed/refractory MM (RRMM) patients. In this trial, the experimental arm showed a significant superiority in terms of ORR (79% vs. 66%), PFS (19.4 vs. 14.9 months, 426

[HR 0.73, 95%CI 0.60-0.89]; P=0.0014) and overall survival (OS) (43.7 vs. 39.6 months, respectively), and a delay of 12 months in the time to next treatment (TNT) (HR=0.62, 95%CI 0.50-0,77).35 Regarding anti-CD38 mAbs, three different compounds, daratumumab, isatuximab and MOR22 (MORO3O87), are currently being investigated.

Daratumumab has shown clear activity as a single agent, as it has recently been updated in a pooled analysis of 148 patients with RRMM having received more than two prior lines of therapy. Daratumumab was given at the standard dose of 16mg/kg.36-38 It should be noted that 86.5% of the patients were double refractory (to PI and IMiD). The ORR was 31.1%, including thirteen very good partial responses (VGPRs), four complete responses (CRs), and three stringent complete responses (sCRs). The median duration of response was 7.6 months. Median PFS and OS were 4.0 months and 20.1 months, respectively. The toxicity profile was very acceptable. No immunogenicity was observed, and the most commonly reported adverse events were infusion-related reactions that occurred predominantly during the first full infusion. Both the FDA and EMA have already approved daratumumab for this indication.

The efficacy markedly increased upon combining daratumumab with Rd, with an ORR of 81%.39 These positive results were the basis of a large randomized phase III POLLUX trial that compared the triple combination of daratumumab plus Rd (DRd) with the standard two drug combination, Rd, in relapsed patients not refractory to lenalidomide.40 The most recently presented results showed a highly significant superiority for the experimental arm in terms of ORR (93% vs. 76%), CR rate (43% vs. 19%), TNT (not reached (NR) vs. 18.4 months), and PFS (NR vs.

18.4 months, HR 0.37 [0.27-0.52], P<0.0001), with an unprecedented estimated median PFS for the triplet DRd arm of 44 months in relapsed MM. Similar positive results have also been presented in the phase III randomized CASTOR trial that compared daratumumab, bortezomib plus dexamethasone (DVd) with bortezomib plus dexamethasone (Vd) in relapsed patients not refractory to bortezomib.41 Again, the triplet combination of DVd was superior to the control arm in terms of ORR (83% vs. 63%), CR rate (20% vs. 9%), median TNT (NR vs. 7.3 months, HR: 0.30 (95% CI, 0.21-0.43); P<0.0001) and also for PFS, as the median PFS for the triplet arm was NR vs. 7.2 months in the Vd arm (HR: 0.39 (95% CI, 0.28-0.53); P<0.0001).

Other combinations with carfilzomib plus dexamethasone and also pomalidomide plus dexamethasone are under investigation in a phase I trial (clinicaltrials.gov Identifier:01998971). Preliminary results for the daratumumab plus pomalidomide plus dexamethasone combination have shown an ORR of 71% with a CR of 9% in a RRMM population.42 As far as isatuximab is concerned, a phase I study identified 10 mg/kg as the optimal dose with a response rate of 29% used as a single agent; infusion reactions were mainly grade 1/2 and only during the first doses.43 The efficacy of isatuximab in combination with Rd was confirmed in a small pilot study with a response rate of 58%, including 6% sCR and 23% VGPR, in patients that were mostly (84%) refractory to lenalidomide.44 In general, the safety profile of these mAbs, both elotuzumab and anti-CD38 mAbs, is manageable, and infusion-related reactions (IRRs) are the more frequent adverse events, present in haematologica | 2017; 102(3)

Immunotherapy strategies in Multiple Myeloma

Figure 3. There are four major targets for cancer immunotherapy. 1. Direct target of surface tumor antigens with monoclonal antibodies; 2. Boost immune effector using adoptive cell therapy; 3. Improve immunity against tumors with vaccines; 4. Overcome immune suppression with checkpoint blockade. Chemo: chemotherapy; HD: high-dose; ASCT: autologous stem cell transplantation; CAR: chimeric antigen receptors; BiTEs: bispecific T-cell engagers; CML: chronic myelogenous leukemia; NHL: non-Hodgkin lymphoma.

around 10% of patients for elotuzumab and 48% of patients for daratumumab. IRRs are more frequent during the first infusion, and most are grade 1-2, with virtually no patients having to be discontinued. Two other issues that need to be stressed regarding mAbs treatment are the interference in both the evaluation of response and in the blood typing. The first problem applies to all mAbs, considering that they are immunoglobulins (most are IgGÎş) that could be detected in both the serum protein electrophoresis (SPEP) and the immunofixation test. The second problem is limited to anti-CD38 antibodies (Abs), since CD38 antigen is also expressed on the surface of red blood cells, thus interfering in blood typing because of a false positive indirect Coombs test.

There is further development in the field of mAbs with the use of bispecific antibodies, such as bispecific T-cell engagers (BiTEs). These drugs combine the specificities of two antibodies; one involves the engagement and activation of T cells via CD3, and the other recognizes the cancer antigen. This class of drugs may overcome the inhibition of an immunosuppressive microenvironment because they activate and bind the effector T cell to the tumor cell, and thereby lead to an increased lytic potential of autologous effector T cells.45 The first BiTE to be generated haematologica | 2017; 102(3)

against myeloma cells was developed by combining single-chain variable fragments (ScFvs) of a mAb that binds normal and malignant PCs (Wue-1).46 Other BiTEs are under development using other antigens, such as B-cell maturation antigen (BCMA).47 Antibodies can also be conjugated with cytotoxic molecules, such as monomethyl auristatin E (e.g., ABBV-838), or radioactive particles.48 Both technologies are also being explored in MM, both in preclinical and clinical studies (clinicaltrials.gov Identifier:02462525).

Boosting immune effectors through adoptive cell therapy A second strategy to improve and/or increase immunity against cancer would be the use of adoptive cell therapy (ACT) either with tumor-infiltrating lymphocytes (TILs), NK cells,49-51 or engineered T cells.52 Natural TILs are typically anergic in vivo by the expression of immunosuppressive molecules, such as PD-1, LAG-3 or CTLA-4.

Removing T cells from the tumor immunosuppressive environment enables their activation and expansion.53,54 The reinfusion of these cells after ex vivo expansion can trigger the eradication of the tumor.55,56 The emergence of neo-antigens is an important factor contributing to the 427

P. Rodriguez-Otero et al.

efficacy of TILs, which explains why this approach has mainly been used in solid tumors (e.g., melanoma) rather than in hematological malignancies.57,58 Clinical experience with TILs in MM is scanty, however, the work from Borrello et al. with marrow-infiltrating lymphocytes (MILs) is encouraging, with twenty-three patients treated with MILs in the setting of ASCT with evidence of antimyeloma immunity, effective trafficking of the MILs to the BM, persistence over time, and correlation between clinical response and myeloma-specific immunity,55 demonstrating the feasibility of, and interest in, the approach.

Progress in gene engineering technologies has simplified the generation of specific antitumor T cells, overcoming many of the practical barriers that have limited wide dissemination of ACT using TIL cells.59,60 Theoretically, gene engineering may well be capable of targeting virtually any cancer histology. Genetically redirecting a T-cell’s specificity toward a patient’s cancer cell can be accomplished in two ways. In one approach a cloned T-cell receptor (TCR) conferring tumor recognition is inserted into circulating lymphocytes. Similarly to the endogenous TCR, genetically inserted TCRs recognized tumor antigens within the groove of a specific MHC molecule. In a second approach, an alternative way to provide specificity to transduced T cells and overcome some of the limitations of TCR engineered T cells, is with the use of chimeric antigen receptors (CARs).52,61 CARs are engineered fusion proteins that contain an extracellular antigen-binding domain composed of a ScFv derived from an Ab, that confers recognition to a tumor-associated antigen, linked in tandem to intracellular signaling motifs capable of T-cell activation, such as CD3z, or costimulatory molecules, like CD28 or CD137.62 By means of retroviral or lentiviral transduction, or by electroporation transfer, patient’s T cells express the CAR. Both CAR and TCR T cells have some advantages and limitations. CAR T cells are not restricted by the human leukocyte antigen of the patient, but selecting the appropriate antigen is critical to prevent off-target toxicity.

Many potential targets used for CAR T-cell immunotherapy have a broad expression across normal cells and tissues, therefore requiring careful evaluation.63 Another advantage of CAR T-cell therapy is the possibility to insert other genes encoding molecules involved in costimulation, survival, proliferation or inflammation, allowing the T cell to avoid inhibitory mechanisms displayed by the tumor.64,65 Differently from CAR T cells, TCR-engineered T cells recognize antigens presented by specific HLA molecules.

Accordingly, only a limited number of individuals presenting such HLA molecules are eligible for this treatment option.66 However, TCR, but not CAR, T cells can recognize intracellular proteins, providing a broader array of potential therapeutic targets.52 There are other considerations concerning both strategies that need to be stressed. Potent antitumor effect without off-target damage can occur if the target is only expressed in the tumor cells, such as NY-ESO TCR T cells.67 However, if the T cells are modified with a receptor that recognizes the antigen both in malignant and nonmalignant cells, like anti-CD19 CAR T cells, normal cells will be equally attacked resulting in the off-target effects seen with these therapies that are actually limiting its availability.68 Moreover, there is risk of a massive release of pro-inflammatory cytokines produced by hyperactive 428

CAR T cells, resulting in cytokine release syndrome (CRS), characterized by fever, hypotension or renal failure.69 CAR T cells with CD19 specificity have been used for the treatment of lymphoid malignancies, with impressive clinical results mainly in refractory B-cell acute lymphocytic leukemia (ALL) patients.70,71 The same anti-CD19 CAR T-cell model has been used in one refractory myeloma patient with intriguing results (CR after ASCT and CAR T-cell infusion with a response duration of longer than 12 months), despite the fact that clonal PCs did not express CD19; 99.95% were negative for this antigen.72 Other targets, such as BCMA, a surface antigen expressed on normal and malignant PCs, have also been used for CAR73,74 development as well as for BiTEs in RRMM patients. In the first-in-human trial using CAR T cells with BCMA specificity, twelve patients were treated, with four out of twelve responders, including one patient achieving a sCR (two PR, one VGPR, one sCR).75 Of note, the patient that achieved sCR experienced a cytokine release syndrome including fever, tachycardia, hypotension, elevated liver enzymes, and elevated creatine kinase, all of which resolved in two weeks or less.

TCR-engineered T cells with NY-ESO-1 and LAGE-1 specificity have also been tested in MM patients. In a phase I/II trial, twenty patients with active MM (six patients at diagnosis and fourteen at relapse) were treated with NY-ESO-1- and LAGE-1-specific TCR-engineered T cells. Cells were infused on day +2 after ASCT conditioned with high-dose melphalan and followed by lenalidomide maintenance. Fourteen out of twenty patients achieved at least near complete response after the planned treatment, although it is difficult to dissect the specific contribution of the TCR-engineered T cells from ASCT or maintenance treatment. Interestingly, affinityenhanced T cells showed extended persistence (more than two years in some patients), and MM progression correlated either with a loss of persisting TCR T cells or the appearance of a negative subclone.

Improving antigen-specific immunity using vaccines Although surface antigens can serve as targets for Abbased therapeutics, most cancer-associated or cancer-specific antigens are derived from intracellular proteins. Thus, another strategy to enhance anticancer immunity would be the use of vaccines to improve the immune response against cancer. There are different vaccination strategies.76 It should be noted that the choice of the antigen, vaccine formulation, delivery system, adjuvant, immunomodulation, treatment schedule and treatment setting can all modify the quality and strength of the T-cell response vaccines are expected to induce.76 Hematological malignancies are an opportunity for vaccination development given the relatively high availability of cellular antigens, the possibility of using whole tumor cell lysates to charge APCs as well as the use of a post-transplant setting as a window of opportunity for vaccination, based on the “resetting” of immune relations seen after this procedure.

Several groups have investigated the value of cancer vaccines in MM.77,78 Two separate vaccination approaches have been developed, one using peptide-based and the other using dendritic cell fusion vaccines. Several trials are ongoing which are evaluating the efficacy of peptide vaccination using different antigens, alone or in combination, such as NY-ESO-1, MAGE-AE, WT-1, and XBP-1, which are broadly expressed in MM cells from selected haematologica | 2017; 102(3)

Immunotherapy strategies in Multiple Myeloma

patients.79-83 Globally, vaccination is able to induce immune response, but so far its clinical efficacy remains modest.

On the other hand, dendritic cell fusion vaccines exploit the ability of dendritic cells to present several tumor antigens to the host immune effector cells. This strategy has also been evaluated in phase I-II trials in two single institutions. In the first phase I trial, patients with active MM with a median of four prior lines of treatment were treated with the DC/MM fusion vaccine. Eleven out of sixteen evaluable patients achieved stabilization of the disease after vaccination. Vaccination was well tolerated, and was capable of inducing immune responses against myeloma cells.84 A second phase II trial using the same DC/MM cell fusion vaccine approach in the context of an ASCT, showed a CR/VGPR rate of 78% early after ASCT, and 24% of patients improved their response from PR to CR/near complete remission (nCR) after vaccination at more than three months after ASCT, thus suggesting a vaccine-mediated effect on residual disease.85 A phase III trial is ongoing to confirm these results (CTN 1401), also in the post-ASCT setting. An innovative vaccination approach has recently been reported using intravenously administered ribonucleic acid (RNA)-lipoplexes (RNALPX) to enhance DCs. RNA-LPX encoding viral or mutant neo-antigens or endogenous self-antigens induce strong effector and memory T-cell responses, and mediate potent IFNÎą-dependent rejection of progressive tumors. A phase I dose-escalation trial is ongoing to further evaluate this strategy.86

Overcoming inhibitory immune suppression with checkpoint inhibitors The limited clinical benefit of vaccines could be related, at least in part, to the inhibition of tumor-specific effector cells through the expression of checkpoint receptors. The activation of T cells includes a two-step process: 1) interaction between the TCR and the antigen presented by MHC molecules, and 2) costimulatory signals to enhance T-cell activation. In the absence of this signal, T cells fail to respond and become inactivated.

Under normal physiological conditions, immune checkpoints are essential for the maintenance of self-tolerance and to protect tissues from the potential damage of an exacerbated T-cell response. There are two types of checkpoint receptors: inhibitory receptors, like CTLA-4, PD-1, TIM-3, or LAG-3, and stimulatory receptors, such as CD28, CD137, or OX40, among others.9,87 T-cell responses can be modulated either using agonist antibodies directed to stimulatory receptors that amplify the immune response, or using antibodies to block inhibitory receptors and release the brakes of immune cells.9 Different checkpoint drugs are under development, targeting both activating receptors and also inhibitory receptors. The latter are more advanced in their clinical development, and can be divided into three groups according to their respective targets; CTLA-4, PD-1, and PD-L1 inhibitors.88 The first checkpoint receptor explored for treatment intervention was the cytotoxic T-Lymphocyte antigen 4 (CTLA-4), and ipilimumab was the first drug administered to target this checkpoint. CTLA-4 is an inhibitory receptor that regulates T cells during the initial activation steps of the immune response; thus it is expressed in the surface of activated and regulatory T cells. When CTLA-4 binds to its ligands (CD80 and CD86), the result is T-cell inhibition interfering with interhaematologica | 2017; 102(3)

leukin-2 (IL-2) secretion and interleukin-2 receptor (IL-2R) expression.9,88,89 The second group of checkpoint inhibitors are the programmed death receptor-1 (PD-1) inhibitors, nivolumab, pembrolizumab, or pidilizumab. PD-1 is, like CTLA-4, an inhibitory receptor that is expressed on the surface of activated T cells to limit their activity at later stages of immune responses. PD-L1 and PD-L2 are the two ligands of PD-1 and they are expressed on the surface of APC and tumor cells. The binding of PD-1 to its ligand induces the inhibition of T cells.9,88 Cancer cells upregulate PD-L1 to take advantage of the PD-1 pathway and create an immunosuppressive milieu. Such upregulation of PD-L1 expression has been described in different cancer types, such as melanoma, non-small cell lung cancer, and also in MM, and this expression has been typically linked to poor clinical outcomes.90-92 Furthermore, TILs have also been shown to express significantly higher levels of PD-1, which could be induced by increased levels of proinflammatory cytokines, such as IFNÎł in the tumor microenvironment. Altogether, cancer cells evade immune responses through higher PD-L1 expression on tumors along with higher PD-1 expression on TILs.54,93-96 In MM, PD-L1 is expressed in clonal PCs across all disease stages, but this expression is significantly higher at relapse and in MRD positive patients. Similarly, expression of PD-1 on T cells was also increased in the BM of patients at relapse and with MRD.16 The first results of PD-1 blockade in hematological malignancies were obtained in a phase I trial with nivolumab.97 ORR reported in diffuse large B-cell lymphoma and follicular lymphoma were 36% and 40%, respectively. In Hodgkin lymphoma, ORR reached 87% in a heavily pretreated and refractory population with 17% CRs. PD-1 blockade in MM patients alone has not induced objective responses, and only 67% of patients with stabilization of the disease were noted.97 Experimental data has shown that lenalidomide reduces PD-L1 and PD-1 expression on MM cells, T cells, and MDSCs, respectively.

Moreover, a clear synergism between lenalidomide and anti-PD-1 and anti-PD-L1 was also observed.98 Therefore, based on this preclinical data showing a synergistic effect between PD-1 blockade and IMiDs, combination strategies with PD-L1 blockade and lenalidomide or pomalidomide are under evaluation. Data on two phase I trials combining pembrolizumab and IMiDs has been recently reported. The phase I KEYNOTE-023 trial is evaluating pembrolizumab in combination with Rd for RRMM patients that received more than two prior lines of therapy, including both IMiDs and a PI. A total of sixty-two patients were included with a median of four prior lines of therapy, and 76% of the patients were refractory to lenalidomide, with 50% being double, triple or quadruple refractory. Efficacy data, recently updated at the 2016 American Society of Clinical Oncology (ASCO) meeting, showed a 50% rate of ORR in the overall population (n=40) and 36% in lenalidomide-refractory patients.

Overall the combination was well tolerated, and the adverse effects were consistent with those observed for pembrolizumab and lenalidomide in their respectively approved indications. It should be noted that immunerelated adverse events, such as pneumonitis, hepatitis or colitis, which are typically described with these types of therapies, have not been observed so far in this trial; nevertheless longer follow up is still needed in order to clearly 429

P. Rodriguez-Otero et al.

assess the frequency and severity of these adverse events.99 Another currently ongoing phase I/II trial is evaluating the efficacy and safety of pembrolizumab in combination with pomalidomide and low-dose dexamethasone in a similar patient population, that of RRMM having received more than two prior lines of therapy, including an IMiD and a PI. A total of thirty-eight patients have been included, with 89% of them being refractory to lenalidomide and 70% double refractory. ORR in the total population was 66% and 65% in lenalidomide-refractory patients, and a median PFS of 14 months was reported at last follow up. The safety profile was acceptable, with 38% of patients suffering immune-related adverse events (IRAEs), and 14% of patients experiencing pneumonitis.100 Overall, both combinations are well tolerated and show promising preliminary efficacy in the heavily pretreated RRMM population, but further studies with a larger series of patients and longer follow up are needed to confirm these results.

Future perspectives The clinical success of checkpoint inhibition, particularly in solid tumors, has reignited the interest in immunotherapy against cancer, and this field is now moving forward very rapidly. Nevertheless, there are still many open questions.

1. It is important to define the target populations that will benefit most from specific immunotherapeutic strategies. So far these therapies have been mainly explored in the relapse setting, but a higher efficacy, specially for checkpoint inhibitors, would probably be expected at stages when a better preserved immune system exists. For this reason combinatorial strategies, including immunotherapeutic agents, should be evaluated in other patient populations, such as in newly diagnosed patients, high-risk SMM and high-risk myeloma patients in early relapse after transplantation, or after consolidation treatment in patients who didn’t achieve CR or VGPR. Even more interesting would be to test their efficacy at the time

References 1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011 Mar 4;144(5):646-674.

2. Howlader N NA, Krapcho M, Miller D, Bishop K, Altekruse SF, Kosary CL, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA (eds). SEER Cancer Statistics Review, 1975-2013, National Cancer Institute. Bethesda, MD; November 2015. [updated April 2016].

Available from: http:// seercancergov/csr/1975_2013/.

Last accessed: 21st September 2016.

3. 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.

4. Chen DS, Mellman I. Oncology meets immunology: the cancer-immunity cycle.

Immunity. 2013;39(1):1-10.

430

of MRD persistence, maintenance treatment or biochemical relapse in order to improve immune surveillance against residual myeloma cells. Accordingly, specific clinical trials would be welcome in these patient cohorts.

2. We need new biomarkers to predict response to certain treatments, such as the expression of PD-1 or PD-L1, the mutational load or microsatellite instability in the case of checkpoint inhibitors.

3. There is also a need for accurate immune profiling at baseline, to try to identify ideal candidates for therapy, and immune monitoring to identify those patients who would benefit the most.

4. Do we need novel immune-related response criteria or new clinical endpoints, such as TNT? The answer is probably “yes”, since some patients may not achieve a CR, or can experience an indolent relapse that may remain under the control of the immune system for longer periods than that currently observed upon using approved antimyeloma agents.

6. Can we be successful with one immune approach or do we need combination therapies? Immune therapeutic strategies targeting only one pathway are often ineffective or short-lived, and scientific rationale supports the hypothesis that the combination of two or three approaches may enhance the clinical activity of immunotherapy strategies. Then, the challenge will be how to combine or sequence multiple drugs.

Combinations will probably need to be based upon the specificity of their mechanism of action and disease stage.

7. Eventually, we have to elucidate how much can we rely on pre-clinical data to guide the design of clinical trials, and how can we possibly improve on this most effectively, with multicenter and multinational activities.

Although there are numerous open questions to address and solve for immunotherapy approaches in MM, this is a fascinating time for myeloma therapy, and current data (supporting the old “immune” experience with donor lymphocyte infusions) already indicate that immunotherapy will be a backbone that might revolutionize treatment, and hopefully improve patient outcomes further.

5. Gigante M, Gesualdo L, Ranieri E. TGFbeta: a master switch in tumor immunity.

Curr Pharm Des. 2012;18(27):4126-4134.

6. Thomas DA, Massague J. TGF-beta directly targets cytotoxic T cell functions during tumor evasion of immune surveillance.

Cancer Cell. 2005;8(5):369-380.

7. Racanelli V, Leone P, Frassanito MA, et al.

Alterations in the antigen processing-presenting machinery of transformed plasma cells are associated with reduced recognition by CD8+ T cells and characterize the progression of MGUS to multiple myeloma. Blood. 2010;115(6):1185-1193.

8. Butte MJ, Keir ME, Phamduy TB, Sharpe AH, Freeman GJ. Programmed death-1 ligand 1 interacts specifically with the B7-1 costimulatory molecule to inhibit T cell responses. Immunity. 2007;27(1):111-122.

9. Mellman I, Coukos G, Dranoff G. Cancer immunotherapy comes of age. Nature.

2011;480(7378):480-489.

10. Lesokhin AM, Hohl TM, Kitano S, et al.

Monocytic CCR2(+) myeloid-derived suppressor cells promote immune escape by limiting activated CD8 T-cell infiltration into the tumor microenvironment. Cancer Res. 2012;72(4):876-886.

11. Rawstron AC, Davies FE, Owen RG, et al.

B-lymphocyte suppression in multiple myeloma is a reversible phenomenon specific to normal B-cell progenitors and plasma cell precursors. Br J Haematol. 1998; 100(1):176-183.

12. Koike M, Sekigawa I, Okada M, et al.

Relationship between CD4(+)/CD8(+) T cell ratio and T cell activation in multiple myeloma: reference to IL-16. Leuk Res. 2002;26(8):705-711.

13. Ogawara H, Handa H, Yamazaki T, et al.

High Th1/Th2 ratio in patients with multiple myeloma. Leuk Res. 2005;29(2):135140.

14. Sze DM, Giesajtis G, Brown RD, et al.

haematologica | 2017; 102(3)

Immunotherapy strategies in Multiple Myeloma

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Clonal cytotoxic T cells are expanded in myeloma and reside in the CD8(+)CD57(+)CD28(-) compartment.

Blood. 2001;98(9):2817-2827.

Dosani T, Carlsten M, Maric I, Landgren O. The cellular immune system in myelomagenesis: NK cells and T cells in the development of myeloma [corrected] and their uses in immunotherapies. Blood Cancer J. 2015;5:e306.

Paiva B, Azpilikueta A, Puig N, et al. PDL1/PD-1 presence in the tumor microenvironment and activity of PD-1 blockade in multiple myeloma. Leukemia. 2015;29(10): 2110-2113.

Brown RD, Pope B, Yuen E, Gibson J, Joshua DE. The expression of T cell related costimulatory molecules in multiple myeloma. Leuk Lymphoma. 1998;31(3-4):379384.

Malek E, de Lima M, Letterio JJ, et al.

Myeloid-derived suppressor cells: The green light for myeloma immune escape.

Blood Rev. 2016;30(5):341-348.

Giallongo C, Tibullo D, Parrinello NL, et al.

Granulocyte-like myeloid derived suppressor cells (G-MDSC) are increased in multiple myeloma and are driven by dysfunctional mesenchymal stem cells (MSC).

Oncotarget. 2016 Mar 7.

Franssen LE, van de Donk NW, Emmelot ME, et al. The impact of circulating suppressor cells in multiple myeloma patients on clinical outcome of DLIs. Bone Marrow Transplant. 2015;50(6):822-828.

Paiva B, Mateos MV, Sanchez-Abarca LI, et al. Immune status of high-risk smoldering multiple myeloma patients and its therapeutic modulation under LenDex: a longitudinal analysis. Blood. 2016;127(9):11511162.

Ladetto M, Ferrero S, Drandi D, et al.

Prospective molecular monitoring of minimal residual disease after non-myeloablative allografting in newly diagnosed multiple myeloma. Leukemia. 2016;30(5):12111214.

Paiva B, Cedena MT, Puig N, et al. Minimal residual disease monitoring and immune profiling in multiple myeloma in elderly patients. Blood. 2016;127(25):3165-3174.

Pessoa de Magalhaes RJ, Vidriales MB, Paiva B, et al. Analysis of the immune system of multiple myeloma patients achieving long-term disease control by multidimensional flow cytometry. Haematologica.

2013;98(1):79-86.

Tura S, Cavo M. Allogeneic bone marrow transplantation in multiple myeloma.

Hematol Oncol Clin North Am. 1992;6(2): 425-435.

Gahrton G, Iacobelli S, Bjorkstrand B, et al.

Autologous/reduced-intensity allogeneic stem cell transplantation vs autologous transplantation in multiple myeloma: longterm results of the EBMT-NMAM2000 study. Blood. 2013;121(25):5055-5063.

Sahebi F, Iacobelli S, Biezen AV, et al.

Comparison of upfront tandem autologous-allogeneic transplantation versus reduced intensity allogeneic transplantation for multiple myeloma. Bone Marrow Transplant. 2015;50(6):802-807.

Auner HW, Szydlo R, van Biezen A, et al.

Reduced intensity-conditioned allogeneic stem cell transplantation for multiple myeloma relapsing or progressing after autologous transplantation: a study by the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant.

haematologica | 2017; 102(3)

2013;48(11):1395-1400.

29. Interferon as therapy for multiple myeloma: an individual patient data overview of 24 randomized trials and 4012 patients. Br J Haematol. 2001;113(4):1020-1034.

30. Minn AJ, Wherry EJ. Combination Cancer Therapies with Immune Checkpoint Blockade: Convergence on Interferon Signaling. Cell. 2016;165(2):272-275.

31. Shuptrine CW, Surana R, Weiner LM. Monoclonal antibodies for the treatment of cancer. Semin Cancer Biol. 2012;22(1):3-13.

32. Coiffier B, Lepage E, Briere J, et al. CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(4):235-242.

33. Hsi ED, Steinle R, Balasa B, et al. CS1, a potential new therapeutic antibody target for the treatment of multiple myeloma.

Clin Cancer Res. 2008;14(9):2775-2784.

34. Richardson PG, Jagannath S, Moreau P, et al. Elotuzumab in combination with lenalidomide and dexamethasone in patients with relapsed multiple myeloma: final phase 2 results from the randomised, open-label, phase 1b-2 dose-escalation study. Lancet Haematol. 2015;2(12):e516527.

35. Lonial S, Dimopoulos M, Palumbo A, et al.

Elotuzumab therapy for relapsed or refractory multiple myeloma. N Engl J Med. 2015;373(7):621-631.

36. Lokhorst HM, Plesner T, Laubach JP, et al.

Targeting CD38 with daratumumab monotherapy in multiple myeloma. N Engl J Med. 2015;373(13):1207-1219.

37. Lonial S, Weiss BM, Usmani SZ, et al.

Daratumumab monotherapy in patients with treatment-refractory multiple myeloma (SIRIUS): an open-label, randomised, phase 2 trial. Lancet. 2016; 387(10027): 1551-1560.

38. Usmani SZ, Weiss BM, Plesner T, et al.

Clinical efficacy of daratumumab monotherapy in patients with heavily pretreated relapsed or refractory multiple myeloma. Blood. 2016;128(1):37-44.

39. Plesner T, Arkenau HT, Gimsing P, et al.

Phase 1/2 study of daratumumab, lenalidomide, and dexamethasone for relapsed multiple myeloma. Blood. 2016.

40. Dimopoulos M, Nahi H, San Miguel J, et al.

An open-label, randomised phase 3 study of daratumumab, lenalidomide, and dexamethasone (DRd) versus lenalidomide and dexamethasone (Rd) in relapsed or refractory multiple myeloma (RRMM): POLLUX. 2016 EHA Meeting Abstract Book Hematologica 2016.

41. Palumbo A, Chanan-Khan A, Weisel K, et al. Daratumumab, bortezomib, and dexamethasone for multiple myeloma. N Engl J Med. 2016;375(8):754-766.

42. Chari A, Lonial S, Suvannasankha A, et al.

Open-label, multicenter, phase 1b study of daratumumab in combination with pomalidomide and dexamethasone in patients with at least 2 lines of prior therapy and relapsed or relapsed and refractory multiple myeloma. Blood. 2015;126(23):508-508.

43. Martin T, Richter J, Vij R, et al. A dose finding phase II trial of isatuximab (SAR650984, anti-CD38 mAb) as a single agent in relapsed/refractory multiple myeloma. Blood. 2015;126(23):509-509.

44. Martin TG, Baz R, Benson DM, et al. A A phase Ib dose escalation trial of SAR650984 (anti-CD-38 mAb) in combination with lenalidomide and dexamethasone in

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

relapsed/refractory multiple myeloma.

Blood. 2014;124(21):83-83.

Klinger M, Benjamin J, Kischel R, Stienen S, Zugmaier G. Harnessing T cells to fight cancer with BiTE(R) antibody constructs-past developments and future directions.

Immunol Rev. 2016;270(1):193-208.

Honemann D, Kufer P, Rimpler MM, et al.

A novel recombinant bispecific single-chain antibody, bscWue-1 x CD3, induces T-cellmediated cytotoxicity towards human multiple myeloma cells. Leukemia.

2004;18(3):636-644.

Seckinger A, Delgado JA, Moreno L, et al.

Target expression, preclinical activity and mechanism of action of EM801: a novel first-in-class Bcma T-cell bispecific antibody for the treatment of multiple myeloma. Blood. 2015;126(23):117-117.

Green DJ, Jones JC, Lin Y, et al. A novel bispecific CD38 antibody eradicates multiple myeloma in a mouse model following Yttrium-90-DOTA capture. Blood. 2015; 126(23):118-118.

Garg TK, Szmania SM, Khan JA, et al.

Highly activated and expanded natural killer cells for multiple myeloma immunotherapy. Haematologica. 2012; 97(9):1348-1356.

Shi J, Tricot G, Szmania S, et al. Infusion of haplo-identical killer immunoglobulin-like receptor ligand mismatched NK cells for relapsed myeloma in the setting of autologous stem cell transplantation. Br J Haematol. 2008;143(5):641-653.

Song W, van der Vliet HJ, Tai YT, et al.

Generation of antitumor invariant natural killer T cell lines in multiple myeloma and promotion of their functions via lenalidomide: a strategy for immunotherapy. Clin Cancer Res. 2008;14(21):6955-6962.

Restifo NP, Dudley ME, Rosenberg SA. Adoptive immunotherapy for cancer: harnessing the T cell response. Nat Rev Immunol. 2012;12(4):269-281.

Baitsch L, Baumgaertner P, Devevre E, et al.

Exhaustion of tumor-specific CD8(+) T cells in metastases from melanoma patients. J Clin Invest. 2011;121(6):23502360.

Ahmadzadeh M, Johnson LA, Heemskerk B, et al. Tumor antigen-specific CD8 T cells infiltrating the tumor express high levels of PD-1 and are functionally impaired. Blood.

2009;114(8):1537-1544.

Noonan KA, Huff CA, Davis J, et al.

Adoptive transfer of activated marrowinfiltrating lymphocytes induces measurable antitumor immunity in the bone marrow in multiple myeloma. Sci Transl Med. 2015;7(288):288ra278.

Besser MJ, Shapira-Frommer R, Itzhaki O, et al. Adoptive transfer of tumor-infiltrating lymphocytes in patients with metastatic melanoma: intent-to-treat analysis and efficacy after failure to prior immunotherapies.

Clin Cancer Res. 2013;19(17):4792-4800.

Stronen E, Toebes M, Kelderman S, et al.

Targeting of cancer neoantigens with donor-derived T cell receptor repertoires.

Science. 2016;352(6291):1337-1341.

Verdegaal EM, de Miranda NF, Visser M, et al. Neoantigen landscape dynamics during human melanoma-T cell interactions.

Nature. 2016.

Morgan RA, Dudley ME, Wunderlich JR, et al. Cancer regression in patients after transfer of genetically engineered lymphocytes.

Science. 2006;314(5796):126-129.

Rosenberg SA, Restifo NP. Adoptive cell

431

P. Rodriguez-Otero et al.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

432

transfer as personalized immunotherapy for human cancer. Science. 2015; 348(6230): 62-68.

Klebanoff CA, Rosenberg SA, Restifo NP. Prospects for gene-engineered T cell immunotherapy for solid cancers. Nat Med. 2016;22(1):26-36.

Eshhar Z, Waks T, Gross G, Schindler DG. Specific activation and targeting of cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding domains and the gamma or zeta subunits of the immunoglobulin and T-cell receptors. Proc Natl Acad Sci USA. 1993;90(2): 720-724.

San Miguel JF, Paiva B, Lasarte JJ. Engineering anti-myeloma responses using affinity-enhanced TCR-engineered T cells.

Cancer Cell. 2015;28(3):281-283.

Cherkassky L, Morello A, Villena-Vargas J, et al. Human CAR T cells with cell-intrinsic PD-1 checkpoint blockade resist tumormediated inhibition. J Clin Invest. 2016; 126(8):3130-3144.

Long AH, Haso WM, Shern JF, et al. 4-1BB costimulation ameliorates T cell exhaustion induced by tonic signaling of chimeric antigen receptors. Nat Med. 2015;21(6):581590.

Kunert A, Straetemans T, Govers C, et al.

TCR-engineered T cells meet new challenges to treat solid tumors: choice of antigen, T cell fitness, and sensitization of tumor milieu. Front Immunol. 2013;4:363.

Robbins PF, Kassim SH, Tran TL, et al. A pilot trial using lymphocytes genetically engineered with an NY-ESO-1-reactive Tcell receptor: long-term follow-up and correlates with response. Clin Cancer Res. 2015;21(5):1019-1027.

Kochenderfer JN, Wilson WH, Janik JE, et al. Eradication of B-lineage cells and regression of lymphoma in a patient treated with autologous T cells genetically engineered to recognize CD19. Blood. 2010;116(20): 4099-4102.

Kochenderfer JN, Dudley ME, Feldman SA, et al. B-cell depletion and remissions of malignancy along with cytokine-associated toxicity in a clinical trial of anti-CD19 chimeric-antigen-receptor-transduced T cells. Blood. 2012;119(12):2709-2720.

Kochenderfer JN, Dudley ME, Kassim SH, et al. Chemotherapy-refractory diffuse large B-cell lymphoma and indolent B-cell malignancies can be effectively treated with autologous T cells expressing an antiCD19 chimeric antigen receptor. J Clin Oncol. 2015;33(6):540-549.

Kochenderfer JN, Rosenberg SA. Treating B-cell cancer with T cells expressing antiCD19 chimeric antigen receptors. Nat Rev Clin Oncol. 2013;10(5):267-276.

Garfall AL, Maus MV, Hwang WT, et al.

Chimeric antigen receptor T cells against CD19 for multiple myeloma. N Engl J Med. 2015;373(11):1040-1047.

Danhof S, Gogishvili T, Koch S, et al. CARengineered T cells specific for the elotuzumab target SLAMF7 eliminate primary myeloma cells and confer selective fratricide of SLAMF7+ normal lymphocyte subsets. Blood. 2015;126(23):115-115.

74. Galetto R, Chion-Sotinel I, Gouble A, Smith J. Bypassing the constraint for chimeric antigen receptor (CAR) development in T-cells expressing the targeted antigen: improvement of anti-CS1 CAR activity in allogenic TCRa/CS1 double knockout T-cells for the treatment of multiple myeloma (MM). Blood. 2015;126(23):116-116.

75. Ali SA, Shi V, Maric I, et al. T cells expressing an anti-B-cell-maturation-antigen chimeric antigen receptor cause remissions of multiple myeloma. Blood. 2016 Jul 13.

76. Melief CJ, van Hall T, Arens R, Ossendorp F, van der Burg SH. Therapeutic cancer vaccines. J Clin Invest. 2015;125(9):3401-3412.

77. Karp Leaf R, Cho HJ, Avigan D. Immunotherapy for multiple myeloma, past, present, and future: monoclonal antibodies, vaccines, and cellular therapies.

Curr Hematol Malig Rep. 2015;10(4):395404.

78. Rosenblatt J, Bar-Natan M, Munshi NC, Avigan DE. Immunotherapy for multiple myeloma. Expert Rev Hematol. 2014; 7(1):91-96.

79. Bae J, Song W, Smith R, et al. A novel immunogenic CS1-specific peptide inducing antigen-specific cytotoxic T lymphocytes targeting multiple myeloma. Br J Haematol. 2012;157(6):687-701.

80. McCann KJ, Godeseth R, Chudley L, et al.

Idiotypic DNA vaccination for the treatment of multiple myeloma: safety and immunogenicity in a phase I clinical study.

Cancer Immunol Immunother. 2015; 64(8):1021-1032.

81. Rapoport AP, Aqui NA, Stadtmauer EA, et al. Combination immunotherapy after ASCT for multiple myeloma using MAGEA3/Poly-ICLC immunizations followed by adoptive transfer of vaccine-primed and costimulated autologous T cells. Clin Cancer Res. 2014;20(5):1355-1365.

82. Szmania S, Gnjatic S, Tricot G, et al.

Immunization with a recombinant MAGEA3 protein after high-dose therapy for myeloma. J Immunother. 2007;30(8):847854.

83. Tsuboi A, Oka Y, Nakajima H, et al. Wilms tumor gene WT1 peptide-based immunotherapy induced a minimal response in a patient with advanced therapy-resistant multiple myeloma. Int J Hematol. 2007;86(5):414-417.

84. Rosenblatt J, Vasir B, Uhl L, et al.

Vaccination with dendritic cell/tumor fusion cells results in cellular and humoral antitumor immune responses in patients with multiple myeloma. Blood. 2011; 117(2):393-402.

85. Rosenblatt J, Avivi I, Vasir B, et al.

Vaccination with dendritic cell/tumor fusions following autologous stem cell transplant induces immunologic and clinical responses in multiple myeloma patients.

Clin Cancer Res. 2013;19(13):3640-3648.

86. Kranz LM, Diken M, Haas H, et al.

Systemic RNA delivery to dendritic cells exploits antiviral defence for cancer immunotherapy. Nature. 2016; 534(7607): 396-401.

87. Melero I, Berman DM, Aznar MA, et al.

Evolving synergistic combinations of tar-

geted immunotherapies to combat cancer.

Nat Rev Cancer. 2015;15(8):457-472.

88. Sharma P, Allison JP. The future of immune checkpoint therapy. Science. 2015; 348(6230):56-61.

89. Buchbinder E, Hodi FS. Cytotoxic T lymphocyte antigen-4 and immune checkpoint blockade. J Clin Invest. 2015;125(9):33773383.

90. Mu CY, Huang JA, Chen Y, Chen C, Zhang XG. High expression of PD-L1 in lung cancer may contribute to poor prognosis and tumor cells immune escape through suppressing tumor infiltrating dendritic cells maturation. Med Oncol. 2011;28(3):682688.

91. Ghebeh H, Mohammed S, Al-Omair A, et al. The B7-H1 (PD-L1) T lymphocyteinhibitory molecule is expressed in breast cancer patients with infiltrating ductal carcinoma: correlation with important highrisk prognostic factors. Neoplasia. 2006; 8(3):190-198.

92. Konishi J, Yamazaki K, Azuma M, et al. B7H1 expression on non-small cell lung cancer cells and its relationship with tumor-infiltrating lymphocytes and their PD-1 expression. Clin Cancer Res. 2004;10(15):50945100.

93. Fourcade J, Sun Z, Benallaoua M, et al.

Upregulation of Tim-3 and PD-1 expression is associated with tumor antigen-specific CD8+ T cell dysfunction in melanoma patients. J Exp Med. 2010;207(10):2175-2186.

94. Wang W, Lau R, Yu D, et al. PD1 blockade reverses the suppression of melanoma antigen-specific CTL by CD4+ CD25(Hi) regulatory T cells. Int Immunol. 2009; 21(9):1065-1077.

95. Akbay EA, Koyama S, Carretero J, et al.

Activation of the PD-1 pathway contributes to immune escape in EGFR-driven lung tumors. Cancer Discov. 2013; 3(12):1355-1363.

96. Zhang Y, Huang S, Gong D, Qin Y, Shen Q. Programmed death-1 upregulation is correlated with dysfunction of tumor-infiltrating CD8+ T lymphocytes in human non-small cell lung cancer. Cell Mol Immunol. 2010; 7(5):389-395.

97. Lesokhin AM, Ansell SM, Armand P, et al.

Nivolumab in patients with relapsed or refractory hematologic malignancy: preliminary results of a phase Ib study. J Clin Oncol. 2016;34(23):2698-2704.

98. GĂśrgĂźn G, Samur MK, Cowens KB, et al.

Lenalidomide enhances immune checkpoint blockade-induced immune response in multiple myeloma. Clin Cancer Res. 2015;21(20):4607-4618.

99. Mateos M-V, Orlowski RZ, Siegel DSD, et al. Pembrolizumab in combination with lenalidomide and low-dose dexamethasone for relapsed/refractory multiple myeloma (RRMM): Final efficacy and safety analysis.

ASCO Meeting Abstracts. 2016;34 (15_suppl):8010.

100.Badros AZ, Kocoglu MH, Ma N, et al. A phase II study of anti PD-1 antibody pembrolizumab, pomalidomide and dexamethasone in patients with relapsed/refractory multiple myeloma (RRMM). Blood.

2015;126(23):506-506.

haematologica | 2017; 102(3)

GUIDELINE ARTICLE

ECIL-6 guidelines for the treatment of invasive candidiasis, aspergillosis and mucormycosis in leukemia and hematopoietic stem cell transplant patients

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Frederic Tissot,1 Samir Agrawal,2 Livio Pagano,3 Georgios Petrikkos,4 Andreas H. Groll,5 Anna Skiada,6 Cornelia Lass-Flörl,7 Thierry Calandra, 1 Claudio Viscoli 8 and Raoul Herbrecht9 Infectious Diseases Service, Centre Hospitalier Universitaire Vaudois and University of Lausanne, Switzerland; 2Division of Haemato-Oncology, St Bartholomew's Hospital and Blizard Institute, Queen Mary University, London, UK; 3Hematology, Catholic University of Sacred Heart, Roma, Italy; 4School of Medicine, European University Cyprus, Engomi, Cyprus; 5Infectious Disease Research Program, Center for Bone Marrow Transplantation and Department of Pediatric Hematology/Oncology, University Children’s Hospital, Münster, Germany; 61st Department of Medicine, University of Athens, Greece; 7Division of Hygiene and Medical Microbiology, Medical University of Innsbruck, Austria; 8 University of Genova (DISSAL), Infectious Disease Division, IRCCS San Martino-IST, Genova, Italy and 9Oncology and Hematology, Hôpitaux Universitaires de Strasbourg and Université de Strasbourg, France 1

Haematologica 2017 Volume 102(3):433-444

ABSTRACT

T

he European Conference on Infections in Leukemia (ECIL) provides recommendations for diagnostic strategies and prophylactic, pre-emptive or targeted therapy strategies for various types of infection in patients with hematologic malignancies or hematopoietic stem cell transplantation recipients. Meetings are held every two years since 2005 and evidence-based recommendations are elaborated after evaluation of the literature and discussion among specialists of nearly all European countries. In this manuscript, the ECIL group presents the 2015-update of the recommendations for the targeted treatment of invasive candidiasis, aspergillosis and mucormycosis. Current data now allow a very strong recommendation in favor of echinocandins for first-line therapy of candidemia irrespective of the underlying predisposing factors. Anidulafungin has been given the same grading as the other echinocandins for hemato-oncological patients. The beneficial role of catheter removal in candidemia is strengthened. Aspergillus guidelines now recommend the use of either voriconazole or isavuconazole for first-line treatment of invasive aspergillosis, while first-line combination antifungal therapy is not routinely recommended. As only few new data were published since the last ECIL guidelines, no major changes were made to mucormycosis recommendations.

Introduction The European Conference on Infections in Leukemia (ECIL) is the result of a collaboration between the European Organization for Research and Treatment of Cancer (EORTC), the European Society for Blood and Marrow Transplantation (EBMT), the European Leukemia Net (ELN), and the International Immunocompromised Host Society (ICSH). First recommendations for the treatment of Candida and Aspergillus infections in hematologic patients were published in 2007 after the first conference (ECIL-1) and have then been updated at ECIL-2 and ECIL-3.1,2 First recommendations for the diagnosis and treatment of mucormycosis have been published after ECIL-3.3 ECIL-4 updates for antifungal therapy were only available as slides on the websites of these participating societies without publication of a manuscript in consideration of the lack of substantial new data and the haematologica | 2017; 102(3)

Correspondence: frederic.tissot@chuv.ch

Received: August 5, 2016.

Accepted: December 20, 2016.

Pre-published: December 23, 2016. doi:10.3324/haematol.2016.152900 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/433 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

433

F. Tissot et al.

limited modifications compared to the latest publication.

With respect to the targeted treatment of fungal infections, the goals for ECIL-5 were to update the recommendations with analysis of the new data for invasive candidiasis, aspergillosis and mucormycosis in hematologic patients. The update was also necessary to change the prior 5-level grading (A to E) used during the ECILs 1 to 4 for the strength of recommendations for Candida and Aspergillus infections into the 3-level grading (A to C) already used during ECIL-3 for the first recommendation for mucormycosis (Table 1).1-3 The grading for quality of evidence has not been modified.

Methods The ECIL-5 meeting was held in September 2013 and involved 57 experts from 21 countries, including 3 non-European countries.

Slides of the conclusions of the ECIL-5 were made available on the websites of the EORTC, EBMT, ELN, and ICHS. The ECIL-6 meeting was held in September 2015 with the presence of 55 experts from 24 countries, including 4 non-European countries (see list of collaborators at the end of this Review).

At both the ECIL-5 and the ECIL-6 meetings, the antifungal therapy working group made a search for new publications regarding treatment of invasive candidiasis, aspergillosis and mucormycosis. The group was divided into three subgroups, each being responsible for one of each fungal infection type. The literature search was performed in Pubmed and Cochrane databases.

Abstracts presented at major congresses during the previous two years were also retrieved and integrated into the ECIL recommendation. All recommendations referring to an abstract, however, were classified as provisional until the publication of the final manuscript.

The working group presented its recommendations during the plenary session at the ECIL-5 meeting and then incorporated the suggestions coming from the assembly. In cases in which full consensus was not obtained, the decision was put to the vote, and the final decision was based on a majority of votes from the full ECIL5 assembly. The updated recommendations were presented on the next day during a second plenary session for final approval.

Recommendations were graded on the basis of the strength of recommendations (3-level scale: A, B, or C) and quality of evidence (3-level scale: I, II, or III), as detailed in Table 1.

The manuscript of the ECIL-5 was put on hold after a debate arose on differences between ECIL and European Society for Clinical Microbiology and Infectious Diseases (ESCMID) / European Confederation of Medical Mycology (ECMM) recommendations on guidelines for prophylaxis and treatment of invasive aspergillosis (draft presented at the ECCMID 2014).4 Two joint meetings were subsequently held (December 2014 and April 2015) to identify the differences and the exact reasons for these differences. The aim was not to modify the recommendations made by each of the two groups but rather to add explanations on the differences in the manuscript. For further clarification, a joint presentation was given at the ECIL-6 by members of the ECIL group and of the ESCMID/ECMM group. This resulted in a delay in publication of the ECIL-5 recommendations and during the ECIL-6 plenary session, the ECIL assembly approved a new search for publications or abstracts until September 2015 with inclusion of all relevant data on aspergillosis, candidiasis and mucormycosis for a full update of the guidelines. Final approval by the majority of the members of the group was obtained in Autumn 2015. The current manuscript includes updates from both the ECIL-5 and the ECIL-6 and is called â&#x20AC;&#x153;ECIL-6 guidelines for the treatment of inva434

sive candidiasis, aspergillosis and mucormycosis in leukemia and hematopoietic stem cell transplant patientsâ&#x20AC;?.

Invasive candidiasis Like previous ECIL recommendations, the current guidelines for invasive candidiasis cover the hematologic population as well as the general population of patients.

Although hematologic patients are the main focus of the recommendation, this distinction is maintained because available data from the original randomized controlled trials mainly include non-neutropenic patients. Chronic infections are not considered. Twenty-two major publications were identified (Tables 2 and 3).5-26 Fifteen reported primary results from clinical trials.5-11,13-17,19,20 One publication analyzed results of a subgroup of cancer patients from a previously published trial.12 One publication reported the analysis of pooled data from 2 trials previously published with a focus on patients with an underlying malignancy.21 All these studies were published before the ECIL-4. Since then, 5 studies have been identified, including one patientlevel quantitative review of 7 published trials on invasive candidiasis, one pooled patient-level data analysis from 5 prospective trials on anidulafungin, one systematic review of 17 randomized clinical trials focusing on invasive candidiasis in neutropenic patients, one prospective non-comparative trial evaluating a strategy of early oral switch from anidulafungin for invasive candidiasis, and one observational study comparing the initial use of echinocandin-based versus azole-based regimen for C. parapsilosis candidemia.22-26 These publications were the reasons for the change in guidelines. Characteristics of these studies and main results are shown in Tables 2 and 3.

The number of neutropenic patients included in each of these studies was low and limited the level of evidence of the recommendation for this group of patients. The review published by Andes et al. showed that, in the univariate analysis, neutropenia was one of the factors significantly and negatively associated both with clinical outcome and with survival.22 In the multivariate analysis, however, the effect of neutropenia disappeared, but there was a significant association of immunosuppressive therapy (including steroids) with lower survival. Other factors significantly associated with lower survival were the APACHE score, infection by C. tropicalis and age, while treatment with an echinocandin [Odds Ratio (OR) 0.65, 95%CI: 0.45-0.94; P=0.02] and catheter removal were both significantly associated with better survival (OR 0.50, 95%CI: 0.35-0.72; P=0.0001).

Based on the patient-level quantitative analysis by Andes et al., echinocandins must be considered as first-line choice for invasive Candida infections before species identification (Table 4).22 The strength of recommendation is the same (A) for anidulafungin, caspofungin and micafungin and is also the same for the overall and the hematologic population. However, the quality of evidence is lower for hematologic patients (II) compared to the overall population, as the number of neutropenic patients recruited in the clinical trials was low. A recent communication on a patient-level pooled analysis of one randomized clinical trial and 4 open label studies focusing on anidulafungin in 46 neutropenic patients with candidemia showed comparable response and survival rates to those observed with caspofungin and micafungin in other studies.25 Therefore, the grading is now similar (A II) for all three echinocandins haematologica | 2017; 102(3)

ECIL-6 guidelines: fungal infections in leukemia and HSCT patients Table 1. Evolution over time of the grading system used for treatment of invasive Candida and Aspergillus infections.

Strength of recommendations ECIL-5 and 6

Grade

ECIL-1 to 4

A

Strong evidence for efficacy and substantial clinical benefit: strongly recommended Strong or moderate evidence for efficacy, but only limited clinical benefit: generally recommended Insufficient evidence for efficacy; or efficacy does not outweigh possible adverse consequences (e.g. drug toxicity or interactions) or cost of chemoprophylaxis or alternative approaches: optional Moderate evidence against efficacy or for adverse outcome: generally not recommended Strong evidence against efficacy or for adverse outcome: never recommended

B C

D E

Good evidence to support a recommendation for use Moderate evidence to support a recommendation for use Poor evidence to support a recommendation for use

Omitted Omitted

Quality of evidence Grade I II III

ECIL-1 to 6 (no change) Evidence from â&#x2030;Ľ 1 properly randomized, controlled trial Evidence from â&#x2030;Ľ 1 well-designed clinical trial, without randomization; from cohort or case-controlled analytical studies (preferably from > 1 center); from multiple time-series; or from dramatic results from uncontrolled experiments Evidence from opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees

ECIL: European Conference on Infections in Leukemia.

for the treatment of invasive candidiasis in hematologic patients.

Liposomal amphotericin B has also been graded A I for the overall population and A II for hematologic patients due to similar efficacy in comparison to micafungin.15,21 However, its safety profile is less favorable and therefore liposomal amphotericin B should be considered as an alternative in case of contraindication to echinocandins.

Fluconazole and voriconazole are potential alternatives for first-line treatment in the overall population provided there is no previous exposure to azoles and the infection is not severe (fluconazole).

After species identification, susceptibility testing should guide the treatment. In general, echinocandins remain the drug of choice, except for C. parapsilosis where fluconazole is more appropriate (Table 5). However, a recent observational study reported no difference in 30-day mortality and persistent candidemia at 72 hours of an echinocandinbased regimen compared to an azole-based therapy for patients with C. parapsilosis candidemia.26 Therefore, the continuing use of echinocandins might be considered in patients with a clinical and microbiological response.

When Candida species is azole-susceptible, step-down to fluconazole can be considered in stable patients after five days of intravenous (iv) therapy.24 In patients with Candida krusei infection, switch to oral voriconazole is an option.

Although the role of catheter removal in the management of candidemia has long been controversial, most recent studies suggest a beneficial effect on outcome.6-8,10,11,15,16,20,26-33 Garnacho-Montero et al. showed in a large number of candidemia that early adequate therapy and removal of central venous line were independently associated with lower mortality.34 The patient-level quantitative analysis by Andes et al. also demonstrated in a haematologica | 2017; 102(3)

multivariate analysis that removal of catheter was associated with a decreased mortality (OR 0.50; 95%CI: 0.350.72; P=0.0001).22 The recommendation is, therefore, to rapidly remove the catheter in the overall population (grade A II) as well as in hematologic patients (grade B II) irrespective of the Candida species. If central venous catheter cannot be removed, treatment should include an echinocandin or a lipid formulation of amphotericin B due to their better activity on Candida biofilms.35-37

Invasive Aspergillus infections Nine prospective trials (only 4 being randomized comparative trials) had been published before the ECIL-4 and provided the basis of the previous guidelines for first-line therapy in invasive aspergillosis (Table 6).38-46 An additional paper reported a post-hoc analysis of the trial comparing standard dose of liposomal amphotericin B to high-dose liposomal amphotericin B.47 This post-hoc analysis comparing outcome in possible versus mycologically documented aspergillosis underscored the limited number of mycologically documented infections but did not lead to any change in the grading for liposomal amphotericin B. A second post-hoc analysis was performed on the voriconazole versus amphotericin B deoxycholate trial.48 Integration of the results of baseline galactomannan detection tests performed after primary analysis and re-categorization according to the 2008 EORTC/MSG definition criteria allowed more mycologically documented cases of invasive aspergillosis to be identified.49 Conclusions of this post-hoc analysis were similar to those of the primary analysis and therefore its results did not affect the grading for voriconazole and for amphotericin B deoxycholate.

At the time of the ECIL-5, results from the comparative study of voriconazole plus anidulafungin versus voricona435

F. Tissot et al.

Table 2. Trials for first-line therapy of invasive candidiasis: critical inclusion and exclusion criteria, treatment and relevant characteristics of the patients.

1st author, year, reference

Type of study and critical inclusion and exclusion criteria

Rex, 19945

RCT; candidemia; pts with neutropenia or hematologic Fluconazole cancer excluded (400 mg) d-AmB (0.5-0.6 mg/kg) Prospective observational; candidemia; d-AmB (mostly 0.5-0.7 mg/kg) any Candida species Fluconazole (50-800 mg)

103

33

22

0

103 227 67

32 107 32

24 NA NA

0 NA NA

RCT; candidemia and other acute invasive candidiasis Fluconazole (400 mg) including urinary tract infections; any Candida species d-AmB (25-50 mg; 0.67 mg/kg for neutropenic pts) Matched cohort study; candidemia; Fluconazole (200-600 mg) any Candida species; only cancer pts d-AmB (0.3-1.2 mg/kg) RCT; candidemia; C. krusei and Fluconazole (800 on day 1 then 400 mg) C. glabrata infections excluded d-AmB (0.6 mg/kg) RCT; candidemia or deep-seated infections; Caspofungin (70 on day 1 then 50 mg) any Candida species; neutropenic pts excluded d-AmB (0.6-1.0 mg/kg)

75 67

43 42

NA NA

16c 20c

45 45 50 53 109 115

45 45 10 12 30 38

NA NA 16 22 28 18

11b 11b 0 0 14 10

Rex, 200311

RCT; candidemia; C. krusei infections excluded; neutropenic pts excluded

107 112

20 21

29 26

0 0

DiNubile, 200512

Invasive candidiasis in cancer pts; subgroup analysis of #6; numbers of pts not consistent with primary manuscript RCT; candidemia; any Candida species; neutropenic pts excluded

Fluconazole (800 mg) Fluconazole (800 mg) + d-AmB (0.6-0.7 mg/kg) Caspofungin (70 on day 1 then 50 mg) d-AmB (0.6-1.0 mg/kg)

41 33

41 33

NA NA

14 10

Voriconazole (12 on day 1 then 6 mg/kg) d-AmB (0.7-1.0 mg/kg) then fluconazole (400 mg) Micafungin (<50->200 mg) Micafungin (>50->200 mg) Micafungin (>50->200 mg) + other agent Micafungin (100 mg) L-AmB (3 mg/kg) Micafungin (100 mg) Micafungin (150 mg) Caspofungin (70 on day 1 then 50 mg) Anidulafungin (200 on day 1 then 100 mg) Fluconazole (800 on day 1 then 400 mg)

248 122

NA NA

NA NA

0 0

72 25 29

NA NA NA

NA NA NA

10 10 9

264 267 191 199 188 127 118

85 90 68 56 52 28 27

111 111 NA NA NA 18 27

34 28 22 17 11 3 4

Nguyen, 19956

Anaissie, 19967

Anaissie, 19968 Phillips, 19979 Mora-Duarte, 200210

Kullberg, 200513

OstroskyZeichner, 200514

Kuse, 200715 Pappas, 200716

Kanji, 201323

436

N of ptsa with

N of ptsa Cancer

Prospective, non-comparative; monotherapy for de novo candidemia (n=72); monotherapy (n=25) or combination (n=29) for salvage therapy RCT; candidemia or deep-seated infections; any Candida species RCT; candidemia or deep-seated infections; any Candida species

Reboli, 200717 RCT; candidemia or deep-seated infections; 18 and Reboli, 2011 C. krusei infections excluded; second publication on factors associated with improved outcome in C. albicans infections Queiroz-Telles, RCT; candidemia or deep-seated infections; 200819 any Candida species; only pediatric pts Betts, 200920 RCT; candidemia or deep-seated infections; safety as primary objective; any Candida species Cornely 201121 Analysis of pooled data from #12 and 13 restricted to cancer pts Andes, 201222

Treatment (daily dose)

A pt-level quantitative review of #1, 6, 7, 8, 11, 12, 13; candidemia and deep-seated infections; any Candida species Systematic review of 17 RCT; focus on candidemia and deep-seated infections in neutropenic pts

Micafungin (2 mg/kg limited to 100 mg) 48 L-AmB (3 mg/kg) 50 Caspofungin (70 on day 1 then 50 mg) 104 Caspofungin (150 mg) 100 Micafungin (100 mg), micafungin (150 mg), 1067 caspofungin (70 on day 1 then 50 mg), L-AmB (3 mg/kg) Fluconazole, d-AmB, L-AmB, d-AmB 1915 + fluconazole, d-AmB then fluconazole, voriconazole, caspofungin anidulafungin, micafungin d-AmB, d-AmB + flucytosine, L-AmB, ABLC, 5675 ketoconazole, fluconazole, voriconazole, caspofungin, micafungin, anidulafungin

IS Neutropenia therapy

NA NA 27 33 359

NA NA 29 29 NA

6 13 7 8 114

410

440

139

NA

NA

342

haematologica | 2017; 102(3)

ECIL-6 guidelines: fungal infections in leukemia and HSCT patients Vasquez, 201424

Prospective, non-comparative, evaluating iv to oral step-down strategy; candidemia or deep-seated infections; any Candida species 25 Herbrecht, 2014 Pooled analysis of an RCT and 4 non-comparative open label studies; candidemia; focus on neutropenic pts treated with anidulafungin Fernandez-Ruis, Prospective non-interventional population 201526 -based study; C. parapsilosis candidemia.

Anidulafungin (200 on day 1 then 100 mg), possible switch to oral fluconazole (400 mg) or voriconazole (200 mg bid) after day 5 Anidulafungin (200 on day 1 then 100 mg)

250

NA

NA

9

46

NA

NA

46

Azole-based (42%), echinocandin-based (24.7%), amphotericin B-based (19%), combination therapy (14.4%). Dose not specified.

194d

61d

72d

7d

a Numbers of patients refer to the modified intent to treat population when available or to the intent to treat population; for this reason and due to some inconsistencies numbers may be different in primary manuscript and in pooled analysis. bNeutropenia defined by less than 1000/mL; cneutropenia defined by less than 500/mL; dnumber of episodes. pts: patients; IS: immunosuppressive (including steroids therapy); iv: intravenous; ABLC: amphotericin B lipid complex; d-AmB: deoxycholate amphotericin B; L-AmB: liposomal amphotericin B; RCT: randomized controlled trial.

zole plus placebo were only available in abstract form.

The results have been discussed with a provisional grading that could be transformed in a definite grading, as no additional data available in the full paper suggested a need for change in provisional recommendations.50 This study failed to reach the primary endpoint of decreased all-cause mortality at week 6 (difference of -8.2% in favor of combination; P=0.087). However, in a subgroup of patients with an invasive aspergillosis documented by positive galactomannan in either serum or bronchoalveolar lavage, 6-week all-cause mortality was lower in patients receiving combination therapy (difference of 11.6% in favor of combination; P=0.037). A large majority of the ECIL members felt that this subgroup analysis, that had not been originally planned, was not sufficient to give a stronger recommendation although this subgroup included 80% of the modified intent-to-treat population.

Therefore, the combination of voriconazole plus anidulafungin was graded C I for primary therapy of invasive aspergillosis while all other combinations were graded C III in the absence of well-designed studies for first-line therapy.

Table 6 summarizes the main characteristics and results of the various studies. Importantly, very few studies had a large number of patients with a mycological documentation.40,41,50 As shown by the 2 post-hoc analyses, survival was substantially lower in mycologically-documented infections compared to possible cases.47,48 Therefore, studies with a limited number of documented cases cannot lead to the strongest recommendations. As no study specifically addressed management of breakthrough aspergillosis after failure of posaconazole or voriconazole prophylaxis, no recommendation could be made on this issue.

The clinical trial comparing the new triazole isavuconazole versus voriconazole for primary therapy of invasive aspergillosis could not be discussed during the ECIL-5 as results were only presented as an abstract in 2014.

However, the group could review the data from these abstracts during the ECIL-6 meeting. Isavuconazole appears to be as effective as voriconazole for the treatment of invasive aspergillosis and has a better safety profile. Therefore, a grade A I similar to the grading for voriconazole has been given to isavuconazole (Table 7).

As the full paper was published shortly after the meeting, and confirms the results, the provisional grading attributed during the meeting has been transformed into a definite grading in this manuscript.51 haematologica | 2017; 102(3)

Currently, amphotericin B deoxycholate is considered to have no role in the treatment of invasive aspergillosis when more effective and less toxic agents are available. Its limited efficacy and its poor safety profile led to a recommendation against its use. No substantial change has been made for second-line therapy in the absence of new data (Table 8).

Mucormycosis Diagnostic and therapeutic strategies were discussed during the ECIL-5 and the ECIL-6. Rhizopus, Mucor, Lichtheimia (previously classified as Absidia), Cunninghamella, Rhizomucor, Apophysomyces, and Saksenaea are the genera most frequently involved in human disease.52 Cunninghamella species is more virulent in experimental models and may be associated with a higher mortality rate in patients.53 So far, there has not been enough evidence that identification of mucormycosis to the genus and/or species level helps guide antifungal treatment.54,55 Species identification remains, nevertheless, important for outbreak investigations.56 However, the differentiation between mucormycosis and other invasive mold infection is of critical importance as it has major therapeutic implications.

While epidemiological aspects and some clinical (sinus disease, concomitant diabetes, occurrence under voriconazole therapy) and radiological (reverse halo sign on chest CT-scan) factors may help to suspect mucormycosis, the diagnosis remains difficult and biopsy of the lesion is often required. Identification of the pathogen most often comes from microscopic, culture and/or histopathological examination of relevant samples. New diagnostic approaches include molecular testing on serum and various other clinical samples including formalin-fixed tissues, MALDI-TOF and Mucorales-specific T-cell detection.57-64 Although these new approaches are very promising for an earlier diagnosis, no grading for their use can be given yet due to the lack of data.

Amphotericin B, posaconazole and isavuconazole are the most potent agents in vitro.65-67 Currently, no validated minimum inhibitory concentration breakpoints for any of the drugs are available and thus determination of susceptibility categories is not possible for the agents of mucormycosis. The ECIL-3 recommendations for the treatment of mucormycosis were mostly based on retrospective studies, registry data and small prospective non-controlled studies.3,68-77 Few new data are available for the treatment of mucormycosis since the ECIL 4 and, therefore, the current recommendations are very similar (Table 9).

437

F. Tissot et al.

Table 3. Trials for first-line therapy of invasive candidiasis: outcomes

1st author, year, reference 5

Rex, 1994

Nguyen, 19956

Treatment Fluconazole d-AmB Fluconazole d-AmB

Anaissie, 19967

Fluconazole d-AmB

Anaissie, 19968

Fluconazole d-AmB

Phillips, 19979

Fluconazole d-AmB Caspofungin d-AmB

Mora-Duarte, 200210

Rex, 200311

Response rate at end of therapy Other efficacy outcomes No difference in response rates Fluconazole as efficacious as d-AmB Similar for fluconazole and d-AmB Similar for fluconazole and d-AmB relapse and survival rates Similar for fluconazole and d-AmB Caspofungin not inferior to d-AmB

Fluconazole Fluconazole + d-AmB

Improved success rate for the combination therapy

DiNubile, 200512

Voriconazole d-AmB then fluconazole

Kullberg, 200513

Caspofungin d-AmB

Voriconazole not inferior to d-AmB/fluconazole cultures Similar for caspofungin and d-AmB

Ostrosky-Zeichner, 200514

Micafungin Micafungin + other agent Anidulafungin

Kuse, 200715

High success rate for first-line and salvage therapy Higher response rate for anidulafungin

Safety profile

No difference in survival No difference in survival

Fluconazole better tolerated Fluconazole better tolerated

No difference in survival

Fluconazole better tolerated

No difference in time to defervescence

Fluconazole better tolerated

No difference in survival rates Similar survival and relapse rate

Fluconazole better tolerated Less clinical and laboratory drug-related adverse events with caspofungin Fluconazole monotherapy better tolerated than combination therapy

Similar time to failure and survival; higher rate of blood culture clearance with combination Similar survival and time to clear blood receiving voriconazole Response rate lower in neutropenic than in non-neutropenic cancer pts High success rate in neutropenic pts

Less all-cause adverse events in patients receiving voriconazole Caspofungin better tolerated than d-AmB No unexpected adverse event

Similar 6-week survival; More drug-related elevation higher microbiological in liver enzymes Fluconazole response rate for anidulafungin; in patients receiving same conclusion for subgroup fluconazole of C. albicans infections Pappas, 200716 Micafungin Micafungin Similar survival and time Less clinical and L-AmB not inferior to L-AmB to clear blood cultures biological drug-related adverse events in pts adverse events in pts receiving micafungin receiving micafungin Micafungin (100 mg) Similar for the No significant Same safety profile Reboli, 200717 and Micafungin (150 mg) three arms in difference in survival; for both doses Reboli, 201118 Caspofungin neutropenic pts similar response rates of micafungin and caspofungin and caspofungin Queiroz-Telles, 200819 Micafungin Similar for both Similar survival; More adverse events L-AmB treatments efficacy independent leading to treatment of the age discontinuation in L-AmB arm discontinuation in L-AmB arm Caspofungin (50 mg) Similar for both Similar Safety not inferior Betts, 200920 Caspofungin (150 mg) doses of caspofungin survival and time for high-dose to clear blood cultures caspofungin Cornely, 201121 Micafungin (100 mg), Similar response rate across Similar survival NA micafungin (150 mg), the two trials and all for all treatments caspofungin, L-AmB treatment arms groups for pts with or for pts with or without malignancy without malignancy Fluconazole, Higher response rate Higher mortality when older age, NA Andes, 201222 d-AmB, L-AmB, when use of greater Apache II score, d-AmB + fluconazole, echinocandin or central immunosuppressive therapy, d-AmB then fluconazole, voriconazole, catheter removed; or C. tropicalis infection; caspofungin anidulafungin, lower response rate lower mortality when micafungin when greater Apache II score use of echinocandin or central venous catheter removed 438

haematologica | 2017; 102(3)

ECIL-6 guidelines: fungal infections in leukemia and HSCT patients Kanji, 201323

Vasquez, 201424

d-AmB, d-AmB + flucytosine, Trends favoring L-AmB, ABLC, ketoconazole, non-polyene compounds fluconazole,voriconazole, caspofungin micafungin, anidulafungin Anidulafungin, then fluconazole Similar success rate for or voriconazole early switch (<7d) and MITT population across all Candida species

Herbrecht, 201425 Fernandez-Ruis, 201526

Anidulafungin Echinocandin-basedNA Azole-based

NA

NA

No difference in survival

Nausea and vomiting as the most frequent drug-related adverse events

Overall 52% success rate, lower when persistent neutropenia NA

24% all-cause mortality at day 28 No difference in clinical failure (all-cause mortality between day 3 and 30 and persistent candidemia > 72 h after start of antifungal therapy)

NA NA

NA: not available; MITT: modified intention-to-treat; d-AmB: deoxycholate amphotericin B; L-AmB: liposomal amphotericin B; pts: patients.

A prospective non-comparative trial assessed the efficacy and safety of first-line therapy with high-dose liposomal amphotericin B given at 10 mg/kg/day combined with surgery when appropriate.78 This trial demonstrated efficacy of high-dose liposomal amphotericin B plus surgery in mucormycosis with a survival rate of 62% at week 12.

The only factor associated with mortality was the presence of hematologic malignancy or cancer (HR: 3.15; 95%CI: 1.12-8.91; P=0.02). Renal impairment of any degree was observed in 40% of the patients but was transient in most of them. These results confirm the beneficial role of liposomal amphotericin B but do not yet allow any recommendation for the administration of such a high dose of 10 mg/kg/day.

A short paper presented data from a retrospective analysis of a combination of posaconazole and a lipid formulation of amphotericin B.79 Thirty-two patients received this combination of posaconazole with liposomal ampho-

Table 4. ECIL-6 recommendations for initial first-line treatment of candidemia.

Antifungal therapy Micafungina Anidulafungin Caspofungin Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholatec Fluconazoled,e Voriconazoled Catheter removalf

Overall population

Hematologic patients

AI AI AI AI B II B II CI AI AI A II

A II A IIb A II A II B II B II C II C III B II B II

See warning box in European label; bprovisional grading; cclose monitoring for adverse event is required; dnot in severely ill unstable patients; enot in patients with previous azole exposure; fif the catheter cannot be removed, use of an echinocandin or a lipid formulation of amphotericin B is recommended.

a

Table 5. ECIL-6 recommendations for first-line treatment of candidemia after species identification.

Candida species C. albicans

C. glabrata

C. krusei

Oral stepdown C. parapsilosis

Overall population a

Echinocandins Fluconazoleb Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholate Echinocandinsa Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholate Echinocandinsa Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholate Voriconazole Fluconazole Echinocandinsc

Hematologic patients AI AI AI A II A II CI AI BI B II B II CI A II BI B II B II CI BI A II B II

Echinocandins Fluconazole Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholate Echinocandins Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholate Echinocandinsa Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Amphotericin B deoxycholate Voriconazole Fluconazole Echinocandins

A II C III B II B II B II C II A II B II B II B II C II A III B II B II B II C II C III A III B III

a Same grading for anidulafungin, caspofungin, micafungin; bnot in severely ill patients; cif echinocandin-based regimen introduced before species identification and patient responding clinically and microbiologically (sterile blood cultures at 72 h), continuing use of echinocandin might be considered.

haematologica | 2017; 102(3)

439

F. Tissot et al.

tericin B (n=27) or amphotericin B lipid complex (n=5).

Only 3 of them were treated with this combination in first line. Overall response rate was 56% but a large proportion of patients (59%) died before day 90. The low number of patients, the retrospective nature of the study, and the high mortality rate at day 90 only allowed for a B III recommendation for this combination for salvage therapy of mucormycosis (Table 10).

Discussion and conclusions An update of the ECIL antifungal treatment recommendations was needed as there were important new data, and also because of necessary changes in the ECIL grading system so as to be in harmony with other ECIL recommendations. The most important data for invasive candidiasis came from a large review of patients included in 7

Table 6. Trials for first-line therapy of invasive aspergillosis: main characteristics and outcome.

1st author, year, reference

Type of study

Patient population

Antifungal agent (daily dose)

Ellis, 199838

RCT

Caillot, 200139

Prospective, non-comparative

Hematologic malignancy, HSCT Hematologic malignancy, HSCT, other IS condition

41 46 31

8 (20%) 12 (26%) 14 (45%)

58% 54% 48%

58%d 51%d 87%

Bowden, 200240

RCT, double blind

L-AmB (1 mg/kg) L-AmB (4 mg/kg) Itraconazole (iv, 2x200 for 2 days then 200 for 12 days then oral 2x200 mg) d-AmB (1-1.5 mg/kg) ABCD (6 mg/kg)

86 88

81 (94%) 75 (85%)

35%d 35%d

45%d 50%d

Herbrecht, 200241

RCT

133 144

84 (63%) 98 (68%)

32% 53%

58% 71%

Candoni, 200542

Prospective, non-comparative RCT, double blind

Hematologic malignancy, HSCT Hematologic malignancy, HSCT, other IS condition

d-AmB (1-1.5 mg/kg) Voriconazole (iv, 2x6 mg/kg on day 1 then 2x4 mg/kg then oral 2x200 mg) Caspofungin

32

NA

56%

53%e

107f 94f

41 (40%)g 36 (39%)g

50% 46%

72% 59%

Prospective, non-comparative Prospective, non-comparative Prospective dose-escalation study Post-hoc analysis of study published in 2002

Hematologic malignancy Allogeneic HSCT

L-AmB (3 mg/kg) L-AmB (10 mg/kg for 14 days then 3 mg/kg) Caspofungin (70 on day 1 then 50 mg) Caspofungin

61

61 (100%)

33%

53%

24

24 (100%)

42%

50%

Caspofungin (70-200 mg)

46

26 (57%)

57%

72%

d-AmB (1-1.5 mg/kg) 164 Voriconazole (iv, 2x6 179 mg/kg on day 1 then 2x4 mg/kg then oral 2x200 mg) Voriconazole (iv, 2x6 142 mg/kg on day 1 then 135 2x4 mg/kg then oral 2x300 mg) Voriconazole (iv, 2x6 mg/kg on day 1 then 2x4 mg/kg then oral 2x300 mg) + Anidulafungin (200 on day 1 then 100 mg) Isavuconazole (2x200 143h on day 1 and 2 then 200 mg) 129h Voriconazole (iv, 2x6 mg/kg on day 1 then 2x4 mg/kg then oral 2x200 mg)

113 (69%) 124 (69%)

19% 51%

55% 70%

142 (100%) 135 (100%)

43% 33%

61% 71%

143 129h

62% 60%

70% 66%

Cornely, 200743

Viscoli, 200944 Herbrecht, 201045 Cornely, 201146

Herbrecht, 201548

Hematologic malignancy, HSCT, other IS condition, COPD Hematologic malignancy, HSCT, other IS conditions

Hematologic malignancy, HSCT, other IS condition Hematologic malignancy, HSCT, other IS conditions

Marr, 201550

RCT, double blind

Hematologic malignancy, HSCT

Maertens, 201551

RCT, double blind

Hematologic malignancy, HSCT, other IS condition

N of Mycological Favorablec 12-week ptsa documentationb response survival rate

Numbers of patients refer to the modified intent to treat population when available or to the intent to treat population; bincludes positive microscopy or culture from relevant sites, positive histopathology, or positive galactomannan in serum, BAL or CSF as defined by EORTC/MSG 2008 criteria49; cfavorable response rate includes only complete and partial responses; dtwo-month survival rates; dintent to treat population; etime point not specified (median follow up 10 months); fincludes also other mold infections (4 and 2 in 3 mg/kg and 10 mg/kg arm, respectively); gafter exclusion of the 6 other mold infections; hincludes also a few non-Aspergillus invasive mold diseases (5 in isavuconazole arm and 6 in voriconazole arm) and non-identified invasive mold disease (14 in isavuconazole arm and 15 in voriconazole arm). ABCD: amphotericin B colloidal dispersion; ABLC: amphotericin B lipid complex; COPD: chronic obstructive pulmonary disease; d-AmB: deoxycholate amphotericin B; HSCT: hematopoietic stem cell transplant; IS: immunosuppressive (including steroids therapy); L-AmB: liposomal amphotericin B; pt(s): patient(s); RCT: randomized controlled trial. a

440

haematologica | 2017; 102(3)

ECIL-6 guidelines: fungal infections in leukemia and HSCT patients Table 7. ECIL-6 recommendations for first-line treatment of invasive aspergillosis.

Grade a

Voriconazole

AI

Isavuconazole Liposomal amphotericin B Amphotericin B lipid complex Amphotericin B colloidal dispersion Caspofungin Itraconazole Combination voriconazolea + anidulafungin Other combinations Recommendation against use Amphotericin B deoxycholate

AI BI B II CI C II C III CI C III AI

Comments Daily dose: 2x6 mg/kg on day 1 then 2x4 mg/kg (initiation with oral therapy: C III) As effective as voriconazole and better tolerated Daily dose: 3 mg/kg Daily dose: 5 mg/kg Not more effective than d-AmB but less nephrotoxic

Less effective and more toxic

Monitoring of serum levels is indicated. In the absence of sufficient data for first-line monotherapy, anidulafungin, micafungin and posaconazole have not been graded.

a

Table 8. ECIL-6 recommendations for salvage therapy of invasive aspergillosis.

Liposomal amphotericin B Amphotericin B lipid complex Caspofungin Itraconazole Posaconazolea Voriconazolea Combination

Grade

Comments

B II B II B II C III B II B II B II

No data on voriconazole failure No data on voriconazole failure No data on voriconazole failure Insufficient data No data on voriconazole failure If not used in first-line Various studies and conflicting results

Monitoring of serum levels is indicated, especially if posaconazole oral suspension is used.

a

major trials.22 The multivariate analysis now allows a very strong recommendation in favor of an echinocandin for the first-line therapy of candidemia irrespective of the underlying predisposing factors. The controversy on the beneficial role of catheter removal can now be considered to be resolved. The most interesting new data were the publication of a first-line combination study in invasive aspergillosis and the results of a randomized comparative trial comparing isavuconazole to voriconazole. Aspergillus guidelines now include the results of these 2 clinical trials and should help clinicians in their treatment decision making. Since few new data have been published since the last ECIL guidelines, no major changes were made to mucormycosis management. Importantly, the posology and indication of antifungal agents reported in the current guidelines do not necessarily reflect those licensed by the European Medicines Agency (EMA), but are the result of a consensus-based analysis of available literature within the ECIL group.

There has been controversy about some discrepancies between the ECIL-5 and the ESCMID recommendations for invasive aspergillosis in hematologic patients. These differences were identified during a joint meeting and an ESCMID representative was invited to discuss them at the ECIL-6 meeting. Most differences were minor and mostly reflected a difference in grading system. The ECIL Aspergillus recommendations are restricted to hematologic patients who represent more than 90% of the patients haematologica | 2017; 102(3)

included in the major clinical trials.41,43,50,51 No subgroup of hematologic patients deserving specific recommendation for Aspergillus infection treatment has been identified by the ECIL group. In contrast, the ESCMID group had a broader approach considering all other conditions predisposing to invasive aspergillosis, grading the diagnostic procedures, and including environmental measures in the prevention, also providing a grade for specific infection sites. In addition, the ESCMID group also segregated the hematologic patients into subgroups and provided specific grading for each of them, with usually weaker recommendations when there was not a sufficient number of patients with these specific underlying conditions included in the clinical studies. Finally, and importantly, some data were not available at the time of the ECIL-5 meeting but were in the public domain when the ESCMID group met. In September 2015, the ECIL-6 group was able to incorporate the new data, and this has helped to reduce the apparent differences with the ESCMID guidelines. Therefore, neither the ECIL group nor the ESCMID group felt any change other than this update was required.

Collaborators: Agrawal Samir, United Kingdom; Akova Murat, Turkey; Alanio Alexandre, France; Aljurf Mahmoud, Saudi Arabia; Averbukh Dina, Israel; Barnes Rosemary, United Kingdom; Blennow Ola, Sweden; Bochud Pierre Yves, Switzerland; Bouza Emilio, Spain; Bretagne Stephane, France; Bruggemann Roger, Netherlands; Calandra Thierry, Switzerland; 441

F. Tissot et al.

Table 9. ECIL-6 recommendations for first-line therapy of mucormycosis.

Grade Management includes antifungal therapy, surgery and control of underlying conditions Antifungal therapy Amphotericin B deoxycholate Liposomal amphotericin B

A II C II B II

Amphotericin B lipid complex Amphotericin B colloidal dispersion Posaconazole

B II C II C III

Combination therapy

C III

Control of underlying condition

A II

Surgery Rhino-orbito-cerebral infection Soft tissue infection Localized pulmonary lesion Disseminated infection

A II A II B III C III

Hyperbaric oxygen Recommendation against use Combination with deferasirox

Comments Multidisciplinary approach is required

Daily dose: 5 mg/kg. Liposomal amphotericin B should be preferred in CNS infection and/or renal failure No data to support its use as first-line treatment. Alternative when amphotericin B formulations are absolutely contraindicated.

Includes control of diabetes, hematopoietic growth factor if neutropenia, discontinuation/tapering of steroids, reduction of immunosuppressive therapy

Surgery should be considered on a case by case basis, using a multi-disciplinary approach

C III A II

CNS: central nervous system.

Table 10. ECIL-6 recommendations for salvage and maintenance therapy of mucormycosis.

Grade Salvage therapy Management includes antifungal therapy, control of underlying disease and surgery Posaconazole Combination of lipid amphotericin B and caspofungin Combination of lipid amphotericin B and posaconazole Maintenance therapy Posaconazole

Comments

A II B II B III B III B III

Overlap of a few days with first-line therapy to obtain appropriate serum levels. Monitoring of serum levels might be indicateda

Both comments apply to the oral solution but may not apply to the solid oral formulation.

a

Carratalà Jordi, Spain; Castagnola Elio, Italy; Cesaro Simone, Italy; Cordonnier Catherine, France; Cornely Oliver, Germany; Dalianis Tina, Sweden; De La Camara Rafael, Spain; Donnelly Peter Joseph, Netherlands; Drgona Lubos, Slovakia; Duarte Rafael, Spain; Einsele Hermann, Germany; Engelhard Dan, Israel; Ferreira Isabelina, Portugal; Fox Christopher, United Kingdom; Girmenia Corrado, Italy; Groll Andreas, Germany; Hauser Philippe, Switzerland; Heldal Dag, Norway; Helweg-Larsen Jannik, Denmark; Herbrecht Raoul, France; Hirsch Hans, Switzerland; Hubacek Petr, Czech Republic; Johnson Elizabeth M, United Kingdom; Kibbler Christopher, United Kingdom; Klyasova Galina, Russian Federation; Koskenvuo Minna, Finland; Kouba Michal, Czech Republic; Kullberg Bart-Jan, Netherlands; Lagrou Katrien, Belgium; Lewis Russel, Italy; Ljungman Per, Sweden; Maertens Johan, Belgium; Mallet Vincent, France; Marchetti Oscar, Switzerland; Martino Rodrigo, Spain; Maschmeyer Georg, Germany; Matos Olga, Portugal; Melchers Willem, Netherlands; Mikulska Malgorzata, Italy; Munoz Patricia, Spain; Nucci Marcio,

442

Brazil; Orasch Christina, Switzerland; Padoin Christophe, France; Pagano Livio, Italy; Pagliuca Antonio, United Kingdom; Penack Olaf, Germany; Petrikkos Georgios, Greece; Racil Zdenek, Czech Republic; Ribaud Patricia, France; Rinaldo Christine H, Norway; Robin Christine, France; Roilides Emmanuel, Greece; Rovira Montserrat, Spain; Schellongowski Peter, Austria; Sedlacek Petr, Czech Republic; Sinko Janos, Hungary; Skiada Anna, Greece; Slavin Monica, Australia; Styczynski Jan,Poland; Tissot Frédéric, Switzerland; Van Boemmel Florian, Germany; Viscoli Claudio, Italy; Von Lilienfeld-Toal Marie, Germany; Ward Katherine, United Kingdom.

Acknowledgments The authors and contributors thank the group GL-Events, Lyon, France, for the organization of the meeting. They also thank Valérie Rizzi-Puechal (Pfizer), France; Markus Rupp (MSD), Germany; Sonia Sanchez (Gilead Sciences), UK; Anne-Therese Witschi (Basilea), Switzerland; Lorraine Tweddle (Astellas), UK.

haematologica | 2017; 102(3)

ECIL-6 guidelines: fungal infections in leukemia and HSCT patients

References 1. Herbrecht R, FlĂźckiger U, Gachot B, Ribaud P, Thiebaut A, Cordonnier C. Treatment of invasive Candida and invasive Aspergillus infections in adult haematological patients.

Eur J Cancer Suppl. 2007;5(2):49-59.

2. Maertens J, Marchetti O, Herbrecht R, et al.

European guidelines for antifungal management in leukemia and hematopoietic stem cell transplant recipients: summary of the ECIL 3--2009 update. Bone Marrow Transplant. 2011;46(5):709-718.

3. Skiada A, Lanternier F, Groll AH, et al.

Diagnosis and treatment of mucormycosis in patients with hematological malignancies: guidelines from the 3rd European Conference on Infections in Leukemia (ECIL 3). Haematologica. 2013;98(4):492-504.

4. Ullmann AJ. ESCMID guidelines for the diagnosis and treatment of Aspergillus diseases. Invasive aspergillosis in haematology and oncology. European Congress of Clinical Microbiology and Infectious Diseases. Barcelona, Spain, 2014: Abstract EW081.

5. Rex JH, Bennett JE, Sugar AM, et al. A randomized trial comparing fluconazole with amphotericin B for the treatment of candidemia in patients without neutropenia.

Candidemia Study Group and the National Institute. N Engl J Med. 1994;331(20):13251330.

6. Nguyen MH, Peacock JE Jr, Tanner DC, et al.

Therapeutic approaches in patients with candidemia. Evaluation in a multicenter, prospective, observational study. Arch Intern Med. 1995;155(22):2429-2435.

7. Anaissie EJ, Darouiche RO, Abi-Said D, et al.

Management of invasive candidal infections: results of a prospective, randomized, multicenter study of fluconazole versus amphotericin B and review of the literature. Clin Infect Dis. 1996;23(5):964-972.

8. Anaissie EJ, Vartivarian SE, Abi-Said D, et al.

Fluconazole versus amphotericin B in the treatment of hematogenous candidiasis: a matched cohort study. Am J Med. 1996; 101(2):170-176.

9. Phillips P, Shafran S, Garber G, et al.

Multicenter randomized trial of fluconazole versus amphotericin B for treatment of candidemia in non-neutropenic patients.

Canadian Candidemia Study Group. Eur J Clin Microbiol Infect Dis. 1997;16(5):337345.

10. Mora-Duarte J, Betts R, Rotstein C, et al.

Comparison of caspofungin and amphotericin B for invasive candidiasis. N Engl J Med. 2002;347(25):2020-2029.

11. Rex JH, Pappas PG, Karchmer AW, et al. A randomized and blinded multicenter trial of high-dose fluconazole plus placebo versus fluconazole plus amphotericin B as therapy for candidemia and its consequences in nonneutropenic subjects. Clin Infect Dis. 2003;36(10):1221-1228.

12. DiNubile MJ, Hille D, Sable CA, Kartsonis NA. Invasive candidiasis in cancer patients: observations from a randomized clinical trial. J Infect. 2005;50(5):443-449.

13. Kullberg BJ, Sobel JD, Ruhnke M, et al.

Voriconazole versus a regimen of amphotericin B followed by fluconazole for candidaemia in non-neutropenic patients: a randomised non-inferiority trial. Lancet.

2005;366(9495):1435-1442.

14. Ostrosky-Zeichner L, Kontoyiannis D, Raffalli J, et al. International, open-label, noncomparative, clinical trial of micafungin

haematologica | 2017; 102(3)

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

alone and in combination for treatment of newly diagnosed and refractory candidemia.

Eur J Clin Microbiol Infect Dis. 2005;24(10):654-661.

Kuse ER, Chetchotisakd P, da Cunha CA, et al. Micafungin versus liposomal amphotericin B for candidaemia and invasive candidosis: a phase III randomised double-blind trial. Lancet. 2007;369(9572):1519-1527.

Pappas PG, Rotstein CM, Betts RF, et al.

Micafungin versus caspofungin for treatment of candidemia and other forms of invasive candidiasis. Clin Infect Dis. 2007;45(7):883-893.

Reboli AC, Rotstein C, Pappas PG, et al.

Anidulafungin versus fluconazole for invasive candidiasis. N Engl J Med. 2007;356(24):2472-2482.

Reboli AC, Shorr AF, Rotstein C, et al.

Anidulafungin compared with fluconazole for treatment of candidemia and other forms of invasive candidiasis caused by Candida albicans: a multivariate analysis of factors associated with improved outcome. BMC Infect Dis. 2011;11:261.

Queiroz-Telles F, Berezin E, Leverger G, et al.

Micafungin versus liposomal amphotericin B for pediatric patients with invasive candidiasis: substudy of a randomized doubleblind trial. Pediatr Infect Dis J. 2008;27(9):820-826.

Betts RF, Nucci M, Talwar D, et al. A Multicenter, double-blind trial of a highdose caspofungin treatment regimen versus a standard caspofungin treatment regimen for adult patients with invasive candidiasis.

Clin Infect Dis. 2009;48(12):1676-1684.

Cornely OA, Marty FM, Stucker F, Pappas PG, Ullmann AJ. Efficacy and safety of micafungin for treatment of serious Candida infections in patients with or without malignant disease. Mycoses. 2011;54(6):e838-847.

Andes DR, Safdar N, Baddley JW, et al.

Impact of treatment strategy on outcomes in patients with candidemia and other forms of invasive candidiasis: a patient-level quantitative review of randomized trials. Clin Infect Dis. 2012;54(8):1110-1122.

Kanji JN, Laverdiere M, Rotstein C, Walsh TJ, Shah PS, Haider S. Treatment of invasive candidiasis in neutropenic patients: systematic review of randomized controlled treatment trials.

Leuk Lymphoma.

2013;54(7):1479-1487.

Vazquez J, Reboli AC, Pappas PG, et al.

Evaluation of an early step-down strategy from intravenous anidulafungin to oral azole therapy for the treatment of candidemia and other forms of invasive candidiasis: results from an open-label trial. BMC Infect Dis. 2014;14:97.

Herbrecht R, Conte U, Biswas P, Capparella MR, Aram J. Efficacy of anidulafungin in the treatment of invasive candidiasis in neutropenic patients: analysis of pooled data from five prospective studies. European Conference on Clinical Microbiology and Infectious Diseases. Barcelona, Spain, 2014: Abstract R692.

Fernandez-Ruiz M, Aguado JM, Almirante B, et al. Initial use of echinocandins does not negatively influence outcome in Candida parapsilosis bloodstream infection: a propensity score analysis. Clin Infect Dis. 2014;58(10):1413-1421.

Chalmers C, Gaur S, Chew J, et al.

Epidemiology and management of candidaemia--a retrospective, multicentre study in five hospitals in the UK. Mycoses.

2011;54(6):e795-800.

Horn DL, Ostrosky-Zeichner L, Morris MI,

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

et al. Factors related to survival and treatment success in invasive candidiasis or candidemia: a pooled analysis of two large, prospective, micafungin trials. Eur J Clin Microbiol Infect Dis. 2010;29(2):223-229.

Nucci M, Anaissie E. Should vascular catheters be removed from all patients with candidemia? An evidence-based review.

Clin Infect Dis. 2002;34(5):591-599.

Rodriguez D, Park BJ, Almirante B, et al.

Impact of early central venous catheter removal on outcome in patients with candidaemia.

Clin Microbiol Infect.

2007;13(8):788-793.

Velasco E, Portugal RD. Factors prompting early central venous catheter removal from cancer patients with candidaemia. Scand J Infect Dis. 2011;43(1):27-31.

Rex JH, Bennett JE, Sugar AM, et al.

Intravascular catheter exchange and duration of candidemia. NIAID Mycoses Study Group and the Candidemia Study Group.

Clin Infect Dis. 1995;21(4):994-996.

Raad I, Hanna H, Boktour M, et al.

Management of central venous catheters in patients with cancer and candidemia. Clin Infect Dis. 2004;38(8):1119-1127.

Garnacho-Montero J, Diaz-Martin A, Garcia-Cabrera E, Ruiz Perez de Pipaon M, Hernandez-Caballero C, Lepe-Jimenez JA. Impact on hospital mortality of catheter removal and adequate antifungal therapy in Candida spp. bloodstream infections. J Antimicrob Chemother. 2013;68(1):206-213.

Kucharikova S, Sharma N, Spriet I, Maertens J, Van Dijck P, Lagrou K. Activities of systemically administered echinocandins against in vivo mature Candida albicans biofilms developed in a rat subcutaneous model. Antimicrob Agents Chemother.

2013;57(5):2365-2368.

Seidler M, Salvenmoser S, Muller FM. Liposomal amphotericin B eradicates Candida albicans biofilm in a continuous catheter flow model. FEMS Yeast Res. 2010;10(4):492-495.

Shuford JA, Rouse MS, Piper KE, Steckelberg JM, Patel R. Evaluation of caspofungin and amphotericin B deoxycholate against Candida albicans biofilms in an experimental intravascular catheter infection model. J Infect Dis. 2006;194(5):710-713.

Ellis M, Spence D, de Pauw B, et al. An EORTC international multicenter randomized trial (EORTC number 19923) comparing two dosages of liposomal amphotericin B for treatment of invasive aspergillosis. Clin Infect Dis. 1998;27(6):1406-1412.

Caillot D, Bassaris H, McGeer A, et al.

Intravenous itraconazole followed by oral itraconazole in the treatment of invasive pulmonary aspergillosis in patients with hematologic malignancies, chronic granulomatous disease, or AIDS. Clin Infect Dis. 2001;33(8):e83-90.

Bowden R, Chandrasekar P, White MH, et al. A double-blind, randomized, controlled trial of amphotericin B colloidal dispersion versus amphotericin B for treatment of invasive aspergillosis in immunocompromised patients. Clin Infect Dis. 2002;35(4):359-366.

Herbrecht R, Denning DW, Patterson TF, et al. Voriconazole versus amphotericin B for primary therapy of invasive aspergillosis. N Engl J Med. 2002;347(6):408-415.

Candoni A, Mestroni R, Damiani D, et al.

Caspofungin as first line therapy of pulmonary invasive fungal infections in 32 immunocompromised patients with hematologic malignancies. Eur J Haematol.

2005;75(3):227-233.

443

F. Tissot et al.

43. Cornely OA, Maertens J, Bresnik M, et al.

Liposomal amphotericin B as initial therapy for invasive mold infection: a randomized trial comparing a high-loading dose regimen with standard dosing (AmBiLoad trial). Clin Infect Dis. 2007;44(10):1289-1297.

44. Viscoli C, Herbrecht R, Akan H, et al. An EORTC Phase II study of caspofungin as first-line therapy of invasive aspergillosis in haematological patients. J Antimicrob Chemother. 2009;64(6):1274-1281.

45. Herbrecht R, Maertens J, Baila L, et al.

Caspofungin first-line therapy for invasive aspergillosis in allogeneic hematopoietic stem cell transplant patients: an European Organisation for Research and Treatment of Cancer study. Bone Marrow Transplant.

2010;45(7):1227-1233.

46. Cornely OA, Vehreschild JJ, Vehreschild MJ, et al. Phase II dose escalation study of caspofungin for invasive Aspergillosis. Antimicrob Agents Chemother. 2011; 55(12):5798-5803.

47. Cornely OA, Maertens J, Bresnik M, et al.

Efficacy outcomes in a randomised trial of liposomal amphotericin B based on revised EORTC/MSG 2008 definitions of invasive mould disease. Mycoses. 2011;54(5):e449455.

48. Herbrecht R, Patterson TF, Slavin MA, et al. Application of the 2008 Definitions for Invasive Fungal Diseases to the Trial Comparing Voriconazole Versus Amphotericin B for Therapy of Invasive Aspergillosis: A Collaborative Study of the Mycoses Study Group (MSG 05) and the European Organization for Research and Treatment of Cancer Infectious Diseases Group. Clin Infect Dis. 2015;60(5):713720.

49. De Pauw B, Walsh TJ, Donnelly JP, et al.

Revised definitions of invasive fungal disease from the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group. Clin Infect Dis. 2008;46(12):1813-1821.

50. Marr KA, Schlamm HT, Herbrecht R, et al.

Combination antifungal therapy for invasive aspergillosis: a randomized trial. Ann Intern Med. 2015;162(2):81-89.

51. Maertens JA, Raad, II, Marr KA, et al.

Isavuconazole versus voriconazole for primary treatment of invasive mould disease caused by Aspergillus and other filamentous fungi (SECURE): a phase 3, randomised-controlled, non-inferiority trial. Lancet.

2016;387(10020):760-769.

52. Kwon-Chung KJ. Taxonomy of fungi causing mucormycosis and entomophthoramycosis (zygomycosis) and nomenclature of the disease: molecular mycologic perspectives. Clin Infect Dis. 2012;54(Suppl 1):S8S15.

53. Petraitis V, Petraitiene R, Antachopoulos C, et al. Increased virulence of Cunninghamella bertholletiae in experimental pulmonary mucormycosis: correlation with circulating

444

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

molecular biomarkers, sporangiospore germination and hyphal metabolism. Med Mycol. 2013;51(1):72-82.

Rodriguez MM, Pastor FJ, Sutton DA, et al.

Correlation between in vitro activity of posaconazole and in vivo efficacy against Rhizopus oryzae infection in mice.

Antimicrob Agents Chemother.

2010;54(5):1665-1669.

Salas V, Pastor FJ, Calvo E, et al. In vitro and in vivo activities of posaconazole and amphotericin B in a murine invasive infection by Mucor circinelloides: poor efficacy of posaconazole. Antimicrob Agents Chemother. 2012;56(5):2246-2250.

Rammaert B, Lanternier F, Zahar JR, et al.

Healthcare-associated mucormycosis. Clin Infect Dis. 2012;54 (Suppl 1):S44-54.

Bernal-Martinez L, Buitrago MJ, Castelli MV, Rodriguez-Tudela JL, Cuenca-Estrella M. Development of a single tube multiplex realtime PCR to detect the most clinically relevant Mucormycetes species. Clin Microbiol Infect. 2013;19(1):E1-7.

Buitrago MJ, Aguado JM, Ballen A, et al.

Efficacy of DNA amplification in tissue biopsy samples to improve the detection of invasive fungal disease. Clin Microbiol Infect. 2013;19(6):E271-277.

Millon L, Herbrecht R, Grenouillet F, et al.

Early diagnosis and monitoring of mucormycosis by detection of circulating DNA in serum: retrospective analysis of 44 cases collected through the French Surveillance Network of Invasive Fungal Infections (RESSIF). Clin Microbiol Infect.

2016;22(9):810.e1-810.e8.

Lass-Florl C, Mutschlechner W, Aigner M, et al. Utility of PCR in diagnosis of invasive fungal infections: real-life data from a multicenter study. J Clin Microbiol. 2013; 51(3):863-868.

Walther G, Pawlowska J, AlastrueyIzquierdo A, et al. DNA barcoding in Mucorales: an inventory of biodiversity.

Persoonia. 2013;30:11-47.

De Carolis E, Posteraro B, Lass-Florl C, et al.

Species identification of Aspergillus, Fusarium and Mucorales with direct surface analysis by matrix-assisted laser desorption ionization time-of-flight mass spectrometry.

Clin Microbiol Infect. 2012; 18(5):475-484.

Schrodl W, Heydel T, Schwartze VU, et al.

Direct analysis and identification of pathogenic Lichtheimia species by matrix-assisted laser desorption ionization-time of flight analyzer-mediated mass spectrometry. J Clin Microbiol. 2012;50(2):419-427.

Potenza L, Vallerini D, Barozzi P, et al.

Mucorales-specific T cells emerge in the course of invasive mucormycosis and may be used as a surrogate diagnostic marker in high-risk patients. Blood. 2011; 118(20): 5416-5419.

Verweij PE, Gonzalez GM, Wiedrhold NP, et al. In vitro antifungal activity of isavuconazole against 345 mucorales isolates collected at study centers in eight countries. J Chemother. 2009;21(3):272-281.

66. Vitale RG, de Hoog GS, Schwarz P, et al.

Antifungal susceptibility and phylogeny of opportunistic members of the order mucorales. J Clin Microbiol. 2012;50(1):66-75.

67. Drogari-Apiranthitou M, Mantopoulou FD, Skiada A, et al. In vitro antifungal susceptibility of filamentous fungi causing rare infections: synergy testing of amphotericin B, posaconazole and anidulafungin in pairs. J Antimicrob Chemother. 2012;67(8):19371940.

68. Greenberg RN, Mullane K, van Burik JAH, et al. Posaconazole as salvage therapy for zygomycosis.

Antimicrob Agents Chemother. 2006;50(1):126-133.

69. Herbrecht R, Letscher-Bru V, Bowden RA, et al. Treatment of 21 cases of invasive mucormycosis with amphotericin B colloidal dispersion. Eur J Clin Microbiol Infect Dis. 2001;20(7):460-466.

70. Oppenheim BA, Herbrecht R, Kusne S. The safety and efficacy of amphotericin B colloidal dispersion in the treatment of invasive mycoses. Clin Infect Dis. 1995; 21(5):11451153.

71. Reed C, Bryant R, Ibrahim AS, et al.

Combination polyene-caspofungin treatment of rhino-orbital-cerebral mucormycosis. Clin Infect Dis. 2008;47(3):364-371.

72. Roden MM, Zaoutis TE, Buchanan WL, et al. Epidemiology and outcome of zygomycosis: a review of 929 reported cases. Clin Infect Dis. 2005;41(5):634-653.

73. Ruping MJ, Heinz WJ, Kindo AJ, et al. Fortyone recent cases of invasive zygomycosis from a global clinical registry. J Antimicrob Chemother. 2010;65(2):296-302.

74. Spellberg B, Ibrahim AS, Chin-Hong PV, et al. The Deferasirox-AmBisome Therapy for Mucormycosis (DEFEAT Mucor) study: a randomized, double-blinded, placebo-controlled trial. J Antimicrob Chemother. 2012; 67(3):715-722.

75. van Burik JA, Hare RS, Solomon HF, Corrado ML, Kontoyiannis DP. Posaconazole is effective as salvage therapy in zygomycosis: a retrospective summary of 91 cases. Clin Infect Dis. 2006;42(7):e61-65.

76. Xhaard A, Lanternier F, Porcher R, et al.

Mucormycosis after allogeneic haematopoietic stem cell transplantation: a French Multicentre Cohort Study (2003-2008). Clin Microbiol Infect. 2012; 18(10):E396-400.

77. Yohai RA, Bullock JD, Aziz AA, Markert RJ. Survival factors in rhino-orbital-cerebral mucormycosis.

Surv Ophthalmol.

1994;39(1):3-22.

78. Lanternier F, Poiree S, Elie C, et al.

Prospective pilot study of high-dose (10 mg/kg/day) liposomal amphotericin B (LAMB) for the initial treatment of mucormycosis. J Antimicrob Chemother. 2015; 70(11):3116-3123.

79. Pagano L, Cornely OA, Busca A, et al.

Combined antifungal approach for the treatment of invasive mucormycosis in patients with hematologic diseases: a report from the SEIFEM and FUNGISCOPE registries.

Haematologica. 2013;98 (10):e127-130.

haematologica | 2017; 102(3)

ARTICLE

Hematopoiesis

IFNα-mediated remodeling of endothelial cells in the bone marrow niche Áine M. Prendergast,1,2 Andrea Kuck,1,2 Mieke van Essen,2 Simon Haas,1,2 Sandra Blaszkiewicz1,2 and Marieke A. G. Essers1,2 Heidelberg Institute for Stem Cell Technology and Experimental Medicine and Hematopoietic Stem Cells and Stress Group, Division of Stem Cells and Cancer, Deutsches Krebsforschungszentrum, Heidelberg, Germany

1 2

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(3):445-453

ABSTRACT

I

n the bone marrow, endothelial cells are a major component of the hematopoietic stem cell vascular niche and are a first line of defense against inflammatory stress and infection. The primary response of an organism to infection involves the synthesis of immune-modulatory cytokines, including interferon alpha. In the bone marrow, interferon alpha induces rapid cell cycle entry of hematopoietic stem cells in vivo.

However, the effect of interferon alpha on bone marrow endothelial cells has not been described. Here, we demonstrate that acute interferon alpha treatment leads to rapid stimulation of bone marrow endothelial cells in vivo, resulting in increased bone marrow vascularity and vascular leakage. We find that activation of bone marrow endothelial cells involves the expression of key inflammatory and endothelial cell-stimulatory markers. This interferon alpha-mediated activation of bone marrow endothelial cells is dependent in part on vascular endothelial growth factor signaling in bone marrow hematopoietic cell types, including hematopoietic stem cells. Thus, this implies a role for hematopoietic stem cells in remodeling of the bone marrow niche in vivo following inflammatory stress. These data increase our current understanding of the relationship between hematopoietic stem cells and the bone marrow niche under inflammatory stress and also clarify the response of bone marrow niche endothelial cells to acute interferon alpha treatment in vivo.

Introduction Tissue vasculature serves as a barrier between sites of inflammation or infection and immune cells.1 Endothelial cells (ECs) are a diverse cell population which line the vasculature. These cells form a cell monolayer and are interconnected by junction molecules, including VE-cadherin and ESAM. This regulatory monolayer is ensheathed by pericytes and forms a selective, semi-permeable barrier that regulates tissue fluid homeostasis and migration of blood cells through the vessel wall.2 Thus, ECs are primary responders to inflammation and infection. During an inflammatory response, ECs proliferate in order to maintain vessel integrity.3 Immune cell loss, as well as interactions between immune cells and ECs, facilitates the emigration of circulating cells across the EC barrier to sites of inflammation. This process can, in turn, lead to EC activation.4,5 Production of pro-inflammatory factors and the interaction between stimulatory cytokines and chemokines is a critical step in the inflammatory response. Interferon α (IFNα) is one of the most prominent immune-modulatory cytokines which is produced to facilitate the response to inflammation. Vascular endothelial growth factor (VEGF) regulates ECs during both homeostasis and inflammation. VEGF regulation is central to vascular dynamics, promoting EC survival, proliferation and migration.4,6 In the bone marrow (BM) microenvironment, the vascular system consists of a network of sinusoids, arterioles, and transition zones. Subtypes of BM vessels are haematologica | 2017; 102(3)

Correspondence: m.essers@dkfz.de

Received: June 20, 2016.

Accepted: October 6, 2016.

Pre-published: October 14, 2016. doi:10.3324/haematol.2016.151209 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/445 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

445

Á.M. Prendergast et al. heterogenous in both properties and functions.7,8 BM ECs form a critical part of the hematopoietic stem cell (HSC) vascular niche. ECs have well-defined roles in HSC function and maintenance, retaining HSCs in culture and contributing to the creation of HSC niches.9-13 In vivo ablation of ECs leads to hematopoietic failure.14 In response to infection, hematopoietic cells mediate an altered expression of adherens molecules on the surface of ECs.13,15 This suggests that HSCs may also directly affect ECs. However, in contrast to the defined roles for ECs on HSCs, the effect of HSCs on ECs in the BM niche remains unclear. In addition, little is known about the influence of inflammatory stress on ECs in the BM, or the interaction between IFNα and VEGF. The stimulatory effect of IFNα on HSCs in vivo is not reflected in vitro. In vitro, IFNα has an inhibitory effect which leads to inhibition of HSC proliferation.16 This suggests a role for the BM niche, including ECs, in inflammation-induced HSC stimulation. However, how crosstalk between HSCs and ECs in the BM is regulated under inflammatory stress remains unknown. To understand how inflammatory stress impacts on ECs in the BM niche, we investigated how BM ECs respond to IFNα in vivo and how the interaction between HSCs and ECs is regulated.

We found that IFNα treatment of mice led to a rapid stimulation of BM ECs in vivo. IFNα stimulation of ECs was both direct and indirect. VEGF signaling, mediated by other BM hematopoietic cell types including HSCs, was a central mediator of the observed EC stimulation. This novel communication between activated hematopoietic cells and ECs in the BM suggests an acute 'emergency' response of the BM niche to primary inflammatory signaling from the hematopoietic system.

A

Methods Animals Eight- to 12-week old female wild-type (WT) mice [C57Bl/6J (CD45.2), Harlan Laboratories or B6.SJL-Ptprca Pepcb/BoyJ (CD45.1), Charles River Laboratories] and IFNAR–/– mice17 were maintained in individually ventilated cages in the Deutsches Krebsforschungszentrum animal facility. All animal protocols were approved by the Animal Care and Use Committees of the German Regierungspräsidium Karlsruhe für Tierschutz und Arzneimittelüberwachung.

In vivo treatments Mice were injected intraperitoneally (i.p.) with PBS, 5 mg/kg polyinosinic-polycytidylic acid (pI:C) (Invitrogen), subcutaneously (s.c) with 5x106U/kg recombinant mouse IFNα (Miltenyi Biotech) or intravenously (i.v.) with 2.5 mg/kg Avastin (Roche).

In vivo vascular labeling In vivo labeling was carried out as described by Kunisaki et al.18

Evans blue assay Evans blue assay was carried out as described by Radu et al.19

Isolation of BM Cells and flow cytometry Mice were sacrificed and BM cells were subsequently prepared and analyzed as described in Haas et al.20 In addition, ECs were stained using antibodies against CD4, CD8 CD11b, B220a, Gr-1 and TER119 as indicated,20 and CD45 (30-F11), CD31 (390), VECadherin (VECD1), VEGFR2 (Avas12), ESAM (1G8) (Biolegend), and Laminin (Sigma). Cells were stained with anti-VEGF antibody according to the manufacturer's instructions (Abcam).

B

C ** **

*

D

E **

*

Figure 1. Interferon α (IFNα) treatment leads to increased bone marrow (BM) vascularity and vascular permeability. (A) Representative sections of murine femurs, with metaphysis and diaphysis regions indicated, from wild-type (WT) C57Bl/6 mice treated with either PBS or the IFN mimetic, pI:C, (5 mg/kg for 24 h). 8 μm sections of femurs were stained with Laminin (green) and mounted in DAPI containing mountant (blue). Scale bar represents 100 mm. (B) Quantification of Laminin positive vasculature in BM sections. Corrected total cell fluorescence is represented as Arbitrary Units (AU). (C) Laminin expression on ECs (Lin– CD45– CD31+) from WT mice treated with either PBS, pI:C (5 mg/kg for 24 h) or IFNα (5x106U/kg for 24 h) was quantified by flow cytometry. (D) Graph representing the vessel diameter in BM from WT mice treated with either PBS or pI:C (5 mg/kg for 24 h) quantified following in vivo labeling with Alexa 633. (E) Evans blue assay to determine vessel leakiness in WT and IFNAR–/– mice treated with PBS (0 h) or pI:C (5 mg/kg for 24 h). Absorbance was measured at 620 nm. Data are representative of 3 or more independent experiments. Data are presented as mean±Standard Error of Mean (SEM) (n≥3). Statistical analysis was performed using unpaired Student t-test (twotailed). ns: not significant, *P<0.001, **P<0.0001.

446

haematologica | 2017; 102(3)

IFNα mediated remodeling of BM endothelial cells

Vascular endothelial growth factor ELISA

Immunofluorescence of bone sections

Mice were injected i.p. with BrdU (18 mg/kg, Sigma) 16 h prior to harvesting the BM. BM cells were stained as described and BrdU staining was carried out using the BD PharmigenTM BrdU Flow Kit according to the manufacturer’s instructions.

8 mm bone sections of frozen femurs were prepared using the Kawamoto tape method.21 In brief, sections were stained overnight using anti-VEGFR2 (Avas12) and anti-Laminin antibodies, and subsequently with Alexa Fluor 488 secondary antibody (Jackson) for 1 h at room temperature. Images were acquired using an LSM710 microscope and were prepared using FIJI software.

Corrected total cell fluorescence was calculated as: integrated density - (area of selected cell x mean fluorescence of background readings).

Bone marrow transplantations

Statistical analysis

ELISA was carried out on BM supernatant from one crushed tibia and femur per mouse according to the manufacturer's instructions (BD Bioscience).

BrdU incorporation assay

3x106 BM cells were diluted in 200 ml PBS and i.v. injected into lethally irradiated (2x500 rad) WT or IFNAR–/– mice.

A

B

E

F

GraphPad Prism® 6.0 was used for statistical analysis and graphical representation of data. Statistical analysis was per-

C

D

G

H

Figure 2. Bone marrow (BM) endothelial cells (EC) are stimulated following interferon α (IFNα) treatment in vivo. (A) FACS analysis of percentage of BrdU positive ECs (Lin– CD45– CD31+) from wild-type (WT) or IFNAR–/– mice treated with either phosphate-buffered saline (PBS) (0 h) or IFNα (5x106U/kg for up to 24 h) and BrdU (18 mg/kg, 16 h). (B) Percentage of Lin- CD45- CD31+ BM cells in BM from WT mice treated with either PBS, the interferon mimetic, pI:C, (5 mg/kg for 24 h), or IFNα (5x106U/kg for 24 h). (C and D) FACS analysis of the expression of ESAM, VE-Cadherin or Laminin on ECs (Lin- CD45- CD31+) from (C) WT mice treated with either PBS or IFNα (5x106U/kg for 24 h) or (D) IFNAR–/– mice treated with either PBS or pI:C (5 mg/kg for 24 h). (E and F) FACS analysis of the expression of ESAM, VECadherin or Laminin on ECs (Lin– CD45– CD31+) from WT mice treated with either (E) pI:C (0-5 mg/kg for 24 h) or (F) IFNα (0-5x106U/kg for 24 h). (G) FACS analysis of the expression of Laminin on ECs (Lin– CD45– CD31+) from WT mice treated with either PBS (0 h) or pI:C (5 mg/kg for 0-120 h). (H) FACS analysis of the expression of VE-Cadherin and ESAM on ECs (Lin– CD45– CD31+) from WT mice treated with either PBS (0 h) or pI:C (5 mg/kg for 0-120 h). Data are representative of 3 or more (A-C) or 2 or more (D-H) independent experiments. Data are presented as mean±Standard Error of Mean (SEM) (n≥3). Statistical analysis was performed using unpaired Student t-test (two-tailed). ns: not significant, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

A

B

Figure 3. Bone marrow (BM) endothelial cells (EC) are not activated by multiple rounds of treatment with the interferon mimetic, pI:C. (A) Experimental design. 1x: treatment with PBS or interferon, pI:C, for 24 h. 8x: treatment with PBS or pI:C every second day for 16 days. Mice were sacrificed 24 h after final treatment (on day 17). (B) FACS analysis of the expression of ESAM, VE-Cadherin and Laminin on ECs (Lin– CD45– CD31+) from wild-type (WT) mice treated with either 1 round or 8 rounds of PBS or pI:C (5 mg/kg). (C) Data are representative of 2 or more independent experiments and data are presented as mean±Standard Error of Mean (SEM) (n≥3). Statistical analysis was performed using unpaired Student t-test (two-tailed). ns (not significant), *P<0.01, **P<0.0001

haematologica | 2017; 102(3)

447

Á.M. Prendergast et al.

formed using unpaired Student t-test (two-tailed). All data are expressed as mean±Standard Error of Mean (SEM) unless otherwise indicated. Statistical significance is indicated in the individual figures.

signaling. Taken together, the observed increase in BM vascularity, Laminin expression on ECs and compromised vessel integrity suggests that acute inflammatory signaling stimulates the vasculature within the BM.

Results Acute inflammatory stress induces increased BM vascularity and vessel permeability To monitor the response of the BM vasculature to inflammatory stress, we treated WT C57Bl/6 mice with a single dose of the IFNα mimetic, pI:C, to mimic an acute inflammatory response. After 24 h, there was a significant increase in BM vasculature in both the diaphysis and metaphysis regions of the BM in pI:C-treated WT mice in comparison to mice treated with PBS, as visualized and quantified by anti-Laminin staining in frozen BM sections (Figure 1A and B). Increased expression of Laminin on ECs upon injection of either pI:C or IFNα was confirmed by FACS analysis (Figure 1C). To quantify the IFNα-mediated increase in vasculature, BM vessels were directly labeled in vivo by i.v. injection of Alexa Fluor 633 phalloidin18 (Figure 1D). Quantification of BM vessel diameter based on Alexa 633 labeling showed that the BM vasculature became enlarged 24 h following pI:C treatment. The integrity of the BM vasculature was quantified using an Evans blue assay, as previously described.19 Evans blue staining in the BM of PBS-treated mice showed basal efflux of macromolecules over the EC vasculature under homeostasis (0 h, Figure 1E). However, 24 h after pI:C treatment, BM Evans blue staining increased 2-fold in WT mice, but not in mice lacking the IFNα receptor (IFNAR–/–) (Figure 1E). This indicated that increased vessel leakage was the result of IFNα

A

B

D

Acute inflammatory stress induces transient BM EC proliferation and activation in vivo To investigate whether the observed vascular expansion was due to an increased activation of ECs, we next analyzed the proliferative and activation status of ECs following IFNα treatment. BrdU incorporation was increased after 4 h in BM ECs (Lin– CD45– CD31+) from mice treated with IFNα in comparison to PBS-treated mice (Figure 2A and B). This suggested an increase in cells which were in S-phase of the cell cycle. IFNα treatment of IFNAR–/– mice confirmed that the increased BrdU incorporation was due to IFN signaling. To determine the activation status of BM ECs, we analyzed the expression of the key cellular junction proteins ESAM, VE-cadherin and Laminin.22 Twentyfour hours after either IFNα or pI:C treatment of mice, expression of ESAM, VE-cadherin and Laminin were upregulated on the surface of WT BM ECs (Figure 2C) but not on IFNAR–/– BM ECs (Figure 2D). Indeed, increased BM EC activation was detectable even from low-dose treatment. Exposure of mice to low-dose pI:C (0.5 mg/kg) (Figure 2E) or IFNα (0.1 Units/kg) (Figure 2F) led to increased expression of activation markers. These data indicated that BM ECs were activated by IFNα or pI:C treatment in an IFNα-dependent manner, and that BM ECs were activated even in response to low doses of treatment. When mice were allowed to recover after treatment, upregulation of Laminin (Figure 2G), VE-Cadherin

C

E

Figure 4. Bone marrow (BM) endothelial cell (EC) activation can be mediated by the interferon (IFN) mimetic, pI:C, stimulation of hematopoietic or non-hematopoietic cells. (A) Experimental design: 3x106 BM cells from either wild-type (WT) (CD45.1) or IFNAR-/- (CD45.2) mice was transplanted into lethally irradiated IIFNAR-/- or WT mice, respectively. Mice were allowed to recover for 90 days (d) prior to treatment with either PBS or pI:C (5 mg/kg for 24 h). (B) FACS analysis of percentage of BrdU positive HSCs (Lin- ckithi CD150+CD48-) from chimeric mice, as described in (A) following treatment with either PBS or pI:C (5 mg/kg for 24 h) and BrdU (18 mg/kg, 16 h). (C-E) FACS analysis of the expression of (C) Laminin (D) VE-Cadherin and (E) ESAM on ECs (Lin- CD45- CD31+) from chimeric mice, as described in (A) following treatment with either PBS or pI:C (5 mg/kg for 24 h). Data are representative of 3 or more (B) and 2 or more (C-E) independent experiments, and data are presented as mean±Standard Error of Mean (SEM) (n=≥3). Statistical analysis was performed using unpaired Student t-test (two-tailed). ns: not significant, *P<0.05, **P<0.01, ***P<0.0001.

448

haematologica | 2017; 102(3)

IFNα mediated remodeling of BM endothelial cells

and ESAM (Figure 2H) returned to homeostatic levels after 96 h. This indicated that, similar to the response of HSCs,16 the rapid response of ECs to IFNα treatment is transient. Thus, EC proliferation and activation is modulated following acute IFNα treatment. Proliferation and activation are dependent on expression on the IFNα receptor. Taken together with increased vascularity and compromised BM vessel integrity (Figure 1), EC proliferation and activation indicate enhanced BM vessel remodeling occurs.

To test whether chronic IFNα treatment could lead to an accumulation or an exhaustion of BM EC activation, as previously described,23 mice were treated with pI:C every second day for a total of 16 days (Figure 3A). In contrast to the increase in activation markers upon 1 injection (1x), BM ECs expressed homeostatic levels of ESAM, VECadherin and Laminin after multiple injections with pI:C (8x) (Figure 3B). Thus, BM ECs were not continually activated by multiple treatments of pI:C. These data are indicative of the contrast in the response of BM ECs, as well as HSCs, to acute and chronic IFNα treatment.23 This supports the hypothesis of a rapid, acute stimulation of BM ECs following inflammation.

BM EC stimulation by IFNα occurs via both hematopoietic and non-hematopoietic pathways

IFNα has been reported to have heterogenous effects on ECs.24-28 We tested whether the observed stimulatory effect of IFNα on BM ECs was directly or indirectly mediated by IFNα.16 BM chimeras were created where either

A

WT or IFNAR–/– BM cells were transplanted into lethally irradiated IFNAR–/– or WT host mice, respectively. Thus, either the BM (IFNAR–/– BM into a WT niche) or the niche (WT BM into an IFNAR–/– niche) can no longer directly respond directly to IFNα in these mice (Figure 4A). In agreement with our previous data,16 WT HSCs in recipient IFNAR–/– mice (WT BM into an IFNAR–/– niche) proliferated in response to pI:C treatment; IFNAR–/– HSCs in WT recipient mice (IFNAR–/– BM into a WT niche) did not (Figure 4B). This indicated that the response of HSCs to pI:C was dependent on the expression of the IFNα receptor (IFNAR) on HSCs, not on niche cells. In contrast, Laminin (Figure 4C), VE-Cadherin (Figure 4D) and ESAM (Figure 4F) expression was increased on IFNAR–/– ECs with a WT hematopoietic system present (WT BM into an IFNAR–/– niche) and on WT ECs with a hematopoietic system lacking the IFNα receptor (IFNAR–/– BM into a WT niche).

These data indicated that BM ECs can be stimulated by IFNα via a non-hematopoietic effect of IFNα directly on the BM ECs as well as an indirect effect of IFNα via signaling from IFNα-stimulated hematopoietic cells in the BM.

pI:C treatment induces VEGF production and signaling in the BM Bone marrow chimera experiments suggested that IFNα-stimulated hematopoietic cells may produce factors which can stimulate BM ECs (Figure 4C and E). Thus, we next tested the activity of known mediators of EC activation in this setting. Platelet activation and VEGF signaling are fundamental mediators of EC activation during inflam-

B

D

C

E

Figure 5. The IFN mimetic-, pI:C, mediated bone marrow (BM) endothelial cell (EC) stimulation is not affected by platelet abrogation. (A) Experimental design: mice were treated with the anti-platelet antibody R300 (2 mg/g) and either PBS or the IFN mimetic, pI:C, (5 mg/kg) for 24 h. (B) Platelet counts in the peripheral blood of wild-type (WT) mice following treatment as outlined in (A). (C-E) FACS analysis of the expression of (C) ESAM (D) VE-Cadherin and (E) Laminin on ECs (Lin- CD45CD31+) from WT mice treated as outlined in (A). Data are representative of 3 or more independent experiments, and are presented as mean±Standard Error of Mean (SEM) (n≥3). Statistical analysis was performed using unpaired Student t-test (two-tailed). ns: not significant, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

haematologica | 2017; 102(3)

449

Á.M. Prendergast et al. mation,29,30 and megakaryocytes, which give rise to platelets, regulate BM HSC quiescence.31 To test pI:Cmediated EC activation in the absence of platelets, mice were treated with a platelet depletion antibody, antiGPIbα (CD42b) antibody (R300), prior to pI:C administration (Figure 5A). Platelet levels were reduced upon R300 treatment (Figure 5B); however, EC activation markers were not altered following platelet depletion (Figure 5CE). This suggested that platelet activation was not central to IFNα-induced BM EC stimulation. To investigate whether VEGF was regulated by acute pI:C treatment, we carried out a BM ELISA and intracellular staining for VEGF in BM cell types following pI:C treatment. After 24 h there was a significant increase in secreted VEGF in the BM supernatant of pI:C treated mice (Figure 6A). Intracellular VEGF did not increase in BM ECs (Figure 6B and C).

However, there was a significant increase in intracellular VEGF in hematopoietic cells, including progenitors and HSCs (Lin– ckithi CD150+ CD48– CD34–), both after pI:C (Figure 6B) and IFNα (Figure 6C) treatment. Consistent with the transient increase in activation of BM ECs (Figure 2G and H), the increase in intracellular VEGF levels in HSCs upon pI:C treatment was also transient. VEGF production peaked after 24 h and returned to homeostatic levels 72 h after treatment (Figure 6D). In addition, VEGF production was increased in WT HSCs in recipient IFNAR–/– mice (WT BM into an IFNAR–/– niche) and in IFNAR–/– HSCs in WT recipient mice (IFNAR–/– BM into a WT niche) following pI:C treatment (Figure 6E). These data suggested that production of VEGF production in HSCs was stimulated both directly and indirectly by pI:C treatment. To assess whether VEGF signaling was consequently active in the BM,6 the expression of the VEGF receptor, VEGFR2, was analyzed in pI:C treated mice as an indicator of VEGF signaling. VEGFR2 expression was increased in BM sec-

A

F

B

tions (Figure 6F). In addition, VEGFR2 was up-regulated on the surface of ECs, but not on HSCs, in response to both pI:C and IFNα (Figure 6G-I). This suggested that VEGF signaling was active in ECs but differs between ECs and HSCs at this time point. As with VEGF production (Figure 6D), the increase in VEGFR2 expression on BM ECs was transient, peaking 24 h after pI:C treatment (Figure 6J). This time point correlated with the peak of increased expression of activation markers on BM ECs (Figure 2G and H). Taken together, these data indicate that pI:C and IFNα treatment leads to an increase in VEGF production and signaling in the BM. In addition, they suggest that, upon pI:C treatment of mice, BM ECs responded to VEGF, which is produced by other BM cell types including HSCs in response to pI:C. Thus, VEGF may be a mediating factor in the activation of BM ECs by IFNα-stimulated hematopoietic cells.

IFNα-mediated stimulation of ECs in vivo is facilitated

by VEGF

To test whether VEGF signaling was involved in BM EC activation, mice were co-treated with pI:C and the VEGF binding antibody, Avastin (Figure 7A). Avastin treatment did not affect the expression of VEGF or VEGFR2 in comparison to PBS-treated mice (Figure 7B-D). While the expression level of VEGF in ECs was unchanged (Figure 7B), pI:C-induced VEGF expression in HSCs (LK SLAM CD34–) was significantly reduced by co-treatment with Avastin (Figure 7C). In addition, the pI:C-induced expression of VEGFR2 on BM ECs was reduced upon Avastin co-treatment (Figure 7D). In contrast, Avastin treatment did not affect pI:C-mediated proliferation of HSCs (Figure 7E). This suggests that co-treatment with Avastin leads to reduced pI:C-mediated VEGF signaling in the BM. To assess the effect of diminished VEGF signaling on pI:C-mediated EC

C

G

D

H

I

E

J

Figure 6. Bone marrow (BM) vascular endothelial growth factor (VEGF) is modulated by the interferon mimetic, pI:C. (A) ELISA of BM VEGF from wild-type (WT) mice treated with PBS or the IFN mimetic, pI:C, (5 mg/kg for 24 h). (B and C) FACS analysis of intra-cellular staining of VEGF in indicated BM cell types after treatment with either PBS, (B) pI:C (5 mg/kg for 24 h), or (C) IFNα (5x106U/kg for 24 h). Data are presented as fold change in mean fluorescence intensity (MFI). (D) FACS analysis of intra-cellular staining of VEGF in hematopoietic stem cells (HSCs) (Lin–ckithi CD150+CD48– CD34–) after treatment with either PBS (0 h) or pI:C (5 mg/kg, 0-120 h). Data are presented as fold change in MFI. (E) FACS analysis of intra-cellular staining of VEGF in HSCs (Lin– ckithi CD150+CD48–CD34–) from chimeric mice, as described in Figure 4A, following treatment with either PBS or pI:C (5 mg/kg for 24 h). Data are presented as fold change in MFI. (F) Representative bone sections of VEGFR2 expression (Red) and Laminin (Green) after treatment with either PBS or pI:C (5 mg/kg for 24 h). (G) FACS analysis of VEGFR2 expression on (F) Lamininhigh/low ECs (Lin– CD45– CD31+, Lamininhigh/low) treated with PBS or pI:C (5 mg/kg for 24 h). (H and I) FACS analysis of VEGFR2 expression on ECs (Lin- CD45CD31+) and HSCs (Lin– ckithi CD150+CD48– CD34–) from mice treated with either (H) PBS or pI:C (mg/kg for 24 h) or (I) PBS or IFNα (5x106U/kg for 24 h). (J) FACS analysis of the expression of VEGFR2 on ECs from WT mice treated with either PBS or pI:C (5 mg/kg for up to 120 h). Data are representative of 3 or more independent experiments (A, B, D-G) and 2 or more (C and H) independent experiments, and are presented as mean±Standard Error of Mean (SEM) (n≥3). Statistical analysis was performed using unpaired Student t-test (two-tailed). ns: not significant, *P<0.05, **P<0.01, ***P<0.001.

450

haematologica | 2017; 102(3)

IFNα mediated remodeling of BM endothelial cells

activation, the expression of EC activation markers following Avastin treatment was analyzed. While the increased expression of ESAM was not affected, the pI:C-induced expression of both VE-Cadherin and Laminin was significantly diminished upon co-treatment with Avastin (Figure 7F and G). This indicated that VEGF was involved, in part, in pI:C-mediated BM EC activation. Taken together, these data demonstrate that VEGF signaling is important for the stimulation of BM ECs following pI:C treatment.

A

E

B

Discussion Bone marrow ECs are a primary defense against infection, so understanding the effect of inflammation on the BM vasculature is critical. We demonstrate for the first time a rapid, transient activation of the BM vasculature in response to acute inflammatory signaling. We find that there is a direct and indirect effect of IFNα signaling in the BM on ECs, mediated by an activated hematopoietic sys-

C

F

D

G

Figure 7. The IFN mimetic-, pI:C, mediated stimulation of bone marrow (BM) endothelial cells (ECs) is mediated by vascular endothelial growth factor (VEGF) signaling. (A) Experimental design: mice were treated with Avastin (2.5 mg/kg) and either PBS or the interferon mimetic, pI:C, (5 mg/kg) for 24 h. (B and C) FACS analysis of intra-cellular staining of VEGF in (B) ECs (Lin- CD45- CD31+) and (C) HSCs (Lin- ckithi CD150+CD48- CD34-) from wild-type (WT) mice treated as described in (A).

Data are presented as fold change in mean fluorescence intensity (MFI). (D) FACS analysis of VEGFR2 expression on ECs (Lin- CD45- CD31+) from WT mice treated as described in (A). (E) FACS analysis of percentage of BrdU positive HSCs (Lin- ckithi CD150+ CD48-) from WT mice treated as described in (A) and BrdU (18 mg/kg, 16 h). (F) Representative bone marrow section (Laminin in green, DAPI in blue) from WT mice treated as described in (A). (G) FACS analysis of the expression of Laminin, VE-Cadherin and ESAM on ECs (Lin- CD45- CD31+) from WT mice treated as described in (A). Data are representative of 3 or more (A-D, E-G) and 2 or more (E) independent experiments and are presented as mean±Standard Error of Mean (SEM) (n≥3). Statistical analysis was performed using unpaired Student t-test (twotailed). ns (not significant), *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 8. Acute IFNα-stimulation of BM ECs is mediated

by VEGF signaling in both hematopoietic and nonhematopoietic cells. Model depicting BM vasculature remodeling following stimulation of BM hematopoietic or non-hematopoietic cells by acute interferon α (IFNα) treatment.

haematologica | 2017; 102(3)

451

Á.M. Prendergast et al.

tem. Our data suggest a role for VEGF signaling in the BM in IFNα-mediated BM EC activation. This rapid, transient effect may be an emergency response to inflammatory signaling, coming from the hematopoietic system and affecting BM EC niche cells. This response may in turn facilitate the maintenance of BM homeostasis. In this acute setting, the vasculature may be rapidly 'primed' or activated, likely in anticipation of greater insult. In contrast to treatment with an isolated cytokine, initial inflammatory signaling during a full infection is followed by production of other cytokines, and stimulation of additional signaling.32 It is, therefore, likely that the response elicited by an isolated cytokine differs to that elicited by a full infection, particularly with regard to continuation of signaling and recovery from the initial stimulation.

We have found that acute pI:C exposure results in a transient expansion of the vasculature in the BM after 24 h. The integrity of this expanded vasculature was compromised (Figure 1). Increased BM vascular permeability is in keeping with an increase in the transit of immune cells or leakage of plasma during an inflammatory response.33 Acute pI:C treatment induces production of IFNα and mimics an acute inflammatory response.16 Acute IFNα treatment in vitro may reduce apoptosis of endothelial cell lines.27 Whether IFNα has a similar effect on apoptosis of ECs in vivo is unknown.

Reduced BM EC apoptosis, mediated by IFNα during inflammatory stress, may influence vessel integrity.

Permeability of different types of BM vasculature is distinct.7 Therefore, the integrity of specific BM vessels following an acute inflammatory response is likely influenced by vessel permeability under homeostasis. This may be a mechanism to maintain BM homeostasis during the early stages of an inflammatory response.

The effects of IFNα on ECs and other hematopoietic cell types in vivo are in contrast to the in vitro situation26,28,34-36 where cells are isolated from their niche environment. In the BM niche, IFNα treatment rapidly and efficiently stimulates HSC proliferation in vivo, whereas in vitro, HSCs do not undergo increased proliferation.16 We have shown BM EC activation in response to acute IFNα exposure in vivo (Figures 2 and 3), in contrast to the described effect of IFNα on ECs in vitro. In addition to the differential effect of IFNα in vitro and in vivo, IFNα has also been described as being both stimulatory and inhibitory with regard to VEGF signaling.37-39 We find that VEGF production was increased in the BM in response to pI:C (Figure 6). BM EC activation following acute pI:C treatment was dependent in part on VEGF signaling (Figure 7). pI:C mediated HSC proliferation was not affected by inhibition of VEGF, using Avastin treatment (Figure 7E). This corresponds with previous data showing that HSCs are directly activated by IFNα.16 An IFNα-mediated increase of VEGF is in contrast to previous studies, which suggest that VEGF is suppressed by long-term IFN treatment or in combination with chemotherapy.38-41 Together, these data highlight the contrast between chronic versus acute IFN treatments, and between in vitro and in vivo cytokine responses. As inflammatory stress is a complex signaling cascade, the in vivo cytokine response in mice is, therefore, more reflective of the inflammatory response than the in vitro situation.

In the BM niche, signaling between different cell types is imperative for maintenance of cellular homeostasis and a rapid response to inflammation. ECs and the vasculature itself have been ascribed many functions in the BM as regulators of HSCs.8,10,13,18,42-45 Furthermore, there are distinct vessel 452

subtypes within the BM which differentially regulate hematopoiesis.7 In addition, Notch signaling in BM ECs has been shown to expand the HSC niche in vivo.8 As Notch signaling is involved in the inflammatory response of ECs,46 these data may further support a role for inflammatory signaling in BM niche remodeling. Furthermore, IFNα does not lead to mobilization of hematopoietic stem progenitor cells HSPCs that are not in the BM.16,35 However, the percentage of HSCs found within 20 mm of arterioles in sternal BM is significantly reduced following treatment with pI:C in comparison to the control.18 These data suggest that the location of HSCs in the BM is affected by pI:C. Relocating HSPCs can potentially affect many different BM cell types, and the BM vasculature, following treatment with pI:C. Whether pI:Cstimulated BM vasculature affects the location of HSPCs within the BM remains unclear. Many cytokines produced by hematopoietic cells, such as EPO and GCSF, have been shown to affect specific EC functions in isolation.47,48 However, signaling from the hematopoietic system to the ECs in the BM niche has not been examined in detail. To address this question, we have created BM chimeras in which only either hematopoietic cells or niche cells respond directly to IFNα. Using this system, we have demonstrated that inflammatory signaling from an activated hematopoietic system can affect the BM vasculature (Figure 4). Platelets, which can induce EC stimulation,30 do not play a major role in BM EC stimulation in this setting (Figure 5). As the inflammatory response is complex, these data do not exclude the possibility that IFNα-mediated signaling from other cell types within the BM, in addition to hematopoietic cells, may be involved in this response. Further, these data cannot discriminate within the heterogeneous population of BM ECs. Whether there is crosstalk between pI:C-stimulated BM ECs and BM HSPCs within this context, remains to be elucidated.

Importantly, we demonstrate that BM hematopoietic cells, including HSCs, are implicated in the pI:C-mediated BM EC stimulation, and thus in vasculature remodeling. This provides evidence for crosstalk between BM HSPCs and ECs under inflammatory stress conditions (Figure 8).

Our findings demonstrate a novel response of the BM vasculature to primary inflammatory signaling. We have revealed potential crosstalk between the hematopoietic system and the BM vasculature under inflammatory stress.

The transient activation of the BM vasculature represents a novel, emergency response of the BM stem cell niche to inflammation. Future studies will likely uncover other emergency situations in which HSCs influence the BM niche. Understanding this critical cellular relationship under stress conditions such as infection, but also chemotherapy, may reveal new mechanisms for the maintenance and recovery of BM homeostasis.

Acknowledgments The authors would like to thank Drs M. Milsom, H. Augustin and A. Trumpp for helpful discussions, A. Atzberger and Dr S. Schmitt from the DKFZ Flow Cytometry Facility, M. Brom, Dr F. Bestvater and Dr D. Krunic from the DKFZ Imaging Core Facility, the DKFZ Animal Laboratory Services for their expertise and assistance, and L. Prendergast for critical reading of the manuscript.

Funding This work was supported by FOR2033 NicHem and SFB873, both funded by the Deutsche Forschungsgemeinschaft, and the Dietmar Hopp Foundation. haematologica | 2017; 102(3)

IFNÎą mediated remodeling of BM endothelial cells

References 18.

1. Orozco AS, Zhou X, Filler SG. Mechanisms of the proinflammatory response of endothelial cells to Candida albicans infection. Infect Immun. 2000;68(3):1134-1141.

2. Bazzoni G, Dejana E. Endothelial cell-tocell junctions: molecular organization and role in vascular homeostasis. Physiol Rev. 2004;84(3):869-901.

3. Pober JS. Endothelial activation: intracellular signaling pathways. Arthritis Res. 2002; 4(Suppl 3):S109-16.

4. Granger DN, Rodrigues SF, Yildirim A, Senchenkova EY. Microvascular responses to cardiovascular risk factors.

Microcirculation. 2010;17(3):192-205.

5. Muller WA. Getting leukocytes to the site of inflammation. Vet Pathol. 2013;50(1):722.

6. Jones N, Iljin K, Dumont DJ, Alitalo K. Tie receptors: new modulators of angiogenic and lymphangiogenic responses. Nat Rev Mol Cell Biol. 2001;2(4):257-267.

7. Itkin T, Gur-Cohen S, Spencer JA, et al.

Distinct bone marrow blood vessels differentially regulate haematopoiesis. Nature.

2016;532(7599):323-328.

8. Kusumbe AP, Ramasamy SK, Itkin T, et al.

Age-dependent modulation of vascular niches for haematopoietic stem cells.

Nature. 2016;532(7599):380-384.

9. Cardier JE, Barbera-Guillem E. Extramedullary hematopoiesis in the adult mouse liver is associated with specific hepatic sinusoidal endothelial cells.

Hepatology. 1997;26(1):165-175.

10. Ding L, Saunders TL, Enikolopov G, Morrison SJ. Endothelial and perivascular cells maintain haematopoietic stem cells.

Nature. 2012;481(7382):457-462.

11. Gattazzo F, Urciuolo A, Bonaldo P. Extracellular matrix: a dynamic microenvironment for stem cell niche. Biochim Biophys Acta. 2014;1840(8):2506-2519.

12. Li W, Johnson SA, Shelley WC, Yoder MC. Hematopoietic stem cell repopulating ability can be maintained in vitro by some primary endothelial cells. Exp Hematol. 2004; 32(12):1226-1237.

13. Morrison SJ, Scadden DT. The bone marrow niche for haematopoietic stem cells.

Nature. 2014;505(7483):327-334.

14. Avecilla ST, Hattori K, Heissig B, et al.

Chemokine-mediated interaction of hematopoietic progenitors with the bone marrow vascular niche is required for thrombopoiesis. Nat Med. 2004;10(1):64-71.

15. Zhao YD, Huang X, Yi F, et al. Endothelial FoxM1 mediates bone marrow progenitor cell-induced vascular repair and resolution of inflammation following inflammatory lung injury. Stem Cells. 2014;32(7):18551864.

16. Essers MA, Offner S, Blanco-Bose WE, et al.

IFNalpha activates dormant haematopoietic stem cells in vivo. Nature.

2009;458(7240):904-908.

17. Muller U, Steinhoff U, Reis LF, et al.

Functional role of type I and type II interfer-

haematologica | 2017; 102(3)

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

ons in antiviral defense. Science. 1994; 264(5167):1918-1921.

Kunisaki Y, Bruns I, Scheiermann C, et al.

Arteriolar niches maintain haematopoietic stem cell quiescence. Nature. 2013; 502(7473):637-643.

Radu M, Chernoff J. An in vivo assay to test blood vessel permeability. J Vis Exp. 2013;73):e50062.

Haas S, Hansson J, Klimmeck D, et al.

Inflammation-Induced Emergency Megakaryopoiesis Driven by Hematopoietic Stem Cell-like Megakaryocyte Progenitors.

Cell Stem Cell. 2015;17(4):422-434.

Kawamoto T. Use of a new adhesive film for the preparation of multi-purpose freshfrozen sections from hard tissues, wholeanimals, insects and plants. Arch Histol Cytol. 2003;66(2):123-143.

Bentley K, Franco CA, Philippides A, et al.

The role of differential VE-cadherin dynamics in cell rearrangement during angiogenesis. Nat Cell Biol. 2014;16(4):309321.

Pietras EM, Lakshminarasimhan R, Techner JM, et al. Re-entry into quiescence protects hematopoietic stem cells from the killing effect of chronic exposure to type I interferons. J Exp Med. 2014;211(2):245-262.

Wada H, Nagano H, Yamamoto H, et al.

Combination of interferon-alpha and 5-fluorouracil inhibits endothelial cell growth directly and by regulation of angiogenic factors released by tumor cells. BMC Cancer. 2009;9:361.

Sgonc R, Fuerhapter C, Boeck G, Swerlick R, Fritsch P, Sepp N. Induction of apoptosis in human dermal microvascular endothelial cells and infantile hemangiomas by interferon-alpha. Int Arch Allergy Immunol.

1998;117(3):209-214.

Reynolds JA, Ray DW, Zeef LA, O'Neill T, Bruce IN, Alexander MY. The effect of type 1 IFN on human aortic endothelial cell function in vitro: relevance to systemic lupus erythematosus. J Interferon Cytokine Res. 2014;34(5):404-412.

Pammer J, Reinisch C, Birner P, Pogoda K, Sturzl M, Tschachler E. Interferon-alpha prevents apoptosis of endothelial cells after short-term exposure but induces replicative senescence after continuous stimulation.

Lab Invest. 2006;86(10):997-1007.

Cheng X, Liu Y, Chu H, Kao HY. Promyelocytic leukemia protein (PML) regulates endothelial cell network formation and migration in response to tumor necrosis factor alpha (TNFalpha) and interferon alpha (IFNalpha). J Biol Chem. 2012; 287(28):23356-23367.

Lee S, Chen TT, Barber CL, et al. Autocrine VEGF signaling is required for vascular homeostasis. Cell. 2007;130(4):691-703.

Stokes KY, Granger DN. Platelets: a critical link between inflammation and microvascular dysfunction. J Physiol. 2012;590(Pt 5):1023-1034.

Bruns I, Lucas D, Pinho S, et al.

Megakaryocytes regulate hematopoietic stem cell quiescence through CXCL4 secretion. Nat Med. 2014;20(11):1315-1320.

32. Mogensen TH. Pathogen recognition and inflammatory signaling in innate immune defenses. Clin Microbiol Rev. 2009; 22(2):240-273, Table of Contents.

33. Rajendran P, Rengarajan T, Thangavel J, et al. The vascular endothelium and human diseases. Int J Biol Sci. 2013;9(10):10571069.

34. Borden EC, Sen GC, Uze G, et al.

Interferons at age 50: past, current and future impact on biomedicine. Nat Rev Drug Discov. 2007;6(12):975-990.

35. Essers MA, Trumpp A. Targeting leukemic stem cells by breaking their dormancy. Mol Oncol. 2010;4(5):443-450.

36. Rosa R, Monteleone F, Zambrano N, Bianco R. In vitro and in vivo models for analysis of resistance to anticancer molecular therapies. Curr Med Chem. 2014; 21(14):15951606.

37. Bauer EM, Zheng H, Lotze MT, Bauer PM. Recombinant human interferon alpha 2b prevents and reverses experimental pulmonary hypertension. PLoS One. 2014; 9(5):e96720.

38. Merkle M, Ribeiro A, Belling F, et al.

Response of VEGF to activation of viral receptors and TNFalpha in human mesangial cells. Mol Cell Biochem. 2012;370(12):151-161.

39. von Marschall Z, Scholz A, Cramer T, et al.

Effects of interferon alpha on vascular endothelial growth factor gene transcription and tumor angiogenesis. J Natl Cancer Inst. 2003;95(6):437-448.

40. Gambara G, Desideri M, Stoppacciaro A, et al. TLR3 engagement induces IRF-3-dependent apoptosis in androgen-sensitive prostate cancer cells and inhibits tumour growth in vivo. J Cell Mol Med. 2015; 19(2):327-339.

41. Raig ET, Jones NB, Varker KA, et al. VEGF secretion is inhibited by interferon-alpha in several melanoma cell lines. J Interferon Cytokine Res. 2008;28(9):553-561.

42. Boulais PE, Frenette PS. Making sense of hematopoietic stem cell niches. Blood.

2015;125(17):2621-2629.

43. Ding L, Morrison SJ. Haematopoietic stem cells and early lymphoid progenitors occupy distinct bone marrow niches. Nature.

2013;495(7440):231-235.

44. Kunisaki Y, Frenette PS. Influences of vascular niches on hematopoietic stem cell fate. Int J Hematol. 2014;99(6):699-705.

45. Lamagna C, Bergers G. The bone marrow constitutes a reservoir of pericyte progenitors. J Leukoc Biol. 2006;80(4):677-681.

46. Quillard T, Charreau B. Impact of notch signaling on inflammatory responses in cardiovascular disorders. Int J Mol Sci. 2013; 14(4):6863-6888.

47. Ribatti D, Vacca A, De Falco G, Ria R, Roncali L, Dammacco F. Role of hematopoietic growth factors in angiogenesis. Acta Haematol. 2001;106(4):157-161.

48. Ribatti D, Vacca A, Roncali L, Dammacco F. Hematopoiesis and angiogenesis: a link between two apparently independent processes. J Hematother Stem Cell Res. 2000;9(1):13-19.

453

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Red Cell Biology & Its Disorders

Ferrata Storti Foundation

Iron-heme-Bach1 axis is involved in erythroblast adaptation to iron deficiency Masahiro Kobayashi,1,2 Hiroki Kato,1,2 Hiroshi Hada,1 Ari Itoh-Nakadai,1,3,4 Tohru Fujiwara,2,5 Akihiko Muto,1,3 Yukihiro Inoguchi,6 Kenji Ichiyanagi,6,7 Wataru Hojo,8 Naohisa Tomosugi,9 Hiroyuki Sasaki,3,6 Hideo Harigae2,5 and Kazuhiko Igarashi1,3,5 MK and HK contributed equally to this work.

Department of Biochemistry, Tohoku University Graduate School of Medicine, Sendai; Department of Hematology and Rheumatology, Tohoku University Graduate School of Medicine, Sendai; 3AMED-CREST, Japan Agency for Medical Research and Development, Tokyo; 4Department of Experimental Immunology, Institute of Development, Aging and Cancer, Tohoku University; 5Center for Regulatory Epigenome and Diseases, Tohoku University Graduate School of Medicine, Sendai; 6Division of Epigenomics and Development, Medical Institute of Bioregulation, Kyushu University, Fukuoka; 7 Laboratory of Genome and Epigenome Dynamics, Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Aichi; 8 Department of Research and Development, Cellspect Co. Ltd., Morioka and 9Division of Systems Bioscience for Drug Discovery, Medical Research Institute, Kanazawa Medical University, Ishikawa, Japan 1

Haematologica 2017 Volume 102(3):454-465

2

ABSTRACT

I

Correspondence: igarashi@med.tohoku.ac.jp

Received: June 15, 2016.

Accepted: December 1, 2016.

Pre-published: December 7, 2016 doi:10.3324/haematol.2016.151043 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/454 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

454

ron plays the central role in oxygen transport by erythrocytes as a constituent of heme and hemoglobin. The importance of iron and heme is also to be found in their regulatory roles during erythroblast maturation. The transcription factor Bach1 may be involved in their regulatory roles since it is deactivated by direct binding of heme. To address whether Bach1 is involved in the responses of erythroblasts to iron status, low iron conditions that induced severe iron deficiency in mice were established.

Under iron deficiency, extensive gene expression changes and mitophagy disorder were induced during maturation of erythroblasts. Bach1–/– mice showed more severe iron deficiency anemia in the developmental phase of mice and a retarded recovery once iron was replenished when compared with wild-type mice. In the absence of Bach1, the expression of globin genes and Hmox1 (encoding heme oxygenase-1) was de-repressed in erythroblasts under iron deficiency, suggesting that Bach1 represses these genes in erythroblasts under iron deficiency to balance the levels of heme and globin. Moreover, an increase in genome-wide DNA methylation was observed in erythroblasts of Bach1–/– mice under iron deficiency.

These findings reveal the principle role of iron as a regulator of gene expression in erythroblast maturation and suggest that the iron-hemeBach1 axis is important for a proper adaptation of erythroblast to iron deficiency to avoid toxic aggregates of non-heme globin.

Introduction In order to transport oxygen to peripheral tissues, abundant synthesis of the oxygen carrier hemoglobin is required during erythrocyte maturation. Since iron is essential for hemoglobin synthesis, erythrocyte consumes a major part of the body iron in order to synthesize a sufficient amount of hemoglobin.1 As such, elaborate mechanisms are required to identify cellular iron status in order to maintain homeostasis. Several factors have been proposed to function as the sensors of iron, including iron regulatory element (IRE)/iron regulatory protein (IRP) system and ubiquitin E3 ligase FBXL5.2,3 Several lines of evidence suggest that iron affects the gene expression profile in erythroid cells. For instance, it has been reported that iron deficiency (ID) in pregnant mother mice affects the gene expression in definitive erythroid cells in the fetal livers.4 Since iron is a co-factor for enzymes involved in epigenetic regulation, such as the Tet family of DNA demethylase and the Jumonji haematologica | 2017; 102(3)

Role for iron and Bach1 in erythroblast maturation

family of histone demethylases,5,6 iron may directly affect epigenetic regulations. Nevertheless, it has remained unclear how gene expressions in erythroblast in the adult body are affected by iron status.

While heme and globin are central to the function of erythrocyte, excessive levels of globin lead to proteotoxicity, as evidenced by β-thalassemia.7 Non-heme globins (free globins) are easily oxidized and form toxic aggregates.8,9 Therefore there must be a strict regulatory system to balance the intracellular heme and globin. It was previously reported that heme-regulated eIF2a kinase (HRI) represses the translation of globin to maintain their balance, and this function is required for erythrocyte development under ID.4,10 However, such a mechanism that regulates the intracellular heme and globin balance at the transcriptional level has remained elusive.

Bach1 is a transcription factor whose functions are repressed by heme.11-13 Since Bach1 transcriptionally represses globin genes and Hmox1 [the latter encoding heme oxygenase-1 (HO-1)] in erythroid cell lines,14-16 Bach1 might have an important role in erythrocyte development as a balancer of intracellular heme and globins. However, show apparently unperturbed Bach1–/– mice erythropoiesis.16 As heme is abundantly synthesized in erythrocytes, Bach1 knockout may not exhibit an apparent phenotype in an iron sufficient status. Namely, when the heme content is reduced in erythroblasts, Bach1 is presumed to carry out its regulatory roles. This led us to consider that ID may involve alterations in the gene expression regulated by Bach1.

To address these issues, we established low iron conditions (LIC) that induced severe ID in mice. Under LIC, transcriptome and DNA methylome analyses in erythroblasts were performed to reveal the regulatory function of iron. Transcriptomic analyses revealed that LIC inhibited mitophagy in erythroblasts, suggesting a connection between iron and mitophagy in erythroblasts. While Bach1–/– mice showed iron deficiency anemia (IDA) similar to wild-type mice when LIC was initiated after weaning, Bach1–/– mice were more vulnerable to LIC than wild-type mice when LIC was initiated during the embryonic stage.

In addition, DNA methylation was increased in Bach1–/– erythroblasts under LIC and Bach1–/– mice showed slower recovery from IDA. These results indicate that iron plays extremely important roles in tuning the gene expressions in developing erythroblasts and point to a crucial function for Bach1 in erythroblasts under ID. These observations give us new insight into the pathophysiology and molecular biology of IDA, one of the most common disorders affecting the global population.17,18

Methods Mice All mice were from the C57BL/6J genetic background and housed in specific pathogen-free (SPF) conditions. Bach1–/– mice were obtained as described previously.16 All experiments were approved by the Institutional Animal Care and Use Committee of the Tohoku University Environmental and Safety Committee.

Diet-induced iron deficiency In order to achieve a state of ID, mice were fed a low iron diet (LID) containing 3.6 ppm Fe. As control, CLEA Rodent Diet CE-2, which contains 310.2 ppm Fe, was used as normal diet (ND). Both haematologica | 2017; 102(3)

of these were purchased from CLEA Japan, Inc. Diets were provided ad libitum. In the LID-after weaning (LID-W) regimen, mice were fed LID starting from three weeks upon weaning. In the LIDdevelopment and after weaning (LID-DW) regimen, LID was initiated from 0.5 days post coitum via mothers, and newborn mice were also fed LID upon weaning till 12 weeks of age. Technical details of hematologic and serum biochemistry analyses are described in the Online Supplementary Methods.

Flow cytometry and cell sorting analyses Bone marrow (BM) cells and peripheral blood (PB) cells were collected from the bilateral femur and tibia or retro-orbital vein, respectively. Cells were suspended in staining buffer (PBS with 3% FBS) and stained with fluorescent-conjugated antibodies specific for CD71 (clone: C2 or R17217) and Ter119 (clone: Ter119) (BD Bioscience, eBioscience). Mitochondria staining was performed using the Mitotracker Green FM (Mito-G; Thermo Fisher Scientific) according to the manufacturer’s instructions.

Intracellular hydrogen peroxide was measured using 7’-dichlorodihydrofluorescein diacetate (DCFDA; Sigma-Aldrich) as previously described.19 Cells were analyzed and sorted by FACS Aria II (BD Bioscience) according to the manufacturer’s instructions. Data were analyzed using the FlowJo software program (TreeStar).

Quantitative PCR Quantitative PCR (qPCR) analysis was performed as previously described20 with only slight modifications (Online Supplementary Methods and Table S1). Levels of mRNAs were normalized to the β-actin levels.

Chromatin immunoprecipitation A qPCR-based chromatin immunoprecipitation (ChIP) analysis was performed as previously described21 with only slight modifications (Online Supplementary Methods). Sequences of the primers are described in Online Supplementary Table S2.

Expression profiling by microarray analysis and DNA methylation profiling by post-bisulfite conversion adaptor tagging method Technical details of the microarray analysis and post-bisulfite conversion adaptor tagging (PBAT) method are described in the Online Supplementary Methods. The microarray data and the sequencing data have been deposited to Gene Expression Omnibus (GEO) under the accession number, GSE77694 and GSE78955, respectively.

Statistical analysis Data were analyzed by two-tailed t-test using Student t-test or Welch t-test, as appropriate. P<0.05 was considered statistically significant. The numbers of mice analyzed in each experiment are described in the figure legends. For the volcano plot analysis, moderate t-test with Benjamin-Hochberg multiple testing was performed to calculate statistical significance, and P<0.05 and 2-fold change were used as cut-off value.22,23 Calculations of significance in GSEA were performed using the GSEA software.24,25

Results LID induces severe IDA In LID-W, wild-type mice started to exhibit anemia after three weeks and their blood parameters worsened according to the length of time on LID. Conversely, platelet counts and the serum total iron binding capacity were significantly increased (Figure 1A and B). Platelet counts were unaccountably high after seven weeks with LID. There 455

M. Kobayashi et al.

was no evidence of hemolysis according to total bilirubin concentrations (Figure 1C). These observations were similar to the features of human IDA.18 Amounts of iron of erythroblasts were indeed decreased in LID-W (Figure 1D). These results show that LID-W can induce severe IDA within a shorter period of time only by nutritional iron restriction, without using iron chelators; this is in contrast to previous reports.4,26,27

Maturation disorder of erythroblasts upon ID Under LID-W, total BM cell numbers were significantly reduced (Figure 2A). To describe the detail of erythropoiesis, several staining combinations are available for use.

For example, Ter119/CD44/forward scatter (FSC) combination can distinguish erythroblasts into six subsets.28 However, this approach might be inappropriate because of the erythrocyte volume reduction induced by ID.27 On the other hand, a CD71/Ter119 combination makes it pos-

sible to classify erythroblasts into four subsets that reflect maturation steps from subset I to subset IV.29 According to a previous report,30 CD71 (transferrin receptor 1) expression is independent of iron status in differentiating mouse erythroblast because of the altered regulation of IRP. Furthermore, this approach has already been used to analyze ID.4 By this approach, erythroblast cell numbers were significantly decreased from subset II in ID (Figure 2B and C), suggesting that maturation disorder of immature erythroblasts occurred in ID. As expected, subset II to whole Ter119-positive cell ratio was unchanged by ID (Figure 2D). Therefore, an alteration of CD71 expression itself was not the cause of the observed changes in the numbers of BM erythroblasts.

Transcriptome analysis of BM erythroblasts in ID Given the results of the flow cytometry analysis (Figure 2C and D), a microarray analysis of subset II erythroblasts

A

WBC (x109L)

B

Plt (x109L)

C

D

Figure 1. Complete peripheral blood (PB) counts and laboratory parameters of wild-type mice under a normal diet condition or low iron condition (LIC). All mice under LIC were fed a low iron diet after weaning (LID-W). (A) Results of complete blood counts in PB. MCV: mean corpuscular volume; WBC: white blood cells; Plt: platelets; ND: normal diet; LID-W 3w: LID after weaning for three weeks; LID-W 7w: LID after weaning for seven weeks. N.A.: not available. (B) Concentrations of serum Fe and TIBC (total iron binding capacity) (ND: n=9; LID-W: n=7). (C) Concentrations of total bilirubin (T-bil) (ND: n=5; LID-W: n=9). N.S.: not significant. (D) Total amount of iron of Ter119-positive cells in bone marrow (BM) of each condition (ND: n=4; LID-W: n=3). Results shown in bar plot are expressed by average and standard error of mean (s.e.m.). Symbols in beeswarm plots represent sex of mice: male (â&#x2014;?), female (â&#x2014;?). *P<0.05; **P<0.01.

456

haematologica | 2017; 102(3)

Role for iron and Bach1 in erythroblast maturation

sorted from wild-type mice under ND or LID-W was performed to reveal the effects of ID on the gene expressions.

Three samples each from three conditions (ND, LID-W for 3 weeks, LID-W for 7 weeks) were investigated. A clustering analysis clearly separated the samples into three groups corresponding to the experimental conditions (Figure 3A). According to the volcano plots analysis, 1843 genes (represented by 2629 entities) and 1708 genes (3351 entities) were significantly up-regulated in LID-W for three weeks and seven weeks, respectively. Conversely, 2289 genes (4061 entities) and 3023 genes (4631 entities) were significantly down-regulated in LID-W for three weeks and seven weeks, respectively (Figure 3B and Online Supplementary Table S3). The gene expression profiles in erythroblasts change rapidly and pervasively in response to ID. For example, expressions of genes encoding heme biosynthetic enzymes Alas2 (δ-aminolevulinate

A

synthase) and Fech (Ferrochelatase) were up-regulated in LID-W for seven weeks, suggesting the compensatory reactions for lower heme induced by ID (Online Supplementary Figure S1A). In contrast, expressions of cytoplasmic iron homeostasis related genes,31 such as Dmt1 and Ferroportin1, were down-regulated by LID, suggesting a retardation of iron transportation in ID (Online Supplementary Figure S1B). GSEA showed that the expressions of putative GATA-1 target genes were significantly decreased in LID-W (Figure 3C). Therefore, ID may inhibit the GATA-1 function, which is consistent with the findings of a previous report using fetal liver hematopoiesis as a model for IDA.4 While no obvious differences in expression of selected genes related to mitochondrial iron homeostasis were observed (Online Supplementary Figure S1C),31 many gene sets related to mitochondrial biology were significantly down-regulated in LID-W according to GSEA

B

C

D

Figure 2. Flow cytometry analysis of bone marrow (BM) erythroblasts terminal maturation. For this analysis, low iron diet after weaning (LID-W) was continued for 7-12 weeks (for 7 weeks: n=3; for 12 weeks: n=3). (A) Total BM cells of each condition. (B) Representative density plots of freshly isolated wild-type BM (Subset I: proerythroblasts; Subset II: basophilic erythroblasts; Subset III: late basophilic and chromatophilic erythroblasts; Subset IV: orthochromatophilic erythroblasts). (C) Numbers of cells in the indicated erythroblast subsets of each condition in BM. (D) Ratio of BM erythroblast Subset II to total Ter119-positive cells (Subset II + III + IV) of each condition. N.S.: not significant. Results shown in bar plot are expressed by average and standard error of mean (s.e.m.). *P<0.05; **P<0.01.

haematologica | 2017; 102(3)

457

M. Kobayashi et al.

(Figure 3C and Online Supplementary Table S4). These results suggest that iron is required for the proper expression of genes related to GATA-1 and mitochondrial biology.

Mitophagy disorder occurs under iron deficiency One of the unique features of erythroblast maturation is that erythroblasts undergo mitophagy, which removes mitochondria in a manner dependent upon GATA-1.32 Since results suggested GATA-1 function and mitochondrial biology were expected to be impaired in ID, we hypothesized that mitophagy disorder would occur in erythroblasts under ID. To address this, total mitochondrial mass in erythroblasts was measured by staining erythroblasts by Mito-G. This analysis showed that the mitochondrial mass gradually decreased alongside erythroblast maturation under ND, which was consistent with an initiation of mitophagy (Figure 4A). While there was no significant difference in the mitochondrial mass of BM erythroblasts between the ND and LID-W (Figure 4A), there was a significant increase in the mitochondrial mass under ID in PB erythrocyte (Figure 4B and C). However, ID did not affect enucleation of PB erythrocytes (Figure 4D).

These findings suggest that ID specifically impaired removal of mitochondria during erythroblast maturation.

Intriguingly, the erythrocytes with higher mitochondrial mass also showed higher expression of CD71 (Figure 4B and C) whereas BM erythroblasts did not show alterations

A

in the level of CD71 expression (see above). This altered frequency of CD71 expressing erythrocytes is consistent with the phenotype of the mice with mitophagy disorder in erythrocytes.33 Therefore, downregulation of CD71 during erythroblast maturation might be controlled by a similar regulatory mechanism that removes mitochondria in erythroblasts.

Since GATA-1 has been suggested to play a regulatory role in erythroblast autophagy,32 gene expression levels of Gata1, its co-factor gene Lmo2, and autophagy-related gene Atg4d of BM erythroblasts were analyzed. The expression levels of Gata1, Lmo2 and Atg4d were significantly decreased in subset III (Figure 4E), suggesting that iron is required to maintain the expressions of these genes during maturation. This mitochondria accumulation in erythrocytes was recovered by returning the mice to the ND for another eight weeks (Online Supplementary Figure S2A and B). Therefore, defective mitophagy observed in ID is reversible.

Bach1 plays pivotal roles in ID during the juvenile period Since Bach1 represses several iron-related genes, such as Hmox1 and globin genes,34 we hypothesized that Bach1 acts to balance intracellular heme and globin levels in erythroblast. However, no clear defect of erythropoiesis in Bach1–/– mice was observed under ND (Figure 5A and data not shown). Considering that Bach1 is negatively regulated

B

C

Figure 3. Transcriptome analysis of erythroblasts under iron deficiency. Clustering analysis of microarray results in erythroblasts Subset II from wild-type (WT) bone marrow (BM) under normal diet (ND), low iron diet after weaning (LID-W) for three weeks and LID-W for seven weeks. (B) Volcano plot analysis in microarray results of LID-W for three weeks or for seven weeks compared to ND. Genes with fold change ≥2 and P<0.05 are highlighted in red. (C) Results of the GSEA are shown.

Geneset data were referenced by Molecular Signatures Database of Broad Institute. Geneset V$GATA1_05 is composed of the genes that have GATA-1 binding motifs in their promoter regions [-2kb, 2kb]. Geneset MITOCHONDRIAL_MEMBRANE is composed of the genes related to the bilayers of the mitochondria. NES: normalized enrichment score; P: nominal P-value; q: false discovery rate as implemented in GSEA. vs: versus.

458

haematologica | 2017; 102(3)

Role for iron and Bach1 in erythroblast maturation

by heme, the function of Bach1 may become apparent under a low heme state, which is induced by ID. We, therefore, compared the endurance for ID between Bach1– /– mice and wild-type mice using the LID-W regimen.

However, there was no significant difference in severity of anemia under LID-W (Figure 5A), suggesting that Bach1 was not involved in the severity of IDA under the condition of LID-W. Since pregnant mother and developing child are vulnerable to ID because of their high iron demands,18 it was presumed that Bach1 function would become more apparent under ID in that developmental period. To address this,

A

C

we next compared the endurance for ID under LID-DW. Under this regimen, Bach1–/– mice exhibited more severe anemia than wild-type mice regarding the hemoglobin concentration and hematocrit level (Figure 5B and C).

These observations suggest that Bach1 facilitates adaptation of erythropoiesis to ID in embryo and/or in the juvenile period.

Bach1 controls heme and globin balance by directly repressing its targets The balance of heme and globin levels is critical for proper erythroid development.4 Therefore, we considered

B

D

E

Figure 4. Mitophagy disorder was induced by low iron condition in wild-type (WT) mice. (A) Mean fluorescent intensity (MFI) of Mito-G staining in each bone marrow (BM) erythroblast subset of WT mice under normal diet (ND) or low iron diet after weaning (LID-W) for 12 weeks (n=3). (B) Representative results of flow cytometry (FCM) analysis of peripheral blood (PB) samples in WT mice under ND or LID-W for 12 weeks. (Lower panels) Ter119-positive gated cells. (C) Cumulative results from more than 3 times experiments of Mito-G and CD71 double positive populations in PB of WT mice under ND or LID-W for 7-12 weeks. (D) May-Giemsa stained PB smear samples obtained from wild-type mice under ND or LID-W for 12 weeks. Images were obtained with a BX53 microscope and DP26 camera (Olympus, Tokyo, Japan); an eyepiece, WHN (Olympus); objective lens, MPlanApoN (Olympus). Scale bars: 20 μm. (E) Quantitative PCR (qPCR) analysis of indicated genes of each BM erythroblast subset in WT mice under ND or LID-W for 12 weeks (n=3). Relative gene expressions normalized by β-actin expressions were shown. Results shown in bar plot are expressed by average and standard error of mean (s.e.m.). *P<0.05; **P<0.01.

haematologica | 2017; 102(3)

459

M. Kobayashi et al.

whether the globin gene expression might be altered in Bach1–/– mice under LID-DW. We sorted BM erythroblasts (subsets I and II) from wild-type and Bach1–/– mice under ND or LID-DW, and performed a qPCR analysis of Hba-a (α-globin) and Hbb-b (β-globin). The expression levels of the globin genes were increased in subset II compared to subset I in both wild-type and Bach1–/– mice under ND (Figure 6A), indicating that erythroid maturation is accompanied by an induction of the globin expression, as reported previously.35 In erythroblasts from wild-type mice under LID-DW, the expression levels of globin genes were decreased in both subsets I and II (Figure 6A), indicating that the iron supply to erythroblasts leads to globin gene induction, and are consistent with the hypothesis that a shortage of iron and/or heme down-regulates the transcription of globin genes. In contrast, the expression of globin genes was not down-regulated in subset II erythroblasts from Bach1–/– mice under LID-DW (Figure 6A).

Intriguingly, the Hmox1 expression was significantly increased only in both subsets I and II of Bach1–/– mice under LID-DW (Figure 6A). Taken together, Bach1 represses the expression of Hba-a, Hbb-b and Hmox1 in erythroblasts, at least under a sustained ID. There are several Maf recognition elements (MARE) (the target sequence of Bach1 and MafK heterodimer) in the locus control regions

of Hba-a and Hbb-b and in the enhancer regions E1 and E2 of Hmox1 (Figure 6B), to which direct binding of Bach1 has been reported using MEL cells and non-erythroid cells.14,16,36-38 Using primary Ter119-positive BM cells isolated from wild-type mice and a chromatin immunoprecipitation assay, we found enrichments of Bach1 and MafK on MARE-containing regions of these genes (Figure 6C), suggesting that Bach1 directly represses these genes in mouse primary erythroblasts. In addition, Bach1 enrichments on its target regions were actually increased in ID (Figure 6D), indicating an increased activity of Bach1 in ID. These results substantially support the notion that the repression of Bach1 target genes at the transcriptional level is required under severe and/or prolonged ID to maintain the balance of heme and globin for the prevention of accumulations of free globin. Consistent with this interpretation, ROS accumulation of Bach1–/– erythroblasts was observed under LID-DW (Online Supplementary Figure S3A), which might be induced by the accumulations of free globin.39 On the other hand, no obvious difference was observed in the mitochondria amount in PB between wild-type and Bach1–/– mice even in ID (Online Supplementary Figure S3B). Therefore, Bach1 may not be directly involved in the regulation of erythroblast mitophagy in ID.

A

B

C

Figure 5. Bach1–/– mice showed severe iron deficiency anemia (IDA) under low iron diet development and after weaning (LID-DW) regimen. (A) Peripheral blood (PB) hemoglobin concentrations of wild-type (WT) and Bach1–/– (B1KO) mice under normal diet (ND), low iron diet after weaning (LID-W) for three weeks and LID-W for seven weeks. (B and C) PB hemoglobin and hematocrit concentrations of wild-type (WT) and Bach1–/– mice under ND or low iron diet development and after weaning (LID-DW). Only the significance between WT and Bach1–/– mice under each condition is highlighted in quotes. Symbols represent sex of mice in beeswarm plots: male (●), female (●). *P<0.01.

460

haematologica | 2017; 102(3)

Role for iron and Bach1 in erythroblast maturation

Iron and Bach1 control the DNA methylation status in erythroblasts To investigate how extensive transcriptomic changes were induced under ID and whether there are any other roles for Bach1 in the adaptation of erythropoiesis under ID, we compared the DNA methylation status of erythroblasts. After seven weeks of LID-W or ND, BM erythroblast subset II from wild-type or Bach1–/– mice were

isolated. DNA methylation levels were compared genome-wide using the PBAT method. Hemoglobin concentrations of wild-type mice under the ND or LID-W and Bach1–/– mice under the ND or LID-W were 13.1, 11.7, 14.1 and 4.0 g/dL, respectively. Therefore, Bach1–/– mouse in LID-W showed more severe anemia in this set of the experiment. There was no clear difference in the overall DNA methylation status between erythroblasts from

A

B

C

D

Figure 6. Bach1 regulates the essential genes in hemoglobin synthesis. (A) Quantitative PCR (qPCR) analysis of indicated gene expressions of each bone marrow (BM) erythroblast subset in wild-type (WT) and Bach1–/– (B1KO) mice under normal diet (ND) or low iron diet development and after weaning (LID-DW). Relative gene expressions normalized by β-actin expressions are shown. Data are cumulative results of 3 or more biological replicates. Bar plots are expressed by average and standard error of mean (s.e.m.). (B) Schematic representation of the mouse β-globin gene, α-globin gene and Hmox1 locus. The arrow indicates Maf recognition elements (MAREs). HS: DNase I-hypersensitive site; E1 and E2: enhancer regions of the Hmox1. (C) A chromatin immunoprecipitation analysis was performed by using antibodies against Bach1 and MafK. MAREs are found in the HS2, HS26, E1 and E2 site as shown in Figure 6B. The promoter regions (pro) of indicated genes do not contain MAREs. NRS: normal rabbit serum. A similar result was obtained from another independent experiment. *P<0.05; **P<0.01. (D) A chromatin immunoprecipitation analysis was performed by using antibodies against Bach1 of WT BM Ter119-positive erythroblast under ND or low iron diet after weaning (LID-W) (n=2).

Alas int8 and Hmox1 promoter regions used as negative controls. LCR (HS2) and Hmox1 E2 regions are Bach1 targets.

haematologica | 2017; 102(3)

461

M. Kobayashi et al.

A

E

B

C

D

F

G

H

Figure 7. DNA methylation analysis of bone marrow (BM) erythroblasts. (A) Whole genome methylated DNA percentages of BM erythroblast Subset II from each condition. Bar plots are expressed by average and standard error of mean (s.e.m.) (*P<0.01). (B-D) DNA methylation levels (%) of 100-kb non-overlapping tiles across the genome of erythroblast Subset II from each condition were compared. These data are shown by Hexbin plots as in the graphs. Dot plots in green zones were the genomic regions their methylation levels had changed 25% or more between compared conditions. (E) Hexbin plot of DNA methylation difference in promoters (from 2 kb upstream of transcription start sites to 0.5 kb downstream) determined by PBAT family-based association tests and gene expression difference determined by microarray analysis [difference between wild-type (WT) LID-W for 7 weeks (7w) and WT normal diet (ND)]. The correlation coefficient (R) was -0.024. (F) Peripheral blood (PB) hemoglobin concentrations of WT and Bach1–/– (B1KO) mice under LID-W for seven weeks followed by four weeks or eight weeks ND. Symbols represent sex of mice in beeswarm plots: male (●), female (●). (G and H) Schematic models for the Bach1 function in erythropoiesis under iron deficiency. (G) In iron deficient WT mice erythroblasts, Bach1 maintains the proper balance between heme and globin by repressing the globin genes and Hmox1. (H) In iron deficient Bach1–/– mice erythroblasts, the expressions of globin genes and Hmox1 are not repressed adequately resulting in an excess of free globin which can be the toxic aggregate to disturbance of erythroblast maturation.

462

haematologica | 2017; 102(3)

Role for iron and Bach1 in erythroblast maturation

wild-type mice under the ND compared with that of LIDW (Figure 7A). However, there were apparent differences in certain genomic regions (Figure 7B). (Dot plots within green zones indicate DNA regions where the methylation status changed ≥ 25% by ID). In addition, Bach1–/– erythroblasts under LID-W showed a higher DNA methylation level compared with wild-type erythroblasts under LID-W or Bach1–/– erythroblasts under the ND (Figure 7A, C and D). Although these results need to be interpreted with care due to the small number of samples and variation in the anemic severity, these results suggest that iron and Bach1 may be involved in the determination of DNA methylation status in erythroblasts and that these alterations affect the pathophysiology of IDA. When DNA regions with higher methylation levels were analyzed for possible correlations with the expression levels of nearby genes in WT mice, there was no clear correlation (Figure 7E). Such a relatively poor correlation between the DNA methylation status and the gene expression profile was consistent with previous reports.40,41 Therefore, further studies are needed to uncover the details of this mechanism involving DNA methylation and its physiological significance under ID.

Bach1 facilitates recovery from IDA Considering that DNA methylation alterations seen in Bach1–/– erythroblasts in LID-W, we compared recovery processes from IDA. After being fed LID for seven weeks according to the LID-W, wild-type or Bach1–/– mice were re-fed the ND. Initial recovery of the anemia was rapid and similar for both genotypes (Figure 7F). However, while the hemoglobin content reached 14.85±0.34 g/dL in wild-type mice after eight weeks, that in Bach1–/– mice remained around 13.23±0.33 g/dL (Figure 7F), suggesting that Bach1 is involved in the recovery process from IDA. The overall increase in the DNA methylation may be related to this slower recovery seen in Bach1–/– mice.

Discussion In the present study, we established that LIC induced severe ID and performed a comprehensive transcriptional analysis of BM erythroblasts. Unexpectedly large numbers of gene expressions in erythroblasts were altered under LID-W. In addition, not only maturation disorder, but also mitophagy disorder occurred in erythrocytes under LIDW. Importantly, the data suggest that iron plays a pivotal role in controlling GATA-1 function and expression of genes related to GATA-1 and/or GATA-1 downstream genes, including those related to mitophagy. Since it has been reported that heme is required for the cellular differentiation governed by GATA-1 in erythroid cells,42 impaired heme synthesis induced by ID can be the cause of repression of GATA-1 downstream genes. Therefore, a loss of iron induces the discordance of the erythroblast maturation program, which may be the fundamental cause of IDA. Our results raise the possibility that iron drives the process of mitophagy during maturation of erythroblasts.

The reduced function of GATA-1 under ID may be the primary cause of mitophagy disorder. These results also suggest that the sustained presence of mitochondria and the expression of CD71 in erythrocyte under IDA may provide differential diagnostic parameters for anemia. haematologica | 2017; 102(3)

Although we did not directly observe reticulocyte in PB, mito-G high erythrocyte can be considered to be reticulocyte because reticulocyte still possesses subcellular organelle such as mitochondria. Therefore, mice in ID might have a high percentage of reticulocyte in PB that is not induced by increased erythropoiesis but by defective mitophagy; this might be part of the reason why reticulocyte counts do not always decrease in human IDA.43 Since we did not observe nucleated erythrocyte under ID, terminal enucleation of erythroblasts is likely driven by a mechanism distinctive from that driving mitophagy.

Bach1–/– mice showed severe anemia under the LID-DW regimen. Because the embryonic period and early life stage are known as periods at high risk of IDA, it is speculated that Bach1 plays an important role in regulating erythropoiesis under ID especially during the early life stage. Because Bach1 is a unique transcription factor whose repressor function is inhibited by heme, its repressor activity is expected to increase under ID. In wild-type erythroblast, the heme concentration is decreased when iron is low, leading to the activation of Bach1, and thereby an inhibition of globin genes and Hmox1 transcription (Figure 7G). Therefore, both heme and globin are maintained at mutually stoichiometric levels by Bach1. In contrast, in the absence of Bach1, globin genes and Hmox1 are not down-regulated when the heme concentration decreases due to ID (Figure 7H). In this case, excess free globin may disturb erythroblast maturation. Therefore, Bach1 plays a critical role in sustained ID by working to balance heme and globin at the transcriptional level. This novel mechanism might be more specific than the function of HRI at the translational level, which phosphorylates eIF2a and therefore affects essentially global protein synthesis.26 The existence both of transcriptional and translational mechanisms governing the heme and globin balance indicates the importance of an appropriate maintenance of this balance in erythroblast maturation.

Massive alteration of the transcriptome in ID erythroblast may involve additional mechanisms. Although the details of a mechanism by which iron controls the gene expression levels remain an open question, the results of the PBAT analysis suggest that iron is involved in regulating the DNA methylation status. Although DNA methylation was not grossly affected by ID in wild-type mice, it is still possible that the observed DNA region-specific alterations affected the expression of nearby genes and alternated gene expressions induce massive transcriptomic modification as a whole, as seen in the result of transcriptomic analysis. Conversely, Jeffrey et al. previously reported that global DNA methylation of erythroblasts gradually decreases during maturation,44 suggesting the importance of the DNA demethylation in maturation. Histone methylation may also be affected since many histone demethylases are dependent on iron, such as non-heme irondependent histone demethylases (JmjC family).6 In addition, other regulators that are under the control of iron may participate in the transcriptomic response under ID, such as the IRE/IRP system.2,3 Very intriguingly, the alterations in DNA methylation in Bach1-deficient cells induced by ID were significant, indicating that Bach1 maintain the appropriate DNA methylation status especially under ID. It is also possible that such changes in DNA methylation contributed to the severe IDA observed in Bach1–/– mice. These possibilities are consistent with the observation that Bach1–/– mice showed slower recovery 463

M. Kobayashi et al.

from IDA induced by LID-W. Bach1 may pre-condition genes that are affected by ID to return to normal regulation when iron stores are repleted. Further studies are required to address how Bach1 controls DNA methylation status under ID. In our experimental approach, there is a limitation as to how far Bach1 function can be totally excluded in non-erythroid cells. In other words, it is also possible that Bach1 regulates iron absorption and/or trafficking as well. To answer this question, additional studies, such as those using a tissue-specific conditional knockout system,45 are needed. Nonetheless, it is now clear that erythrocytes in ID reflect an active adaptation to iron depletion by changing gene expressions in erythroblasts according to available iron levels.

In conclusion, this study provides new insight into the function of iron and Bach1 in erythroblast maturation by using efficacious regimens that can induce severe ID. Iron is required for the activation of erythroblast maturation program and for maintaining the proper epigenetic DNA methylation status. The pathophysiology of IDA is not merely due to the material shortage, but a consequence of an altered DNA methylation status and gene regulatory

References 12.

1. Andrews NC. Disorders of iron metabolism.

N Engl J Med. 1999;341(26):1986-1995.

2. Muckenthaler MU, Galy B, Hentze MW. Systemic iron homeostasis and the ironresponsive element/iron-regulatory protein (IRE/IRP) regulatory network. Annu Rev Nutr. 2008;(28):197-213.

3. Moroishi T, Nishiyama M, Takeda Y, Iwai K, Nakayama KI. The FBXL5-IRP2 axis is integral to control of iron metabolism in vivo. Cell Metab. 2011;14(3):339-351.

4. Liu S, Bhattacharya S, Han A, et al. Haemregulated eIF2alpha kinase is necessary for adaptive gene expression in erythroid precursors under the stress of iron deficiency. Br J Haematol. 2008;143(1):129-137.

5. Tahiliani M, Koh KP, Shen Y, et al.

Conversion of 5-methylcytosine to 5hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science.

2009;324(5929):930-935.

6. Culhane JC, Cole PA. LSD1 and the chemistry of histone demethylation. Curr Opin Chem Biol. 2007;11(5):561-568.

7. Rund D, Rachmilewitz E. Beta-thalassemia.

N Engl J Med. 2005;353(11):1135-1146.

8. Kong Y, Zhou S, Kihm AJ, et al. Loss of alpha-hemoglobin-stabilizing protein impairs erythropoiesis and exacerbates beta-thalassemia. J Clin Invest. 2004; 114(10):1457-1466.

9. Arlet JB, Ribeil JA, Guillem F, et al. HSP70 sequestration by free alpha-globin promotes ineffective erythropoiesis in beta-thalassaemia. Nature. 2014;514(7521):242-246.

10. Chen JJ. Regulation of protein synthesis by the heme-regulated eIF2alpha kinase: relevance to anemias. Blood. 2007;109(7):26932699.

11. Ogawa K, Sun J, Taketani S, et al. Heme mediates derepression of Maf recognition element through direct binding to transcrip-

464

13.

14.

15.

16.

17.

18.

19.

20.

21.

network elicited by the loss of iron. Furthermore, Bach1 is necessary for the adaptation of erythroblast to ID. These findings reveal new aspects of IDA and might be informative for developing a solution for this worldwide healthcare problem.

Acknowledgments The authors would like to thank members of the Departments of Biochemistry and of Hematology and Rheumatology for insightful discussions and supports. We also thank Metallogenics Co. Ltd. for performing ICP-OES method analysis.

This work was partly performed in the Cooperative Research Project Program of the Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan.

Funding This work has been supported in part by Grants-in-aid (15H02506, 24390066, 25670156, 23116003, 21249014 and 26293225) from the Japan Society for the Promotion of Science.

Additional initiative supports were from the Uehara Memorial Foundation, Takeda Foundation and Astellas Foundation for Research on Metabolic Disorders.

tion repressor Bach1. EMBO J. 2001;20 (11):2835-2843.

Zenke-Kawasaki Y, Dohi Y, Katoh Y, et al.

Heme induces ubiquitination and degradation of the transcription factor Bach1. Mol Cell Biol. 2007;27(19):6962-6971.

Suzuki H, Tashiro S, Hira S, et al. Heme regulates gene expression by triggering Crm1dependent nuclear export of Bach1. EMBO J. 2004;23(13):2544-2553.

Tahara T, Sun J, Nakanishi K, et al. Heme positively regulates the expression of betaglobin at the locus control region via the transcriptional factor Bach1 in erythroid cells. J Biol Chem. 2004;279(7): 5480-5487.

Reichard JF, Sartor MA, Puga A. BACH1 is a specific repressor of HMOX1 that is inactivated by arsenite. J Biol Chem. 2008; 283(33):22363-22370.

Sun J, Hoshino H, Takaku K, et al.

Hemoprotein Bach1 regulates enhancer availability of heme oxygenase-1 gene.

EMBO J. 2002;21(19):5216-5224.

Kassebaum NJ, Jasrasaria R, Naghavi M, et al. A systematic analysis of global anemia burden from 1990 to 2010. Blood. 2014; 123(5):615-624.

Lopez A, Cacoub P, Macdougall IC, PeyrinBiroulet L. Iron deficiency anaemia. Lancet.

2016;387(10021):907-916.

Ota K, Brydun A, Itoh-Nakadai A, Sun J, Igarashi K. Bach1 deficiency and accompanying overexpression of heme oxygenase-1 do not influence aging or tumorigenesis in mice. Oxid Med Cell Longev. 2014; 2014:757901.

Tanaka H, Muto A, Shima H, et al.

Epigenetic Regulation of the Blimp-1 Gene (Prdm1) in B Cells Involves Bach2 and Histone Deacetylase 3. J Biol Chem. 2016; 291(12):6316-6330.

Fujiwara T, O'Geen H, Keles S, et al.

Discovering hematopoietic mechanisms through genome-wide analysis of GATA factor chromatin occupancy. Mol Cell.

2009;36(4):667-681.

22. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3.

23. Fu J, Khaybullin R, Zhang Y, Xia A, Qi X. Gene expression profiling leads to discovery of correlation of matrix metalloproteinase 11 and heparanase 2 in breast cancer progression. BMC Cancer. 2015;15:473.

24. Mootha VK, Lindgren CM, Eriksson KF, et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34(3):267-273.

25. 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):1554515550.

26. Han AP, Yu C, Lu L, et al. Heme-regulated eIF2alpha kinase (HRI) is required for translational regulation and survival of erythroid precursors in iron deficiency. EMBO J. 2001;20(23):6909-6918.

27. Knight ZA, Schmidt SF, Birsoy K, Tan K, Friedman JM. A critical role for mTORC1 in erythropoiesis and anemia. Elife. 2014;3: e01913.

28. Liu J, Zhang J, Ginzburg Y, et al.

Quantitative analysis of murine terminal erythroid differentiation in vivo: novel method to study normal and disordered erythropoiesis. Blood. 2013;121(8):e43-49.

29. Socolovsky M, Nam H, Fleming MD, Haase VH, Brugnara C, Lodish HF. Ineffective erythropoiesis in Stat5a(-/-)5b(-/-) mice due to decreased survival of early erythroblasts.

Blood. 2001;98(12):3261-3273.

30. Schranzhofer M, Schifrer M, Cabrera JA, et al. Remodeling the regulation of iron metabolism during erythroid differentiation to ensure efficient heme biosynthesis. Blood.

2006;107(10):4159-4167.

haematologica | 2017; 102(3)

Role for iron and Bach1 in erythroblast maturation

31. Chen C, Paw BH. Cellular and mitochondrial iron homeostasis in vertebrates. Biochim Biophys Acta. 2012;1823(9):1459-1467.

32. Kang YA, Sanalkumar R, O'Geen H, et al.

Autophagy driven by a master regulator of hematopoiesis. Mol Cell Biol. 2012; 32(1):226-239.

33. Mortensen M, Ferguson DJ, Edelmann M, et al. Loss of autophagy in erythroid cells leads to defective removal of mitochondria and severe anemia in vivo. Proc Natl Acad Sci USA. 2010;107(2):832-837.

34. Igarashi K, Watanabe-Matsui M. Wearing red for signaling: the heme-bach axis in heme metabolism, oxidative stress response and iron immunology. Tohoku J Exp Med. 2014;232(4):229-253.

35. Fukuda Y, Fujita H, Garbaczewski L, Sassa S. Regulation of beta-globin mRNA accumulation by heme in dimethyl sulfoxide (DMSO)-sensitive and DMSO-resistant murine erythroleukemia cells. Blood. 1994; 83(6):1662-1667.

36. Brand M, Ranish JA, Kummer NT, et al.

haematologica | 2017; 102(3)

37.

38.

39.

40.

Dynamic changes in transcription factor complexes during erythroid differentiation revealed by quantitative proteomics. Nat Struct Mol Biol. 2004;11(1):73-80.

Sun J, Brand M, Zenke Y, Tashiro S, Groudine M, Igarashi K. Heme regulates the dynamic exchange of Bach1 and NF-E2related factors in the Maf transcription factor network. Proc Natl Acad Sci USA. 2004; 101(6):1461-1466.

Tahara T, Sun J, Igarashi K, Taketani S. Heme-dependent up-regulation of the alpha-globin gene expression by transcriptional repressor Bach1 in erythroid cells.

Biochem Biophys Res Commun. 2004; 324(1):77-85.

De Franceschi L, Bertoldi M, Matte A, et al. Oxidative stress and beta-thalassemic erythroid cells behind the molecular defect. Oxid Med Cell Longev. 2013;2013: 985210.

Challen GA, Sun D, Jeong M, et al. Dnmt3a is essential for hematopoietic stem cell differentiation. Nat Genet. 2012;44(1):23-31.

41. Luo M, Jeong M, Sun D, et al. Long non-coding RNAs control hematopoietic stem cell function. Cell Stem Cell. 2015; 16(4):426438.

42. Tanimura N, Miller E, Igarashi K, et al.

Mechanism governing heme synthesis reveals a GATA factor/heme circuit that controls differentiation. EMBO Rep. 2016; 17(2):249-265.

43. Parodi E, Giraudo MT, Ricceri F, Aurucci ML, Mazzone R, Ramenghi U. Absolute Reticulocyte Count and Reticulocyte Hemoglobin Content as Predictors of Early Response to Exclusive Oral Iron in Children with Iron Deficiency Anemia. Anemia.

2016;2016:7345835.

44. Shearstone JR, Pop R, Bock C, Boyle P, Meissner A, Socolovsky M. Global DNA demethylation during mouse erythropoiesis in vivo. Science. 2011; 334(6057):799-802.

45. Orban PC, Chui D, Marth JD. Tissue- and site-specific DNA recombination in transgenic mice. Proc Natl Acad Sci USA. 1992; 89(15):6861-6865.

465

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Red Cell Biology & Its Disorders

Ferrata Storti Foundation

Haematologica 2017 Volume 102(3):466-475

Small-molecule factor D inhibitors selectively block the alternative pathway of complement in paroxysmal nocturnal hemoglobinuria and atypical hemolytic uremic syndrome Xuan Yuan,1* Eleni Gavriilaki,1* Jane A. Thanassi,2 Guangwei Yang,2 Andrea C. Baines,1 Steven D. Podos,2 Yongqing Huang,2 Mingjun Huang2 and Robert A. Brodsky1 Division of Hematology, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD and 2Achillion Pharmaceuticals, New Haven, CT, USA

1

*XY and EG contributed equally to this work.

ABSTRACT

P

Correspondence: rbrodsky@jhmi.edu

Received: July 26, 2016.

Accepted: October 24, 2016.

Pre-published: November 3, 2016. doi:10.3324/haematol.2016.153312 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/466 Š2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

466

aroxysmal nocturnal hemoglobinuria and atypical hemolytic uremic syndrome are diseases of excess activation of the alternative pathway of complement that are treated with eculizumab, a humanized monoclonal antibody against the terminal complement component C5.

Eculizumab must be administered intravenously, and moreover some patients with paroxysmal nocturnal hemoglobinuria on eculizumab have symptomatic extravascular hemolysis, indicating an unmet need for additional therapeutic approaches. We report the activity of two novel small-molecule inhibitors of the alternative pathway component Factor D using in vitro correlates of both paroxysmal nocturnal hemoglobinuria and atypical hemolytic uremic syndrome. Both compounds bind human Factor D with high affinity and effectively inhibit its proteolytic activity against purified Factor B in complex with C3b. When tested using the traditional Ham test with cells from paroxysmal nocturnal hemoglobinuria patients, the Factor D inhibitors significantly reduced complementmediated hemolysis at concentrations as low as 0.01 mM. Additionally the compound ACH-4471 significantly decreased C3 fragment deposition on paroxysmal nocturnal hemoglobinuria erythrocytes, indicating a reduced potential relative to eculizumab for extravascular hemolysis.

Using the recently described modified Ham test with serum from patients with atypical hemolytic uremic syndrome, the compounds reduced the alternative pathway-mediated killing of PIGA-null reagent cells, thus establishing their potential utility for this disease of alternative pathway of complement dysregulation and validating the modified Ham test as a system for pre-clinical drug development for atypical hemolytic uremic syndrome. Finally, ACH-4471 blocked alternative pathway activity when administered orally to cynomolgus monkeys. In conclusion, the small-molecule Factor D inhibitors show potential as oral therapeutics for human diseases driven by the alternative pathway of complement, including paroxysmal nocturnal hemoglobinuria and atypical hemolytic uremic syndrome.

Introduction The complement system provides an important defense against bacteria, fungi, and viruses.1,2 Mutations (e.g. PIGA, factor H, factor I, etc.) and autoantibodies (e.g.

Factor H autoantibodies) that lead to unregulated complement activation are implicated in the pathogenesis of many disorders including paroxysmal nocturnal hemohaematologica | 2017; 102(3)

Factor D inhibitors in PNH and aHUS

globinuria (PNH) and atypical hemolytic uremic syndrome (aHUS).3-5 Complement activation is initiated by: 1) the lectin pathway, 2) the classical pathway, and 3) the alternative pathway of complement (APC) which converge on the common central component C3. The APC also serves as an amplification loop for the lectin and classical pathways.

It has been estimated that the APC may account for 80% of complement activation products even when the complement cascade is initiated by the classical and lectin pathways.6 The APC is constitutively activated at low levels via slow spontaneous hydrolysis of an internal thioester within C3 that generates C3(H2O). This activated C3(H2O) in solution phase binds Factor B to generate the proconvertase C3(H2O)B, which is processed by the serine protease Factor D to the APC C3 convertase, C3(H2O)Bb. This C3 convertase then cleaves additional C3 molecules to generate C3a and C3b, the latter of which can covalently attach to available surfaces.7 Deposited C3b can then elicit a rapid localized amplification, especially when complement regulation is impaired, described as follows in a process designated the amplification loop. Deposited C3b can pair with Factor B, which is cleaved by Factor D to generate a second form of the APC C3 convertase, C3bBb. Membrane-bound C3bBb then cleaves additional C3 to generate further C3b deposits, which pair with additional Factor B molecules to repeat the cycle. The end result of this amplification is C3b opsonization, release of the anaphylatoxins C3a and C5a, and assembly of the terminal complement complex (also known as the membrane attack complex, MAC) on the target surface. The entire process is depicted in Figure 1A. Paroxysmal nocturnal hemoglobinuria is caused by a somatic PIGA mutation that prevents expression of glycosylphosphatidylinositol (GPI) anchored proteins on the surfaces of affected cells, among them the complement regulators CD55 and CD59 which normally protect host cells from complement-mediated lysis. CD55 interferes with C3 convertase formation following C3b attachment, preventing spurious amplification of initial activation events,8 whereas CD59 prevents localized MAC assembly following amplification.9 Due to the absence of both regulators from erythrocyte membranes, PNH manifests with chronic hemolytic anemia primarily due to APC activation on mutant erythrocyte membranes. PNH is effectively treated with the humanized monoclonal antibody, eculizumab.10,11 Eculizumab binds C5 and blocks its cleavage to C5a and C5b and thus protects mutant erythrocytes from MAC formation and lysis. However, due to the

absence of CD55, the mutant erythrocytes in PNH patients on eculizumab treatment continue to be opsonized by C3b and thus are susceptible to extravascular hemolysis.12 In most PNH patients on eculizumab, this leads to asymptomatic chronic extravascular hemolysis; however, a subset of PNH patients on eculizumab have symptomatic hemolysis and require red blood cell transfusions.10,13 Moreover, rare patients (approx. 3.5% of the Japanese population) carry a genetic C5 variant (2654G>A, Arg885His) that prevents eculizumab binding and causes eculizumab resistance.14 Another limitation of eculizumab is that it must be administered intravenously indefinitely every two weeks to block intravascular hemolysis and prevent thrombosis. Thus, novel complement inhibitors are needed to address these limitations.

Atypical hemolytic uremic syndrome is also caused by APC dysregulation. aHUS presents with signs and symptoms of thrombotic microangiopathy (TMA) including thrombocytopenia, non-immune hemolytic anemia, peripheral blood schistocytes, and often end-organ damage to the kidneys and central nervous system. Most cases are caused by mutations in APC genes or autoantibodies directed against APC regulatory proteins.15-17 These mutations either disable the regulatory proteins that help degrade cell surface C3b, including haploinsufficient mutations of Factor H (most commonly), Factor I, MCP, and thrombomodulin, or enhance the function of proteins that drive the APC, such as gain-of-function mutations in C3 or factor B.15,17-19 The result is activation of terminal complement on various target tissues, including renal and vascular endothelium, and cell death mediated by the MAC. Distinguishing aHUS from other TMAs such as thrombotic thrombocytopenic purpura (TTP) is challenging. Recently, we described a serum-based assay that distinguishes most cases of aHUS from TTP.20 This assay (modified Ham test) measures the ability of human serum to kill PIGA-null reagent cells. Due to APC dysregulation in the serum of most aHUS patients, the PIGA-null cells that fail to express CD55 and CD59 are more readily killed by aHUS serum than by serum from healthy individuals or TTP patients.

Complement Factor D is a promising therapeutic target to treat PNH and aHUS. Factor B is its only natural substrate, and of all complement proteins, Factor D has the lowest abundance in serum and is the rate-limiting component of the APC.21 Its crystal structure has been elucidated in a series of studies.22-24 Here, we describe two novel, small-molecule inhibitors that bind Factor D and inhibit its cleavage of Factor B. We demonstrate in vitro that these inhibitors block PNH cell hemolysis, mitigate C3 fragment

Table 1. Clinical data on paroxymal nocturnal hemoglobinuria patients studied.

Patient ID

Age, sex

Erythrocyte clone size Type II/ III (%)

Granulocyte clone size (%)

LDH (U/L)

1 2 3*

51, F 58, M 72, F

3/64 1/98 1/13

96 99 81

266 281 842

Hemoglobin Direct Direct Absolute Months on (g/dL) Coombs Coombs reticulocyte eculizumab at C3 IgG count (K/cu mm) q14 days dosing at start of study 9.5 10.3 9.7

Pos Pos Neg

Neg Neg Neg

251.1 321.6 138.3

57 107 0

LDH: lactate dehydrogenase; q14 days: every 14 days; M: male; F: female;; Pos: positive; Neg: negative. *Transfused within 60 days.

haematologica | 2017; 102(3)

467

X. Yuan et al.

Table 2. Clinical data on atypical hemolytic uremic syndrome patients studied.

Patient ID

Age, sex

1

47, F

2a 2b 3

4

Disease ADAMTS13 ADAMTS13 PLT LDH Cr phase activity (%) inhibitor count (mg/dL) (mg/dL) (weeks) (%) (x109/L) Acute

Remission (20 w) 59, M Acute Remission (12 w) 35, F Acute Remission (82 w) 23, F Acute Remission (60 w)

Genetic mutation

Previous Current treatments treatment

NA

NA

314

228

CFHR3-CFHR1 homozygous deletion; CFH PEx (x5) heterozygous Eculizumab c.2850G>T, (x8) p.Gln950His 1.9 None

15 NA

67 NA

39 330

676 170

2.3 1.4

None

18 NA

76 NA

10 108

1270 75

2.4 1.1

None

102

NA NA

62 NA

2820 205

6.3 178

None 3.5

75

NA

9

2505

2.3

-

Eculizumab (x12) Eculizumab (x77) Eculizumab Dialysis (x37) PEx (x4)

Response to eculizumab

Serum tested

-

No

Yes

Yes

Yes

Yes Yes

Yes

No Yes

Yes

No Yes

F: female; M: male; w: weeks; CFHR: complement Factor H-related proteins; CFH: complement Factor H; ADAMTS13: a disintegrin and metalloprotease with thrombospondin type 1 motif, 13; NA: not available; PEx: plasma exchange.

accumulation on PNH cell surfaces, and block dysregulated APC activity in aHUS serum. We also show that the modified Ham test can be used to evaluate novel complement inhibitors in aHUS, and, finally, we demonstrate APC suppression in serum obtained from non-human primates following oral administration of one inhibitor.

inhibition of hemolysis by the classical pathway was assessed with 0.5% normal human serum and antibody-sensitized sheep erythrocytes. Methods for these assays are described in the Online Supplementary Appendix.

Methods

Inhibition of hemolytic assay by Factor D inhibitors was assessed using PNH erythrocytes at 1% hematocrit in GVB0/MgEGTA (pH 6.4) and 20% acidified human serum (Ham test), as previously described.27 Based on the same principle, the modified Ham test was performed to assess the efficacy of Factor D inhibitors in APC-mediated killing caused by aHUS patient serum.20 Both assays are described in the Online Supplementary Appendix.

Human samples Three PNH patients and 4 aHUS patients were enrolled in this study (Tables 1 and 2). Patients were diagnosed with aHUS as previously defined.25,26 All patients gave written informed consent.

This study was approved by the Institutional Review Board and conducted in accordance with the Declaration of Helsinki. Blood was collected in EDTA and serum separation tubes (Online Supplementary Appendix).

Factor D inhibitors Compounds ACH-3856 and ACH-4471 were synthesized by Achillion Pharmaceuticals and fully characterized by proton nuclear magnetic resonance (1H-NMR), high performance liquid chromatography (HPLC), and mass spectrometry.

Assays for Factor D binding, Factor D enzymatic activity, and hemolysis mediated by APC and the classical pathway Binding kinetics and affinities of compounds to human Factor D were determined by surface plasmon resonance. Inhibition of Factor D enzymatic activity was evaluated with 80 nM purified human Factor D and the small synthetic thieoster substrate Z-LysSBzl, or with 0.8 nM purified human Factor D and its natural substrate C3bB. Inhibition of APC-mediated hemolytic activity was assessed with 8% normal human serum and rabbit erythrocytes; 468

Inhibition of APC-mediated cell killing using PNH and aHUS patient samples: Ham test and modified Ham test

Inhibition of C3 fragment deposition C3 fragment deposition on PNH erythrocytes from PNH Patient 2 (blood group O) by 60% acidified C5-depleted human serum was assessed by flow cytometry. Erythrocytes were washed and prepared as in the hemolytic assay. Sample preparation and C3 fragment deposition measurement by flow cytometry are described in detail in the Online Supplementary Appendix.

APC activity in cynomolgus monkeys The in-life portion of these studies was performed at Ricerca Laboratories (Concord, OH, USA) in accordance with Care and Use of Laboratory Animals guidelines and the United States Department of Agriculture Animal Welfare Act. Experimental procedures were reviewed and approved by Ricercaâ&#x20AC;&#x2122;s Institutional Animal Care and Use Committee (IACUC) prior to initiation.

ACH-4471 was formulated in solution at 80 mg/mL in PEG400:water (70:30) (w:w). Three male cynomolgus monkeys were dosed by oral gavage with ACH-4471 at 200 mg/kg twice, 12 haematologica | 2017; 102(3)

Factor D inhibitors in PNH and aHUS

A

B

C

Figure 1. Inhibition of Factor D proteolytic activity. (A) Depiction of Factor D inhibitors in the alternative complement pathway (APC). Factor D (fD) participates in C3 convertase generation by cleavage of Factor B (fB) at two steps in the APC cascade: generation of the initial C3 convertase [C3(H2O)Bb] following spontaneous APC activation (tickover); and the production of surface-bound C3 convertase (C3bBb) which mediates dramatic amplification of the initial activation (amplification loop) and the consequent opsonization of target surfaces by C3b, formation of the terminal complement complex (aka, membrane attack complex, MAC), and release of the anaphylatoxins C3a and C5a. An erythrocyte is shown to depict the membrane-bound events. Additional regulatory proteins not shown here can promote (properdin) or attenuate (Factor H, Factor I, multiple membrane-bound proteins) APC activity. (B) Factor D proteolytic activity against Factor B in complex with C3b. Two stained gels from a single representative experiment show dose-dependent inhibition by compounds of Factor B (fB) cleavage to its products Ba and Bb. Control reactions included one omitting C3b and fB (labeled “fD”) and one omitting fD (labeled “No fD”). (C) Dose-response curves and IC50 values from the representative experiment of (B). Average IC50 values ± standard deviation for ACH-3856 and ACH-4471 were 0.0058 ± 0.0005 mM and 0.015 ± 0.003 mM in 4 independent experiments.

Table 3. Compound binding to and inhibition of Factor D enzyme.

Compound On ka (M-1 s-1) ACH-3856 ACH-4471

1.7e7 ± 7e6 2.6e6 ± 4e5

Off kd (s-1)

fD binding affinity KD (mM)

N

C3bB IC50 (mM)

2 2

0.0058 ± 0.005 0.015 ± 0.003

0.0086 ± 0.0061 0.00048 ± 0.00016 0.0015 ± 0.0001 0.00057 ± 0.00004

fD activity N 4 4

Thioester IC50 (mM)

N

0.036 ± 0.003 0.035 ± 0.001

83 3

fD: Factor D; N: number. Mean ± standard deviation from N independent experiments. fD binding parameters were assessed by surface plasmon resonance; representative sensorgrams are shown in Online Supplementary Figure S2. fD activity was measured both against its natural substrate C3bB and against a small synthetic thioester substrate.

hours apart. Serial blood samples were collected at specified time points through 30 hours. Plasma ACH-4471 concentration was determined by LC-MS/MS with a lower limit of quantitation of 2.44 ng/mL. Pharmacokinetic parameters were calculated with non-compartmental analysis using Phoenix 6.2 WinNonlin (Pharsight, Princeton, NJ, USA) from individual plasma concentration versus time using the linear-trapezoidal method. Serum APC activity was measured in duplicate using the APC-specific Wieslab assay (Euro Diagnostica, Malmö, Sweden). Activity at each time point was normalized to pre-dosing activity in the same animal.

Serum Factor D concentrations were determined using the Human Complement Factor D Quantikine ELISA Kit (R&D Systems).

Results Inhibition of Factor D proteolytic activity The potency of the small-molecule Factor D inhibitors ACH-3856 and ACH-4471 was investigated by biophysical and biochemical methods. Both compounds showed high binding affinity to human Factor D, shown by surface plasmon resonance studies with recombinant enzyme (Table 3 and Online Supplementary Figure S1), with respective KD values of 0.00036 mM and 0.00054 mM. Additionally, the compounds inhibited the proteolytic activity of purified Factor D against its natural substrate Factor B in complex with C3b, blocking production of Bb haematologica | 2017; 102(3)

fragment in a dose-dependent manner with respective IC50 values of 0.0058±0.0005 mM and 0.015±0.003 mM (Table 3 and Figure 1B and C). The compounds act as catalytic inhibitors, established by inhibition of Factor D protease activity against a small synthetic substrate (N-αCbz-L-lysine thiobenzyl ester) (Table 3 and Online Supplementary Figure S2), distinguishing them from the anti-Factor D monoclonal antibody inhibitor lampalizumab which binds an exosite and sterically blocks Factor B access to the active site.28 Additionally, the compounds showed selectivity as they inhibited in a standard assay for rabbit erythrocyte hemolysis upon APC activation but not in a counterpart assay for the complement classical pathway (Table 4), and moreover did not inhibit a panel of 12 human serine proteases (Online Supplementary Table S1), thus establishing their limited potential for off-target effects.

PNH and aHUS patients’ characteristics Paroxymal nocturnal hemoglobinuria and aHUS are diseases of complement dysregulation that could be amenable to Factor D inhibitors. As proof of concept, we assessed the Factor D inhibitors in vitro using erythrocytes and serum obtained from PNH patients and serum from aHUS patients. Patients’ characteristics are shown in Table 1 (PNH) and Table 2 (aHUS). Erythrocytes were recovered 469

X. Yuan et al.

from 3 PNH patients: 2 on long-term eculizumab treatment (Patients 1 and 2) and 1 treatment-naïve (Patient 3), and were used for the hemolytic assay and the C3 fragment deposition assay. Serum was also recovered from PNH Patient 1 and Patient 3 (this time after Patient 3 was started on eculizumab) following eculizumab administration for use in the C3 fragment deposition assay. Five serum samples used in the modified Ham assay were recovered from 4 aHUS patients: 1 in acute phase before treatment and later in remission while on eculizumab treatment (Patient 2, 2 samples), 1 in remission while on eculizumab treatment (Patient 3, 1 sample), and 2 in remission after discontinuation of eculizumab (Patients 1 and 4, one sample each).

Inhibitory effects on hemolysis of PNH erythrocytes The Ham test identifies PNH erythrocytes by their hemolytic susceptibility to acidified serum. We evaluated ACH-3856 and ACH-4471 by incubating erythrocytes from PNH patients with acidified ABO blood group-com-

patible normal human serum containing the inhibitors in half-log dilution series (Figure 2). As C3 convertase formation is Mg++-dependent, EDTA inhibition was used to quantitate complement-independent background in each experiment. Above this background, the acidified serum in the presence of Mg++ mediated the complement-dependent hemolysis of 80% (Patient 1), 48% (Patient 2), and 25% (Patient 3) of patient erythrocytes (Figure 2A). The Factor D inhibitors potently inhibited hemolysis with IC50 values ranging from 0.0029 mM to 0.016 mM for ACH-3856 (IC90 values from 0.0082 mM to 0.064 mM) and from 0.0040 mM to 0.027 mM for ACH-4471 (IC90 values from 0.015 mM to 0.11 mM) (Figure 2B and C). The Factor D inhibitors, therefore, have the potential to inhibit the APC-mediated hemolysis that underlies much of the morbidity and mortality in PNH patients.

C3 fragment deposition on PNH erythrocytes As eculizumab blocks terminal complement activity only, the continued accumulation of C3 fragments on

A

B

C

Figure 2. Factor D inhibitors block hemolysis of erythrocytes from paroxymal nocturnal hemoglobinuria (PNH) patients. (A) Hemolysis of erythrocytes obtained from the 3 PNH patients in 20% acidified normal human serum (Ham test). The control samples “Buffer”, “Serum + EDTA”, and “H2O”, and the patient-specific samples without inhibition, “Serum”, were as defined in the Methods section. (B) Inhibition by ACH-4471 (○) and ACH-3856 (□) of hemolysis reactions using erythrocytes from the 3 PNH patients. Each inhibitor concentration was tested in duplicate with mean±standard deviation (SD) shown. (C) IC50 and IC90 values of ACH-4471 and ACH3856 from the hemolysis reactions in (B).

470

haematologica | 2017; 102(3)

Factor D inhibitors in PNH and aHUS

mutant erythrocytes in treated patients often results in extravascular hemolysis. As Factor D inhibitors should inhibit this opsonization, we tested the ability of ACH4471 to prevent C3 fragment deposition on PNH erythrocytes as detected by flow cytometry. Erythrocytes from PNH Patient 2 were incubated with acidified human serum depleted of C5 to prevent hemolysis, yet allow continued C3 fragment deposition during the incubation period (Figure 3). While serum acidification led to C3 fragment deposition on CD59-negative PNH erythrocytes, EDTA inhibited this event, confirming that it was a complementspecific activity. ACH-4471 inhibited this deposition in a dose-dependent manner, with IC50 and IC90 values of 0.031 mM and 0.089 mM, respectively. Inhibition was also observed with erythrocytes from PNH Patients 1 and 3 when assessed using their own serum collected immediately after eculizumab infusion; the percentage of C3-positive erythrocytes was 37.6% (Patient 1) and 5.5% (Patient 3) without ACH-4471, decreasing to 11.9% (Patient 1) and 1.8% (Patient 3) with 0.1 mM ACH-4471 (data not shown).

Furthermore, in addition to the anti-C3c antibody used above, we also examined the phenomenon with anti-C3b and anti-C3d antibodies. We observed similar results on erythrocytes from PNH Patient 2 except that approximately 10% of erythrocytes stained positive for C3d in the presence of EDTA, indicating that these C3 fragments were deposited in vivo before harvesting from patients (Online Supplementary Figure S3). These results suggest that the extravascular hemolysis that occurs in PNH patients on eculizumab treatment might be blocked by Factor D inhibition.

Factor D inhibitors block dysregulated APC in aHUS patient serum An important distinction between aHUS and other TMAs is the underlying APC dysregulation observed in aHUS. The modified Ham test can distinguish aHUS from other TMAs based on the sensitivity of a PIGA-null reagent cell line to APC activation; cell killing greater than 20% by patient serum has been shown to serve as a sen-

A

B

Figure 3. Factor D inhibitor ACH-4471 blocks C3 fragment deposition on erythrocytes from a paroxymal nocturnal hemoglobinuria (PNH) patient. (A) Representative flow cytometric scattergrams of erythrocytes from PNH Patient 2 treated with acidified C5-depleted serum (ACH-4471-treated or EDTA-treated). C3 fragment deposition was assessed using anti-human C3c antibody (x-axis). PNH mutant erythrocytes were identified by the absence of cell-surface CD59 expression (y-axis). Respective gating thresholds are indicated by vertical and horizontal lines.

Note, a fraction of CD59-negative cells (approx. 10%) were positive when stained with anti-C3d antibody despite staining negative for anti-C3b and anti-C3c antibodies when tested in the presence of EDTA to prevent in vitro complement activation; this result indicates that some CD59-negative cells had been coated with C3 fragments but were transformed into the terminal opsonin C3dg in vivo before harvesting, a process known to be rather rapid (below and Online Supplementary Figure S3).43 (B) ACH-4471 dose-response curve from the experiment in (A); IC50 and IC90 values were 0.031 mM and 0.089 mM, respectively.

haematologica | 2017; 102(3)

471

X. Yuan et al.

A

B

Figure 4. Factor D inhibitors block cell killing by sera from atypical hemolytic uremic syndrome (aHUS) patients. Significant killing of PIGA-null reagent cells by the 5 aHUS patientsâ&#x20AC;&#x2122; sera in the modified Ham test is shown (0.000 mM ACH-3856, left; 0.000 mM ACH-4471, right). The addition of Factor D inhibitors caused a dosedependent reduction of cell killing at the indicated concentrations (ACH-3856, left; ACH-4471, right). Activity was also abrogated by heat inactivation of serum complement (data not shown). Horizontal dashed line indicates the 20% threshold that is considered indicative of cell killing due to alternative pathway of complement (APC) dysregulation in aHUS serum. The 5 patient samples are shown for each condition with mean Âą standard deviation.

sitive and specific indicator of aHUS.5,20 We evaluated ACH-3856 and ACH-4471 in the modified Ham test by incubating PIGA-null cells with aHUS patient sera containing the Factor D inhibitors (Figure 4). In the absence of inhibitor (0.000 mM), all 5 aHUS serum samples promoted a degree of cell killing similar to previously described levels.5,20 Notably, the presence of therapeutic eculizumab in 2 of the 5 sera (Patient 2, sample b, and Patient 3) did not minimize the observed killing; this finding is consistent with previous observations and has been attributed to the eculizumab dilution that accompanies the serum dilution necessary for appropriate assay conditions.20 PIGA-null cell killing activity in aHUS patient sera was heat-sensitive and therefore complement dependent.

ACH-3856 (Figure 4A) and ACH-4471 (Figure 4B) both blocked the APC-mediated killing by aHUS patient sera at sub-micromolar concentrations. These results suggest that small-molecule Factor D inhibitors have the potential to mitigate complement-mediated damage in aHUS patients, and support the utility of the modified Ham assay for preclinical aHUS drug assessment.

Efficacy of ACH-4471 following oral delivery in non-human primates To profile the effects of continuous Factor D inhibition over a period of at least 24 hours in vivo, we administered ACH-4471 orally to 3 monkeys in two serial 200 mg/kg doses separated by 12 hours. This deliberately high dose was chosen based on an initial assessment of pharmacokinetic properties in monkeys which revealed lower oral exposure due to higher clearance than in other animal species, including rats and dogs (data not shown). The monkeys tolerated the compound well and showed no clinical abnormalities. Figure 5A shows ACH-4471 concentrations in plasma at time points from 0 hours (pre-dosing) to 30 hours (18 hours after the second dose). In parallel, serum was collected for determination of APC activity by Wieslab assay and of Factor D concentrations. APC activity was suppressed continuously by more than 95% continuously through the 30-hour time period (Figure 5B) with no significant increase in Factor D concentrations 472

(Figure 5C). The observed stability of serum Factor D concentrations was expected because, as a small molecule, ACH-4471 should not interfere with renal Factor D clearance, in marked contrast to the effect of systemic delivery of the humanized monoclonal anti-Factor D antibody lampalizumab.29 These results demonstrate the potential utility of oral delivery of ACH-4471 for therapeutic inhibition of APC activation.

Discussion The present study describes the novel compounds ACH-3856 and ACH-4471 that potently inhibit Factor D proteolytic activity. Factor D is a highly specific serine protease with Factor B as its only natural substrate,30 making it a favorable target for the largely APC-driven manifestations of both PNH and aHUS. We demonstrate that these inhibitors effectively block APC-mediated cell killing using erythrocytes from PNH patients (Ham test) or serum from aHUS patients (modified Ham test). Furthermore, ACH-4471 inhibited the elevated deposition of C3 fragments that proceeds on PNH cells when terminal complement activity is blocked, such as by therapeutic eculizumab. In addition, ACH-4471 demonstrated good oral bioavailability and inhibited APC activity in serum samples recovered following oral administration to nonhuman primates.

Eculizumab improves life for PNH patients by eliminating intravascular hemolysis, decreasing the need for blood transfusions, and reducing the risk of thrombosis.10,11 Yet, the Factor D inhibitors present a potential advantage as some eculizumab-treated PNH patients experience symptomatic extravascular hemolysis due to dysregulated C3 fragment deposition that proceeds even with inhibition of terminal complement activity.31 Several upstream complement inhibitors under investigation may address this limitation, including the engineered complement receptor 2/Factor H fusion protein TT30,32 peptidic C3 inhibitors (such as Cp40 and APL-2),33 and the C1 esterase inhibitor C1INH (Cinryze).27 Suitable systemic exposure, however, may be difficult to achieve with some of these investigahaematologica | 2017; 102(3)

Factor D inhibitors in PNH and aHUS

A

B

Figure 5. Factor D inhibitor inhibits alternative pathway of complement (APC) activity following oral dosing in cynomolgus monkeys. (A) ACH-4471 concentrations at the indicated time points in plasma samples collected from 3 cynomolgus monkeys following oral dosing with ACH-4471 at 200 mg/kg at 0 and 12 hours (arrows). Key pharmacokinetic parameters were calculated as follows: maximum concentration (Cmax), 4020±1480 ng/mL; time at Cmax (Tmax), 15.3±1.2 hours (3.3±1.2 hours following the second dose), and 30-hour exposure level (AUC0-30), 48,300±19,100 ng/mL/h. Parameters were calculated as mean ± standard deviation (SD) from the 3 animals. (B) APC activity assessed by Wieslab assay in serum samples collected at the indicated time points. Activity in each sample was normalized to the activity in the same animal at 0 h (pre-dosing).

Mean±SD are shown from duplicate assay values. (C) Serum FD concentrations at the indicated time points. Circle, square, and triangle shapes; animals 1, 2, and 3, respectively.

C

Table 4. Alternative pathway of complement-specific inhibition of complement hemolytic activity.

Compound

IC50 (mM)

APC hemolysis IC90 (mM)

N

CPC hemolysis IC50 (mM)

N

ACH-3856 ACH-4471

0.0087 ± 0.0039 0.017 ± 0.011

0.022 ± 0.0089 0.070 ± 0.023

40 6

>200 >200

1 2

APC: alternative pathway of complement; CPC: classical pathway of complement; N: number. Mean ± standard deviation from N independent experiments.

tional compounds, and unlike the Factor D inhibitors these molecules will not be available in oral formulation.

APC-mediated hemolysis of erythrocytes from all 3 PNH patients was sensitive to Factor D inhibitors, albeit over an approximately 7-fold range in IC50 values. The reason for this variation remains unknown, yet the inhibitors showed excellent potency against even the least susceptible PNH patient cells. Also of note, C3 fragment opsonization proved less susceptible than hemolysis, with respective IC50 values of 0.031 mM and 0.004 mM for ACH-4471 against cells from Patient 2. This distinction is likely derived in part from operational differences between the assays, as PNH cell opsonization requires substantially more serum (60%) than hemolysis (20%); moreover, the detection thresholds for APC activity could differ widely if, for example, less activity is required for hemolysis than for C3 fragment detection.

Terminal complement inhibition with eculizumab is haematologica | 2017; 102(3)

also highly effective for treating aHUS. Until recently, there was no functional assay with adequate sensitivity and specificity to distinguish aHUS from other TMA syndromes, or to evaluate potential therapeutic compounds.

The experiments presented here use the recently reported modified Ham test to demonstrate that Factor D inhibitors effectively overcome the characteristic APC activation on target membranes by serum of aHUS patients. Moreover, although the modified Ham test does not replicate the complex pathophysiology of aHUS observed in vivo, this assay may serve as a useful pre-clinical system for testing novel complement inhibitors in aHUS. Importantly for systemic delivery, Factor D levels remained unchanged in the monkeys following ACH4471 administration. The Factor D inhibitors thus differ from the humanized monoclonal anti-Factor D antibody lampalizumab, which is in clinical study as an intravitreal treatment for age-related macular degeneration.34 473

X. Yuan et al.

Lampalizumab delivered intravenously to cynomolgus monkeys elicited nearly 10-fold increases in serum Factor D levels within five hours, effectively neutralizing its inhibitory activity.29 Similar increases have been reported for other therapeutic antibodies against proteins that, like Factor D, have high turnover rates mediated by processes including renal filtration.35 In contrast, the stability of serum Factor D levels in the present study indicates that the small molecule inhibitors will likely maintain their potency following systemic delivery.

A significant concern with complement inhibition is the potential for increased susceptibility to certain bacterial infections. Complete or partial Factor D deficiency in humans is associated with reduced bacterial phagocytosis in vitro,36,37 and complete Factor D deficiency is associated with increased risk for recurrent infections with Neisseria or other encapsulated bacteria which is comparable to the lifetime risk observed in the setting of terminal complement deficiencies.36-41 Hence, vaccination for subjects on a Factor D inhibitor will be warranted as it is for subjects on eculizumab. We anticipate that immunological protection following vaccination against Neisseria will be better preserved by Factor D inhibitors than by eculizumab. Studies have demonstrated, for example, that opsonophagocytosis of Neisseria by granulocytes depends on antibody-mediated activation of the complement classical pathway, and that serum bactericidal activity depends additionally on the complement terminal pathway.42 As Factor D inhibitors selectively target the APC and preserve the classical, lectin,

References 1. Varela JC, Tomlinson S. Complement: An Overview for the Clinician. Hematol Oncol Clin North Am. 2015;29(3):409-427.

2. Walport MJ. Complement. First of two parts. N Engl J Med. 2001;344(14):10581066.

3. Brodsky RA. Paroxysmal nocturnal hemoglobinuria. Blood. 2014;124(18):2804-2811.

4. Cataland SR, Wu HM. Diagnosis and management of complement mediated thrombotic microangiopathies. Blood Rev. 2014; 28(2):67-74.

5. Brodsky RA. Complement in hemolytic anemia. Blood. 2015;126(22):2459-2465.

6. Harboe M, Mollnes TE. The alternative complement pathway revisited. J Cell Mol Med. 2008;12(4):1074-1084.

7. Pangburn MK, Muller-Eberhard HJ. Initiation of the alternative complement pathway due to spontaneous hydrolysis of the thioester of C3. Ann N York Acad Sci. 1983;421:291-298.

8. Lublin DM, Atkinson JP. Decay-accelerating factor: biochemistry, molecular biology, and function. Ann Rev Immunol. 1989; 7:35-58.

9. Rollins SA, Sims PJ. The complementinhibitory activity of CD59 resides in its capacity to block incorporation of C9 into membrane C5b-9. J Immunol. 1990; 144(9):3478-3483.

10. Hillmen P, Young NS, Schubert J, et al. The complement inhibitor eculizumab in paroxysmal nocturnal hemoglobinuria. N Engl J Med. 2006;355(12):1233-1243.

11. Brodsky RA, Young NS, Antonioli E, et al.

474

12.

13.

14.

15.

16.

17.

18.

and terminal pathways, the protection elicited by antibodies through vaccination should be at least partially preserved in the presence of Factor D inhibition. In contrast, eculizumab is expected to block the serum bactericidal activity conferred by vaccination, leading to diminished protection. Furthermore, the higher clearance of small molecules relative to therapeutic biologics will allow for rapid restoration of complement activity if dosing needs to be ceased in the event of infection. Nevertheless, the safety of ACH-4471 will need to be monitored closely, especially given the contribution of the APC-dependent amplification loop, which by one estimate may account for up to 80% of complement classical pathway activation.6 In conclusion, Factor D is a promising target for oral therapy of diseases driven by APC dysregulation. Based on results presented here, and guided by additional assessments of its pharmacology, pharmacokinetic properties, and safety and toxicology, ACH-4471 has been selected for clinical development in PNH and is currently in phase I clinical study.

Funding This work was supported by a grant from the Aplastic Anemia and MDS International Foundation and R01HL133113 (RAB).

We thank the Chemistry, Complement Biology, and ADME groups at Achillion Pharmaceuticals (New Haven, CT, USA) who provided Factor D inhibitors and characterized their pharmacological activities and pharmacokinetics. EG was a Johns Hopkins-Libra fellow during the performance of these studies.

Multicenter phase 3 study of the complement inhibitor eculizumab for the treatment of patients with paroxysmal nocturnal hemoglobinuria. Blood. 2008; 111(4):1840-1847.

Risitano AM, Notaro R, Marando L, et al.

Complement fraction 3 binding on erythrocytes as additional mechanism of disease in paroxysmal nocturnal hemoglobinuria patients treated by eculizumab. Blood.

2009;113(17):4094-4100.

DeZern AE, Dorr D, Brodsky RA. Predictors of hemoglobin response to eculizumab therapy in paroxysmal nocturnal hemoglobinuria. Eur J Haematol. 2013; 90(1):16-24.

Nishimura J, Yamamoto M, Hayashi S, et al. Genetic variants in C5 and poor response to eculizumab. N Engl J Med. 2014;370(7):632-639.

Maga TK, Nishimura CJ, Weaver AE, Frees KL, Smith RJ. Mutations in alternative pathway complement proteins in American patients with atypical hemolytic uremic syndrome. Hum Mutat. 2010; 31(6):E1445-1460.

Frimat M, Tabarin F, Dimitrov JD, et al.

Complement activation by heme as a secondary hit for atypical hemolytic uremic syndrome. Blood. 2013;122(2):282-292.

Noris M, Caprioli J, Bresin E, et al. Relative role of genetic complement abnormalities in sporadic and familial aHUS and their impact on clinical phenotype. Clin J Am Soc Nephrol. 2010;5(10):1844-1859.

Goicoechea de Jorge E, Harris CL, EsparzaGordillo J, et al. Gain-of-function mutations in complement factor B are associated

19.

20.

21.

22.

23.

24.

25.

26.

with atypical hemolytic uremic syndrome.

Proc Natl Acad USA. 2007;104(1):240-245.

Delvaeye M, Noris M, De Vriese A, et al.

Thrombomodulin mutations in atypical hemolytic-uremic syndrome. N Engl J Med. 2009; 361(4):345-357.

Gavriilaki E, Yuan X, Ye Z, et al. Modified Ham test for atypical hemolytic uremic syndrome. Blood. 2015;125(23):3637-3646.

Volanakis JE, Barnum SR, Giddens M, Galla JH. Renal filtration and catabolism of complement protein D. New Engl J Med. 1985;312(7):395-399.

Narayana SV, Carson M, el-Kabbani O, et al. Structure of human factor D. A complement system protein at 2.0 A resolution. J Mol Biol. 1994;235(2):695-708.

Kim S, Narayana SV, Volanakis JE. Catalytic role of a surface loop of the complement serine protease factor D. J Immunol. 1995;154(11):6073-6079.

Cole LB, Chu N, Kilpatrick JM, Volanakis JE, Narayana SV, Babu YS. Structure of diisopropyl fluorophosphate-inhibited factor D. Acta Crystallogr D Biol Crystallogr.

1997;53(Pt 2):143-150.

Coppo P, Schwarzinger M, Buffet M, et al.

Predictive features of severe acquired ADAMTS13 deficiency in idiopathic thrombotic microangiopathies: the French TMA reference center experience. PloS One. 2010;5(4):e10208.

Cataland SR, Yang S, Wu HM. The use of ADAMTS13 activity, platelet count, and serum creatinine to differentiate acquired thrombotic thrombocytopenic purpura from other thrombotic microangiopathies.

Br J Haematol. 2012;157(4):501-503.

haematologica | 2017; 102(3)

Factor D inhibitors in PNH and aHUS

27. DeZern AE, Uknis M, Yuan X, et al.

Complement blockade with a C1 esterase inhibitor in paroxysmal nocturnal hemoglobinuria. Exp Hematol. 2014;42(10):857-861.

28. Katschke KJ Jr, Wu P, Ganesan R, et al.

Inhibiting alternative pathway complement activation by targeting the factor D exosite.

J Biol Chem. 2012;287(16):12886-12892.

29. Loyet KM, Good J, Davancaze T, et al.

Complement inhibition in cynomolgus monkeys by anti-factor d antigen-binding fragment for the treatment of an advanced form of dry age-related macular degeneration. J Pharmacol Exp Ther. 2014; 351(3):527-537.

30. Kim S, Narayana SV, Volanakis JE. Mutational analysis of the substrate binding site of human complement factor D. Biochemistry. 1994;33(48):14393-14399.

31. Hill A, Rother RP, Arnold L, et al.

Eculizumab prevents intravascular hemolysis in patients with paroxysmal nocturnal hemoglobinuria and unmasks low-level extravascular hemolysis occurring through C3 opsonization. Haematologica. 2010; 95(4):567-573.

32. Risitano AM, Notaro R, Pascariello C, et al.

The complement receptor 2/factor H fusion protein TT30 protects paroxysmal noctur-

haematologica | 2017; 102(3)

33.

34.

35.

36.

37.

nal hemoglobinuria erythrocytes from complement-mediated hemolysis and C3 fragment. Blood. 2012;119(26):6307-6316.

Risitano AM, Ricklin D, Huang Y, et al.

Peptide inhibitors of C3 activation as a novel strategy of complement inhibition for the treatment of paroxysmal nocturnal hemoglobinuria. Blood. 2014;123(13):20942101.

Do DV, Pieramici DJ, van Lookeren Campagne M, et al. A phase ia dose-escalation study of the anti-factor D monoclonal antibody fragment FCFD4514S in patients with geographic atrophy. Retina. 2014; 34(2):313-320.

Davda JP, Hansen RJ. Properties of a general PK/PD model of antibody-ligand interactions for therapeutic antibodies that bind to soluble endogenous targets. MAbs. 2010; 2(5):576-588.

Weiss SJ, Ahmed AE, Bonagura VR. Complement factor D deficiency in an infant first seen with pneumococcal neonatal sepsis. J Allergy Clin Immunol. 1998; 102(6 Pt 1):1043-1044.

Biesma DH, Hannema AJ, van Velzen-Blad H, et al. A family with complement factor D deficiency. J Clin Invest. 2001;108(2):233240.

38. Kluin-Nelemans HC, van Velzen-Blad H, van Helden HP, Daha MR. Functional deficiency of complement factor D in a monozygous twin. Clin Exp Immunol.

1984;58(3):724-730.

39. Hiemstra PS, Langeler E, Compier B, et al.

Complete and partial deficiencies of complement factor D in a Dutch family. J Clin Invest. 1989;84(6):1957-1961.

40. Sprong T, Roos D, Weemaes C, et al.

Deficient alternative complement pathway activation due to factor D deficiency by 2 novel mutations in the complement factor D gene in a family with meningococcal infections. Blood. 2006;107(12):4865-4870.

41. Densen P. Complement deficiencies and meningococcal disease. Clin Exp Immunol.

1991;86 (Suppl 1):57-62.

42. Hellerud BC, Aase A, Herstad TK, et al.

Critical roles of complement and antibodies in host defense mechanisms against Neisseria meningitidis as revealed by human complement genetic deficiencies.

Infect Immun. 2010;78(2):802-809.

43. Lin Z, Schmidt CQ, Koutsogiannaki S, et al.

Complement C3dg-mediated erythrophagocytosis: implications for paroxysmal nocturnal hemoglobinuria. Blood.

2015;126(7):891-894.

475

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Blood Transfusion

Ferrata Storti Foundation

Superior survival of ex vivo cultured human reticulocytes following transfusion into mice

Sabine Kupzig,1 Stephen F. Parsons,1 Elinor Curnow, 2 David J. Anstee 1 and Allison Blair 1,3

NIHR Blood and Transplant Research Unit, Bristol Institute for Transfusion Sciences, National Health Service Blood and Transplant; 2Statistics and Clinical Studies, National Health Service Blood and Transplant, Bristol and 3School of Cellular and Molecular Medicine, University of Bristol, UK 1

Haematologica 2017 Volume 102(3):476-483

ABSTRACT

T

Received: August 11, 2016.

Accepted: November 29, 2016.

Pre-published: December 1, 2016.

he generation of cultured red blood cells from stem cell sources may fill an unmet clinical need for transfusion-dependent patients, particularly in countries that lack a sufficient and safe blood supply. Cultured red blood cells were generated from human CD34+ cells from adult peripheral blood or cord blood by ex vivo expansion, and a comprehensive in vivo survival comparison with standard red cell concentrates was undertaken. Significant amplification (>105-fold) was achieved using CD34+ cells from both cord blood and peripheral blood, generating high yields of enucleated cultured red blood cells. Following transfusion, higher levels of cultured red cells could be detected in the murine circulation compared to standard adult red cells. The proportions of cultured blood cells from cord or peripheral blood sources remained high 24 hours post-transfusion (82±5% and 78±9%, respectively), while standard adult blood cells declined rapidly to only 49±9% by this time. In addition, the survival time of cultured blood cells in mice was longer than that of standard adult red cells. A paired comparison of cultured blood cells and standard adult red blood cells from the same donor confirmed the enhanced in vivo survival capacity of the cultured cells. The study herein represents the first demonstration that ex vivo generated cultured red blood cells survive longer than donor red cells using an in vivo model that more closely mimics clinical transfusion. Cultured red blood cells may offer advantages for transfusion-dependent patients by reducing the number of transfusions required.

doi:10.3324/haematol.2016.154443

Introduction

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/476

National Blood Services are an essential part of healthcare, playing key roles in treating patients following trauma, surgery and transplants as well as providing life saving products for patients with blood disorders. Unfortunately, in many countries there are supply shortages of red blood cell (RBC) concentrates for transfusions, and concerns about the safety of the blood supply. The majority of units transfused globally each year are used to treat individuals from developed countries that represent only around 15% of the world population.1 Pressure on blood supply in developed countries is likely to intensify in the longer term with increasing life expectancy, concomitant with greater numbers of surgical procedures in an ageing population and notable rises in the prevalence of cancer.2 Whilst blood transfusions are a life saving procedure for many, as evidenced by the dramatic fall (~99%) in the number of women dying in childbirth from 1920 to 1950,3,4 they can pose significant risks. Individuals who require regular transfusions are at risk of adverse reactions following transfusion of mismatched blood. Patients with chronic transfusion-dependent anemia, such as β-thalassemia or sickle cell disease, are at particular risk of iron overload,5 and aged stored RBCs, which contain a heterogeneous mix of cells at various ages, may have adverse clinical effects in critically ill patients.2 A source of exclusively young RBCs, as found in cultured RBCs, could help address the above challenges for transfusion by increasing the transfusion

Correspondence: allison.blair@bristol.ac.uk

©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

476

haematologica | 2017; 102(3)

Superior in vivo survival of cultured reticulocytes

intervals and reducing iron overload, particularly in patients that depend on regular transfusions.6 Considerable effort has been made to generate cultured red blood cells (cRBCs) ex vivo from CD34+ hemopoietic stem cells (HSCs), human embryonic stem cells or induced pluripotent stem cells (iPSCs).7-16 Many of the published techniques entail multi-phase culture systems, over a period of 18-38 days, and some include co-culture on stroma.

The first group to produce cRBCs in the absence of stroma reported an extrapolated yield of 1.4 units of cRBCs from one cord blood (CB) unit.8 The only clinical study to date used autologous mobilized CD34+ cells from a healthy volunteer as starting material.11 The donor was reinfused with 2ml cRBCs,17 and around 50% of the cultured reticulocytes could be detected 26 days after reinfusion, providing evidence for the feasibility of transfusion of ex vivo generated red cells. Despite these advances, problems associated with enucleation, large-scale generation and financial costs are hurdles that need to be overcome prior to clinical use.18 We have previously described an ex vivo erythroid expansion method for CD34+ cells derived from adult peripheral blood (PB). Using this method it was possible to achieve significant expansion of CD34+ cells to yield 5ml (2.8x 1010) of packed enucleated RBCs. This was the largest yield reported to date from PB, and represented a major advance in developing a product that is suitable for clinical use.13,14 However, few of the reported studies have conducted any in vivo evaluation of the ex vivo generated cells. This is an important consideration that must be addressed to prove that the ex vivo generated cells are suitable for transfusion. Some studies have used sublethally irradiated immune-deficient mice with intraperitoneal (IP) injection of cells following saturation with ABO type O cells, then retro-orbital sampling over a period of 5 days to detect human cells.7,11,15,19 Hu et al. reported that the depletion of macrophages was essential in order to achieve human RBC chimerism in NOD/SCID mice inoculated with CD34+ fetal liver cells alone or with implanted human fetal thymic tissue.20 Whilst these in vivo studies are informative, they do not accurately represent a clinical transfusion where matched red cells are administered intravenously and without additional cell products to saturate the patient.

We have developed a biologically representative in vivo model using NOD/LtSz-scid IL-2Rγc null (NSG) mice, which are more permissive hosts for the engraftment of normal and malignant human blood cells.21,22 In the study herein, we have conducted a comprehensive in vivo evaluation of cRBCs generated from CB and adult PB, and demonstrate that ex vivo generated cRBCs are superior to donor RBCs using a model that more closely mimics clinical transfusion.

Methods Donor samples Blood donor mononuclear cells and CB were provided with informed consent (National Health Service National Research Ethics Committee, reference number 08/H0102/26).

three-stage ex vivo expansion procedure. Briefly, CD34+ HSCs isolated from human PB or thawed cryopreserved CB units were seeded into tissue culture flasks at a density of 2x105 cells/ml, and maintained in the range 2-5x105 cells/ml by division and the addition of first-stage medium until day 10. On days 11-13, cells were maintained at 5-15x105 cells/ml by the addition of second-stage medium. From day 14, cells were maintained in third-stage medium at 10-40x105 cells/ml. Once the total volume reached 200ml, cells were transferred from static flasks to 1.5 liter, stirred (15rpm), vessels. Cells were filtered using a standard leucofilter (Pall WBF, Haemonetics Ltd, Coventry, UK) prior to inoculation into NSG mice.

Microscopy Cytospin preparations of cultured cells were stained using Leishman’s Staining Solution (VWR International, Lutterworth, UK), imaged using a Leica DM750 microscope (Leica Microsystems, Milton Keynes, UK) and photographed using a Pixera Penguin 600CL camera (Digital Imaging Systems, Bourne End, UK).

For live cell confocal microscopy, cRBCs or PB aspirates from transfused mice were stained with fluorescein isothiocyanate (FITC)-conjugated BRIC 256 (mouse monoclonal anti-human CD235a; IBGRL, Bristol, UK) and imaged at 22°C using a Leica SP5 confocal imaging system.

In vivo studies NSG mice were bred and maintained at the University of Bristol, Animal Services Unit. Adult mice were macrophage depleted by intravenous (IV) inoculation of liposome-encapsulated clodronate (dichloromethylene diphosphonate, CI2MDP, The Netherlands) on day -3 (100ml) and day -1 (50ml).

Mice were transfused with 2x108 cRBCs or 2x108 washed adult donor RBCs via the left lateral tail vein. PB aspirates were taken from the right lateral tail vein at 10, 20, 40, 60, 120, 240 and 480 minutes after inoculation, and once daily thereafter up to 9 days. Cells were counted and stained with anti-human CD235a (glycophorin A) antibody and analyzed by flow cytometry. Non-clodronate treated mice were also transfused and analyzed to assess the effects of murine macrophages on the inoculated cells.

For a direct paired comparison, washed RBCs from a standard red cell pack (ABO type, O RhD positive) were transfused 5 days after blood donation. CD34+ HSCs isolated from the same donation were transfused once they had been cultured for 21 days to generate reticulocytes. At the same time, a further aliquot of unmodified donor RBCs was transfused into a separate group of mice; these cells were now 26 days old.

Flow cytometry Cells were stained with BRIC256-FITC for in vivo survival studies, with anti-mouse F4/80-phycoerythrin (PE) for macrophage depletion studies or with BRIC256-FITC and mouse anti-human CD71-RPE (Bio-Rad, Hemel Hempstead, UK) for maturation studies. Samples were analyzed using a Beckman Coulter FC 500 flow cytometer (Beckman Coulter, High Wycombe, UK), the gating strategy is shown in the Online Supplementary Figure S1.

For nucleic acid staining, 6x106 BRIC256-PE stained cells were washed and labelled with 0.1mg/ml thiazole orange (SigmaAldrich, Poole, UK).

Statistical analyses Cell culture See the Online Supplementary Information for full details of the

haematologica | 2017; 102(3)

Full details of statistical analyses are provided in the Online Supplementary Information.

477

S. Kupzig et al.

Results Yields and morphology of cRBCs Using the three-stage culture technique, it was possible to achieve >105-fold amplification of cRBCs from 1x106 CD34+ HSCs. On average, ex vivo cultured adult cells were ~60% enucleated on day 20/21 of culture (Figure 1A). In contrast, CB cells were only ~38% enucleated. The enucleated cells were separated from free nuclei, nucleated precursors and debris by leucocyte filtration, yielding >99% pure fractions (Figure 1B). Filtration yields ranged from 30.5% to 94.9% (average 60.5±7.2%) for adult cRBCs and from 40.5% to 80.5% (average 62.1±8.5%) for cord blood cRBCs. The mean corpuscular volume (MCV) of adult and cord cRBCs was 135mm3 (range 125–142), while that of donor RBCs was 89mm3 (range 83–96).

Mouse RBCs had a MCV of 48mm3 (range 45-52).

Confocal analyses of the leucofiltered cells showed that glycophorin A (GPA) is expressed on the surface, and the morphology of the majority of cells at the end of culture was that of reticulocytes rather than mature biconcave RBCs (Figure 1C), a finding which is also supported by the cells having a larger MCV.

Macrophage depletion of NSG mice is required to allow uptake of RBCs Prior to the commencement of clodronate treatment, macrophage levels in murine PB ranged from 23.7-26.8% (median 25.8%, Figure 2A). Macrophage levels declined significantly to 4.2±1.4% within 24 hours of the second dose of liposomes, remained low over the following 24 hour period (P<0.00003), before gradually increasing over the next 6 days (8 days after the last liposome dose) to reach similar levels to those observed in untreated controls.

To investigate the effects of murine macrophages on transfused cells, separate groups of mice were either pretreated with clodronate liposomes or left untreated prior to inoculation of donor RBCs. Cells were inoculated by IV injection 24 hours after animals received the second dose of clodronate liposomes, when murine macrophages were at their lowest levels; control animals were injected at the same time. In the control, non-treated mice, most of the human RBCs had been removed from circulation within 10 minutes of transfusion, and the levels of human cells detected were almost 6-fold reduced compared to those in macrophage depleted mice (Figure 2B, P=0.0005). The

A

B

remaining human cells were rapidly cleared and were practically undetectable after 1 hour. In contrast, clearance of human cells was significantly slower in mice that had been pre-treated with clodronate liposomes (Figure 2C, P=0.0004). Clodronate-treated animals were used for all results reported below.

In vivo maturation of cultured reticulocytes Macrophage depleted NSG mice were randomized to receive cRBCs or washed donor RBCs 24 hours after receiving the second dose of clodronate liposomes. Live cell confocal imaging of PB samples from mice transfused with adult or cord cRBCs or with donor RBCs showed only human cells stained positive for GPA, and were easily distinguished from the background of mouse RBCs (Figure 3A-C). Human cells were visibly larger than the mouse RBCs, and a significant number of cRBCs appeared to have adopted the biconcave shape of mature red cells.

These cells may have undergone maturation in vivo, most likely in the mouse spleen, when compared to the morphology of the cells that were initially transfused (Figure 1C).

Measurements of cell diameter, taken at various time points, indicated a gradual decrease in the size of human cells transfused into NSG mice (Figure 4A,B). Prior to injection (0 minutes), cord and adult cRBCs had an average diameter of 9.1±0.09mm (n=145) and 9.9±0.09mm (n=64), respectively. In comparison, standard RBCs had a mean diameter of 7.9±0.07mm (n=60) while mouse RBCs measured 5.7±0.07mm (n=60). After 48 hours in the mouse circulation, diameters had decreased to 6.4±0.11mm (n=42) and 6.1±0.19mm (n=16) for cord and adult cRBCs, respectively, indicative of a possible maturation of the cultured human reticulocytes in circulation.

Further evidence for cRBCs maturation in vivo was demonstrated by decreasing CD71 expression (Figure 4C).

In aspirates taken 10 minutes after inoculation, 21.4±7.3% (median fluorescence intensity (MFI) 116±26) of cord cRBCs and 16.2±3.2% (MFI 159±15) of adult cRBCs expressed CD71. After 3 days this had decreased to 7.5±3.4% (MFI 93±36) and 6.4±1.5% (MFI 53±4) for cord and adult cRBCs, respectively, (P≤0.03). By comparison, CD71 expression on donor RBCs did not change significantly over the 3-day time course, ranging from 2.9±0.3% - 5.1±0.7% (MFI 61±5 – 47±5, P=0.12).

The amount of residual nucleic acid remaining in transfused human cells from the different sources was also

C

Figure 1. Cytospin and confocal analyses of day 21 cultured adult reticulocytes. Cultured reticulocytes were processed for morphological analyses by cytospin and stained with Leishman’s solution before (A) and after leucofiltration (B). Leucofiltration resulted in >99% pure population of cultured reticulocytes. (C) Leucofiltered cultured reticulocytes were also processed for live cell confocal imaging using anti-CD235a-FITC antibody.

478

haematologica | 2017; 102(3)

Superior in vivo survival of cultured reticulocytes

assessed by thiazole orange staining. The proportion of thiazole-positive cells decreased from 42.2% and 92.9% at 10 minutes to 9.3% and 8.16% after 3 days in cord and adult cRBCs, respectively. Such a decrease was not observed in donor RBCs (range 1.7%-11.0%, Online Supplementary Figure S2).

Survival of cultured reticulocytes and donor RBCs in vivo Human cells were detectable from 10 minutes after inoculation in mice receiving adult cRBCs (median 0.51% of total circulating murine blood cells, range 0.09-1.37%, n=29), cord cRBCs (median 0.87%, range 0.38-1.57%, n=26) and adult red cells (median 0.91%, range 0.081.61%, n=16, Online Supplementary Figure S3A). Human cells could be detected in murine blood for up to 9 days.

On average, the levels of human cells in animals that received cRBCs peaked 60 mins to 4 hours post-inoculation and then gradually declined. In contrast, standard RBCs peaked during the first 10 minutes followed by a sharp decline in the first 2 hours post-inoculation.

When the proportion of human cells was normalized, with the levels detected 10 minutes after inoculation set to 100%, there were significant differences in the levels of cRBCs and adult RBCs in the murine circulation over the entire 6 day evaluation period (Figure 5A, P≤0.02). Higher levels of human cells were detected in recipients of cRBCs, regardless of source, compared to recipients of adult RBCs for the duration of the experiment. The levels of adult and cord cRBCs detected in the murine circulation were not significantly different (Figure 5A, P=0.7). The levels of both adult and cord cRBCs were significantly higher than that of adult RBCs over the first 3 days (adult cRBCs vs. adult RBCs, P=0.03; cord cRBCs vs. adult RBCs, P=0.01), but particularly so over the initial 8 hours (P<0.0001). At this time point, the proportion of CB cRBCs and adult cRBCs were relatively unchanged (112±6% and 103±7%, respectively). In contrast, the proportion of human cells detected in mice inoculated with RBCs had reduced considerably to 63±7%.

The distribution of half-life, determined by experiment, is shown (Figure 5B). Data for 5 mice receiving cord cRBCs, 6 mice receiving adult cRBCs and 1 mouse receiving adult red cells were excluded from the half-life analysis, since the experiments were terminated prior to cell survival reducing to 50%. There was considerable intraexperimental variation in the half-life of transfused cells (i.e., variation between individual mice transfused with the same source of cells) and inter-experimental variation, likely due to the variation between individual donors.

After taking this into account, some evidence remained of a difference in half-lives across the blood sources, although it was not significant (P=0.1). The mean half-life for cord cRBCs was 56±4 hours, compared with 38±5 hours and 26±8 hours for adult cRBCs and RBCs, respectively. A gradual recovery of murine macrophages was observed 24-48 hours post-transfusion (Figure 2A).

Macrophages are likely to remove all human cells, indiscriminate of source, and this recovery coincides with the observed decline of human cells in the mouse circulation after 24 hours.

To directly compare the survival of cRBCs with donor RBCs in vivo, a matched comparison was undertaken using cells from the same donor (Figure 6A). Overall, significantly higher levels of cRBCs were detected in the murine cirhaematologica | 2017; 102(3)

culation, and the extent of the difference varied over time (P<0.0001). The mean levels of human cells detected remained above 83% in animals transfused with adult cRBCs over the first 8 hours, whilst there was a sharp decline to <53% and <60% in the groups that received day 5 and day 26 adult cells, respectively. The levels of cRBCs detected were significantly higher than day 5 and day 26 adult RBCs (P≤0.01) over this period. There was no significant difference in levels of day 5 and day 26 adult RBCs detected in the murine circulation (P=0.99). From 24 hours after transfusion, similar levels of human cells were detected, regardless of cell source. The proportions of human cells detected in murine circulation are depicted in

A

B

C

Figure 2. Validation of macrophage depletion for transfusion model. Murine peripheral blood samples were collected at designated time points and cells were labelled with PE-conjugated anti-mouse F4/80 (A) or FITC-conjugated antihuman CD235a (B & C) and analyzed by flow cytometry. (A) Circulating levels of murine macrophages measured in liposome treated (n=8) and untreated NSG mice (n=10) over a 10 day period. (B & C) NSG mice that had either been treated with clodronate liposomes to remove macrophages at day -3 and day -1 (n=3) or left untreated (n=2) were inoculated with RBCs from a single donor on day 0. (B) Percentage of human RBCs in the mouse circulation. (C) Clearance rates of human cells in untreated and macrophage depleted mice. Human cells were normalized to 100% at 10 minutes after injection. Data shown as mean±SE. ***P<0.00003. RBCs: red blood cells; PB: peripheral blood.

479

S. Kupzig et al.

the Online Supplementary Figure S3B. The mean half-life for cRBCs was 47±6 hours compared with 46±16 hours and 56±4 hours for adult red cells on day 5 and day 26, respectively (P=0.74, Figure 6B). Again, the recovery of murine macrophages is likely to be responsible for the removal of human cells after 24 hours.

Discussion There is currently a global imbalance between the supply and demand for red blood cells for transfusion. We have previously shown it is possible to generate large numbers of enucleated cRBCs from donor PB CD34+ HSCs.13,14 In the study herein, we used a good manufacturing practice (GMP) compliant procedure to achieve a >105-fold amplification from 106 CD34+ HSCs, with 63% being enucleated, yielding ~10ml packed cells. To the best of our knowledge this represents the greatest yield of enucleated cRBCs reported to date. In addition, we demonstrate that the same method also permits the generation of cRBCs from CD34+ HSCs isolated from cord blood samples, and have conducted a thorough in vivo survival assessment of these ex vivo generated reticulocytes.

CD34+ HPCs were cultured in the presence of cytokines over a 20-21 day period in the absence of a feeder layer.

During this time the cells mature to produce a mixed population of nucleated precursors, free nuclei and around 3095% enucleated reticulocytes. Filtration, using a leucodepletion filter, removes nucleated cells, free nuclei and debris, resulting in a homogeneous suspension of reticulocytes (Figure 1B,C) for transfusion. Following enucleation, a reticulocyte needs to loose 20-30% of its plasma membrane to become a mature biconcave red cell.23,24 We have previously shown that excess plasma membrane is internalized by maturing reticulocytes. These membrane vesi-

cles fuse with autophagosomes, which are subsequently expelled by the cells.13,14,25 Evidence suggests that the final maturation step occurs in the spleen, since it has long been known that splenectomised patients show an increased number of circulating reticulocytes containing autophagosomes.26 In order to fully evaluate the functional capacity of the cRBCs, we assessed their maturation and survival in clodronate depleted NSG mice. We initially tested the method of macrophage depletion reported by Hu et al. in NOD/SCID mice,20 and found similar results with NSG mice, in that clearance of human RBCs was prevented in animals that had been pre-treated with clodronate liposomes. This model more closely mimics a clinical transfusion than those previously reported,7,11,15,19 demonstrating non-toxic survival and maturation of ex vivo generated cRBCs, confirming they are suitable for transfusion in vivo.

Following transfusion into NSG mice, the majority of reticulocytes appear to mature into biconcave RBCs (compare Figure 1C with Figure 3A,B). We were able to measure a reduction in cell diameter over time for transfused cord and adult cRBCs in the mouse circulation. It is likely that this final maturation step takes place in the mouse spleen. Macrophages, which have largely been removed in clodronate-treated mice, do not seem to be required for the RBCs to achieve their final biconcave shape.

However, they may be required to remove any membrane vesicles extruded by the maturing reticulocytes in order to reduce the amount of plasma membrane, cytoplasm and residual organelles.13,14,25 In addition to these morphological changes, we also observed a reduction in the amount of CD71 expression on the surface of transfused human reticulocytes and in the amount of thiazole orange staining over a 3-day time course, suggesting that in vivo maturation of human cRBCs was occurring in the murine system. Given that human cRBCs (MCV=135mm3) and

A

B

C Figure 3. Live images of mouse blood samples containing GPA-positive human cRBCs and adult RBCs. Macrophage depleted mice were injected with 2x108 cord blood cultured reticulocytes (half-life = 54.8 hours) (A), adult cultured reticulocytes (half-life = 48.2 hours) (B) or with donor RBCs (half-life = 29.5 hours) (C). Live confocal imaging of peripheral blood aspirates taken at 10 minutes, 8, 24, 48 and 72 hours postinjection was performed on cells labelled with anti-CD235a-FITC antibody. RBCs: red blood cells; cRBCs: cultured red blood cells; CB: cord blood.

480

haematologica | 2017; 102(3)

Superior in vivo survival of cultured reticulocytes

human donor RBCs (MCV =89mm3) are significantly larger than mouse RBCs (MCV =48mm3), it is possible that increased shear stress in the mouse capillaries contributed to the observed maturation from reticulocytes to red cells.

Survival comparisons of cRBCs and donor RBCs in vivo revealed that higher proportions of human cells were detected in animals transfused with cRBCs from either CB or from adult PB. In contrast, a sharp decline in the proportion of human cells was observed in animals inoculated

A

with adult RBCs. This is most likely due to the heterogeneous cell population present in a standard adult RBC donation, while the ex vivo generated cRBCs comprise a much more homogeneous population of younger red cells.

The dramatic decline observed following transfusion of RBCs mirrors the situation in humans where ~25% of cells are cleared within 24 hours following transfusion of packed RBCs,27 with most cleared during the first hour.28 Despite inoculating up to 25-fold fewer cells in the study

B

C

Figure 4. Phenotypic maturation of transfused cells. (A) The diameters of antiCD235a-FITC labelled cord and adult cRBCs and human donor RBCs were measured before injection into mice (0 minutes) and at various time points following transfusion. Diameters of murine RBCs were also determined. Data represent mean±SE measurements of cRBCs from 1 CB sample (half-life = 54.8 hours) and 2 adult donors (half-life = 48.2 hours) and RBCs from 1 donor (half-life = 29.5 hours). (B) Confocal images of representative CD235a labelled adult cRBCs at 0 minutes and 48 hours (half-life = 48.2 hours) with corresponding bright field images. (C) Proportion of CD71-positive cells in the GPA-positive human cell population. Numbers in key represent the number of mice inoculated with cells from a single donor of each cell source. Data represent mean±SE. GPA: glycophorin A. CB: cord blood: RBCs: red blood cells; cRBCs: cultured red blood cells.

A

B

Figure 5. Detection and survival of human cRBCs and adult RBCs in transfused NSG mice. cRBCs or adult blood cells were inoculated into the lateral tail vein of macrophage depleted NSG mice. Human cells were detected by measuring expression of CD235a in murine blood by flow cytometry. (A) Proportion of human cells surviving in mouse circulation. Data represent mean±SE of 8 independent experiments for the adult cultured reticulocytes, 5 independent experiments for standard donor cells and 4 independent experiments for cord blood cultured reticulocytes. The mice in each independent experiment were injected with cells from the same donor. The proportion of human cells were normalized, with the levels detected 10 minutes after inoculation set to 100%. Survival of both adult and cord cRBCs in NSG mice was significantly greater than adult red cells at all measured time points over the entire time course of the experiment (P≤0.02). (B) Half-life of human cells in transfused NSG mice, by experiment and by source. Data from mice where the levels of human cells decreased to ≤50% by the end of the experiment were used to calculate the half-life. Each point represents an individual mouse, results from mice inoculated with cells from the same donor are stacked vertically, lines represent mean±SE. GPA: glycophorin A; CB: cord blood: RBCs: red blood cells; cRBCs: cultured red blood cells.

haematologica | 2017; 102(3)

481

S. Kupzig et al.

A

B

Figure 6. cRBCs demonstrate better survival than RBCs from the same donor. (A) Direct paired comparison of cRBCs and RBCs from the same donor. Five day old adult red cells were transfused into NSG mice (n=5). Twenty-one days later cRBCs generated from this sample were transfused into a separate group of mice (n=6).

At the same time a third group of mice was transfused with unmodified red cells from the same donor that were now 26 days old (n=4). ANOVA showed that overall survival of cRBCs was significantly better than day 5 and day 26 adult red cells (P<0.0001). (B) Half-life of human cells in transfused NSG mice, by source. Each point represents an individual mouse, meanÂąSE are shown. RBCs: red blood cells; cRBCs: cultured red blood cells; GPA: glycophorin A.

herein, the levels of human red cells detected were comparable with previous reports where cRBCs derived from granulocyte-colony stimulating factor (G-CSF) primed leukaphereses or normal donor PB were inoculated into humanized NOD/SCID mice.7,11 Moreover, we demonstrated that survival of ex vivo generated reticulocytes was significantly superior to that of donor RBCs. The only other published in vivo comparison of cRBCs and native RBCs reported survival from both sources was similar over a 3-day evaluation period, but data on half-lives were not provided.7 Our findings also represent the first report on in vivo survival of cRBCs generated from CB cells.

There was no difference in the levels of human cells detected or the in vivo survival of cRBCs from CB or PB, demonstrating that both are suitable sources for ex vivo reticulocyte generation. The variation observed between batches of cRBCs from CB and PB is a known issue in the field.29 As the starting material is comprised of CD34+ cells at different stages of maturation, such variation is not unexpected. Close monitoring of individual cultures, to optimize production of enucleated RBCs, will be necessary for clinical applications. In the study herein, the median half-life of donor RBCs was 30 hours while those of cRBCs from adult PB or CB were 40 and 58 hours, respectively. The larger variation in half-lives observed using adult cRBCs and RBCs can be attributed to variation in individual donor samples. cRBCs derived from CB were more uniform in this respect. Regardless of source, cRBCs remained detectable for 6-9 days while adult RBCs were largely undetectable after 72 hours. It has been previously reported that the levels of circulating human red cells in mouse models decline within a few days of the last clodronate liposome treatment, possibly due to recovery of murine macrophages.20 We demonstrated that murine macrophages begin to recover 48-72 hours after the last clodronate treatment, 24-48 hours following transfusion, and this recovery coincides with the decline in the levels of human cells detected in murine circulation. This has a direct impact on the half-life of human cells, since all xenogeneic cells, irrespective of source, will be removed by the macrophages. Consequently, it is possible that dif482

ferences in macrophage depletion may also contribute to the observed variation in half-lives.

To further evaluate the observed enhanced survival of cRBCs in vivo, a direct comparison of cRBCs and adult RBCs from the same donor was undertaken. Donor cells were used 5 days after leukapheresis, a typical age of RBC units issued for transfusion and, for the purposes of direct comparison with cRBCs, at day 26. Higher levels of transfused cRBCs were detected compared to both day 5 and day 26 donor red cells. In NSG mice that received cRBCs, high levels of human cells were maintained over the first 2 hours and then a gradual decline was observed. In contrast, the levels of human cells in mice that received adult RBCs declined rapidly and there was no difference in either the levels of human cells detected or the survival of day 5 and day 26 stored cells. This latter finding, showing no significant difference in survival using fresh and stored blood, was not confirmed in a subsequent experiment using day 8 and day 29 adult RBCs (Online Supplementary Figure S4), suggesting deterioration during storage may vary between donors.30 Further work will be required to address this. The lack of a significant difference in the halflives of cRBCs or day 5 and day 26 donor RBCs can be attributed to the recovery of murine macrophages. To the best of our knowledge, this is the first report of paired comparisons of cRBCs and native RBCs from the same donor.

Our rationale for generating red cells ex vivo is that they provide a cohort of younger cells compared to donated blood, and as such may offer clinical advantages by surviving longer and possessing superior functional characteristics. The study herein represents the first demonstration that cRBCs generated from either CB or adult PB have prolonged survival in vivo compared to adult RBCs. We have previously shown adult cRBCs are comparable to donor RBCs in terms of their deformability, oxygen-binding capacity and serology.13,25 Further work is required to determine the quantity of cRBCs that constitute a therapeutic dose, and provide a cost effective manufacturing process.

A logical progression of this work will be an allogeneic haematologica | 2017; 102(3)

Superior in vivo survival of cultured reticulocytes

survival and recovery trial in man to compare the half-life of cRBCs with that of donor RBCs. A cRBC product with increased survival will offer several advantages over current red cell products for certain patient groups. Examples are reduction in donor exposure and iron overload in chronically transfused patients, such as those with β-thalassaemia, and difficult to transfuse patients, such as some patients with sickle cell disease. Cellular reprogramming may also be an important approach to generate stem cells for therapeutic purposes, albeit very difficult to produce on a large scale.31 Erythroid progenitor cells that undergo enucleation and hemoglobin switching in vivo have been generated from human iPSCs.19,32,33 However, such cell lines are not representative of adult hemopoiesis. We anticipate that the creation of immortalised erythroid progenitor cell lines34 and their genetic manipulation could provide a very valuable source of cRBCs with rare blood

References 1. World Health Organization Global Database on Blood Safety Summary Report.

Available from: http:// www.who.int/bloodsafety/global_database/GDBS_Summary_Report_2011.pdf?u a=1. 2011. Last accessed: 29th November 2017.

2. Williamson LM, Devine DV. Challenges in the management of the blood supply.

Lancet. 2013;381(9880):1866-1875.

3. Chamberlain G. British maternal mortality in the 19th and early 20th centuries. J R Soc Med. 2006;99(11):559-563.

4. Goldenberg RL, McClure EM. Maternal mortality. Am J Obstet Gynecol. 2011; 205(4):293-295.

5. Anstee DJ, Gampel A, Toye AM. Ex-vivo generation of human red cells for transfusion. Curr Opin Hematol. 2012;19(3):163169.

6. Triadou P, Girot R, Rebibo D, et al.

Neocytopheresis: a new approach for the transfusion of patients with thalassaemia major. Eur J Pediatr. 1986;145(1-2):10-13.

7. Giarratana MC, Kobari L, Lapillonne H, et al. Ex vivo generation of fully mature human red blood cells from hematopoietic stem cells. Nat Biotechnol. 2005;23(1):69-74.

8. Miharada K, Hiroyama T, Sudo K, Nagasawa T, Nakamura Y. Efficient enucleation of erythroblasts differentiated in vitro from hematopoietic stem and progenitor cells. Nat Biotechnol. 2006;24(10):12551256.

9. Baek EJ, Kim HS, Kim S, Jin H, Choi TY, Kim HO. In vitro clinical-grade generation of red blood cells from human umbilical cord blood CD34+ cells. Transfusion. 2008; 48(10):2235-2245.

10. Boehm D, Murphy WG, Al-Rubeai M. The potential of human peripheral blood derived CD34+ cells for ex vivo red blood cell production. J Biotechnol. 2009; 144(2):127-134.

11. Giarratana MC, Rouard H, Dumont A, et al. Proof of principle for transfusion of in vitro-generated red blood cells. Blood.

2011;118(19):5071-5079.

12. Kim HO, Baek EJ. Red blood cell engineering in stroma and serum/plasma-free conditions and long term storage. Tissue Eng. Part A. 2012;18(1-2):117-126.

13. Griffiths RE, Kupzig S, Cogan N, et al.

haematologica | 2017; 102(3)

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

group phenotypes that could particularly benefit immunized, difficult to transfuse sickle cell patients.

Acknowledgments The authors would like to thank Drs Nicola Cogan and Rebecca Griffiths for assistance in filtering reticulocytes, and Charlotte Cox and Dr Paraskevi Diamanti for assistance with animal studies.

Funding This research was funded by grants from the Department of Health (England), the Wellcome Trust and the National Institute for Health Research Blood and Transplant Unit (NIHR BTRU) in Red Blood Cell Products at University of Bristol in Partnership with NHS Blood and Transplant (NHSBT). NIHR did not fund the animal work described in this study. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, the Department of Health or NHSBT.

Maturing reticulocytes internalize plasma membrane in glycophorin A-containing vesicles that fuse with autophagosomes before exocytosis.

Blood.

2012; 119(26):6296-6306.

Griffiths RE, Kupzig S, Cogan N, et al. The ins and outs of human reticulocyte maturation: autophagy and the endosome/exosome pathway. Autophagy. 2012; 8(7):1150-1151.

Kobari L, Yates F, Oudrhiri N, et al. Human induced pluripotent stem cells can reach complete terminal maturation: in vivo and in vitro evidence in the erythropoietic differentiation model. Haematologica. 2012; 97(12):1795-1803.

Jin H, Kim HS, Kim S, Kim HO. Erythropoietic potential of CD34+ hematopoietic stem cells from human cord blood and G-CSF-mobilized peripheral blood. BioMed Res Int. 2014;2014:435215.

Rousseau GF, Giarratana MC, Douay L. Large-scale production of red blood cells from stem cells: what are the technical challenges ahead? Biotechnol J. 2014; 9(1):28-38.

Kim HO. In-vitro stem cell derived red blood cells for transfusion: are we there yet? Yonsei Med J. 2014;55(2):304-309.

Hirose S, Takayama N, Nakamura S, et al.

Immortalization of erythroblasts by cMYC and BCL-XL enables large-scale erythrocyte production from human pluripotent stem cells. Stem Cell Reports. 2013; 1(6):499-508.

Hu Z, Van Rooijen N, Yang Y-G. Macrophages prevent human red blood cell reconstitution in immunedeficient mice.

Blood. 2011;118(22):9.

Diamanti P, Cox CV, Blair A. Comparison of childhood leukemia initiating cell populations in NOD/SCID and NSG mice.

Leukemia. 2012;26(2):376-380.

Diamanti P, Cox CV, Moppett JP, Blair A. Parthenolide eliminates leukemia-initiating cell populations and improves survival in xenografts of childhood acute lymphoblastic leukemia. Blood. 2013;121(8):13841393.

Wiley JS, Shaller CC. Selective loss of calcium permeability on maturation of reticulocytes. J Clin Invest. 1977;59(6):1113-1119.

Waugh RE, McKenney JB, Bauserman RG, Brooks DM, Valeri CR, Snyder LM. Surface area and volume changes during matura-

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

tion of reticulocytes in the circulation of the baboon. J Lab Clin Med. 1997;129(5):527535.

Mankelow TJ, Griffiths RE, Trompeter S, et al. Autophagic vesicles on mature human reticulocytes explain phosphatidylserinepositive red cells in sickle cell disease.

Blood. 2015;126(15):1831-1834.

Connor J, Pak CC, Schroit AJ. Exposure of phosphatidylserine in the outer leaflet of human red blood cells. Relationship to cell density, cell age, and clearance by mononuclear cells. J Biol Chem. 1994;269(4):23992404.

Hod EA, Brittenham GM, Billote GB, et al.

Transfusion of human volunteers with older, stored red blood cells produces extravascular hemolysis and circulating non-transferrin-bound iron. Blood. 2011; 118(25):6675-6682.

Luten M, Roerdinkholder-Stoelwinder B, Schaap NP, de Grip WJ, Bos HJ, Bosman GJ. Survival of red blood cells after transfusion: a comparison between red cells concentrates of different storage periods.

Transfusion. 2008;48(7):1478-1485.

Douay L. In vitro generation of red blood cells for transfusion: a model for regenerative medicine. Regen. Med. 2014;7(1):2.

Hess JR, Sparrow RL, van der Meer PF, Acker JP, Cardigan RA, Devine DV. Red blood cell hemolysis during blood bank storage: using national quality management data to answer basic scientific questions. Transfusion. 2009;49(12):2599-2603.

Giani FC, Fiorini C, Wakabayashi A, et al.

Targeted application of human genetic variation can improve red blood cell production from stem cells. Cell Stem Cell.

2016;18(1):73-78.

Doulatov S, Vo LT, Chou SS, et al.

Induction of multipotential hematopoietic progenitors from human pluripotent stem cells via respecification of lineage-restricted precursors. Cell Stem Cell. 2013;13(4):459470.

Huang X, Shah S, Wang J, et al. Extensive ex vivo expansion of functional human erythroid precursors established from umbilical cord blood cells by defined factors. Mol Ther. 2014;22(2):451-463.

Trakarnsanga K, Wilson M, Griffiths R, et al. The first human immortalised cell line generated from adult erythroid cells. Vox Sanguinis. 2015;109(Suppl. 1):197-197.

483

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Patelet Biology & Its Disorders

Ferrata Storti Foundation

Gfi1b controls integrin signaling-dependent cytoskeleton dynamics and organization in megakaryocytes Hugues Beauchemin,1 Peiman Shooshtarizadeh,1 Charles Vadnais,1 Lothar Vassen,1 Yves D Pastore2 and Tarik Möröy1,3,4

Institut de Recherches Cliniques de Montréal, IRCM, QC; 2Département de Pédiatrie, Service d'Hématologie et Oncologie, CHU Ste-Justine, Montréal, QC; 3Département de Microbiologie, Infectiologie et Immunologie, Université de Montréal, QC and 4Division of Experimental Medicine, McGill University, Montréal, QC, Canada 1

Haematologica 2017 Volume 102(3):484-497

ABSTRACT

M

Pre-published: January 12, 2017.

utations in GFI1B are associated with inherited bleeding disorders called GFI1B-related thrombocytopenias. We show here that mice with a megakaryocyte-specific Gfi1b deletion exhibit a macrothrombocytopenic phenotype along a megakaryocytic dysplasia reminiscent of GFI1B-related thrombocytopenia. GFI1B deficiency increases megakaryocyte proliferation and affects their ploidy, but also abrogates their responsiveness towards integrin signaling and their ability to spread and reorganize their cytoskeleton. Gfi1b-null megakaryocytes are also unable to form proplatelets, a process independent of integrin signaling. GFI1B-deficient megakaryocytes exhibit aberrant expression of several components of both the actin and microtubule cytoskeleton, with a dramatic reduction of α-tubulin. Inhibition of FAK or ROCK, both important for actin cytoskeleton organization and integrin signaling, only partially restored their response to integrin ligands, but the inhibition of PAK, a regulator of the actin cytoskeleton, completely rescued the responsiveness of Gfi1b-null megakaryocytes to ligands, but not their ability to form proplatelets. We conclude that Gfi1b controls major functions of megakaryocytes such as integrin-dependent cytoskeleton organization, spreading and migration through the regulation of PAK activity whereas the proplatelet formation defect in GFI1B-deficient megakaryocytes is due, at least partially, to an insufficient α-tubulin content.

doi:10.3324/haematol.2016.150375

Introduction

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/484

Platelets are small circulating cell fragments essential for blood clotting. The normal platelet count in humans is 150-400x109/L. The short life span of platelets is compensated by the continuous process of thrombopoiesis (ongoing platelet production) by highly specialized polyploid cells called megakaryocytes that form protrusions named “proplatelets” from which platelets are released into the bloodstream.1 Thrombocytopenia is a disorder defined by low platelet counts caused by different conditions ranging from autoimmune destruction of platelets to the complete absence of platelet-producing megakaryocytes.2,3 Three recent studies in patients exhibiting macrothrombocytopenia associated with megakaryocytic dysplasia have identified the transcription factor GFI1B as being responsible for a novel platelet syndrome4-6 named “GFI1B-related thrombocytopenia” (GFI1B-RT)7 or “Bleeding Disorder, Platelet-Type, 17” (BDPLT17) as listed in the OMIM database (http://www.omim.org/entry/187900).

Previous reports of a potential role for Gfi1b in thrombopoiesis showed that GFI1B deficiency led to a severe impairment of both erythropoiesis and thrombopoiesis in mice8,9 and that acute ablation of Gfi1b in adult animals is lethal due to erythropoietic failure, confirming the role of Gfi1b in late erythropoiesis.10-12 To circumvent this limitation and to analyze how Gfi1b regulates both megakarypoiesis and thrombopoiesis, we crossed mice carrying conditional Gfi1b alleles (Gfi1bfl/fl)13

Correspondence: tarik.moroy@ircm.qc.ca

Received: May 31, 2016.

Accepted: January 11, 2017.

©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

484

haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

either with animals expressing a megakaryocyte-specific Pf4-Cre transgene14 or with Rosa-Cre-ERT transgenic mice that enable a tamoxifen-inducible deletion.11 We showed before that by inducing Cre activity in Rosa-Cre-ERT Gfi1bfl/fl mice using a suboptimal dose of tamoxifen, the block in erythroid maturation can be prevented by allowing the formation of erythroblasts that retain non-deleted Gfi1b alleles.11 Here we show that this strategy allows the mice to survive and induces a strong expansion of megakaryocytes with fully excised Gfi1b alleles. This permitted the analysis of GFI1B-deficient megakaryocytes, and allowed us to demonstrate precisely how Gfi1b acts in both early and late stages of megakaryocytic differentiation. At early stages of megakaryocyte maturation, Gfi1b controls megakaryocyte polyploidization and motility. At later stages, the loss of GFI1B disrupts cytoskeleton organization and blocks platelet formation. Here, we present evidence that Gfi1b exerts its function in cellular spreading and motility by controlling integrin signaling pathways principally through the inhibition of p21-activated kinases (PAKs), whereas the defect of proplatelet formation can be explained by a microtubule defect due to an almost complete absence of α-tubulin in GFI1B-deficient cells.

Methods Mice The protocol for this research was reviewed and assessed by the Animal Care Committee (ACC, #2013-04) of the Clinical Research Institute in Montréal and all the animals used in this experiment were cared for in compliance with the Canadian Council on Animal Care (www.ccac.ca) guidelines and principles.

In vitro culture of megakaryocytes Bone marrow cells were lineage (B220–Mac-1–Gr-1–CD16/32–) depleted on an AutoMACS Pro system (Miltenyi) and suspended in StemSpan SFEM (StemCell Technologies) supplemented with 2.6% fetal bovine serum, 1% L-glutamine and stem cell factor (20 ng/mL) then cultured for 2 days at 37°C in 5% CO2. Cells were transferred into fresh medium containing thrombopoietin (50 ng/mL) and cultured for 4 more days. Mature megakaryocytes were enriched on a bovine serum albumin gradient as previously described15 and plated in 12-well m-Chamber glass slides (ibidi) coated with fibronectin (500 mg/mL; Life Technology), fibrinogen (100 mg/mL; Hyphen BioMed) or collagen (1:20; StemCell Technologies). Cells were then allowed to spread for 3-8 h. Cells were fixed in m-Chamber slides with 4% formaldehyde/phosphate-buffered saline, permeabilized with 0.1% TritonX100/phosphate-buffered saline and blocked with FcBlock (1:500; BD Biosciences), then labeled with FITC-CD41 (BD Biosciences) and AF555-β-tubulin (Cell Signaling) antibodies or AF555Phalloidin (Molecular Probes). All conditions for both controls and knockouts were applied to the same m-Chamber slide to minimize variation between samples.

Results Severe thrombocytopenia associated with the loss of Gfi1b expression We generated a megakaryocyte-specific knockout of Gfi1b using Pf4-Cre and Rosa-Cre-ERT mice (Figure 1). Pf4haematologica | 2017; 102(3)

Cre deletes only in mature Gfi1bwt/fl megakaryocytes (>70%) but in Gfi1bfl/fl megakaryocyte deletion starts earlier and with higher efficiency (>99%) (Figure 1D) probably due to a de-repression of the Pf4 promoter in the absence of GFI1B. Pf4-Cre Gfi1bfl/fl mice were viable, but showed internal bleeding and a 1,000-fold reduction of platelets owing to a failure of platelet production rather than a clearance of circulating platelets (Figure 2A-C). Since the Pf4-Cre transgene is active only from late megakaryocyte progenitors to mature megakaryocytes (Figure 1D), we also used the Rosa-Cre-ERT transgene, enabling a tamoxifen-inducible acute ablation of Gfi1b in adult mice. The two-color RosamT/mG reporter strain shows efficient ablation from lin–cKit+Sca1+ (LSK) cells to mature megakaryocytes (Figure 1D). By reducing tamoxifen doses, we were able to obtain a good excision rate of the floxed Gfi1b alleles in megakaryocytes, while allowing erythroid cells to escape arrest and the mice to survive (Figure 1B-E). In this model, tamoxifen provoked a reduction of circulating platelets by day 4 after the first injection, which was followed by rapid clearance leading to a minimal platelet level 3 days later (Figure 2D, E). A reduction of reticulated platelets was observed as early as 2-3 days after the first tamoxifen injection (Figure 2D-F), suggesting that platelet production stops shortly after Gfi1b ablation. A thrombocytopenic state is maintained in GFI1B-deficient mice even several months after tamoxifen injections, excluding the possibility of a transient effect (Figure 2E).

MGG-stained blood smears confirmed the reduction of circulating platelets in the Rosa-Cre-ERT Gfi1bfl/fl mice and an almost complete absence of platelets in the Pf4-Cre Gfi1bfl/fl mice. Some rare remaining platelets detected in both models showed a larger perimeter and lower α-granule content (Figure 2G), reminiscent of the platelets seen in GFI1B-RT. A cytospin from plasma confirmed this and showed large and highly vacuolar platelets which resembled megakaryocytic fragments (Figure 2H) similar to those seen in GPIbα knockout mice.16 Platelets from both mice were also globally larger than those from controls (Figure 2I).

Loss of GFI1B drives megakaryocyte expansion In Pf4-Cre Gfi1bfl/fl mice, the number of megakaryocytes defined as lin–CD41+CD61+ was increased 3- to 4-fold compared to controls (Figure 3A). This was accompanied by an equivalent increase of megakaryocyte precursors able to form colonies (CFU-Mk), which were otherwise indistinguishable from controls with regards to cell number or morphology (Figure 3B). Megakaryocyte expansion was even more pronounced in Rosa-Cre-ERT Gfi1bfl/fl mice, in which an already detectable increased cellularity 3 days after tamoxifen injection gradually reached a >25-fold increase after 2 months (Figure 3C,D and Online Supplementary Figure S1A). This megakaryocyte expansion was independent of the thrombocytopenic state since it started 1 day before the loss of platelets became apparent and thrombopoietin levels were not affected in the knockout animals (data not shown). Colony-forming assays from Rosa-Cre-ERT Gfi1bfl/fl mice showed a proportional increase in CFU-Mk numbers (Figure 3E and Online Supplementary Figure S1B). In contrast to cells from Pf4-Cre Gfi1bfl/fl mice, cells from Rosa-Cre-ERT Gfi1bfl/fl mice gave rise to colonies with fewer and smaller cells than control CFU-Mk and reminiscent of an immature state (Figure 3E and Online Supplementary Figure S1C). A colony assay was 485

H. Beauchemin et al.

A

C

B

D

E

Figure 1. Strategies and efficiencies of Gfi1b deletion in two mouse models. (A) The Pf4-Cre transgene that drives Cre expression only in megakaryocytes was introduced by breeding into either Gfi1bfl/fl (carrying two floxed alleles) and Gfi1bwt/fl (carrying one floxed allele and one wild-type allele) mice, the Gfi1bfl/EGFP (carrying one floxed allele and one EGFP-knock-in allele) and Gfi1bwt/EGFP (carrying one wild-type allele and one EGFP-knock-in allele) that allowed for visualization of megakaryocytes and progenitor cells, or Gfi1bfl/fl and Gfi1bwt/fl mice also carrying the RosamT/mG reporter gene to monitor Cre activity. (B) The Rosa-Cre-ERT locus that drives Cre-ERT expression in all cell types upon tamoxifen treatment was introduced into Gfi1bfl/fl and Gfi1bwt/fl, Gfi1bfl/EGFP and Gfi1bwt/EGFP, and Gfi1bfl/fl and Gfi1bwt/fl RosamT/mG mice as in (A). Suboptimal doses of tamoxifen were injected at 100 mg/kg on day 0 followed by 50 mg/kg the day after and mice analyzed either from day 2-8 for short-term experiments or after 2 or 9 months for long-term experiments. (C) Efficiency of tamoxifen-induced Gfi1b-depletion in the mice from (B) 2 months after injection was assessed by polymerase chain reaction amplification of the Gfi1b-locus in some hematopoietic cellular subsets to see the conversion of the floxed (Flox) allele toward the knockout (KO) allele. GMP: granulocyte-macrophage progenitor; MEP: megakaryocyte-erythroid progenitor; Mks: megakaryocytes; Ery: erythroid cells. (D) The specificity of the two Cre models shown in (A) and (B) was also assessed using RosamT/mG reporter mice in which cells become EGFP positive upon Cre activation. In Pf4-Cre mice, the Cre was active mainly in the later stages of megakaryocyte maturation, although a little earlier in the knockout but was not present in PreMegE or LSK. In the Rosa-Cre-ERT, the Cre was active in all hematopoietic lineages but more prominent in maturing megakaryocytes. (E) Visualization of the alignment tracks obtained through the UCSC genome browser of the Gfi1b locus taken from the RNA-Seq analyses (Figure 7) of Rosa-Cre-ERT Gfi1bfl/fl [KO (ROSA)] and Gfi1bwt/fl [Ctrl (ROSA)] or Pf4-Cre Gfi1bfl/fl [KO (PF4)] Gfi1bwt/fl [Ctrl (PF4)] showing the absence of exons 2-4 in the knockout, confirming its efficiency, the first exon being on the right. TSS: transcription start site.

done with sorted lin–Kit+ (LK) cells from Rosa-Cre-ERT Gfi1bfl/fl also carrying the RosamT/mG two-color Cre reporter allele expressing red fluorescence prior to Cre exposure and green fluorescence in Cre-expressing cells having actively excised floxed genomic regions. GFP+ (Gfi1b deleted) LK produced CFU-Mk containing small acetylcholinesterase-positive cells whereas Tomato+ (Gfi1b not deleted) LK from the same mice produced phenotypically normal CFU-Mk (Figure 3F), demonstrating that the aberrant megakaryocyte phenotype is intrinsic to Gfi1b-deleted cells.

Abnormal morphology and endomitosis in Gfi1b-null megakaryocytes Compared to controls, GFI1B-deficient megakaryocytes had a hyperlobulated nucleus with an abnormal cellular localization, forming a ring around the cell’s inner mem486

brane instead of occupying a more central localization as seen in controls and were reminiscent of the staghornshaped nucleus of megakaryocytes observed in patients with essential thrombocythemia.17 The ultrastructure of GFI1B-deficient megakaryocytes was also poorly defined with an absence of distinct marginal and intermediate zones despite the presence of a few granules that tended to remain gathered (Figure 4A,B).

A culture in suspension of megakaryocytes from RosaCre-ERT Gfi1bEGFP/fl mice, which express GFP under the control of the Gfi1b promoter mainly in megakaryocytes and precursors, contained high numbers of smaller green cells that were not present in the control (Figure 4C). To investigate whether these smaller cells were megakaryocytes or megakaryocyte precursors, we stained mature megakaryocytes (lin–CD9highCD41highcKit+CD61high) from both Pf4-Cre and Rosa-Cre-ERT Gfi1bfl/fl mice with haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

B

A

C

D

E

F

G

H

I

Figure 2. The loss of GFI1B causes a severe thrombocytopenia. (A) FACS analysis of circulating platelets from Pf4-cre Gfi1bwt/fl control [Ctrl (PF4)] and Pf4-cre Gfi1bfl/fl knockout mice [KO (PF4)]. Platelets are first identified based on their FSC/SSC profile, then true platelets are identified as being CD41+ CD61+. Thiazole orange staining identifies reticulated platelets. (B) Quantitation of circulating platelets in the Pf4-cre Gfi1bfl/fl knockout mice (n = 9) compared to the controls (n = 13; Student t-test: P=0) and bleeding time [WT (n = 9) and KO (n = 5); Mann-Whitney U-test: P<0.0001]. All knockout animals reached the maximum bleeding time allowed without demonstrating signs of blood clotting.

(C) Platelet lifespan was assessed in vivo in Pf4-cre Gfi1bfl/fl (n = 5) and Pf4-cre Gfi1bwt/fl (n = 4) mice after biotin labeling of platelets through intravenous injection of SulfoNHS-biotin. Blood was analyzed as in (A) during 5 days after biotin labeling and the percentage of biotinylated platelets assessed by FACS. (D-F) Kinetics of platelet disappearance upon tamoxifen injection in Rosa-Cre-ERT Gfi1bfl/fl and Rosa-Cre-ERT Gfi1bwt/fl animals (n = 3 per time point). Animals received 100 mg/kg tamoxifen on day 0 and 50 mg/kg on day 1. Blood samples were analyzed on an Advia Hematology System and by flow cytometry (D). (E) The number of platelets given as mean count per liter ± SD. (F) The number of reticulated platelets given as mean count per liter ± SD. LT: a long-term analysis done 2 months after tamoxifen injection. Student t-test: (*) marks statistical significance (P<0.05). (G) Representative MGG-stained blood smears of Pf4-cre Gfi1b controls and knockouts [Ctrl (PF4) and KO (PF4)] and Rosa-Cre-ERT Gfi1b controls and knockouts [Ctrl (ROSA) and KO (ROSA)] showing in the knockouts a reduced number of platelets, the presence of large platelets with a lower granularity (LPT) and erythrocytes exhibiting Howell-Jolly bodies (HJ). (H) MGG-stained platelet cytospin showing the presence of very large vacuolar platelets specifically in the knockouts. (I) Comparison of the mean platelet volume measured on an Advia 120 cell analyzer (left) and the mean platelet forward scatter measured by FACS on CD41+CD61+ cells (FSC; right) in the Pf4-Cre Gfi1bfl/fl (PF4) and the Rosa-Cre Gfi1bfl/fl (ROSA) mice and their respective controls. the red lines represent the median ± interquartile rage. Kruskal-Wallis test: P<0.0001. * indicates statistical differences as determined using the post hoc Dunn multiple comparison test.

haematologica | 2017; 102(3)

487

H. Beauchemin et al.

Hoechst to assess their ploidy (Online Supplementary Figure S2A). In control mice, most of these cells were polyploid (>4N) with a 16N population being predominant (~60% of all megakaryocytes) followed by the 8N and 32N populations, each representing 10-20%, and a very low number of 64N cells (Figures 4D and Online Supplementary Figure S2A), which is in agreement with previous reports. 18,19 Megakaryocytes from Pf4-Cre Gfi1bfl/fl mice were also polyploid, but had a different distribution: the 32N subset was not affected by Gfi1b loss, the 16N megakaryocytes were notably decreased

A

C

E

and the 2N, 4N and 8N fractions were slightly increased compared to controls (Figure 5D and Online Supplementary Figure S2A). Gating these subsets into a rainbow plot based on their ploidy showed that the increase in smaller megakaryocytes in the Gfi1b knockout mice was primarily due to an increase in cells of the 2N and 4N subsets (Figure 4E). Conversely, the 16N and 32N fractions were decreased in megakaryocytes from Rosa-Cre-ERT Gfi1b fl/fl mice, while low-ploidy megakaryocytes represented the vast majority of the smaller cells seen in the Gfi1b knockout mice (Figure 4D,

B

D

F

Figure 3. Megakaryocytic dysplasia associated with the loss of GFI1B in the total bone marrow or specifically in megakaryocytes. (A) Representative FACS plot and quantitation of the bone marrow lin– cKit+CD41+CD61+ megakaryocyte expansion in Pf4-cre Gfi1bfl/fl animals (n = 11) compared to controls (n = 23). Values were normalized to matching controls and reported as mean ± SD. Student t-test with Welch correction: P<0.0001. (B) Collagen-based CFU-Mk assay of lineage-negative bone marrow cells: 5,000 to 10,000 lineage-depleted cells from Pf4-Cre Gfi1bfl/fl knockout (n = 3) or 15,000 lineage-depleted cells from Pf4-Cre Gfi1bwt/fl control mice (n = 3) were plated on collagen and allowed to grow for 7 days. Colonies were stained for acetylcholinestrase (AChE) activity and counted. Colonies staining AChE-positive were counted as CFU-Mk whereas AChE-negative colonies were counted as non-megakaryocyte CFU. In addition, the CFU-Mk were scored according to the number of cells they contained and sorted in three categories (3-20 cells; 21-50 cells; and >50 cells). Nb: number; Mk: megakaryocytes. (C-D) Representative FACS plot and quantitation of the bone marrow megakaryocyte expansion in Rosa-Cre-ERT Gfi1bfl/fl (n = 5) compared to controls (n = 4) 2 months after tamoxifen injection (C; Student t-test with Welch correction: P=0.0024) and kinetics of megakaryocyte expansion upon tamoxifen injection (D). In both experiments, values were normalized to sex- and age-matched controls and reported as mean ± SD. (E) Collagen-based CFU-Mk assay of lineage-negative bone marrow cells from Rosa-Cre-ERT Gfi1bfl/fl mice (n = 3) and control mice (n = 3) were plated on collagen and colonies were assessed as in (B). (F) Collagen-based CFU-Mk assay done on sorted lin-cKit+ (LK) GFP+ and LK GFP- cells from RosamT/mG Rosa-Cre-ERT Gfi1bwt/fl (Ctrl) and RosamT/mG Rosa-Cre-ERT Gfi1bfl/fl (KO) animals treated with tamoxifen 2 months prior to the experiment. In this setting, GFP+ cells are derived from progenitors/stem cells in which the Cre was activated through tamoxifen treatment, whereas the GFP- cells are cells deriving from progenitor/stem cells that escaped Cre activation upon tamoxifen treatment.

488

haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

A

D

C

B

E

F

Figure 4. Abnormal nuclear organization, cell size and ploidy of Gfi1b-null megakaryocytes. (A) Hematoxylin and eosin staining of bone section from Pf4-Cre Gfi1bwt/fl (Ctrl) and Pf4-Cre Gfi1bfl/fl (KO) mice shown at 40X and 100X showing abnormal staghorn-shaped polylobulated nuclei in the Gfi1b-null megakaryocytes compared to controls.

Arrows indicate megakaryocytes. (B) Transmission electron microscopy of a bone section from Pf4-Cre Gfi1bwt/fl (Ctrl; magnification: 1400x) and Pf4-Cre Gfi1bfl/fl (KO; 2900x) mice. In control megakaryocytes, the marginal zone (mz), the demarcation membrane (dm) and the intermediate zone (I) rich in granules are well defined, whereas in knockout megakaryocytes, these structures are poorly defined with the granules (g) often restricted to a single zone of the cytoplasm (in this case within the pocket formed by the nucleus (n). (C) Superimposition of a dark field and fluorescence microscopy of a culture in suspension of primary megakaryocytes derived from Rosa-Cre-ERT Gfi1bwt/EGFP (Ctrl) and Rosa-Cre-ERT Gfi1bfl/EGFP (KO) mice showing the presence of a large number of smaller green cells in the knockout that are completely absent in the control culture which shows only larger green cells. (D) Quantification of the ploidy in the most mature megakaryocyte population (defined as being Lin- cKit+ CD9hi CD41hi CD61hi) in both Pf4-Cre and Rosa-Cre-ERT-driven Gfi1b knockouts 2 months after Cre induction (n = 4 for each group) and given as the mean percentage of the total megakaryocytes ± SD. (E) Megakaryocyte rainbow plots produced by back-gating the cells based on their ploidy on the forward scatter (FSC)/side scatter (SSC) plot. (F) Comparison of the FSC (cell size) and SSC (cell granularity) of the polyploid cell populations (8N, 16N and 32N) of the cells analyzed in (D). The bar graphs present the mean ± SD. The means were compared using a Student t-test corrected for multiple comparisons using the Sidak-Bonferroni method. (*) P≤0.01.

haematologica | 2017; 102(3)

489

H. Beauchemin et al.

B

A

D

C

E

F

G

Figure 5. Impaired response to integrin receptors’ ligands in GFI1B-deficient megakaryocytes. (A) Primary megakaryocyte cultures derived from Pf4-Cre Gfi1bfl/fl mice and control mice were plated on slides coated with collagen, fibronectin or fibrinogen, allowed to spread for 5 h, then fixed and stained with DAPI and FITC-CD41 antibody. (B) To assess the responsiveness of megakaryocytes that spread on coated slides more quantitatively, we first measured their individual periphery (P) and area (A), then divided these peripheries by the circumference of perfect circles of the same area. This ratio, or Roundness Index (RI), gives a quantitative value that is directly proportional to the structural complexity of the cells regardless of their cytoplasmic mass. As visual comparative examples, the shapes of cells with defined RI are presented next to the scale for clarity. (C) RI of Pf4-Cre Gfi1bfl/fl knockout- or Pf4-Cre Gfi1bwt/fl control-derived megakaryocytes spread on fibronectin and plotted as bean plots as well as against their cellular area showing that the RI of Gfi1b-null megakaryocytes remains low even as cell size increases. Mann-Whitney U-test: P<0.0001. (D) Assessment of the capacity of Pf4Cre Gfi1bfl/fl knockout- (KO) or Pf4-Cre Gfi1bwt/fl control-derived (WT) megakaryocytes to produce proplatelet formation and platelet release in vitro. Proplatelet-forming megakaryocytes were identified by light microscopy and counted along non-proplatelet megakaryocytes (rounded shape) and plotted as a percentage of total megakaryocytes. The presence of platelet-like structure (white arrows) was also assessed but not quantified. While proplatelets and platelets were present in all wild-type cultures, they were never observed in GFI1B-deficient megakaryocytes (KO). The bar graph presents the mean ± SD of four independent experiments (n = 4 for each condition). (E) Still images extracted from a time-lapse microscopy in phase contrast recorded over several hours (see Online Supplementary Videos S1 and S2) of Pf4-Cre Gfi1bfl/fl knockoutor Pf4-Cre Gfi1bwt/fl control-derived megakaryocytes plated on fibronectin. (F-G) The motility of Pf4-Cre Gfi1bfl/fl knockout- or Pf4-Cre Gfi1bwt/fl control-derived megakaryocytes was assessed by time-lapse microscopy over several hours upon plating on fibrinogen (Ctrl: n = 11; KO: n = 16), fibronectin (Ctrl: n = 25; KO: n = 9) and collagen (Ctrl: n = 6; KO: n = 5). Single megakaryocytes were manually tracked using the cell centroid position and their path recorded (F), allowing calculation of their average velocity during the course of the experiment. The results are presented as mean velocity ± SD in μm per minute (G). Student t-test: (*) marks statistical significance (P<0.0001). Videos from (F) can be seen in the Online Supplementary Material (Videos S3-S8).

490

haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

E and Online Supplementary Figure S2). Megakaryocytes with the same ploidy from Gfi1b-null mice had the same size (forward scatter) as controls and a tendency to a have lower mean side scatter (Figures 4F and Online Supplementary Figure S2B).

Defective integrin signaling and cellular spreading of Gfi1b-null megakaryocytes When grown on integrin-specific ligands, megakaryocytes start to spread by forming large lamellipodia that depend on the interaction between extracellular matrix components and integrins.20 To assess how Gfi1b affects this process, we cultured megakaryocytes from either Pf4Cre Gfi1bfl/fl or control mice on fibronectin, fibrinogen or collagen. Control megakaryocytes were able to form lamellipodia and filopodia-like protrusions on collagen (a ligand for the α2β1 integrin), fibronectin (a ligand for αIIbβ3 and αVβ1) and fibrinogen (a ligand for αIIbβ3 and αVβ3),21 whereas GFI1B-deficient megakaryocytes were unable to do so and maintained a spherical shape (Figure 5A).

To exclude the possibility that this low response of GFI1B-deficient cells to integrin ligands was simply due to their smaller size, we established a method to measure response to integrin ligands independently of cell size. By comparing the periphery of each cell that attached to a ligand in culture to the circumference of a perfect circle (2πr) with the same area, we calculated a "roundness index" (RI) resulting in a numerical value that increases as the complexity of a spreading cell increases, starting at RI=1 for a perfectly round cell (Figure 5B). A comparison of the RI of cultured cells clearly showed that GFI1B-deficient megakaryocytes maintained significantly low RI and were unable to spread, suggesting a poor response to integrin ligands. This defect was independent of their size (Figure 5C) or their state of maturity (Online Supplementary Figure S3A,B), ruling out that the poor responsiveness could be due to a smaller cytoplasm or to an immature state. Timelapse video microscopy on live cells confirmed a severe impairment of motility of GFI1B-deficient megakaryocytes in response to fibronectin and fibrinogen and to a lesser extent to collagen (Figure 5E-G; Online Supplementary Videos S1-S8) also independently of the state of maturity (Online Supplementary Figure S3C).

The lower ploidy of most Rosa-Cre-ERT Gfi1bfl/fl megakaryocytes suggests an immature state or, at the very least, defective endomitosis and could provide an explanation for the absence of platelet production. However, the observation that, in Pf4-Cre Gfi1bfl/fl, megakaryocytes can fully mature but still cannot produce platelets (Figure 5D) indicates that loss of GFI1B inhibits proplatelet formation in mature megakaryocytes.

GFI1B controls actin polymerization and microtubule organization We next examined how GFI1B deficiency affected events downstream of integrin receptors, such as actin polymerization in megakaryocytes that were allowed to spread on fibronectin. In contrast to control megakaryocytes, which became highly positive for F-actin, Gfi1bdeficient megakaryocytes displayed lower levels of actin filaments with a 3-fold reduction of fluorescence intensity despite the fact that the expression of un-polymerized actin was not affected in the knockout (Figure 6A,C,G).

Assessment by immunofluorescence of β-tubulin in megakaryocytes spread on fibronectin revealed that Gfi1bhaematologica | 2017; 102(3)

null megakaryocytes seemed unable to produce fibers of microtubules (Figure 6B), although quantitation of fluorescence intensity showed only a mild, albeit significant, reduction of β-tubulin in GFI1B-deficient megakaryocytes compared to controls (Figure 6D). Measurements of fluorescence intensity across a horizontal section passing through the center of each cell showed that the intracellular localization of β-tubulin was altered in Gfi1b-null megakaryocytes (Figure 6E). In controls, the tubulin network was found to be positioned both at the center of the cell and at the cellular cortex as a ring of clearly defined microtubule fibers characteristic of later stages of maturation.22 In contrast, in GFI1B-deficient megakaryocytes, βtubulin was retained in the center of the cells, within a pocket formed by the ring-shaped nucleus, seemingly failing to polymerize into microtubules, and this, independently of the maturation state (Figure 6B,E and Online Supplementary Figure S3D and Online Supplementary Video S9), indicating either a profound defect in microtubule polymerization or a failure of GFI1B-deficient megakaryocytes to get activated by fibronectin. However, analysis of both α- and β-tubulin by western blot revealed that while the β-tubulin content was not altered significantly in the knockout, α-tubulin, on the other hand, was dramatically reduced to almost undetectable levels, including its acetylated form and even upon phorbol myristate acetate activation (Figure 6G). Importantly, the lack of responsiveness to integrin-specific ligands observed in Gfi1b-null megakaryocytes was not due to the absence of the proper integrin receptors, which were either unaffected or increased in the knockout when measured by FACS (Figure 6D and Online Supplementary Figure S4) or by immunoblotting (Figure 6G).

Interestingly, subpopulations of megakaryocytes defined by their integrin signature were only marginally affected in the Gfi1b knockout mice compared to the wild-type ones (Online Supplementary Figure S5).

Expression profiling of megakaryocytes from both Pf4Cre- and Rosa-Cre-ERT-driven Gfi1b-knockout and control mice by RNA-Seq showed a strong correlation between the two models and with a previously published data set obtained from a different Gfi1b knockout model12 (Figure 7A and Online Supplementary Figure S6). Genes related to the actin cytoskeleton were found to be deregulated in cells from Pf4-Cre Gfi1bfl/fl mice but not in megakaryocytes from Rosa-Cre Gfi1bfl/fl mice (Figure 7B and Online Supplementary Figure S6C). However, in both models, genes with deregulated expression in GFI1B-deficient megakaryocytes were enriched in categories defined by signaling through Rho GTPases and regulation of microtubules (Figure 8B,C and Online Supplementary Figure S7A,B). Notable exceptions are the α-tubulin genes, which were downregulated in both knockouts (Figure 7C). Also, many progenitor-specific and megakaryocytespecific genes, such as Kitlg,23 Il6,24 Itgb3,20 CD9,25 and Mpl,26 were up-regulated in Gfi1b-null megakaryocytes (Online Supplementary Figure S7C). However, genes from the “hematopoietic lineage” gene set were not enriched in cells from Pf4-Cre Gfi1bfl/fl mice, but showed a strong negative regulation in Rosa-Cre-ERT Gfi1bfl/fl animals (Online Supplementary Figure S7C), suggesting that megakaryocytes from Rosa-Cre-ERT Gfi1bfl/fl mice have a more immature/progenitor-like phenotype. Conversely, a “megakaryocyte differentiation” gene set based on the paper by Lim et al.27 showed a strong enrichment in cells from the Pf4-Cre-driven Gfi1b knockout mice but not in 491

H. Beauchemin et al.

B

A

C

F

D

E

G

Figure 6. Defect in cytoskeleton organization in Gfi1b-null megakaryocytes. (A) Pf4-Cre Gfi1bfl/fl knockout megakaryocytes (KO; n = 188) or control megakaryocytes (Ctrl; n = 30) were plated on fibronectin in a single 12-m-Chambers slide to minimize variation between samples and, after fixing, labeled with a FITC-conjugated anti-CD41 antibody and AF555-conjugated phalloidin to measure actin fibers. Pictures were taken by fluorescent microscopy and the mean fluorescence intensity was measured and plotted against the cell size defined as the area covered on the coated slide. Megakaryocyte enrichment on a bovine serum albumin gradient also allows for some non-megakaryocyte CD41– cells to be present. Therefore, to avoid confusion because of the relative small cell size, CD41+ megakaryocytes are identified with arrows in the knockout. (B) Cells from Pf4-Cre Gfi1bfl/fl knockout (KO; n = 70) or control (Ctrl; n = 41) mice were plated on fibronectin as in (A) and were stained with an AF555-conjugated anti-β-tubulin antibody to show general intracellular distribution of β-tubulin, which showed poor polymerization in the knockout compared to controls. (C) The mean fluorescence intensity (MFI) of F-actin from (A) was measured and the ratio (MFI/cell area) was calculated and presented as mean ± SD. Student t-test with Welch correction: P<0.0001. (D) The mean fluorescence intensity of β-tubulin from (B) was measured and the ratio (MFI/cell area) was calculated and presented as mean ± SD. Student t-test with Welch correction: P<0.0001. (E) Analysis of a cross-section passing through the center of 30 megakaryocytes (n = 30) per genotype [Pf4-Cre Gfi1bflox/flox (KO) and Pf4-Cre Gfi1bflox/wt (Ctrl) mice coming from (B and D)]. The diameter was subdivided into 100 sections, each representing 1% of the cell width and the AF555 intensity was measured at each of these sections with section 0 starting at one side and section 100 ending at the other side. This intensity was then plotted as mean ± CI along the whole cell section. (F) Integrin receptors present at the surface of lin-cKit+Cd41+CD61+ megakaryocytes from Pf4-Cre Gfi1bfl/fl knockout (KO) mice were assessed by FACS and their MFI was compared to that of Pf4-Cre Gfi1bwt/fl control (Ctrl) animals and presented as mean ± SD (n = 2 for both genotypes). (G) Immunoblot detection of megakaryocyte cytoskeleton proteins (actin, β-tubulin, α-tubulin and the acetylated form of α -tubulin) and the two subunits of the megakaryocyte-specific fibrinogen/fibronectin integrin receptor αIIbβ3 in megakaryocytes derived from the bone marrow from Pf4-Cre Gfi1bflox/flox (KO) and Pf4-Cre Gfi1bwt/wt (WT) mice with or without phorbol myristate acetate (PMA) activation. HPRT is used as a loading control. This experiment was repeated at least three times with consistent results.

492

haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

cells from the Rosa-Cre-ERT deleted animals (Online Supplementary Figure S7D), suggesting that megakaryocytes were more differentiated in the Pf4-Cre model than in the Rosa-Cre model. Interestingly, many genes from the microtubule pathways that were up-regulated in

A

D

B

Gfi1b-null cells belonged to the Rho guanine nucleotide exchange factors (ArhGEF) gene family (Figure 7D and Online Supplementary Figure S7B). Western analysis revealed a dramatic increase of both the total and the phosphorylated form of PAK4, a key player in the regula-

C

E

Figure 7. Gene expression profiling in Gfi1b-null megakaryocytes reveals defects in cytoskeleton regulation pathways. (A) Correlation between the gene expression profiles from the RNA-Seq on Pf4-Cre- and Rosa-Cre-ERT-driven Gfi1b knockout. The Pearson coefficient is shown. The P value was determined using a permutation test with 1x106 re-samplings. (B-D) Gene set enrichment analysis (GSEA) of cytoskeleton-related functions on RNA-Seq expression profiles from both Pf4-Cre Gfi1bfl/fl and Rosa-Cre-ERT Gfi1bfl/fl (+ tamoxifen) megakaryocytes. Data from knockout cells were compared with their respective Pf4-Cre Gfi1bwt/fl and Rosa-Cre Gfi1bwt/fl (+ tamoxifen) controls. Above, genes are plotted based on their expression fold change (log2 scale) and for each of these gene sets, the top over-expressed and under-expressed genes are identified on the dot plot. (B) Actin cytoskeleton regulation pathway with integrin genes highlighted with a yellow color . (C) Microtubule growth pathway in which the tubulin genes are highlighted in blue (β-tubulin family) and in green (α-tubulin family). (D) Signaling by RHO GTPase gene set. Normalized enrichment scores (NES) and nominal P-values (P) are given for each GSEA plot. (E) Immunoblot detection of the phosphorylated form of PAK4 (also recognizes the other type II PAKs, PAK5 and PAK6) in megakaryocytes derived from the bone marrow of Pf4-Cre Gfi1bflox/flox (KO) and Pf4-Cre Gfi1bwt/wt (WT) mice (top) and Rosa-Cre-ERT Gfi1bflox/flox (KO) and Rosa-Cre-ERT Gfi1bwt/wt (WT) mice (bottom) with or without phorbol myristate acetate (PMA) activation and using β-tubulin and/or laminin as loading controls. This experiment was repeated at least three times with consistent results.

haematologica | 2017; 102(3)

493

H. Beauchemin et al.

tion of both actin and tubulin cytoskeletons, suggesting that hyperactivation of the integrin signaling through PAK might be responsible for the phenotype.

The spreading defects of Gfi1b-null megakaryocytes can be partially rescued by inhibitors of integrin signaling and fully rescued by a pan-p21 activated kinase inhibitor in vitro To functionally verify the results obtained from the RNA-Seq analysis, which suggested that integrin signaling might be hyper-activated in the knockout, we used small molecule inhibitors against FAK (PF573228), PAK (PF3758309 and FRAX486), ROCK (Y27632), CDC42 (ML141), RAC1-GEF interaction (NSC 23766 and W56) and myosin II (Blebbistatin). Megakaryocytes were allowed to spread for 5 h on fibronectin in the presence or absence of the inhibitors and the RI was measured on CD41+ cells (Figure 8A and Online Supplementary Figure S8A,B). A partial rescue of the Gfi1b null phenotype was observed in the presence of either PAK (FRAX486), FAK, ROCK and RAC1 (W56) inhibitors. The strongest effect and a full rescue was obtained with the pan-PAK inhibitor (PF3758309). The RI for the PF3758309-treated Gfi1b knockout megakaryocytes were undistinguishable from those of PF3758309-treated wild-type megakaryocytes, which themselves spread better than the untreated wildtype control megakaryocytes (Figure 8A,B). These cells also exhibited an improved β-tubulin distribution at the apex of the cells (Figure 8C) and their motility improved to a level undistinguishable from that of the controls (Figure 8D).

However, none of the inhibitors used was able to rescue the in vitro proplatelet production defect of Gfi1b-null megakaryocytes; rather, inhibition of PAK with either PF3758309 or FRAX486 led to the death of megakaryocytes (both wild-type and knockout) over the time-course necessary for cells to produce proplatelets (Figure 8D).

Discussion Deleting Gfi1b by Pf4-Cre at later stages of megakaryocyte differentiation or by Rosa-Cre-ERT in the entire hematopoietic lineage and thus also before megakaryocyte commitment produced a phenotype that resembles the one seen in patients with GFI1B-related thrombocytopenia with rare, larger platelets.4,5 Upon deletion of Gfi1b

by Cre induction in adult bone marrow cells via tamoxifen, we observed defects in the earliest stages of megakaryocyte maturation, with a majority of cells that failed to undergo proper endomitosis or to form large CFU-Mk colonies, suggesting a block in megakaryocyte maturation/differentiation. In contrast, inactivation of Gfi1b at later stages of megakaryocyte differentiation via a Pf4-Cre transgene resulted in only mild defects in polyploidization or CFU-Mk colony formation, but still abrogated platelet production completely, suggesting that in this case, the defect lies in the terminal process of proplatelet formation and platelet release, demonstrating that Gfi1b plays a role in both megakaryocyte differentiation and platelet production. The role of Gfi1b in this terminal process is also supported by the observation that platelet release is arrested very rapidly after tamoxifen injection in Rosa-Cre-ERT Gfi1bfl/fl, before the block in megakaryocyte differentiation could lead to exhaustion of mature megakaryocytes. It has been shown that large platelets 494

and circulating megakaryocyte fragments can be consequences of insufficient membrane- or aberrant organelle structure of megakaryocytes,16 which could provide an explanation for the presence of these features in Gfi1bknockout mice. The presence of rare, large platelets despite the incapacity of megakaryocytes to form proplatelets suggests the existence of an alternative albeit less efficient mechanism such as fragmentation of megakaryocyte cytoplasm in pulmonary capillaries.28 Previous studies have shown that several integrins are deregulated in Gfi1b-null cells5,12,13 and our RNA-Seq data confirm these observations. Expression of the megakaryocyte-specific integrin β3 (Itgb3) is up-regulated in our models and this higher expression leads to the expected increase of the megakaryocyte-specific integrin αIIb/β3 complex at the cell surface.29 Because the αIIb/β3 complex is the main receptor for fibrinogen30 and fibronectin,31 it was surprising that the Gfi1b-null megakaryocytes were not responsive to these two ligands. Conversely, the main collagen receptor (α2β1 integrin)32 is present on Gfi1b-null megakaryocytes at lower levels than controls, but the cells still had a defective response. Hence, the deregulation of integrin receptors alone cannot explain the poor responsiveness of Gfi1b-null megakaryocytes, although it has been reported that abnormal activation of the αIIb/β3 complex can lead to defective proplatelet formation.33 It is rather likely that elements of the intracellular integrin signaling pathways responsible for cell spreading and motility and for megakaryocyte activation are disrupted in the absence of GFI1B. The poor response to integrin activation in Gfi1b-null megakaryocytes supports this notion.

Remodeling of cytoskeleton components is involved in the two cellular processes that are most affected in Gfi1b knockout megakaryocytes: endomitosis and platelet release. Polyploidization that occurs through endomitosis is dependent on the regulation of an atypical multipolar mitotic spindle that, in addition to a failed cleavage furrow formation in late anaphase, leads to the abortion of both cytokinesis and karyokinesis.34,35 Similarly, regulation of microtubules is essential for proplatelet formation22,36,37 and given that megakaryocytes from Pf4-Cre Gfi1bfl/fl mice, which exhibit an almost normal polyploidization, are unable to produce platelets, a failure to assemble, organize or stabilize microtubules, as was observed with megakaryocytes with disorganized β-tubulin that were unable to form visible microtubule fibers, is likely one of the main defects in Gfi1b-null megakaryocytes. We have shown that this profound defect can be explained by the incapacity of β-tubulin to polymerize into microtubules due to the lack of α-tubulin.

On the other hand, cellular motility and spreading, which are also profoundly defective in Gfi1b-null megakaryocytes, rely essentially on the actin cytoskeleton dynamics through integrin signaling. Interestingly, the fact that these defects are seen regardless of the level of maturation in the Pf4-driven Gfi1b knockout argues against the possibility that these defect could be solely due to a block of maturation.

The expression of genes coding for several components involved in the organization of both actin and tubulin cytoskeleton, for instance several small GTPases of the Rho family as well as Rho guanine nucleotide exchange factors (ArhGEF), was deregulated in Gfi1b-null megakaryocytes. Our experiments with small molecule inhibitors have identified integrin signaling pathway haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

A

B

D

E

C

Figure 8. Cytoskeleton rescue by a PAK inhibitor. (A) Megakaryocytes from Pf4-Cre Gfi1bwt/fl (Ctrl) and Pf4-Cre Gfi1bfl/fl (KO) mice were allowed to spread on fibronectin for 3 h in the absence (UT; KO: n = 430; Ctrl: n = 361) or presence of different inhibitory small molecules: PF3758309 (PF309; PAK4 inhibitor; KO: n = 184; Ctrl: n = 147); FRAX486 (FRAX; PAK1-3 inhibitor; KO: n = 220; Ctrl: n = 93); PF573288 (PF228; FAK inhibitor; KO: n = 405; Ctrl: n = 353); Y27632 (Y632; ROCK inhibitor; KO: n = 396; Ctrl: n = 223); ML141 (CDC42 inhibitor; KO: n = 169; Ctrl: n = 123); NSC23766 (NSC; RAC1 inhibitor; KO: n = 153; Ctrl: n = 136); F56 (control peptide; KO: n = 142; Ctrl: n = 107); W56 (RAC1 inhibitor; KO: n = 146; Ctrl: n = 152); and blebbistatin (Blebb; myosin II inhibitor; KO: n = 424; Ctrl: n = 395). The roundness index (RI) of these cells was scored as in Figure 6B and presented as beanplots comparing the median for all conditions. A Kruskal-Wallis one-way analysis of variance done on these data and followed by a Dunn post-hoc multiple comparison test revealed a highly significant difference (P<0.0001) between the samples' median; the Dunn test identifies significant differences between the untreated KO megakaryocytes and the KO megakaryocytes treated with the PAK inhibitors PF3758309 and FRAX486, as well as with the RAC1 inhibitor peptide W56. KO megakaryocytes treated with PF3758309 show no statistical difference with untreated or PF3758309-treated control megakaryocytes. (B) Representative microscopy of PF3758309-treated (PF309) control (Ctrl) and knockout (KO) megakaryocytes from (A) stained with DAPI (blue), CD41 (green) and β-tubulin (red) and (top) overlay of all cells present in the field to highlight their spreading (middle). The RI were plotted against their corresponding cell size (bottom). The blue and red dashed lines show the median RI (horizontal) and cell size (vertical) for the control and knockout megakaryocytes, respectively. (C) Cellular distribution of microtubule in knockout megakaryocytes treated with the PAK4 inhibitor PF3758309 (PF309) was analyzed as in Figure 6E and plotted (red line) against the control cells presented in Figure 6E (shaded gray). (D) The motility of Pf4-Cre Gfi1bfl/fl knockout- or Pf4-Cre Gfi1bwt/fl control-derived megakaryocytes was assessed by time-lapse microscopy as in Figure 5 in the presence (Ctrl: n = 29; KO: n = 34) or absence (Untreated; Ctrl: n = 31; KO: n = 39) of the PAK4 inhibitor PF3758309 (PF309). The results are presented as mean velocity ± SD in mm per minute. Student t-test: (*) marks statistical significance (P<0.0001). (E) Megakaryocytes were allowed to form proplatelet-like structures in vitro and assessed by light microscopy (as in Figure 5D) in the absence (DMSO) or presence of inhibitors as in (A). Conditions in which cells did not survive over the period of 12 h upon treatment with the inhibitors are labeled as "dead" in the graph. Bar graphs present the mean ± SD of three separate experiments (n = 3).

haematologica | 2017; 102(3)

495

H. Beauchemin et al.

components as well as PAK, which is targeted by RAC, RHO and CDC42, as two of the critical components affected by GFI1B-deficiency in megakaryocytes.

Inhibiting FAK or RHO kinase (ROCK), both important for actin cytoskeleton and integrin signaling,38,39 led to moderate but significant improvements in the response of Gfi1b-null megakaryocytes to integrin ligands and in their capacity to spread on fibronectin. This suggests that aberrant integrin signaling can partially explain the defects of GFI1B-deficient megakaryocytes, an idea supported by a recent study showing a role of GFI1B in the regulation of the integrin-bound proteins Talin1 and Kindlin3.40 However, the inhibition of PAK, which regulates both actin cytoskeleton and microtubules through stathmin and GEF-H1,41,42 by the pan-PAK inhibitor PF-3758309 had the strongest effect on GFI1B-deficient megakaryocytes, completely rescuing their responsiveness to ligands, as shown by their capacity to spread efficiently and their increased motility upon inhibition of PAK. Importantly, even though this PAK inhibitor also had a stimulatory effect on normal megakaryocytes, which spread more efficiently on fibronectin than did the untreated control cells, the responsiveness of the Gfi1b-null megakaryocytes was just as good. Because PF-3758309 preferentially inhibits type II PAKs (e.g. PAK4) but lacks specificity,43 we also used the type I PAK inhibitor FRAX486, which targets PAK1-3,43 which only mildly rescued the spreading defect. This suggests that GFI1B-deficiency mainly affects signaling through type II PAKs, which is supported by the observation that both the total and the phosphorylated forms of the PAK4 protein are constitutively increased in Gfi1b knockout cells. Hence, dysregulation of cytoskeleton

References 1. Machlus KR, Thon JN, Italiano JE Jr. Interpreting the developmental dance of the megakaryocyte: a review of the cellular and molecular processes mediating platelet formation. Br J Haematol. 2014;165(2):227-236.

2. Nurden AT, Freson K, Seligsohn U. Inherited platelet disorders. Haemophilia. 2012;18 (Suppl 4):154-160.

3. Freson K, Wijgaerts A, van Geet C. Update on the causes of platelet disorders and functional consequences. Int J Lab Hematol.

2014;36(3):313-325.

4. Stevenson WS, Morel-Kopp MC, Chen Q, et al. GFI1B mutation causes a bleeding disorder with abnormal platelet function. J Thromb Haemost. 2013;11(11):2039-2047.

5. Monteferrario D, Bolar NA, Marneth AE, et al. A dominant-negative GFI1B mutation in the gray platelet syndrome. N Engl J Med. 2014;370(3):245-253.

6. Kitamura K, Okuno Y, Yoshida K, et al.

Functional characterization of a novel GFI1B mutation causing congenital macrothrombocytopenia. J Thromb Haemost. 2016;14 (7):1462-1469.

7. Noris P, Biino G, Pecci A, et al. Platelet diameters in inherited thrombocytopenias: analysis of 376 patients with all known disorders.

Blood. 2014;124(6):e4-e10.

496

dynamics could be the main reason for the absence of the cellular response to several integrin ligands in Gfi1b-null megakaryocytes. This is consistent with the up-regulation of ArhGEF in Gfi1b-null cells seen in our RNA-Seq experiments, since ArhGEF activate PAK. Because Pak4 expression is strongly increased in Gfi1b knockout cells and since the inhibition of type II PAK restored cellular spreading in the absence of GFI1B, it is conceivable that GFI1B normally inhibits Pak. This is further supported by another study that showed that PAK2-deficient megakaryocytes are characterized by an increased polyploidy and an alteration of microtubule organization.44 Studies with other transcription factors that are important for proper megakaryopoiesis, such as GATA-1/FOG1,45,46 NF-E2 p45,47 RUNX148 and FLI1,49 have highlighted their multifarious effects that could not all be explained by the deregulation of single pathways. Similarly, we showed that GFI1B is essential for proplatelet formation, but none of the inhibitors used rescued this cell-autonomous process in Gfi1b-null megakaryocytes. Confirming previous findings,50 we also showed that proplatelet formation occurs in the absence of integrin-specific ligands. In this case, the loss of α-tubulin, whose expression might be controlled at least indirectly by GFI1B, and independently of PAK provides an elegant explanation for this phenotype.

Acknowledgments We thank Mathieu Lapointe and Damien Grapton for their technical assistance. This work was supported by grants from the Canadian Institutes of Health Research (MOP – 111247) and the Canadian Hemophilia Society. TM holds the Canada Research Chair (Tier 1) in Hematopoiesis and Immune Cell Differentiation.

8. Saleque S, Cameron S, Orkin SH. The zincfinger proto-oncogene Gfi-1b is essential for development of the erythroid and megakaryocytic lineages. Genes Dev. 2002; 16(3):301-306.

9. Vassen L, Okayama T, Moroy T. Gfi1b:green fluorescent protein knock-in mice reveal a dynamic expression pattern of Gfi1b during hematopoiesis that is largely complementary to Gfi1. Blood. 2007;109(6): 2356-2364.

10. Garçon L, Lacout C, Svinartchouk F, et al.

Gfi-1B plays a critical role in terminal differentiation of normal and transformed erythroid progenitor cells. Blood. 2005;105(4): 1448-1455.

11. Vassen L, Beauchemin H, Lemsaddek W, Krongold J, Trudel M, Moroy T. Growth factor independence 1b (Gfi1b) is important for the maturation of erythroid cells and the regulation of embryonic globin expression.

PloS One. 2014;9(5):e96636.

12. Foudi A, Kramer DJ, Qin J, et al. Distinct, strict requirements for Gfi-1b in adult bone marrow red cell and platelet generation. J Exp Med. 2014;211(5):909-927.

13. Khandanpour C, Sharif-Askari E, Vassen L, et al. Evidence that growth factor independence 1b regulates dormancy and peripheral blood mobilization of hematopoietic stem cells. Blood. 2010;116(24):5149-5161.

14. Tiedt R, Schomber T, Hao-Shen H, Skoda

15.

16.

17.

18.

19.

20.

RC. Pf4-Cre transgenic mice allow the generation of lineage-restricted gene knockouts for studying megakaryocyte and platelet function in vivo. Blood. 2007;109(4):1503-1506.

Schulze H. Culture of murine megakaryocytes and platelets from fetal liver and bone marrow. Methods Mol Biol. 2012;788: 193203.

Poujol C, Ware J, Nieswandt B, Nurden AT, Nurden P. Absence of GPIbα is responsible for aberrant membrane development during megakaryocyte maturation: ultrastructural study using a transgenic model. Exp Hematol. 2002;30(4):352-360.

Wilkins BS, Erber WN, Bareford D, et al.

Bone marrow pathology in essential thrombocythemia: interobserver reliability and utility for identifying disease subtypes.

Blood. 2008;111(1):60-70.

Ebbe S, Boudreaux MK. Relationship of megakaryocyte ploidy with platelet number and size in cats, dogs, rabbits and mice.

Comp Hematol Int. 1998;8(1):21-25.

Meinders M, Kulu DI, van de Werken HJ, et al. Sp1/Sp3 transcription factors regulate hallmarks of megakaryocyte maturation and platelet formation and function. Blood.

2015;125(12):1957-1967.

Larson MK, Watson SP. Regulation of proplatelet formation and platelet release by integrin αIIbβ3. Blood. 2006;108(5):15091514.

haematologica | 2017; 102(3)

Gfi1b controls megakaryocyte functions

21. Humphries JD, Byron A, Humphries MJ. Integrin ligands at a glance. J Cell Sci. 2006;119(Pt 19):3901-3903.

22. Italiano JE, Jr., Lecine P, Shivdasani RA, Hartwig JH. Blood platelets are assembled principally at the ends of proplatelet processes produced by differentiated megakaryocytes. J Cell Biol. 1999;147(6): 1299-1312.

23. Avraham H, Vannier E, Cowley S, et al.

Effects of the stem cell factor, c-kit ligand, on human megakaryocytic cells. Blood.

1992;79(2):365-371.

24. Ishibashi T, Kimura H, Uchida T, Kariyone S, Friese P, Burstein SA. Human interleukin 6 is a direct promoter of maturation of megakaryocytes in vitro. Proc Natl Acad Sci USA. 1989;86(15):5953-5957.

25. Clay D, Rubinstein E, Mishal Z, et al. CD9 and megakaryocyte differentiation. Blood.

2001;97(7):1982-1989.

26. Debili N, Wendling F, Cosman D, et al. The Mpl receptor is expressed in the megakaryocytic lineage from late progenitors to platelets. Blood. 1995;85(2):391-401.

27. Lim CK, Hwang WY, Aw SE, Sun L. Study of gene expression profile during cord bloodassociated megakaryopoiesis. Eur J Haematol. 2008;81(3):196-208.

28. Zucker-Franklin D, Philipp CS. Platelet production in the pulmonary capillary bed: new ultrastructural evidence for an old concept.

Am J Pathol. 2000;157(1):69-74.

29. Duperray A, Troesch A, Berthier R, et al.

Biosynthesis and assembly of platelet GPIIbIIIa in human megakaryocytes: evidence that assembly between pro-GPIIb and GPIIIa is a prerequisite for expression of the complex on the cell surface. Blood.

1989;74(5):1603-1611.

30. Niewiarowski S, Kornecki E, Budzynski AZ, Morinelli TA, Tuszynski GP. Fibrinogen interaction with platelet receptors. Ann N Y Acad Sci. 1983;408:536-555.

31. Huynh KC, Stoldt VR, Scharf RE. Contribution of distinct platelet integrins to binding, unfolding, and assembly of

haematologica | 2017; 102(3)

32.

33.

34.

35.

36.

37.

38.

39.

40.

fibronectin. Biol Chem. 2013;394(11):14851493.

Jokinen J, Dadu E, Nykvist P, et al. Integrinmediated cell adhesion to type I collagen fibrils. J Biol Chem. 2004;279(30):31956-31963.

Bury L, Malara A, Gresele P, Balduini A. Outside-in signalling generated by a constitutively activated integrin ιIIbβ3 impairs proplatelet formation in human megakaryocytes. PloS one. 2012;7(4):e34449.

Baatout S, Chatelain B, Staquet P, Symann M, Chatelain C. Augmentation of the number of nucleolar organizer regions in human megakaryocyte cell lines after induction of polyploidization by a microtubule inhibitor.

Eur J Clin Invest. 1998;28(2):138-144.

Geddis AE, Fox NE, Tkachenko E, Kaushansky K. Endomitotic megakaryocytes that form a bipolar spindle exhibit cleavage furrow ingression followed by furrow regression. Cell Cycle. 2007;6(4):455-460.

Thon JN, Montalvo A, Patel-Hett S, et al.

Cytoskeletal mechanics of proplatelet maturation and platelet release. J Cell Biol.

2010;191(4):861-874.

Kunishima S, Nishimura S, Suzuki H, Imaizumi M, Saito H. TUBB1 mutation disrupting microtubule assembly impairs proplatelet formation and results in congenital macrothrombocytopenia. Eur J Haematol.

2014;92(4):276-282.

Chang Y, Aurade F, Larbret F, et al.

Proplatelet formation is regulated by the Rho/ROCK pathway. Blood. 2007;109(10): 4229-4236.

Hitchcock IS, Fox NE, Prevost N, Sear K, Shattil SJ, Kaushansky K. Roles of focal adhesion kinase (FAK) in megakaryopoiesis and platelet function: studies using a megakaryocyte lineage specific FAK knockout. Blood. 2008;111(2):596-604.

Singh D, Upadhyay G, Sengupta A, et al.

Cooperative stimulation of megakaryocytic differentiation by Gfi1b gene targets Kindlin3 and Talin1. PloS One. 2016;11 (10):e0164506.

41. Wittmann T, Bokoch GM, Waterman-Storer CM. Regulation of microtubule destabilizing activity of Op18/stathmin downstream of Rac1. J Biol Chem. 2004;279(7):6196-6203.

42. Meiri D, Marshall CB, Mokady D, et al.

Mechanistic insight into GPCR-mediated activation of the microtubule-associated RhoA exchange factor GEF-H1. Nat Commun. 2014;5:4857.

43. Rudolph J, Crawford JJ, Hoeflich KP, Wang W. Inhibitors of p21-activated kinases (PAKs). J Med Chem. 2015;58(1):111-129.

44. Kosoff RE, Aslan JE, Kostyak JC, et al. Pak2 restrains endomitosis during megakaryopoiesis and alters cytoskeleton organization. Blood. 2015;125(19):2995-3005.

45. Muntean AG, Crispino JD. Differential requirements for the activation domain and FOG-interaction surface of GATA-1 in megakaryocyte gene expression and development. Blood. 2005;106(4):1223-1231.

46. Stachura DL, Chou ST, Weiss MJ. Early block to erythromegakaryocytic development conferred by loss of transcription factor GATA-1. Blood. 2006;107(1):87-97.

47. Fock EL, Yan F, Pan S, Chong BH. NF-E2mediated enhancement of megakaryocytic differentiation and platelet production in vitro and in vivo. Exp Hematol.

2008;36(1):78-92.

48. Okada Y, Nagai R, Matsuura E, et al.

Suppression of RUNX1 by siRNA in megakaryocytic UT-7/GM cells. Nucleic Acids Symp Ser (Oxf). 2006;(50):261-262.

49. Kawada H, Ito T, Pharr PN, Spyropoulos DD, Watson DK, Ogawa M. Defective megakaryopoiesis and abnormal erythroid development in Fli-1 gene-targeted mice. Int J Hematol. 2001;73(4):463-468.

50. Lecine P, Villeval JL, Vyas P, Swencki B, Xu Y, Shivdasani RA. Mice lacking transcription factor NF-E2 provide in vivo validation of the proplatelet model of thrombocytopoiesis and show a platelet production defect that is intrinsic to megakaryocytes.

Blood. 1998;92(5):1608-1616.

497

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Myelodysplastic Syndromes

Ferrata Storti Foundation

Haematologica 2017 Volume 102(3):498-508

Progression in patients with low- and intermediate-1-risk del(5q) myelodysplastic syndromes is predicted by a limited subset of mutations Christian Scharenberg,1,2 Valentina Giai,1 Andrea Pellagatti,3 Leonie Saft,4 Marios Dimitriou,1 Monika Jansson,1 Martin Jädersten,1 Alf Grandien,1 Iyadh Douagi,1 Donna S. Neuberg,5 Katarina LeBlanc,1 Jacqueline Boultwood,3 Mohsen Karimi,1 Sten Eirik W. Jacobsen,1,6 Petter S. Woll1 and Eva Hellström-Lindberg1

Department of Medicine, Center for Hematology and Regenerative Medicine, Karolinska University Hospital Huddinge, Karolinska Institutet, Stockholm, Sweden; 2 Department of Medicine, Division of Hematology, Skaraborgs Hospital, Skövde, Sweden; 3Bloodwise Molecular Haematology Unit, Nuffield Division of Clinical Laboratory Sciences, Radcliffe Department of Medicine, University of Oxford, and NIHR Biomedical Research Centre, Oxford, UK; 4Department of Pathology, Karolinska University Hospital, Karolinska Institutet, Stockholm, Sweden; 5Department of Biostatistics and Computational Biology, Dana–Farber Cancer Institute, Boston, MA, USA and 6Department of Cell and Molecular Biology, Karolinska Institutet, Stockholm, Sweden 1

ABSTRACT

A

Correspondence: eva.hellstrom-lindberg@ki.se

Received: July 4, 2016.

Accepted: November 15, 2016.

Pre-published: November 24, 2016. doi:10.3324/haematol.2016.152025 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/498 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

498

A high proportion of patients with lower-risk del(5q) myelodysplastic syndromes will respond to treatment with lenalidomide.

The median duration of transfusion-independence is 2 years with some long-lasting responses, but almost 40% of patients progress to acute leukemia by 5 years after starting treatment. The mechanisms underlying disease progression other than the well-established finding of small TP53-mutated subclones at diagnosis remain unclear. We studied a longitudinal cohort of 35 low- and intermediate-1-risk del(5q) patients treated with lenalidomide (n=22) or not (n=13) by flow cytometric surveillance of hematopoietic stem and progenitor cell subsets, targeted sequencing of mutational patterns, and changes in the bone marrow microenvironment. All 13 patients with disease progression were identified by a limited number of mutations in TP53, RUNX1, and TET2, respectively, with PTPN11 and SF3B1 occurring in one patient each.

TP53 mutations were found in seven of nine patients who developed acute leukemia, and were documented to be present in the earliest sample (n=1) and acquired during lenalidomide treatment (n=6). By contrast, analysis of the microenvironment, and of hematopoietic stem and progenitor cells by flow cytometry was of limited prognostic value. Based on our data, we advocate conducting a prospective study aimed at investigating, in a larger number of cases of del(5q) myelodysplastic syndromes, whether the detection of such mutations before and after lenalidomide treatment can guide clinical decision-making.

Introduction One salient feature of malignant hematopoiesis is clonal dominance, i.e., the suppression of normal hematopoiesis by the neoplastic clone. In myelodysplastic syndromes (MDS) associated with deletion of the long arm of chromosome 5 [del(5q)], clonal dominance leads to the expansion of del(5q) hematopoietic stem cells (HSC) at the expense of normal HSC.1 In a recent study we demonstrated that rare HSC carrying del(5q) are necessary and sufficient to propagate the disease.2 Furthermore, we found that del(5q) HSC are selectively resistant to lenalidomide at the time of complete clinical and cytogenetic remission,3 potentially enabling the continual haematologica | 2017; 102(3)

Progression prediction in lower-risk del(5q) MDS

accrual of mutations and disease progression. At diagnosis, the majority of MDS patients carry recurrent mutations in a number of myeloid candidate genes, several of which are strongly associated with outcome.4-7 We observed this pattern also in patients with del(5q) MDS, in whom 64% had additional mutations detected in the HSC compartment. However, del(5q) seemed to precede the identified driver mutations in most cases, arguing that deletion of 5q is sufficient for a clonal advantage.2 The immunomodulatory drug lenalidomide has a specific effect in patients with lower-risk del(5q) MDS, abrogating the need for transfusions in around 50% of patients.8,9 The corresponding incidence of complete cytogenetic remissions varies between 16% and 26%.9 While a small subgroup of patients may maintain complete remissions for years even after the withdrawal of lenalidomide,10 the median response duration is 2 years9 and approximately 40% of the patients in the MDS 004 study had progressed to acute myeloid leukemia at 5 years.11 As the molecular mechanisms underlying disease progression in del(5q) MDS remain to be elucidated, we do not know how to predict disease progression or how to monitor patients during lenalidomide treatment. We previously reported that small TP53-mutated subclones predict for an unfavorable outcome in del(5q) patients, and that these subclones expand with disease progression.12 However, whether or not other somatic mutations or factors related to the bone marrow microenvironment also contribute to disease progression has not been comprehensively assessed. In this longitudinal study, we show for the first time that all patients with disease progression were identified by a limited subset of mutations. Based on our data, we therefore advocate that mutational profiling should be used before and during treatment of del(5q) MDS patients in order to guide individual clinical decisions.

Methods Clinical characteristics of the patients Between 2004 and 2015 we included consecutive patients at the Karolinska University Hospital in Stockholm, Sweden based on the following criteria: MDS with a low- or intermediate-1-risk score at diagnosis according to the International Prognostic Scoring System and standard cytogenetics including del5q31 without or with one additional abnormality. All patients were followed until April 2016 for survival, disease progression, and treatment. In total, 35 patients were analyzed by targeted sequencing for frequently mutated genes with known or putative implications in the pathogenesis of myeloid diseases, and material from the vast majority of patients was available for flow cytometric analysis of hematopoietic stem- and progenitor cells (HSPC). In six patients we were able to purify HSPC, by fluorescence activated cell sorting (FACS), from two or more consecutive visits for further studies. The above studies of MDS patients, as well as a study of healthy individuals were approved by the institutional review board at Karolinska Institute and both patients and healthy individuals provided written informed consent. As a definition for clinical progression, we used the same definition as that used by Jadersten et al.12 Besides defining “progression” as the development of acute myeloid leukemia (9 cases), four cases were defined as having progression based on the acquisition of additional karyotypic abnormalities (n=3) and an increase of marrow blasts from <5% to 11% (n=1) in combination with worsening cytopenias. haematologica | 2017; 102(3)

Flow cytometry, cell sorting and analysis of gene expression HSC, multipotent progenitors (MPP), lymphoid-primed multipotent progenitors (LMPP) and three subsets of myeloid progenitors, including common myeloid progenitors (CMP), granulocyte–macrophage progenitors (GMP), and megakaryocyte– erythroid progenitors (MEP) were identified using a panel of antibodies based on the following surface markers:13,14 HSC (lin– CD34+CD38–CD90+CD45RA–), MPP (lin–CD34+CD38–CD90– CD45RA–), LMPP (lin–CD34+CD38– CD90–CD45RA+), CMP (lin– CD34+CD38+CD123+CD45RA–), GMP (lin– CD34+CD38+CD123+CD45RA+); and MEP (lin– + + – – CD34 CD38 CD123 CD45RA ). Cell populations were isolated from CD34-enriched normal and MDS mononuclear cells by FACS on a FACS Aria and used for subsequent analyses. Gene expression was analyzed by Fluidigm Dynamic Arrays as previously described3 (Online Supplementary Figure S5).

Fluorescence in situ hybridization Flow-sorted cell populations were spun onto glass slides.

Slides were subsequently treated with pepsin and fixed with formaldehyde/MgCl2. The LSI EGR1/D5S721, D5S23 Dual Color Probe (Abbott-Vysis, Downers Grove, IL, USA) was used to detect deletions of 5q31; LSI EGR1 detects deletions of 5q31, and LSI D5S721, D5S23 detects 5p15.2 and serves as an internal control. Probes were applied as recommended by the manufacturer. As fluorescence in situ hybridization (FISH) analysis does not detect additional cytogenetic changes, standard cytogenetic studies were performed on mononuclear cells taken at the same time-points.

Bone marrow morphology and immunohistochemistry Sequential bone marrow samples were assessed by routine morphology and immunohistochemistry at each time-point. Bone marrow samples were assessed at diagnosis in all patients while five were analyzed prior to and at various time-points during treatment with lenalidomide (MDS063, 094, 106, 110, and 143). Bone marrow cellularity and fibrosis were assessed according to European consensus guidelines.15 Immunohistochemistry was performed for different markers including p53 DO-1 (Santa Cruz, Biotechnology, Inc., USA),11 CD34, CD68, Nestin, CD271, CD146 (Novocastra, UK), using the automated BondTM and Ventana Bench Mark XT systems according to the manufacturers’ instructions.

Microvascular density was quantified as the number of blood vessels per high power field, using regular light microscopy at high (400x) magnification as previously described.16 Blood vessels were identified as CD34+ endothelial cells forming a structure with a clearly discernable lumen. The frequency of CD34+ mononuclear cells and the tendency of CD34+ cells to form clusters were assessed as previously described.17

Mesenchymal stromal cell cultures and RNA isolation Mesenchymal stromal cells (MSC) were isolated from six untreated del(5q) cases and six healthy volunteers using a previously published standard procedure18 and expanded as detailed elsewhere, while uniformly fulfilling the minimal MSC criteria.19 Cell lysates were harvested with lysis buffer (Qiagen, Hilden, Germany), RNA was extracted using a Qiagen RNeasy minikit, and then stored in RNase-free water at -80 °C.

Affymetrix gene expression of mesenchymal stromal cells Gene expression profiling of MSC was performed as was previously described for CD34+ cells.20,21 Briefly, for each sample 100 ng of total RNA were amplified and labeled with the 3' IVT Express Kit (Affymetrix, Santa Clara, CA, USA) following the manufactur499

C. Scharenberg et al.

er’s recommendations. Biotin-labeled fragmented cRNA was hybridized to GeneChip Human Genome U133 Plus2.0 arrays (Affymetrix), covering over 47,000 transcripts. Hybridization was performed at 45 °C for 16 h in a Hybridization Oven 640 (Affymetrix). Chips were washed and stained in a Fluidics Station 450 (Affymetrix) and scanned using a GeneChip Scanner 3000 (Affymetrix). Affymetrix CEL files were pre-processed using the robust multiarray average algorithm.22 Data were analyzed using GeneSpring 12.6 (Agilent Technologies). Quality control results obtained for scale factors, background levels, percentages of present calls, 3’/5’ GAPDH ratio, and intensities of spike hybridization controls were within the acceptable ranges for all samples.23

DNA sequencing and bioinformatics analyses Haloplex target enrichment for Illumina (Agilent) was applied for mutation screening in panels of either 42 or 74 frequently mutated genes (Online Supplementary Table S1) according to the manufacturer’s instructions. Of note, the 42-gene panel covers most genes reported to be recurrently mutated in MDS4-6 and were included in both kits. Briefly, bone marrow mononuclear cells were separated by density gradient centrifugation using Lymphoprep (Nycomed, Oslo, Norway). Genomic DNA was extracted using a GeneElute DNA extraction kit (Sigma) and quantified by Qubit. All samples were individually barcoded using 96 barcoding oligos by Agilent during enrichment and the quality of individual libraries was checked by Tape Station D1K assays (Agilent). Sequencing was performed on pooled samples using either a HiSeq 2000 (Illumina) sequencer through paired–end, 100 bp reads or a MiSeq sequencer through paired-end, 150 bp reads.

Illumina Sequencing adapters were removed using Cutadapt (v.0.9.5) and reads were aligned to the hg19 using MosaikAligner (v.2.1.33). Sequence variants were identified using VarScan 2 in mpileup2cns mode. Variants were annotated using ANNOVAR.24 An ECD DNA control included in the Haloplex kit was used to filter out sequencing errors. Variants were selected for further analysis if they met the following criteria: (i) minimum coverage of 100X, (ii) minimum of 20 variant reads, (iii) having a variant allele frequency of >0.03 for all genes except for TP53 for which the limit of detection was set at ≥0.01 based on our previous studies that demonstrated these detection levels to be clinically relevant,11,12,25 (iv) not present in the 1000 Genome database, (v) not listed in dbSNP unless listed in the COSMIC 65 database, (vi) truncating or damaging based on SIFT if not present in COSMIC 65.

Sequencing results for case MDS019 were obtained by exome capture performed using SureSelect Human All Exon 50 Mb (Agilent), with sequencing done on an Illumina Genome Analyzer IIx platform. Sequencing results for case MDS110 have already been published.2

Statistical analysis Survival and time to progression (defined as blast increase to >10% or acquisition of a complex karyotype) were updated in April 2016 and were measured from the time of diagnosis.

Continuous variables were compared using the Mann-Whitney Utest or Student t test, as appropriate. Categorical variables were analyzed by the Fisher exact test. All statistical calculations were performed using Graph Pad Prism 6.0 (GraphPad Software, Inc., La Jolla, CA, USA).

Results Patients’ outcome We analyzed 35 patients with del(5q) MDS, with lowor intermediate-1-risk disease according to the 500

International Prognostic Scoring System, at one or more time-points by targeted sequencing (Online Supplementary Tables S1 and S2). Patients were allocated to receive lenalidomide treatment (n=22) or not (n=13) based on the severity of their anemia, co-morbidities and availability of other therapeutic options, such as allogeneic stem cell transplantation. After a median observation period of 61 months (range, 0.5-187) from diagnosis, 17 patients remained alive. The median survival of all patients (median age 82 years, range, 45-96) was 102 months (range, 0.5-187). These patients included six who underwent allogeneic stem cell transplantation at 145, 135, 116, 42, 12 and 6 months after diagnosis, with five of these six transplanted patients still surviving (median age at transplantation 63 years; range, 46-73). The median survival for the 29 patients who did not undergo transplantation was 70 months. Further demographic and clinical characteristics of the patients are detailed in Online Supplementary Figure S1.

Disease progression is associated with the emergence of new mutations In total, 84% of patients had a recurrent mutation in at least one gene in our panels. The summarized results of the targeted sequencing are shown in Online Supplementary Table S2. We found no differences in other clinical parameters (e.g., age, blood counts, or additional cytogenetic abnormalities) between patients in whom we detected recurrent mutations and those in whom we did not.

Considering all time-points for a patient, the most frequently mutated genes were TP53 (n=11 patients), DNMT3A (n=8), TET2 (n=7), ASXL1 (n=6) and RUNX1 (n=3) (Figure 1B). Interestingly, the mutational landscape seemed to differ from that of lower-risk MDS in general, as described in earlier reports: while mutations in genes involved in splicing were less frequent, the spectrum of mutations in this pure del(5q) cohort was more similar to that seen in high-risk MDS patients.4-6,26 Diagnostic or pre-treatment samples were available for 14 of 22 patients treated with lenalidomide (the ‘LEN’ cohort), and all (13/13) patients who did not receive lenalidomide (the ‘no LEN’ cohort), and there were no significant differences in the number or type of mutations between these two groups (P>0.99). Only two out of 22 patients failed to respond to lenalidomide treatment and although both of these patients harbored mutations, meaningful statistical analysis of these two patients was not possible.

Of the 35 patients, 13 (37%) progressed to high-risk MDS (refractory anemia with excess blasts-1, n=3 and refractory anemia with excess blasts-2, n=1) or leukemia (n=9) at a median of 85 months (range, 31-184) after diagnosis. Of the 27 patients for whom diagnostic or pre-treatment samples were available, nine (33%) showed no mutations, while 18 (67%) had one or more mutation (Online Supplementary Table S1). The presence of any recurrent mutation covered by these MDS panels early in the patients’ disease-course and prior to treatment did not predict progression (P=0.68). However, when considering the 20 patients whose samples were neither from diagnosis nor pre-treatment, an absence of mutations was suggestive of freedom from progression (P=0.073).

For 16 patients, material was available from more than one time-point, enabling longitudinal assessment of allelic burden in relation to treatment with lenalidomide or stem haematologica | 2017; 102(3)

Progression prediction in lower-risk del(5q) MDS

cell transplantation. Progression was associated with the detection of a restricted subset of new recurrent mutations, either alone or in combination (Figure 1B): TP53 (n=9, P=0.0004), TET2 (n=6, P=0.006), RUNX1 (n=3, P=0.044). In addition, we observed mutations in SF3B1 and PTPN11 in two single cases. Longitudinal samples were available for all nine patients with leukemic transformation (MDS019, 038, 063, 075, 106, 110, 143, 155, 175).

Interestingly, we detected TP53 mutations in seven of these nine patients (MDS038, 063, 075, 106, 143, 155, 175) (Figure 2A, C-F), confidently detected already in the earli-

est sample in only one case (MDS038) and acquired in six cases (MDS063, 075, 106, 143, 155, 175). Overall, there was a strong correlation between the detection of a TP53 mutation by targeted sequencing and cells staining strongly positive for TP53 by immunohistochemistry (Online Supplementary Table S4). Importantly, of six patients with evidence of acquisition of mutations in TP53 by targeted sequencing, five had been analyzed by deep sequencing without evidence of mutation at the diagnostic timepoint.12 Of these, four were negative for TP53 by immunohistochemistry. Interestingly, however, the fifth patient

A

B

Figure 1. The mutational spectrum in del(5q) patients differs in untreated versus lenalidomide-treated patients. (A) Study outline and clinical fate of patients untreated (‘no LEN’ cohort) or treated with lenalidomide (‘LEN’ cohort). *denotes two patients who are alive and well after stem cell transplantation (SCT). (B) Spectrum of mutations in relation to clinical outcome in LEN-treated versus untreated patients.

haematologica | 2017; 102(3)

501

C. Scharenberg et al.

who was negative by deep sequencing did actually show positive immunohistochemical staining of 4%.

In the remaining two patients with leukemic transformation, we detected mutations in RUNX1 (n=1, MDS019, Figure 2B) and TET2 (n=1, MDS110), present at both timepoints at which samples were taken from these patients.

Patient MDS110 also acquired a NRAS mutation at the later time-point. Three patients transformed to higher-risk MDS and all carried mutations in TET2 (MDS094, 096, 107). The only patient who progressed to refractory anemia with excess blasts-2 in the cohort not treated with lenalidomide showed mutations in TP53 and EZH2.

Regardless of whether the three mutations (TP53, TET2 and RUNX1) were present in the initial sample or whether they developed subsequently, testing positive for any of

them carried a high probability (13/16, 81%) of progression. Follow-up time after the latest mutation screening was similar between patients who progressed (median of 18 months, range 1-91 months) and those who did not (median of 27 months, range 0.3-76 months).

In 11 out of 13 patients the new mutations were detected prior to the time of clinical progression and the median time from detection of the mutation to clinical evidence of progression was 42 months (range, 0-83.9). Thus, we were able to detect the mutation in the majority of cases well before clinical signs of disease progression (Figure 3).

Surveillance of hematopoietic stem and progenitor cell subsets under lenalidomide therapy In order to investigate the impact of lenalidomide ther-

A

B

C

D

E

F

Figure 2. Longitudinal assessment of mutations during treatment with erythropoietin (shaded in red) and lenalidomide (shaded in gray). (A) Frequency of mutations in relation to the del(5q) clone in a patient who progressed to high-risk disease. (B) Variant allele frequency in a patient who progressed to leukemia, received induction therapy and went into complete remission and was transplanted. *** This patient had trisomy 21, the region in which RUNX1 resides, resulting in a homozygous mutation with amplification via trisomy 21. (C-F) Variant allele frequencies in four patients who progressed to leukemia. The size of the del(5q) clone was estimated with fluorescence in situ hybridization analysis of mononuclear cells. VAF: variant allele frequency; LEN: lenalidomide; MNC: mononuclear cells; HSC: hematopoietic stem cells; HSCT: hematopoietic stem cell transplantation; AML: acute myeloid leukemia; ALL: acute lymphocytic leukemia; CCyR: complete cytogenetic response.

502

haematologica | 2017; 102(3)

Progression prediction in lower-risk del(5q) MDS

apy on distinct HSPC in del(5q) MDS patients, we performed multicolor flow-cytometry. The distribution of sub-populations within the Lin–CD34+CD38– compartments, including HSC, MPP and LMPP,2,13 remained similar in del(5q) MDS patients, both at diagnosis and during lenalidomide treatment, compared to that in healthy controls (Figure 4). However, as previously reported,2,27,28 the GMP frequency within Lin–CD34+CD38+ cells was significantly suppressed in diagnostic del(5q) MDS with a concomitant increase in CMP. Upon lenalidomide treatment, the GMP and CMP distribution reverted to frequencies comparable to those of normal Lin–CD34+CD38+ cells.

Serial samples were available for five patients who progressed to leukemia, including one patient who did not respond to lenalidomide (MDS063). This allowed us to monitor kinetic changes in HSPC subsets within the same patient over time during treatment and disease progression (Figure 4C, Online Supplementary Figure S2 and Online Supplementary Table S3). While in each case there was one predominant HSPC subset that expanded prior to progression, the type of subset varied from patient to patient. We combined cell sorting with FISH analysis to assess the clonal size of distinct del(5q) HSPC subsets (Figure 4C and Online Supplementary Figure S3). Notably, although several of the patients investigated had a complete clinical response to lenalidomide, in all but one of these patients the mononuclear bone marrow cells and to a higher degree stem- and progenitor compartments contained a large fraction of 5q-deleted cells, and were thus not in

complete cytogenetic remission. Interestingly, in the only patient (MDS106) who initially showed a complete cytogenetic response based on FISH analysis of mononuclear bone marrow cells, the myeloid progenitor subsets (CMP, GMP and MEP) also showed minimal clonal involvement (Online Supplementary Figure S3B), whereas as much as 54% of the Lin–CD34+CD38–CD90+CD45RA– HSC compartment remained part of the del(5q) clone, supporting previous studies implicating a selective resistance of del(5q) HSC to lenalidomide treatment.

The microenvironment in del(5q) and effects of lenalidomide treatment To determine whether the failure to produce mature progeny is primarily intrinsic due to compromised HSPC or if extrinsic, microenvironmental factors contribute, we initiated MSC cultures from untreated del(5q) and healthy volunteers and generated gene expression profiles by Affymetrix microarray. While expression values for a variety of hematopoietic genes were minimal or absent, MSC cultures from both healthy volunteers and untreated del(5q) MDS patients expressed typical gene signatures for MSC (Online Supplementary Figure S4). However, we found no genes expressed in a statistically significant different manner (P<0.05, Welch t-test and Benjamini–Hochberg multiple testing correction), even when specifically looking for genes previously implicated in HSC-niche interactions29 (Figure 5A).

We next investigated bone marrow biopsies in healthy

Figure 3. Detection of mutations in advance of clinical signs of progression. The individual fates of 13 patients who progressed to either high-risk MDS (n=4) or leukemia (n=9) are depicted showing the time of diagnosis, time-point at which sequencing was performed and whether a mutation was detected or not (see legend). SCT: stem cell transplantation.

haematologica | 2017; 102(3)

503

C. Scharenberg et al.

A

B

C

Figure 4. Surveillance of hematopoietic stem and progenitor cell subsets and the phenotypic changes induced by lenalidomide. (A) FACS profiles of bone marrow stem and progenitor cells in a normal age-matched control (top row), and a representative case of del(5q) myelodysplastic syndrome at diagnosis (middle row), and del(5q) myelodysplastic syndrome treated with lenalidomide. (B) Relative distribution of stem and progenitor cell subsets within lin-CD34+CD38- and lin-CD34+CD38+ compartments in normal controls and diagnostic/untreated del(5q), and lenalidomide-treated del(5q). Indicated P-values are shown when significant by the MannWhitney test. (C) Frequency within total bone marrow and ratio of del(5q) versus normal HSC in serial samples of four patients (3 responders and 1 non-responder) during lenalidomide treatment and progression to acute myeloid leukemia. NBM: normal bone marrow; Dx: diagnosis; LEN: lenalidomide; TTP: time to progression (months); MNC: mononuclear cells; TD: transfusion-dependent; CR: complete response; LR: loss of response; PR: partial response).

504

haematologica | 2017; 102(3)

Progression prediction in lower-risk del(5q) MDS

controls and in del(5q) MDS patients before and during lenalidomide treatment. Material for longitudinal analysis by immunohistochemistry was available for five patients.

These analyses revealed that microvessel density was significantly higher in del(5q) MDS than in normal controls (microvessel density values of 5.2Âą3.2 versus 2.4Âą1.2, respectively; P=0.02) but decreased during the initial phase of lenalidomide treatment in all five patients analyzed (Figure 5B). Subsequent therapeutic failure was associated with an increase in bone marrow cellularity and microves-

sel density in four of the five patients. The number of CD68+ macrophages was not increased in bone marrow samples from del(5q) MDS patients as compared to the number in controls; however, upon lenalidomide treatment a decrease was noted which was paralleled by a decrease in cellularity (Figure 5B). Surrogate markers for MSC (e.g., nestin, CD271, CD146) demonstrated labeling restricted to perivascular mesenchymal cells including endothelial cells and adventitial sinusoidal cells.

Taken together, these experiments demonstrate that

A

Figure 5. Minor alterations within the microenvironment. (A) Heatmap of 13 genes associated with the hematopoietic stem cell-niche interaction. The left six lanes show the healthy controls (NBM) and the right six the del(5q) cases. (B) Immunohistochemistry for markers associated with niche cells in the bone marrow microenvironment. Representative images from a normal control (normal BM) compared to one patient (MDS143) before lenalidomide-treatment, during complete cytogenetic response (19 months on lenalidomide, CCyR) and when the patient stopped responding to lenalidomide (35 months). LEN: lenalidomide; CCyR: complete cytogenetic response; resp: response.

B

haematologica | 2017; 102(3)

505

C. Scharenberg et al.

despite affecting microvessel density, lenalidomide did not exhibit its effects primarily via alteration of the cellular composition of the microenvironment based on the MSC markers tested.

Discussion In this study we found that all patients with lower-risk MDS and isolated del(5q) who progressed to either higherrisk MDS or transformed to acute leukemia harbored recurrent mutations in TP53, RUNX1, and TET2 in addition to the 5q deletion. Not surprisingly, we found that mutations increased in individual patients over time. While 62% of samples obtained before treatment showed mutations in addition to del(5q), 84% of samples carried mutations at the latest time-point analyzed, and several patients showed increased allele burdens and gains of new mutations during the course of disease and treatment. This suggests that clonal evolution is frequent in patients with lower-risk del(5q) MDS and argues that the del(5q) aberration is associated with marked clonal instability. By contrast, Chesnais et al. reported next-generation sequencing data from 94 nondel(5q) lower-risk patients treated with lenalidomide Âą erythropoietin and found that only about one-third of these patients had more than one genetic event, most often consisting of a SF3B1 mutation plus one additional mutation.

Moreover, response to lenalidomide was associated with a decrease in allelic burdens of the identified mutations, and only two of 18 patients analyzed at a later time-point had acquired new recurrent mutations.30 In our cohort, 13 of 35 patients progressed to either higher-risk MDS (n=4) or leukemia (n=9), 12 of whom were treated with lenalidomide. Seven of the nine patients who developed leukemia carried a TP53 mutation. Based on a median sequencing depth of 370 reads, the mutation was considered present before treatment in one of these patients (MDS038) and to have developed under treatment in the other six. The presence of very small TP53 mutated subclones prior to treatment cannot be excluded, but five of these six patients had previously been analyzed in a study by Jädersten et al. using deep-sequencing analysis (coverage of 1200X) and had been found to be negative,12 and four also proved to be TP53-negative by immunohistochemistry in the present study. Given that normal function of TP53 is a requirement for apoptosis of erythroid cells due to haploinsufficiency of RPS14,31 it is highly possible that TP53 mutations may be selected out as a consequence of the 5q deletion. We observed that disease progression associated with the acquisition of TP53 mutations was relatively common in these lenalidomidetreated del(5q) patients, with some patients even exhibiting more than one TP53 mutation. The marked clonal heterogeneity and instability revealed in this study, is likely to play a role in disease progression of lower-risk del(5q) MDS treated with lenalidomide. While isolated del(5q) in lower-risk MDS has been associated with a relatively low risk for leukemic transformation compared to other MDS subtypes, del(5q) is known to be associated with an adverse prognosis and a high incidence of TP53 mutations in the context of complex karyotypes in newly diagnosed MDS as well as de novo acute myeloid leukemia.32-34 Our data show that although TP53 was the most common molecular event at progression, the emergence of other mutations could be linked to either loss of treatment 506

response or to progression. RUNX1 mutations in our cohort were restricted to patients with disease progression and found in three of 13 patients. In none of these patients did the RUNX1 allele burden suggest the presence of a germ-line mutation. RUNX1 is a well-established marker of poor prognosis in both MDS and acute myeloid leukemia.35-37 Furthermore, we found mutations in TET2 in six of 13 patients with evidence of disease progression.

Although three patients had mutations in both TP53 and TET2, our data do not provide evidence that this was not a result of independent mutational processes. We note that while mutations in TET2 are relatively common in myeloid neoplasms in general,38,39 their impact in MDS is less clear,40,41 although one study reported that TET2 mutations were associated with shorter survival in MDS patients undergoing HSC transplantation.42 Our data on del(5q) patients are in line with recent findings in myeloproliferative neoplasms in which TET2 mutations were associated with disease progression if they were acquired in a JAK2-mutated subclone.43 Our study of the clonal dynamics of all major HSPC in vivo shows that clonal advantage is not a feature restricted only to MDS stem cells but also extends to the myeloid and erythroid progenitor compartments. Using flow cytometry for surveillance of HSPC subsets in lenalidomide-treated patients, we found that neither lenalidomide treatment nor the acquisition of additional mutations led to any uniform, profound changes in the hematopoietic hierarchy unless the patient showed clinical signs of progression. Importantly, among patients who eventually progressed but initially had a complete clinical response, there was no difference between patients who reached a complete cytogenetic response and those who did not.

Although lenalidomide temporarily reduced the size of the del(5q) stem and myeloid progenitor cell compartments, in no case did we observe complete clearance of del(5q) cells, and this was again irrespective of the mutational status of the patient.

Mutations in either tumor-suppressors or oncogenes have the potential to modify the competitive nature of cells, transforming them into either winners or losers with respect to normal cells.44 The relative cell fitness is dependent upon the cellular context and not simply the result of altered cell proliferation. In this regard, the microenvironment is an important regulatory component when cancer cells compete with normal (stem) cells.

However, our data do not support that the microenvironment in del(5q) MDS exerts a dominant constraint towards healthy hematopoiesis. Our studies of MSC grown in vitro confirm previous findings that the stromal component of the marrow microenvironment is not derived from the malignant clone in MDS.45 Microarray analysis exhibited an expression footprint consistent with MSC with high expression of MSC markers and absence of hematopoietic gene signatures. However, we observed only minor differences in gene expression between pre-treatment del(5q) and healthy MSC. While seemingly at odds with recent findings in cohorts of multiple subtypes of MDS,46,47 our studies in a pure del(5q) cohort are in line with earlier studies by other groups who found the stromal abnormalities to be reversible and that MDS stroma is able to support normal in vitro hematopoiesis.48,49 In conclusion, while flow cytometric analysis of HSPC populations or analysis of the microenvironment had limited predictive value in this cohort of lower-risk del(5q) MDS, haematologica | 2017; 102(3)

Progression prediction in lower-risk del(5q) MDS

all patients who progressed to either higher-risk MDS or leukemia were identified by harboring recurrent mutations in a limited number of genes, i.e., TP53, RUNX1, and TET2.

Based on our data, we advocate conducting a prospective study aimed at investigating, in a larger number of del(5q) MDS cases before and after lenalidomide treatment, whether the detection of such mutations can guide clinical decision-making, such as suggesting which patients should undergo hematopoietic cell transplantation.

References 1. Nilsson L, Astrand-Grundström I, Arvidsson I, et al. Isolation and characterization of hematopoietic progenitor/stem cells in 5qdeleted myelodysplastic syndromes: evidence for involvement at the hematopoietic stem cell level. Blood. 2000;96(6):2012–2021.

2. Woll PS, Kjällquist U, Chowdhury O, et al.

Myelodysplastic syndromes are propagated by rare and distinct human cancer stem cells in vivo. Cancer Cell. 2014;25(6):794–808.

3. Tehranchi R, Woll PS, Anderson K, et al.

Persistent malignant stem cells in del(5q) myelodysplasia in remission. N Engl J Med. 2010;363(11):1025–1037.

4. Bejar R, Stevenson K, Abdel-Wahab O, et al.

Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med. 2011;364(26):2496–2506.

5. Haferlach T, Nagata Y, Grossmann V, et al.

Landscape of genetic lesions in 944 patients with myelodysplastic syndromes.

Leukemia. 2014;28(2):241–247.

6. Papaemmanuil E, Gerstung M, Malcovati L, et al. Clinical and biological implications of driver mutations in myelodysplastic syndromes. Blood. 2013;122(22):3616–3627.

7. Karimi M, Nilsson C, Dimitriou M, et al.

High-throughput mutational screening adds clinically important information in myelodysplastic syndromes and secondary or therapy-related acute myeloid leukemia.

Haematologica. 2015;100(6):e223–225.

8. List A, Dewald G, Bennett J, et al.

Lenalidomide in the myelodysplastic syndrome with chromosome 5q deletion. N Engl J Med. 2006;355(14):1456–1465.

9. Fenaux P, Giagounidis A, Selleslag D, et al. A randomized phase 3 study of lenalidomide versus placebo in RBC transfusion-dependent patients with low-/intermediate-1-risk myelodysplastic syndromes with del5q.

Blood. 2011;118(14):3765–3776.

10. Giagounidis AAN, Kulasekararaj A, Germing U, et al. Long-term transfusion independence in del(5q) MDS patients who discontinue lenalidomide. Leukemia.

2012;26(4):855–858.

11. Saft L, Karimi M, Ghaderi M, et al. p53 protein expression independently predicts outcome in patients with lower-risk myelodysplastic syndromes with del(5q).

Haematologica. 2014;99(6):1049.

12. Jädersten M, Saft L, Smith A, et al. TP53 mutations in low-risk myelodysplastic syndromes with del(5q) predict disease progression. J Clin Oncol. 2011;29(15):1971–1979.

13. Majeti R, Park CY, Weissman IL. Identification of a hierarchy of multipotent hematopoietic progenitors in human cord

haematologica | 2017; 102(3)

Acknowledgments EHL is funded through the Swedish Cancer Society, the Scientific Research Council and the Cancer Society in Stockholm.

CS was supported by a research fund at Skaraborgs Hospital, the Skaraborg Research and Development Council and received a PhD fellowship from Karolinska Institutet. AP and JB are supported by Bloodwise (UK). SEWJ is supported by the Tobias Foundation and a grant from the Center for Innovative Medicine (CIMED) at the Karolinska Institute.

blood. Cell Stem Cell. 2007;1(6):635–645.

14. Goardon N, Marchi E, Atzberger A, et al.

Coexistence of LMPP-like and GMP-like leukemia stem cells in acute myeloid leukemia. Cancer Cell. 2011;19(1):138–152.

15. Thiele J, Kvasnicka HM, Facchetti F, et al.

European consensus on grading bone marrow fibrosis and assessment of cellularity.

Haematologica. 2005;90(8):1128–1132.

16. Lundberg LG, Hellström-Lindberg E, KanterLewensohn L, Lerner R, Palmblad J. Angiogenesis in relation to clinical stage, apoptosis and prognostic score in myelodysplastic syndromes. Leuk Res. 2006;30(3): 247–253.

17. Porta Della MG, Malcovati L, Boveri E, et al.

Clinical relevance of bone marrow fibrosis and CD34-positive cell clusters in primary myelodysplastic syndromes. J Clin Oncol.

2009;27(5):754–762.

18. Moll G, Rasmusson-Duprez I, Bahr von L, et al. Are therapeutic human mesenchymal stromal cells compatible with human blood?

Stem Cells. 2012;30(7):1565–1574.

19. Krampera M, Galipeau J, Shi Y, et al.

Immunological characterization of multipotent mesenchymal stromal cells--the International Society for Cellular Therapy (ISCT) working proposal. Cytotherapy.

2013;15(9):1054–1061.

20. Pellagatti A, Cazzola M, Giagounidis AA, et al. Gene expression profiles of CD34+ cells in myelodysplastic syndromes: involvement of interferon-stimulated genes and correlation to FAB subtype and karyotype. Blood.

2006;108(1):337–345.

21. Pellagatti A, Cazzola M, Giagounidis A, et al. Deregulated gene expression pathways in myelodysplastic syndrome hematopoietic stem cells. Leukemia. 2010;24(4):756–764.

22. Irizarry RA, Hobbs B, Collin F, et al.

Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4(2):249–264.

23. Pellagatti A, Benner A, Mills KI, et al.

Identification of gene expression-based prognostic markers in the hematopoietic stem cells of patients with myelodysplastic syndromes. J Clin Oncol. 2013;31(28): 3557– 3564.

24. Koboldt DC, Zhang Q, Larson DE, et al.

VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res. 2012;22(3): 568– 576.

25. Jädersten M, Saft L, Pellagatti A, et al. Clonal heterogeneity in the 5q- syndrome: p53 expressing progenitors prevail during lenalidomide treatment and expand at disease progression. Haematologica. 2009;94 (12):1762–1766.

26. Yoshida K, Sanada M, Shiraishi Y, et al.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Frequent pathway mutations of splicing machinery in myelodysplasia. Nature.

2011;478(7367):64–69.

Will B, Zhou L, Vogler TO, et al. Stem and progenitor cells in myelodysplastic syndromes show aberrant stage-specific expansion and harbor genetic and epigenetic alterations. Blood. 2012;120(10):2076–2086.

Pang WW, Pluvinage JV, Price EA, et al.

Hematopoietic stem cell and progenitor cell mechanisms in myelodysplastic syndromes.

Proc Natl Acad Sci USA. 2013;110(8): 3011– 3016.

Méndez-Ferrer S, Michurina TV, Ferraro F, et al. Mesenchymal and haematopoietic stem cells form a unique bone marrow niche.

Nature. 2010;466(7308):829–834.

Chesnais V, Renneville A, Toma A, et al.

Effect of lenalidomide treatment on clonal architecture of myelodysplastic syndromes without 5q deletion. Blood. 2016;127(6): 749–760.

Barlow JL, Drynan LF, Hewett DR, et al. A p53-dependent mechanism underlies macrocytic anemia in a mouse model of human 5q- syndrome. Nat Med. 2010;16 (1):59–66.

Zemanova Z, Michalova K, Buryova H, et al. Involvement of deleted chromosome 5 in complex chromosomal aberrations in newly diagnosed myelodysplastic syndromes (MDS) is correlated with extremely adverse prognosis. Leuk Res. 2014;38(5): 537–544.

Sebaa A, Ades L, Baran-Marzack F, et al.

Incidence of 17p deletions and TP53 mutation in myelodysplastic syndrome and acute myeloid leukemia with 5q deletion.

Genes Chromosomes Cancer. 2012;51(12): 1086–1092.

Milosevic JD, Puda A, Malcovati L, et al.

Clinical significance of genetic aberrations in secondary acute myeloid leukemia. Am J Hematol. 2012;87(11):1010–1016.

Chen C-Y, Lin L-I, Tang J-L, et al. RUNX1 gene mutation in primary myelodysplastic syndrome--the mutation can be detected early at diagnosis or acquired during disease progression and is associated with poor outcome. Br J Haematol. 2007;139(3):405–414.

Steensma DP, Gibbons RJ, Mesa RA, Tefferi A, Higgs DR. Somatic point mutations in RUNX1/CBFA2/AML1 are common in highrisk myelodysplastic syndrome, but not in myelofibrosis with myeloid metaplasia. Eur J Haematol. 2005;74(1):47–53.

Tang J-L, Hou H-A, Chen C-Y, et al.

AML1/RUNX1 mutations in 470 adult patients with de novo acute myeloid leukemia: prognostic implication and interaction with other gene alterations. Blood.

2009;114(26):5352–5361.

Delhommeau F, Dupont S, Valle Della V, et

507

C. Scharenberg et al.

39.

40.

41.

42.

508

al. Mutation in TET2 in myeloid cancers. N Engl J Med. 2009;360(22):2289–2301.

Langemeijer SMC, Kuiper RP, Berends M, et al. Acquired mutations in TET2 are common in myelodysplastic syndromes. Nat Genet.

2009;41(7):838–842.

Abdel-Wahab O, Mullally A, Hedvat C, et al. Genetic characterization of TET1, TET2, and TET3 alterations in myeloid malignancies. Blood. 2009;114(1):144–147.

Smith AE, Mohamedali AM, Kulasekararaj A, et al. Next-generation sequencing of the TET2 gene in 355 MDS and CMML patients reveals low-abundance mutant clones with early origins, but indicates no definite prognostic value. Blood. 2010;116 (19): 3923–3932.

Bejar R, Stevenson KE, Caughey B, et al.

Somatic mutations predict poor outcome in patients with myelodysplastic syn-

43.

44.

45.

46.

drome after hematopoietic stem-cell transplantation. J Clin Oncol. 2014;32(25): 2691–2698.

Schaub FX, Lehmann T, Looser R, et al.

Transition to homozygosity does not appear to provide a clonal advantage to hematopoietic progenitors carrying mutations in TET2.

Blood. 2011;117(6):2075–2076.

Ballesteros-Arias L, Saavedra V, Morata G. Cell competition may function either as tumour-suppressing or as tumour-stimulating factor in Drosophila. Oncogene.

2014;33(35):4377–4384.

Ramakrishnan A, Awaya N, Bryant E, Torok-Storb B. The stromal component of the marrow microenvironment is not derived from the malignant clone in MDS. Blood. 2006;108(2):772–773.

Medyouf H, Mossner M, Jann J-C, 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.

47. Geyh S, Oz S, Cadeddu RP, et al. Insufficient stromal support in MDS results from molecular and functional deficits of mesenchymal stromal cells. Leukemia. 2013;27(9): 1841– 1851.

48. Soenen-Cornu V, Tourino C, Bonnet M-L, et al. Mesenchymal cells generated from patients with myelodysplastic syndromes are devoid of chromosomal clonal markers and support short- and long-term hematopoiesis in vitro. Oncogene. 2005;24 (15):2441–2448.

49. Deeg HJ, Beckham C, Loken MR, et al.

Negative regulators of hemopoiesis and stroma function in patients with myelodysplastic syndrome. Leuk Lymphoma.

2000;37(3-4):405–414.

haematologica | 2017; 102(3)

ARTICLE

Myeloproliferative Disorders

Aberrant let7a/HMGA2 signaling activity with unique clinical phenotype in JAK2-mutated myeloproliferative neoplasms

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Chih-Cheng Chen,1,3 Jie-Yu You,4,5 Jrhau Lung,2 Cih-En Huang,1,3 Yi-Yang Chen,1 Yu-Wei Leu,6 Hsing-Ying Ho,1 Chian-Pei Li,1 Chang-Hsien Lu,1,3 Kuan-Der Lee,1,3 Chia-Chen Hsu1,* and Jyh-Pyng Gau5,7*

Division of Hematology and Oncology, Department of Medicine, Chang Gung Memorial Hospital, Chiayi; 2Division of Pulmonary and Critical Care Medicine, Department of Medicine, Chang Gung Memorial Hospital, Chiayi; 3College of Medicine, Chang Gung University, Tao-Yuan; 4Division of Hematology and Oncology, Department of Medicine, Lotung Poh-Ai Hospital, Yilan; 5School of Medicine, National Yang-Ming University, Taipei; 6 Department of Life Science and Institute of Molecular Biology, National Chung Cheng University, Chiayi; 7Division of Hematology, Department of Medicine, Taipei Veterans General Hospital, Taiwan 1

*C-CH and J-PG contributed equally to this work.

Haematologica 2017 Volume 102(3):509-518

ABSTRACT

H

igh mobility group AT-hook 2 (HMGA2) is an architectural transcription factor that is negatively regulated by let-7 microRNA through binding to it’s 3’-untranslated region. Transgenic mice expressing Hmga2 with a truncation of its 3’-untranslated region has been shown to exhibit a myeloproliferative phenotype. To decipher the let-7-HMGA2 axis in myeloproliferative neoplasms, we employed an in vitro model supplemented with clinical correlation. Ba/F3 cells with inducible JAK2V617F expression (Ton.JAK2.V617F cells) showed upregulation of HMGA2 with concurrent let-7a repression. Ton.JAK2.V617F cells treated with a let-7a inhibitor exhibited further escalation of Hmga2 expression, while a let-7a mimic diminished the Hmga2 transcript level.

Hmga2 overexpression conferred JAK2-mutated cells with a survival advantage through inhibited apoptosis. A pan-JAK inhibitor, INC424, increased the expression of let-7a, downregulated the level of Hmga2, and led to increased apoptosis in Ton.JAK2.V617F cells in a dose-dependent manner. In samples from 151 patients with myeloproliferative neoplasms, there was a modest inverse correlation between the expression levels of let-7a and HMGA2. Overexpression of HMGA2 was detected in 29 (19.2%) of the cases, and it was more commonly seen in patients with essential thrombocythemia than in those with polycythemia vera (26.9% vs. 12.7%, P=0.044). Patients with upregulated HMGA2 showed an increased propensity for developing major thrombotic events, and they were more likely to harbor one of the 3 driver myeloproliferative neoplasm mutations in JAK2, MPL and CALR. Our findings suggest that, in a subset of myeloproliferative neoplasm patients, the let-7-HMGA2 axis plays a prominent role in the pathogenesis of the disease that leads to unique clinical phenotypes.

Introduction Philadelphia chromosome-negative myeloproliferative neoplasms (MPNs), which include polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF), are a group of clonal disorders of the hematopoietic system characterized by excessive production of differentiated myeloid cells. With the discoveries of underlying driver mutations in JAK2, MPL, and calreticulin (CALR) that together account for 90% of patients with myeloproliferative neoplasm (MPN), it is now clear these conditions are characterized by dysregulated JAKhaematologica | 2017; 102(3)

Correspondence: jpgau@vghtpe.gov.tw/ chiachen.hsu1008@gmail.com Received: August 10, 2016.

Accepted: December 30, 2016.

Pre-published: January 5, 2017.

doi:10.3324/haematol.2016.154385 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/509 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

509

C.-C. Chen et al.

STAT (signal transducer and activator of transcription) signaling pathways.1 However, none of these mutations have been proven to be specific to disease subtype. As a result, they cannot be used in the molecular classification of MPNs. In addition, it remains unclear why the same acquired mutation in one of these genes causes similar clinical entities with distinct phenotypes.

Some studies allude to the fact that JAK2V617F confers only a weak growth advantage to hematopoietic stem cells.2,3 An emerging hypothesis suggests that several cooperating genetic hits might be required to induce disease and allow progression, as mutations in signaling molecules are not sufficient for disease development in humans.4 There are quite a few other genetic alterations identified in MPN;5 among them, several are implicated in epigenetic regulation, either in histone modifications or in DNA methylation control.6 Studies have shown that these epigenetic regulator genes and JAK2 mutations are synergistic by combining an early and late amplification, with mutation of the former mainly expanding the hematopoietic progenitor cells, whereas JAK2V617F mainly expands the mature fraction.4 Therefore, it is widely postulated that epigenetic regulation might also play an important role in the pathogenesis of MPN. The high mobility group A2 (HMGA2) gene codes for a non-histone protein that has no intrinsic transcriptional activity, but can modulate transcription by altering the chromatin architecture.7,8 The HMGA2 protein, highly expressed during embryogenesis but diminished in normal adult tissues, is thought to play an essential role in selfrenewal and the control of differentiation of embryonic stem cells.9 However, high levels of HMGA2 are found in various tumors, especially those of mesenchymal origin.10 The 3’-untranslated region (3’-UTR) of HMGA2 contains 7 sequences complementary to the let-7 microRNA (miRNA), which negatively regulates HMGA2 expression.11 In some tumors, rearrangement around the region of chromosome 12q14-15, the location of the HMGA2 gene, can lead to a deletion of the HMGA2 3’UTR and loss of let-7 binding sites. This results in overexpression of a full-length or truncated HMGA2 protein which promotes tumor formation.2 Guglielmelli et al. first reported the association between HMGA2 upregulation and MPNs. In their seminal work studying the molecular profiling of CD34+ cells in PMF, they found that abnormal expression of HMGA2 was dependent on the presence of JAK2V617F mutation.12 Subsequently, Ikeda and colleagues demonstrated that transgenic mice expressing 3’UTR-truncated Hmga2 (ΔHmga2) cDNA exhibited a myeloproliferative phenotype.13 Enforced expression of ΔHmga2 led to a proliferative advantage in hematopoietic stem and progenitor cells. However, in spite of these studies, there are only scarce data available on the frequencies of dysregulated HMGA2 signaling activity in MPN patients, which severely limits the kinds of conclusions one can draw. Moreover, it remains unclear how HMGA2 and JAK2V617F interact with each other, and whether upregulation of HMGA2 plays specific roles in the pathogenesis of JAK2-mutated MPNs. In the study herein, we aimed to address some of these issues. An in vitro model was employed to elucidate the correlation between JAK2V617F and HMGA2 expression.

Furthermore, the phenotypic influences of HMGA2 overexpression on JAK2-mutated MPNs and the cause of HMGA2 upregulation were also explored.

510

Methods Study population and mutational analysis Relevant information on the patient enrollment, diagnosis,14 treatment,15 definition of events,16,17 and measurement of survival are listed in the Online Supplementary File. All participants provided informed consent in accordance with the Declaration of Helsinki.

The study was approved by our Institutional Review Board. The detection of JAK2V617F, JAK2 Exon 12, CALR, and MPL mutations in clinical samples was performed as previously described.18

Cell lines and doxycycline induction Interleukin-3 (IL-3)-dependent Ba/F3 cells with inducible expression of JAK2V617F (Ton.JAK2.V617F) or wild-type (WT) JAK2 (Ton.JAK2.WT) were kindly provided by Professor Gregor Hoermann and Professor Matthias Mayerhofer (Medical University of Vienna, Austria). The expression of JAK2 was induced by the addition of doxycycline (1mg/ml). The cells were maintained in IL-3 throughout the experiments until 3 hours

Table 1. Clinical and laboratory features of 151 patients with myeloproliferative neoplasm, stratified by expressional status of the HMGA2 gene.

Age, years Female, no. (%) WBC count, x 109/L Hemoglobin, g/L PLT count, x 109/L LDH, U/L Marrow Cellularity, % Splenomegaly, no. (%) Major thrombosis, no. (%) Diagnosis PV (n=63)

HMGA2(-) N=122

HMGA2(+) N=29

P

59.5 ± 16.1 50 (41.0%) 13.6 ± 9.1 15.3 ± 4.1 576 ± 414 287 ± 145 72.9 ± 20.5 49 (45.8%) 19 (15.6%)

65.7 ± 14.4 19 (65.5%) 13.9 ± 6.2 14.5 ± 3.7 789 ± 445 322 ± 264 68.5 ± 21.0 17 (58.6%) 10 (34.5%)

0.058 0.017 0.873 0.345 0.016 0.368 0.449 0.220 0.021

55 (87.3%)

8 (12.7%)

ET (n=67)

49 (73.1%)

PMF (n=17)

14 (82.4%)

MPN, unclassifiable (n=4) 4 (100%) Disease risk^ in PV/ET patients Low-risk 45 High-risk 59 MF Transformation*, no. (%) 5 (4.8%) Driver mutation# Present 94 (77.0%) Absent 28 JAK2 mutation versus triple negative JAK2 mutated, no. (%) 78 (73.6%) Triple negative 28

(PV vs. ET) 0.044 18 (26.9%) (ET vs. PMF) 0.542 3 (17.6%) (PV vs. PMF) 0.693 0 0.026 5 21 1 (3.8%) 0.875 0.051 27 (93.1%) 2 0.040 22 (91.7%) 2

^Disease risk stratified according to ELN (European LeukemiaNet) guideline.15 *MF transformation: Transformation to post-PV or post-ET myelofibrosis in PV & ET patients. # Driver mutations included JAK2V617F, JAK2 Exon12, CALR, and MPL mutation. Information regarding thrombosis included events at diagnosis and during follow up. MF transformation was assessed during follow up. Other hematological and clinical parameters were collected at diagnosis. Values are reported as mean ± SD unless otherwise indicated. P values with statistically significant differences are shown in bold. WBC: white blood cell; PLT: platelet; LDH: lactate dehydrogenase; PV: polycythemia vera; ET: essential thrombocythemia; PMF: primary myelofibrosis; MPN: myeloproliferative neoplasms; MF: myelofibrosis.

haematologica | 2017; 102(3)

let-7a/HMGA2 axis in JAK2-mutated MPNs

before they were subjected to real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and western blot analysis. Sources of other cells used are listed in the Online Supplementary File.

Gene silencing by small interfering RNA transfection, miRNA ectopic expression and inhibition The small interfering RNA (siRNA) oligos that directed against the mouse Hmga2 messenger RNA (mRNA, siHmga2) or a scrambled sequence (siScr) were purchased from Sigma-Aldrich. Small oligos that either mimic or inhibit endogenous let-7a were purchased from ABI (mirVana, Thermo Fisher Scientific Inc.). All the transfection was performed using X-tremeGENE siRNA Transfection Reagent (Roche) according to the manufacturer's specifications. The efficiency of various siRNA oligos is demonstrated in the Online Supplementary Figure S1. The si-1 Hmga2 siRNA (siHmga2) was considered an ideal option and chosen for further experiments. The oligo concentrations used for let-7a inhibition were 0.2 and 0.5 nM according to the manufacturer's suggestion, whereas 0.5 nM was used for the ectopic expression of

let-7a. All the cells were harvested and validated by qRT-PCR or western blotting after 48 hours of transfection.

Cell viability and apoptosis assays Cell survival was measured by the XTT assay according to the manufacturer's instructions (Biological Industries). The percentages of apoptotic cells were determined using a FACS Canto II flow cytometer (Becton, Dickinson and Company) after the cells were treated with a 7-AAD and PE-Annexin V Apoptosis Detection Kit (BD Pharmingen). Each data result and its respective error bar were measured by 3 independent experiments run in triplicate.

Growth inhibition assay The pan-JAK inhibitor INC424 (ruxolitinib) was kindly provided by Novartis Pharmaceuticals. INC424 was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 10 mM as the stock solution. 0.5% DMSO was added to the culture medium as control. Candidate cells were exposed to various concentrations of hydroxyurea (Bristol-Myers Squibb) or INC424 for 72 h. The drug

B

Hmga2

A

D

Hmga2

C

Figure 1. The levels of HMGA2 expression in cells with various JAK-STAT signaling activity. (A) Quantitative RT-PCR analysis of Hmga2 transcript levels in parental Ba/F3 cells, stable Ba/F3 cells co-transfected with MPL and either type 1 (deletion; DEL) or type 2 (insertion; INS) CALR mutant, and stable, inducible Ton.JAK2.V617F cells. The Ton.JAK2.V617F cells were treated with doxycycline (1 mg/ml) for at least 6 days before being subjected to analysis. Representative data from three independent experiments are presented. The error bars show the standard deviation (Âą SD) of three independent experiments. Asterisk indicates statistical significance (t-test; *P<0.01; ***P<0.0001). (B) Western blot analysis of JAK2, phosphorylated JAK2 (p-JAK2), and HMGA2 levels in 4 human acute myeloid leukemia cell lines: JAK2V61F-carrying HEL and UKE-1 cells, and JAK2-unmutated KG-1a and HL-60 cells. (C) The expression levels of JAK2, p-JAK2, STAT3, p-STAT3, STAT5, p-STAT5, and HMGA2 proteins in stable, inducible Ton.JAK2.WT and Ton.JAK2.V617F cells. Cells were treated with doxycycline (1 mg/ml; for at least 6 days) before being harvested concurrently with untreated controls for subsequent analysis. (D) Quantitative RT-PCR analysis of Hmga2 transcripts in parental Ba/F3, Ton.JAK2.WT and Ton.JAK2.V617F cells at baseline as well as 2, 4, and 6 days after addition of doxycycline (1 mg/ml), respectively. The error bars show the standard deviation of three independent experiments. Asterisk indicates statistical significance (t-test; *P<0.01; **P<0.001). WT: wild-type.

haematologica | 2017; 102(3)

511

C.-C. Chen et al.

concentrations that inhibited cell growth by 50% (IC50) and 90% (IC90) were determined by the XTT assay.

Others Protocols on western blotting, qRT-PCR, and fluorescence in situ hybridization (FISH) are all listed in the Online Supplementary File.19

Statistical analysis All calculations were performed using the Statistical Package for Social Sciences software (version 17.0; SPSS, Inc.). The level of statistical significance was set at 0.05 for all tests.

Results Mutated JAK2 activates JAK-STAT pathway and up-regulates HMGA2 expression We hypothesized that HMGA2 upregulation could be seen in cells with JAK-STAT signaling pathway activation, and chose to check its expressional status in MPN cells harboring either one of the two most common driver mutations (JAK2 and CALR). As demonstrated in Figure 1A, there was an 8-fold increase in Hmga2 levels in Ton.JAK2.V617F cells. The increment, however, was only around 2-fold in both Ba/F3 cells co-transduced with wildtype MPL and either type I (deletion) or type II (insertion) CALR mutants. Knowing that both mutated JAK2 and CALR activated JAK-STAT signaling,20-22 and considering the fact that a rise in Hmga2 expression was more prominent in JAK2-mutated cells, we used mutated JAK2 as our model of current investigation, but did not further explore CALR-mutated cells. To further validate the rise of HMGA2 expression in JAK2-mutated cells, we screened four different human acute myeloid leukemia (AML) cell lines. Western blotting showed that JAK2V617F-carrying HEL and UKE-1 cells exhibited increased JAK2 phosphorylation and enhanced HMGA2 expression (Figure 1B). On the contrary, HMGA2 expression was not increased in either JAK2-unmutated KG-1a or HL-60 cells. We next examined the signaling activity in cells with induced expression of JAK2. As demonstrated in Figure 1C, upon addition of doxycycline, Ton.JAK2.V617F cells exhibited enhanced JAK-STAT signaling as shown by increased phosphorylation of JAK2, STAT3, and STAT5. The level of HMGA2 protein was also increased significantly. On the other hand, the phosphorylation of JAK2 protein was rather weak in Ton.JAK2.WT cells. Although there was HMGA2 upregulation in these wild-type cells, the increment in HMGA2 expression was not as substantial as that seen in Ton.JAK2.V617F cells. Increased Hmga2 transcripts could be observed at 2 days after induction of JAK2V617F expression (Figure 1D), and the levels of transcripts continued to rise after extended exposure to doxycycline in Ton.JAK2.V617F cells. These data demonstrate that mutated JAK2 could activate the JAK-STAT pathway and upregulate Hmga2 expression in Ba/F3 cells.

culate the relative HMGA2 expression in our patients.

Among the healthy individuals, the distribution of HMGA2 transcript levels ranged between 0.02–3.0 times the mean level of the whole control population.

Therefore, we defined the patients with a 2−ΔΔCT value of HMGA2 transcript greater than 3 as HMGA2(+), whereas the remainders were designated as HMGA2(-). Overall, 29 (19.2%) patients exhibited overexpression of HMGA2 in their peripheral blood (PB) granulocytes. Table 1 demonstrates the clinical and laboratory characteristics of the 151 MPN patients subcategorized into the two groups.

Patients whose granulocytes harbored increased HMGA2 expression had a higher probability of carrying a driver mutation than those without, though the comparison was of borderline significance (93.1% vs. 77.0%, P=0.051).

When taking only JAK2-mutated and triple negative (TN) patients into consideration, we had 22 (91.7%) JAK2-mutated and 2 (8.3%) TN patients in the HMGA2(+) group, and 78 (73.6%) and 28 (26.4%) in the HMGA2(-) group, respectively. The comparison regarding their difference was more significant (P=0.040, Table 1). Based on our in vitro data, it’s plausible that JAK2V617F induced a higher degree of Hmga2 upregulation (Figure 1A). Therefore, the difference in the frequencies of JAK2 mutation between the HMGA2(+) and (-) groups became more prominent when we included only JAK2 -mutated and TN patients as the denominator. Female patients constituted the main population of HMGA2(+) patients (65.5%), in contrast with that seen in the HMGA2(-) group (41.0%, P=0.017).

Strikingly, compared to those without HMGA2 overexpression, HMGA2-upregulated patients presented with a higher platelet count (789 ± 445 x 109/L vs. 576 ± 414 x 109L, P=0.016). This could be the result of a higher prevalent rate of HMGA2 overexpression among ET patients (26.7%) as compared to that seen in PV patients (12.7%, P=0.044). Overall, HMGA2(+) MPN patients were more likely to suffer from major thrombotic events (34.5% vs.

15.6%, P=0.021). Consistently, Kaplan-Meier estimates showed that HMGA2-overexpressing MPN patients had a significantly inferior thrombosis-free survival (P=0.049, Figure 2). The cumulative incidences of thrombosis at 10 years were 31.0% and 13.9% in HMGA2(+) and

HMGA2-overexpressing MPN patients exhibit unique clinical characteristics To assess whether increased expression of HMGA2 could be detected in clinical samples, we checked the levels of HMGA2 transcripts in the granulocytes of 151 patients with MPN and 27 adult healthy controls. The cycle numbers (Ct) in all healthy individuals were greater than 40, indicating the expression of HMGA2 was rather low in this population. We used the 2−ΔΔCT method to cal512

Figure 2. Impacts of HMGA2 overexpression on clinical outcome of patients with MPN. Kaplan–Meier estimates of thrombosis-free survival (TFS) in 151 MPN patients stratified by the expression status of HMGA2 showed those with upregulated HMGA2 had significantly inferior TFS than those with lower expression (P=0.049, log-rank test).

haematologica | 2017; 102(3)

let-7a/HMGA2 axis in JAK2-mutated MPNs

HMGA2(-) patients, respectively. Coupled with the finding that HMGA2-overexpressing MPN patients were slightly older (65.7 ± 14.4 years old, vs. 59.5 ± 16.1, P=0.058), it’s not surprising that HMGA2(+) ET/PV patients were more likely to carry high-risk disease (21/26; 80.8%), as defined by the European LeukemiaNet criteria,15 than their HMGA2(-) counterparts (59/104; 56.7%; P=0.026). These data suggest that upregulation of HMGA2 confers unique characteristics in patients with MPN, including a heightened probability of harboring either one of the 3 driver mutations, an increased propensity for the ET phenotype, and a higher likelihood of major thrombotic events.

Hmga2 overexpression confers JAK2V617F-mutated cells a survival advantage To delineate if Hmga2 overexpression altered cellular

A

phenotypes, candidate cells were treated with either Hmga2 siRNA or a scramble control to assess their respective characteristics (Figure 3). Figure 3A demonstrates the efficacy of Hmga2 downregulation in both Ton.JAK2.WT and Ton.JAK2.V617F cells. Compared with a mock control or scrambled siRNA-treated control, Ton.JAK2.V617F cells treated with Hmga2 siRNA exhibited decreased viability by 50-60% (Figure 3B). The apoptosis assay revealed that the siHmga2-treated Ton.JAK2.V617F population had more cells in apoptosis (Figure 3C,D), a finding further confirmed by increased cleaved poly (ADP-ribose) polymerase (PARP) on western blotting (Figure 3E). Noticeably, the Hmga2 knockdown efficiency was also confirmed by western blot analysis (Figure 3E). The data illustrate that Hmga2 overexpression confers JAK2V617F-mutated cells with a survival advantage by rendering them more resistant to apoptotic death.

C siHmga2

siHmga2

siHmga2

B siHmga2

D

E

siHmga2

siHmga2

Figure 3. Effects of Hmga2 downregulation in JAK2-mutated cells. Stable, inducible Ton.JAK2.WT and Ton.JAK2.V61F cells were pretreated with doxycycline (1 mg/mL) for 4 days and then transfected with either scrambled (siScr) or Hmga2 siRNA (siHmga2). (A) Quantitative RT-PCR confirmed the knockdown efficiency. (t-test; **P<0.001; ***P<0.0001). (B) Viability assay was assessed at 48 hours posttreatment with mock, siScr or siHmga2 in Ton.JAK2.WT and Ton.JAK2.V61F cells. The error bars show the standard deviation of 3 independent experiments run in triplicate. Asterisks indicate t-test statistical significance (*P<0.01; ***P<0.0001). (C) Cell apoptotic rates were detected by flow cytometric analysis using PE conjugated-Annexin V and 7-AAD double staining. Cells positively stained with Annexin V only were defined as being in early apoptosis, whereas those with positive stains for both Annexin V and 7-AAD were considered as being late apoptotic cells. (D) Histogram of apoptotic percentages (*P<0.01; **P<0.001; ***P<0.0001, statistical significance by t-test). The data were collected from three individual experiments. (E) To detect the discrepancy of apoptotic rates between siHmga2- and siScr-treated cells, western blotting was performed to assess the amount of full and cleavage form of poly (ADP-ribose) polymerase (PARP). Arrowhead indicates the full-length of PARP protein (116kDa) and the asterisk represents the cleavage form of PARP (89kDa). WT: wild-type.

haematologica | 2017; 102(3)

513

C.-C. Chen et al.

HMGA2 overexpression is associated with let-7a downregulation instead of chromosomal rearrangement around the 12q15 region To clarify the mechanism that leads to abnormal HMGA2 expression, we employed interphase FISH to detect potential cytogenetic abnormalities around this region, but found no abnormal rearrangement in all the cases evaluated (data not shown). We next assessed whether let-7a microRNA might play a role in the upregulation of HMGA2 in MPN patients. As illustrated in Figure 4A, there was a modest inverse correlation between the expression levels of let-7a and HMGA2 transcripts in the granulocytes of our patients (Pearson’s correlation coefficient r=-0.291, P=0.0002). In vitro experiments disclosed that the gradual decline of the let-7a expression could be observed in Ton.JAK2.V617F cells upon the addition of doxycycline in a time-dependent manner (Figure 4B), as opposed to the progressive increment of Hmga2 expression seen in similar assays (Figure 1D). To further validate whether the expression of HMGA2 could be regulated by let-7a, we first treated parental Ba/F3 cells with two different concentrations of a let-7a inhibitor. The result demonstrated that upon let-7a suppression, the expression of

B

E

let-7a

A

HMGA2 was increased in a dose-dependent manner at the protein (Figure 4C, upper panel) and mRNA (Figure 4C, lower panel) levels. Ectopic expression of let-7a in Ton.JAK2.V617F cells resulted in the downregulation of Hmga2 mRNA (Figure 4D, right panel) and protein (Figure 4D, left panel), whereas cells treated with a let-7a inhibitor had increased HMGA2 expression. There was an inverse correlation between the expression levels of Hmga2 and let-7a in those treated Ton.JAK2.V617F cells (r=-0.9281, P=0.002; Pearson’s correlation). Overexpression of let-7a in Ton.JAK2.V617F cells also led to increased apoptosis, as shown by increased cleaved PARP and decreased phosphorylation of the pro-apoptotic protein Bcl-2-associated death promoter (BAD) (Figure 4D, left panel). Conversely, treatment with the let-7a inhibitor endued the Ton.JAK2.V617F cells with more resistance to apoptosis, as the cleaved PARP was reduced but the phosphorylated BAD was increased (Figure 4D). Importantly, the let-7a inhibitor significantly increased the cellular viability in both parental Ba/F3 cells and inducible Ton.JAK2.V617F cells, whereas overexpressed let-7a suppressed these JAK2-mutated cells’ viability (Figure 4E). It was noteworthy that the levels of p-JAK2 were not significantly affect-

let-7a

C

let-7a

D Hmga2

let-7a

let-7a

let-7a

let-7a

let-7a

Hmga2 let-7a

NC = Scramble Control Inc: = let-7a inhibitor Mic: = let-7a mimic Figure 4. Correlation between the expression levels of let-7a and HMGA2 transcripts in MPN cells. (A) Scatter plot illustrating the correlation between the expression levels of let-7a and HMGA2 transcripts in the granulocytes of 151 patients with MPN (hollow circles; correlation coefficient r=-0.291, Pearson’s correlation, P=0.0002). Note that data from healthy individuals (solid gray circles) mostly fell on the y-axis, and many of them were very close to the origin. Data are shown in ΔΔCT. (B) Quantitative RT-PCR was performed to detect the levels of let-7a miRNA transcripts in parental Ba/F3, Ton.JAK2.WT and Ton.JAK2.V61F cells after doxycycline treatment (1 mg/ml) for indicated periods of time. (C) Effects on a let-7a inhibitory oligo on the expression of let-7a and Hmga2 in parental Ba/F3 cells. Two oligo concentrations (0.2 and 0.5 nM) were used for let-7a inhibition. All cells were harvested after 48 hours of transfection. Upper panel: western blot analysis of HMGA2; Lower panel: qRT-PCR analysis of let-7a and Hmga2 transcript levels. (t-test; *P<0.01; **P<0.001). (D) Effects of let-7a inhibitory and mimic oligos on Ton.JAK2.V617F cells. The working concentration was 0.5 nM for both oligos. Right panel: qRT-PCR analysis of let-7a and Hmga2 transcript levels. Note the inverse correlation between the expression levels of let-7a and Hmga2 (correlation coefficient r=-0.9281, Pearson’s correlation, P=0.002). Left panel: western blot analysis on the expression levels of JAK2, p-JAK2, PARP, cleaved PARP, BAD (Bcl-2-associated death promoter protein), p-BAD, and HMGA2. NC: scramble control; “-“: treatment with a let-7a inhibitor; “+”: treatment with a let-7a mimic. (E) The viability of parental Ba/F3 and inducible Ton.JAK2.V617F cells treated with indicated let-7a oligos. The error bars show the standard deviation of three independent experiments. Asterisks indicate statistical significance (t-test; *P<0.01; **P<0.001; ***P<0.0001).

514

haematologica | 2017; 102(3)

let-7a/HMGA2 axis in JAK2-mutated MPNs

ed by either ectopically expressed let-7a or let-7a inhibition (Figure 4D), indicating that alteration in the let-7a-HMGA2 axis would not lead to a change in JAK-STAT signaling activity. These results imply that in MPN, HMGA2 overexpression is mediated through let-7a downregulation, whereas chromosomal rearrangement around the 12q15 region might not play a potential role in this aspect. The data also further complement our earlier experiments showing that HMGA2 overexpression confers JAK2-mutated cells with a survival advantage (Figure 3).

Previous reports have demonstrated that LIN28A expression hinders the maturation of let-7.23 Therefore, we also explored the potential roles of LIN28A in the JAK2V617F-mediated upregulation of Hmga2. We found that the expression levels of LIN28A in Ton.JAK2.V617F

A

C

B

D

cells were also increased (Online Supplementary Figure S2).

Nevertheless, when clinical samples were tested, we could not identify any association between the expression levels of LIN28A and let-7a (Online Supplementary Figure S2C).

Both hydroxyurea and pan-JAK inhibitor INC424 ameliorate JAK-STAT activation and alleviate HMGA2 upregulation Hydroxyurea (HU) is considered an integral part in the treatment of patients with MPN.15 To appraise the potential effects of HU on JAK-STAT signaling activity and HMGA2 expressional status, we first identified its estimated IC50 concentration (5.5mg/ml) in Ton.JAK2.V617F cells in a growth inhibition assay (Figure 5A).

Subsequently, the cells were treated with hydroxyurea at

E let-7a

Hmga2

Figure 5. Influence of hydroxyurea and pan-JAK inhibitor INC424 on the phenotypes of MPN cells. Stable, inducible Ton.JAK2.V617F and Ton.JAK2.WT cells were treated with 1 mg/mL doxycycline for 6 days and then exposed to various concentrations of hydroxyurea or INC424 for 72 h. The drug concentrations that inhibited cell growth by 50% (IC50) and 90% (IC90) were determined by the XTT assay. (A) The viability of Ton.JAK2.V617F cells plotted against various concentrations of hydroxyurea. (B) Western blot analysis on various proteins of the Ton.JAK2.V617F cells treated with or without hydroxyurea at the concentration of 5.55 mg/ml (the IC50 concentration for HU in these cells) for 3 days. (C) The viability of Ton.JAK2.WT (left panel) and Ton.JAK2.V617F (right panel) cells plotted against various concentrations of INC424. (D) Western blot analysis on various proteins of the Ton.JAK2.WT and Ton.JAK2.V617F cells treated with or without INC424 at the concentration of either IC50 or IC90 for 3 days. (E) Effects of various concentrations of INC424 on the expression of let-7a and Hmga2 in Ton.JAK2.WT and Ton.JAK2.V617F cells.

Cells were harvested for qRT-PCR analysis at 3 days after treatment with the drug at indicated concentrations. The error bars show the standard deviation of three independent experiments. Asterisks indicate statistical significance (t-test; *P<0.01; **P<0.001; ***P<0.0001). WT: wild-type; HU: hydroxyurea; Conc: concentration.

haematologica | 2017; 102(3)

515

C.-C. Chen et al.

the concentration for 3 days and then subjected to western blot analysis. As shown in Figure 5B, hydroxyurea abolished JAK2 phosphorylation almost completely, which in turn ameliorated downstream STAT3/STAT5 activity and significantly suppressed HMGA2 expression. To be more specific, we also selected INC424 (ruxolitinib, a pan-JAK inhibitor used in the treatment of MPN24,25) to assess the influence of JAK-STAT activity on the expression of HMGA2. Upon obtaining the IC50 and IC90 concentrations of INC424 in Ton.JAK2.WT and Ton.JAK2.V617F cells (Figure 5C), we treated the cells with the indicated drug concentrations for 3 days. Both sets of cells showed decreased JAK/STAT activity, abolished HMGA2 expression, and enhanced apoptosis in a dose-dependent manner (Figure 5D). The decreased Hmga2 expression, as well as upregulated let-7a in INC424-treated cells, was also confirmed by qRT-PCR (Figure 5E). The results provide complementary evidence supporting a non-redundant role of JAK/STAT activation in let-7a inhibition and HMGA2 upregulation in MPN cells.

Discussion In the study, we used Ba/F3 cells with conditional expression of JAK2V617F or WT JAK2 as a working model to recapitulate the phenotypic comparison between JAK2-mutated and -unmutated MPN cells. We demonstrate that JAK2-mutated cells exhibit upregulation of HMGA2, and the expression of HMGA2 could be affected by the level of let-7a miRNA. HMGA2 overexpression confers JAK2-mutated cells with a survival advantage through inhibited apoptosis, and MPN patients harboring upregulated HMGA2 show an increased propensity for ET phenotype as well as a higher likelihood of developing thrombosis.

The finding, in the study herein, that HMGA2-overexpressing MPN patients are more likely to belong to the ET subtype and have higher platelet counts is actually supported by several in vivo studies. Oguro et al. disclosed that overexpression of Hmga2 in hematopoietic stem cells induced a myeloproliferative state with enhanced megakaryopoiesis in mice,26 whereas Yang et al. similarly demonstrated that Hmga2 significantly increased megakaryocytic colonies in the bone marrow of JAK2V617F mice.27 Interestingly, Ikeda et al. showed that overexpression of Hmga2 could lead to an increased level of Jak2 transcripts and a rise in STAT3 phosphorylation.13 As a result, the proliferation of hematopoietic stem cells was expanded. In the study herein, we further illustrated that HMGA2 knockdown in JAK2-mutated cells resulted in growth inhibition and a significant increase in apoptosis, a finding consistently seen in other cancer types.28-30 Coupled with the fact that Hmga2-overexpressing BM cells have a growth advantage over control cells in mice competitive repopulation and serial BM transplantation models,13 it is conceivable that the upregulated HMGA2 “turns on” a state of proliferative hematopoiesis as well as inhibited apoptosis, and the predilection for megakaryocytic expansion26,27 may account for a higher platelet count and ET phenotype in HMGA2-overexpressing patients.

However, the stronger association between HMGA2 overexpression and ET subtype in the work herein contrasted with previous reports showing a higher prevalence 516

of upregulated HMGA2 in patients with PMF.31,32 There are several possibilities that lead to such a discrepancy. Most likely, HMGA2-overexpressing ET and PMF might represent a continuum of disease entities at different stages, in which the overexpressed HMGA2 collaborates with coexisting driver mutations or other pathognomonic mutations to cultivate the phenotypic expression. The argument can be supported by the fact that overexpression of Hmga2 leads to both increased megakaryopoiesis and accelerated progression of myelofibrosis in animal models.26,27 Secondly, patients from various studies have different ethnical and environmental backgrounds, which may contribute to the discrepancy in the HMGA2-overexpression rates in these reports, including ours. Lastly, the composition of the enrolled patients and the sample sizes across studies are different. Although currently available data suggest that overexpressed HMGA2 might play a critical role in the pathogenesis of PMF,12,13,26,27,31-34 in some studies the reproducibility of clinical data is hampered by either the low number of PMF patients evaluated or insufficient inclusion of PV and ET patients.31-33 The Italian study12 and ours enrolled the largest cohorts of patients, with the PMF population being the largest representative subgroup [n=88 (out of 158); 55.7%] in the former study.

Supplemented by other smaller series, investigators from that study convincingly showed a high frequency of HMGA2 overexpression in patients with PMF.12 It is very likely that we had too few PMF patients (17 cases only) to make our estimation justifiable. On the other hand, our study enrolled by far the largest number of ET (n=67) and PV (n=63) patients for the assessment of HMGA2 expression. We revealed that about one-fourth of ET patients harbored upregulated HMGA2, a finding mirrored by Harada-Shirado et al.34 Although PMF patients comprised the major subgroup that showed upregulated HMGA2 in the Italian study (as compared to ET and PV patients as well as normal controls), there was still a specific portion of ET patients who harbored overexpressed HMGA2.12 Supported by the findings from these two studies,12,34 our work provides further proof of evidence endorsing the notion that HMGA2 overexpression not only plays an essential role in the pathogenesis of PMF, but also exerts specific effects on the phenotypic presentation of ET. Although case reports suggested that chromosomal translocation involving the 12q14-15 region led to overexpression of HMGA2 in patients with MPN,35-38 our data nonetheless disclosed an essential role of let-7a miRNA regulation on the expression of HMGA2 in patients with MPN. These findings are in line with a recent report demonstrating that decreased let-7 miRNA expression, instead of generating the loss of 3’-UTR of the HMGA2 gene, is the major cause of dysregulated HMGA2 mRNA expression in MPN.34 The aberrancy in microRNA expression that leads to the genetic complexity of MPN is also supported by some other studies, as the reduction of a wide variety of miRNAs, including let-7a,31 miR-149,32 miR150,32 let-7f,39 and let-7g39 has been reported, and the former three are also associated with upregulated HMGA2 mRNA in these reports.31,32 The correlation between aberrant let-7a-HMGA2 activity and JAK2V617F mutation has been partially dissected by several investigators. Through transcriptome comparative microarray analysis, Guglielmelli et al. revealed that abnormal HMGA2 expression in the granulocytes of patients with PMF was dependent on the presence of haematologica | 2017; 102(3)

let-7a/HMGA2 axis in JAK2-mutated MPNs

JAK2V617F mutation.12 Potential alteration in the expression of microRNAs, however, was not assessed in this study. Bruchova et al. performed gene expression profiling and found that JAK2V617F frequency was inversely correlated with let-7a expression in PV granulocytes.31 Nevertheless, upregulated HMGA2 was detected in PMF (but not in PV) patients, and no correlation between the expression levels of let-7a and HMGA2 were identified.31 With only 35 patients included, the study was probably hampered by small sample size. Although an inverse correlation between the expression levels of let-7a and HMGA2 was found, Harada-Shirado et al. could not identify any potential relationship between HMGA2 expressional status and JAK2V617F mutation.34 Our study, enrolling significantly more patients, clearly demonstrates that HMGA2-overexpressing MPN patients are more likely to carry either one of the 3 driver mutations (Table 1).

Coupled with our in vitro findings showing an apparent correlation between JAK-STAT activation and the let-7aHMGA2 axis, we comprehensively elaborated the dependence of aberrant let-7a-HMGA2 axis activity on the presence of JAK2V617F mutation. A major flaw of our study might lie in the fact that we were unable to identify how mutant JAK2 regulates the let-7a-HMGA2 axis activity. Further exploratory work would be important to delineate the missing link between them.

In conclusion, we have demonstrated that enhanced Hmga2 expression could be seen in Ba/F3 cells with

References 1. Nangalia J, Green TR. The evolving genomic landscape of myeloproliferative neoplasms. Hematology Am Soc Hematol Educ Program. 2014;2014(1):287-296.

2. James C, Mazurier F, Dupont S, et al. The hematopoietic stem cell compartment of JAK2V617F-positive myeloproliferative disorders is a reflection of disease heterogeneity. Blood. 2008;112(6):2429-2438.

3. Stein BL, Williams DM, Rogers O, Isaacs MA, Spivak JL, Moliterno AR. Disease burden at the progenitor level is a feature of primary myelofibrosis: a multivariable analysis of 164 JAK2 V617F-positive myeloproliferative neoplasm patients. Exp Hematol. 2011;39(1):95-101.

4. Vainchenker W, Delhommeau F, Constantinescu SN, Bernard OA. New mutations and pathogenesis of myeloproliferative neoplasms. Blood. 2011; 118(7): 1723-1735.

5. Tefferi A. Novel mutations and their functional and clinical relevance in myeloproliferative neoplasms: JAK2, MPL, TET2, ASXL1, CBL, IDH and IKZF1. Leukemia.

2010;24(6):1128-1138.

6. Shih AH, Abdel-Wahab O, Patel JP, Levine RL. The role of mutations in epigenetic regulators in myeloid malignancies. Nature Rev Cancer. 2012;12(9):599-612.

7. Sgarra R, Rustighi A, Tessari MA, et al.

Nuclear phosphoproteins HMGA and their relationship with chromatin structure and cancer. FEBS letters. 2004;574(1-3):1-8.

8. Fedele M, Battista S, Kenyon L, et al.

Overexpression of the HMGA2 gene in transgenic mice leads to the onset of pituitary adenomas. Oncogene. 2002;21(20): 3190-3198.

haematologica | 2017; 102(3)

enforced JAK2V617F expression. HMGA2-overexpressing patients exhibit a trend of a higher likelihood of carrying one of the 3 MPN-relevant driver mutations. In vitro data confirm that downregulation of let-7 miRNA plays an essential role in the dysregulated expression of Hmga2; a result supplemented by the inverse correlation between the expression levels of let-7 and HMGA2 in clinical samples. Strikingly, expression of HMGA2 confers JAK2-mutated cells with a survival advantage and endues MPN patients with a unique clinical phenotype. Our findings suggest that, in a subset of MPN patients, the let-7-HMGA2 axis plays a prominent role in the pathogenesis of the disease that leads to unique clinical phenotypes.

Acknowledgments The authors thank Professor Gregor Hoermann and Professor Matthias Mayerhofer (both of the Medical University of Vienna, Austria) for kindly providing Ba/F3 cells with inducible expression of JAK2V617F (Ton.JAK2.V617F) or wild-type (WT) JAK2 (Ton.JAK2.WT). We also thank Miss Hsing-Yi Tsou and Miss I-Shan Chen for their assistance with data collection.

Funding The study was supported by the Medical Ministry of Science and Technology (Taiwan, R.O.C) grant to CC Chen (MOST 103-2314-B-182-051-MY3) and Chang Gung Memorial Hospital grants to CC Chen (CMRPG6B0223, CORPG6B0373 and CORPG6F0031).

9. Zhou X, Benson KF, Przybysz K, et al.

Genomic structure and expression of the murine Hmgi-c gene. Nucleic Acids Res. 1996;24(20):4071-4077.

10. Fusco A, Fedele M. Roles of HMGA proteins in cancer. Nature Rev Cancer. 2007;7 (12):899-910.

11. Young AR, Narita M. Oncogenic HMGA2: short or small? Genes Dev. 2007;21 (9):1005-1009.

12. Guglielmelli P, Zini R, Bogani C, et al.

Molecular profiling of CD34+ cells in idiopathic myelofibrosis identifies a set of disease-associated genes and reveals the clinical significance of Wilms' tumor gene 1 (WT1). Stem Cells. 2007;25(1):165-173.

13. Ikeda K, Mason PJ, Bessler M. 3'UTR-truncated Hmga2 cDNA causes MPN-like hematopoiesis by conferring a clonal growth advantage at the level of HSC in mice. Blood. 2011;117(22):5860-5869.

14. Swerdlow SH CE, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues.

Lyon, France: International Agency for Research on Cancer, 2008.

15. Barbui T, Barosi G, Birgegard G, et al.

Philadelphia-negative classical myeloproliferative neoplasms: critical concepts and management recommendations from European LeukemiaNet. J Clin Oncol.

2011;29(6):761-770.

16. Rotunno G, Mannarelli C, Guglielmelli P, et al. Impact of calreticulin mutations on clinical and hematological phenotype and outcome in essential thrombocythemia. Blood.

2014;123(10):1552-1555.

17. Vannucchi AM, Antonioli E, Guglielmelli P, et al. Clinical profile of homozygous JAK2 617V>F mutation in patients with polycythemia vera or essential thrombo-

cythemia. Blood. 2007;110(3):840-846.

18. Chen CC, Gau JP, Chou HJ, et al.

Frequencies, clinical characteristics, and outcome of somatic CALR mutations in JAK2unmutated essential thrombocythemia. Ann Hematol. 2014;93(12):2029-2036.

19. Chen CC, You JY, Gau JP, et al. Favorable clinical outcome and unique characteristics in association with Twist1 overexpression in de novo acute myeloid leukemia. Blood Cancer J. 2015;5:e339.

20. Kralovics R, Passamonti F, Buser AS, et al. A gain-of-function mutation of JAK2 in myeloproliferative disorders. N Engl J Med. 2005;352(17):1779-1790.

21. Chachoua I, Pecquet C, El-Khoury M, et al.

Thrombopoietin receptor activation by myeloproliferative neoplasm associated calreticulin mutants. Blood. 2016; 127(10):1325-1335.

22. Marty C, Pecquet C, Nivarthi H, et al.

Calreticulin mutants in mice induce an MPL-dependent thrombocytosis with frequent progression to myelofibrosis. Blood.

2016;127(10):1317-1324.

23. Viswanathan SR, Daley GQ, Gregory RI. Selective blockade of microRNA processing

by Lin28. Science. 2008;320(5872):97-100.

24. Vannucchi AM, Kiladjian JJ, Griesshammer M, et al. Ruxolitinib versus standard therapy for the treatment of polycythemia vera.

N Engl J Med. 2015;372(5):426-435.

25. Verstovsek S, Mesa RA, Gotlib J, et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med. 2012;366(9):799-807.

26. Oguro H, Yuan J, Tanaka S, et al. Lethal myelofibrosis induced by Bmi1-deficient hematopoietic cells unveils a tumor suppressor function of the polycomb group genes. J Exp Med. 2012;209(3):445-454.

517

C.-C. Chen et al.

27. Yang Y, Akada H, Nath D, Hutchison RE, Mohi G. Loss of Ezh2 cooperates with Jak2V617F in the development of myelofibrosis in a mouse model of myeloproliferative neoplasm. Blood. 2016;127(26):34103423.

28. Pentimalli F, Dentice M, Fedele M, et al.

Suppression of HMGA2 protein synthesis could be a tool for the therapy of well differentiated liposarcomas overexpressing HMGA2. Cancer Res. 2003;63(21):74237427.

29. Malek A, Bakhidze E, Noske A, et al.

HMGA2 gene is a promising target for ovarian cancer silencing therapy. Int J Cancer. 2008;123(2):348-356.

30. Cai J, Shen G, Liu S, Meng Q. Downregulation of HMGA2 inhibits cellular proliferation and invasion, improves cellular apoptosis in prostate cancer. Tumour Biol. 2016;37(1):699-707.

31. Bruchova H, Merkerova M, Prchal JT.

518

32.

33.

34.

35.

Aberrant expression of microRNA in polycythemia vera. Haematologica. 2008; 93(7):1009-1016.

Guglielmelli P, Tozzi L, Pancrazzi A, et al.

MicroRNA expression profile in granulocytes from primary myelofibrosis patients.

Exp Hematol. 2007;35(11):1708-1718.

Andrieux J, Bilhou-Nabera C, Lippert E, et al. Expression of HMGA2 in PB leukocytes and purified CD34+ cells from controls and patients with Myelofibrosis and myeloid metaplasia. Leuk Lymphoma. 2006; 47(9):1956-1959.

Harada-Shirado K, Ikeda K, Ogawa K, et al.

Dysregulation of the MIRLET7/HMGA2 axis with methylation of the CDKN2A promoter in myeloproliferative neoplasms. Br J Haematol. 2015;168(3):338-349.

Martin SE, Sausen M, Joseph A, Kingham BF, Martin ES. Identification of a HMGA2EFCAB6 gene rearrangement following next-generation sequencing in a patient with

36.

37.

38.

39.

a t(12;22)(q14.3;q13.2) and JAK2V617F-positive myeloproliferative neoplasm. Cancer Genet. 2012;205(6):295-303.

Aliano S, Cirmena G, Garuti A, et al.

HMGA2 overexpression in polycythemia vera with t(12;21)(q14;q22). Cancer Genet Cytogenet. 2007;177(2):115-119.

Andrieux J, Demory JL, Dupriez B, et al.

Dysregulation and overexpression of HMGA2 in myelofibrosis with myeloid metaplasia. Genes Chromosomes Cancer.

2004;39(1):82-87.

Storlazzi CT, Albano F, Locunsolo C, et al. t(3;12)(q26;q14) in polycythemia vera is associated with upregulation of the HMGA2 gene. Leukemia. 2006;20(12): 2190-2192.

Zhan H, Cardozo C, Yu W, et al.

MicroRNA deregulation in polycythemia vera and essential thrombocythemia patients. Blood Cells Mol Dis. 2013; 50(3):190-195.

haematologica | 2017; 102(3)

ARTICLE

Chronic Myeloid Leukemia

Phase 1 dose-finding study of rebastinib (DCC-2036) in patients with relapsed chronic myeloid leukemia and acute myeloid leukemia

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Jorge Cortes,1 Moshe Talpaz,2 Hedy P. Smith,3 David S. Snyder,4 Jean Khoury,5 Kapil N. Bhalla,1 Javier Pinilla-Ibarz,6 Richard Larson,7 David Mitchell,8 Scott C. Wise,9 Thomas J. Rutkoski,9 Bryan D. Smith,9 Daniel L. Flynn,9 Hagop M. Kantarjian,1 Oliver Rosen9 and Richard A. Van Etten10

University of Texas, MD Anderson Cancer Center, Houston, TX; 2University of Michigan, Ann Arbor, MI; 3Tufts Medical Center, Boston, MA; 4City of Hope Cancer Center, Duarte, CA; 5Winship Cancer Institute of Emory University, Atlanta, GA; 6H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL; 7University of Chicago, IL; 8Mitchell Pharmaceutical Consulting, Lafayette, CO; 9Deciphera Pharmaceuticals LLC, Waltham, MA and 10UC Irvine Chao Family Comprehensive Cancer Center, Orange, CA, USA 1

Haematologica 2017 Volume 102(3):519-528

ABSTRACT

A

vailable tyrosine kinase inhibitors for chronic myeloid leukemia bind in an adenosine 5'-triphosphate-binding pocket and are affected by evolving mutations that confer resistance. Rebastinib was identified as a switch control inhibitor of BCR-ABL1 and FLT3 and may be active against resistant mutations. A Phase 1, first-in-human, single-agent study investigated rebastinib in relapsed or refractory chronic or acute myeloid leukemia. The primary objectives were to investigate the safety of rebastinib and establish the maximum tolerated dose and recommended Phase 2 dose. Fifty-seven patients received treatment with rebastinib. Sixteen patients were treated using powder-in-capsule preparations at doses from 57 mg to 1200 mg daily, and 41 received tablet preparations at doses of 100 mg to 400 mg daily. Dose-limiting toxicities were dysarthria, muscle weakness, and peripheral neuropathy. The maximum tolerated dose was 150 mg tablets administered twice daily.

Rebastinib was rapidly absorbed. Bioavailability was 3- to 4-fold greater with formulated tablets compared to unformulated capsules. Eight complete hematologic responses were achieved in 40 evaluable chronic myeloid leukemia patients, 4 of which had a T315I mutation. None of the 5 patients with acute myeloid leukemia responded.

Pharmacodynamic analysis showed inhibition of phosphorylation of substrates of BCR-ABL1 or FLT3 by rebastinib. Although clinical activity was observed, clinical benefit was insufficient to justify continued development in chronic or acute myeloid leukemia. Pharmacodynamic analyses suggest that other kinases inhibited by rebastinib, such as TIE2, may be more relevant targets for the clinical development of rebastinib (clinicaltrials.gov Identifier:00827138).

Correspondence: jcortes@mdanderson.org

Received: July 15, 2016.

Accepted: November 29, 2016.

Pre-published: December 7, 2016. doi:10.3324/haematol.2016.152710 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/519 Š2017 Ferrata Storti Foundation

Introduction Genetic changes that alter tyrosine kinase function are found in multiple malignancies, including chronic myeloid leukemia (CML) and acute myeloid leukemia (AML). In CML, the BCR-ABL1 translocation is the hallmark of the disease, leading to constitutive kinase signaling.1 Tyrosine kinase inhibitors (TKIs) have altered the natural history of CML.2,3 In AML, FLT3 internal tandem duplication (ITD) also leads to constitutive activation of FLT3, and likewise, FLT3 inhibitors have shown clinical efficacy in the treatment of FLT3-mutated AML.4,5-8 Although TKIs are very effective treatments for CML, resistance and disease progression eventually occur in some patients. Second generation ABL1 TKIs (dasahaematologica | 2017; 102(3)

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.

519

J. Cortes et al.

tinib, nilotinib, and bosutinib) are not effective against the T315I BCR-ABL1 mutation, which is the most prevalent mutation found in patients progressing on these agents.9,10 The recent approval of ponatinib has added a treatment effective against the T315I mutation to the therapeutic armamentarium.11 However, resistance to ponatinib has been observed in some patients, making the development of new T315I-targeted inhibitors a necessity.

All TKIs currently approved by the United States Food and Drug Administration inhibit their target kinases by competing with adenosine 5'-triphosphate (ATP) for binding. Rebastinib (DCC 2036) is a “switch control” inhibitor of several tyrosine kinases, including ABL1, FLT3, and TIE2.12 Based on crystallographic analyses, rebastinib binds as a type II inhibitor that additionally penetrates into and binds the switch control pocket of the ABL1 kinase domain, key amino acid residues that mediate the change from inactive to active conformation. Through occupancy of this switch control pocket, rebastinib blocks the conformational process required for kinase activation. This unique binding mode enables rebastinib to maintain the ABL1 kinase domain in the inactive conformation, independent of the state of phosphorylation of the regulatory Tyr 393. Additionally, for certain TKI resistance mutations that also dysregulate and activate c-ABL1 (e.g., T315I and Y253F), rebastinib binding forces these mutant variants of BCR-ABL1 to adopt inactive conformations, durably blocking downstream cellular signaling pathways.12-16 Rebastinib inhibits the BCR-ABL1 T315I mutant kinase in an ATP noncompetitive manner, even at the high millimolar concentrations of ATP found intracellularly. In vivo, rebastinib is effective in a mouse model of CML-like myeloproliferative neoplasia induced by BCR-ABL1 T315I.12 Potential advantages of switch pocket inhibitors compared to ATP-competitive inhibitors include longer off-rates compared to ATP-binding site inhibitors due to their binding into the deeply embedded switch pocket, and the fact that switch pockets are more varied in sequence from kinase to kinase than are ATP-binding pockets. The latter could potentially lead to inhibition of fewer off-target kinases.

This first-in-human study with rebastinib was performed in order to determine the maximum tolerated dose (MTD) when administered daily, and to evaluate preliminary efficacy in patients with relapsed CML or AML.

Methods

for patients with CML and ≤1 for those with AML was required.

After the MTD was established, only patients with chronic phase chronic myeloid leukemia (CP CML) and accelerated phase chronic myeloid leukemia (AP CML) were enrolled into a CML expansion portion. After treating 13 patients in the CML expansion portion, enrollment was further limited to patients with CP-CML with the T315I mutation. Detailed inclusion and exclusion criteria are available in the Online Supplementary Information.

Rebastinib was administered orally daily in continuous 28 day cycles. Dosing started with powder-in-capsules (PIC) at 57 mg once daily (QD), escalating in 11 consecutive cohorts up to 1200 mg QD. Doses between the planned initial single-patient cohorts were increased by 100%. Subsequent cohorts enrolled at least 3 patients each in an accelerated 3+3 dose-escalation design, whereupon occurrence of grade ≥2 adverse events (AEs) accelerated titration stopped and a standard 3+3 design came into effect. During the trial, formulated tablets were introduced to seek more predictable pharmacokinetics. These were administered to patients in 3 cohorts from 100 mg to 200 mg. Patients receiving PIC were transferred to tablets after the safety of tablets had been confirmed. Intra-patient dose-escalation was permitted to a higher dose that had been evaluated in other patients for at least 28 days.

Safety and efficacy assessments Safety was assessed using the National Cancer Institiute Common Toxicity Criteria (NCI CTC), Version 3.0.17 During the course of the study, serial ophthalmologic examinations, echocardiograms, and measurements of N-terminal pro-hormone brain natriuretic peptide (NT-proBNP) were added. Disease response was evaluated locally every 3 cycles using standard criteria.18,19

Pharmacokinetic and pharmacodynamic analyses Plasma samples were collected on Days 1, 8, 15, and 22 of Cycle 1 and analyzed by liquid chromatography–mass spectrometry (LC MS)/mass spectrometry (MS) after single and multiple doses of either PIC or tablets. Pharmacokinetic samples were collected at predose, and up to 24 hr postdose on Days 1 and 8, and single predose samples were collected on Days 15 and 22. Pharmacokinetic parameters were derived using non-compartmental methods and based on actual blood sampling and dosing times for each individual. Whole blood samples for pharmacodynamic evaluation were collected at baseline and Days 1 and 8 of Cycle 1. In whole blood assays, mononuclear cells were fixed and permeabilized after incubation. pCRKL was analyzed by flow cytometry. For assays using isolated white blood cells, phospho-protein levels were determined by immunoblot. Target inhibition and rebastinib plasma concentrations were used for determination of pharmacokinetic/pharmacodynamic relationships.

Trial design This first-in-human, multi-center, single arm study evaluated the safety, tolerability, pharmacokinetics (PK), efficacy and pharmacodynamic (PD) effects of rebastinib in patients with CML or AML. The study was conducted in accordance with the Declaration of Helsinki and approved by the institutional review boards at each participating institution.

Patients

Patients ≥18 years old were eligible if they had a diagnosis of CML in chronic, accelerated or blast phase resistant to ≥2 TKIs or with T315I mutation, or had resistant or refractory AML with FLT3 ITD and had not received prior FLT3 inhibitors, or were ≥60 years old and not candidates for standard chemotherapy. An Eastern Cooperative Oncology Group performance status of ≤2 520

Results Patient population Sixty-nine patients were enrolled between March 2009 and February 2012, and 57 patients received rebastinib (CML:52 and AML:5). At data cutoff on 17 January 2013, all patients were off study (Online Supplementary Figure S1).

The median age was 58.0 years (range: 25 to 80) and 59.6% of patients were male. All of the 52 patients with CML had received at least 1 prior TKI, 92% received 2 prior TKIs, and 62% received 3 prior TKIs. More than half (52%) of the CML patients had previously received imatinib, dasatinib, and nilotinib. Twenty (40%) of the CML patients harbored a T315I mutation (Table 1). haematologica | 2017; 102(3)

First-in-human study of rebastinib in relapsed CML

Table 1. Baseline Patient Characteristics.

Characteristic

CML N=52

Median Age, years (range) 56.5 (25-80) Sex, Male, n (%) 31 (59.6) Median BSA, m2 (range) 1.9 (1.5-2.6)a Disease Phase: CP / AP / BP, n (%) 42 / 8 / 2 (81/15/4) No. of Prior TKI, n (%) ≥ 1 TKI 52 (100) ≥ 2 TKI 48 (92.3) ≥ 3 TKI 32 (61.5) Imatinib, dasatinib & nilotinib 27 (51.9) Known resistance or intolerance to ≥2 TKI, n (%) 43 (83) Mutation analysis at screening N=50 (%) No mutation 20 (40) 1 mutation 27 (54) 2 mutations 3 (6) >2 mutations 0 T315I 20 (40) Other 10 (20) M244V/E459K 1 (2) L248R/F359I 1(2) E255K/V299L 1 (2) L298V 1 (2) V299L 1 (2) F317I 1 (2) F317L 1 (2) F359V 1 (2) H396R 1 (2) E450K 1 (2)

AML N=5

All Patients N=57

64.0 (23-73) 3 (60.0) 1.7 (1.4-2.3)b

58.0 (23-80) 34 (59.6) 1.9 (1.4-2.6)c NA

NA NA

AML: acute myeloid leukemia; AP: accelerated phase; BP: blast phase; BSA: body surface area; CML: chronic myeloid leukemia; CP: chronic phase; NA: not applicable; TKI: tyrosine kinase inhibitor. an = 51; bn = 4; cn = 55.

Maximum tolerated dose and safety Fifty-seven patients received rebastinib at a total of 15 dosing regimens using the two formulations, PIC and tablet, and were analyzed in 11 dose groups (Table 2). A group represented different formulations with similar amounts of active ingredients. A total of 306 completed cycles were administered to 57 patients. Intra-patient dose-escalation was permitted and 20 patients received one or more dose levels in addition to their initial level.

Dose-limiting toxicities (DLTs) were defined as the occurrence of one of the following AEs, which were considered to be at least possibly related to rebastinib during the first treatment cycle of the dose-escalation portion: 1) Grade 4 hematologic toxicity sustained for ≥4 weeks (≥6 weeks in patients with prior bone marrow transplantation), 2) ≥Grade 3 non-hematologic toxicity (except for nausea, vomiting or diarrhea that could be managed to ≤Grade 2 severity), and 3) any rebastinib-attributed AE that resulted in a treatment delay ≥28 days. Twenty seven patients were treated in the dose-escalation portion. Three patients experienced a DLT in the first cycle of treatment (Table 3). After exceeding the MTD at 200 mg tablets twice daily (BID), an additional 30 patients were enrolled in the expansion portion at 150 mg tablets BID. In the dose-escalation portion, three patients experienced a DLT in the first cycle of treatment at doses of 1200 mg QD (PIC, n=1) or 200 mg BID (tablet, n=2) (Table 3).

One patient in the 1200 mg QD dose level (PIC) experihaematologica | 2017; 102(3)

enced 2 DLTs (Grade 3 dysarthria and muscular weakness). This dose level was not expanded to 6 patients because of a planned change to tablet formulation. Two patients experienced a DLT in the first cycle of the 200 mg BID dose level with tablets; one experienced Grade 3 muscle weakness of their upper and lower extremities and the other experienced Grade 3 peripheral neuropathy.

Consequently, the tablet dose of 150 mg BID was determined to be the MTD, and this dose cohort was expanded with an additional 27 patients to further characterize the safety profile. Of these 27 patients, 18 experienced a Grade 3 adverse event of special interest (AESI) considered possibly or probably related.

Analyses of all 57 (100%) patients showed that all experienced at least one AE of any grade during the entire treatment. Fifteen patients (26%) reported an AE with a maximum intensity of Grade 1 or 2, 34 patients (60%) reported a Grade 3 or 4 AE and 8 patients experienced an AE that resulted in death. Forty-nine patients (86%) reported AEs that were considered study drug related; related Grade 1 or 2 AEs by 23 patients (40%), related Grade 3 AEs by 25 patients (44%), and 1 patient experienced an AE that resulted in death, and the relationship to treatment is unknown; no related Grade 4 AE was reported (Table 3).

Adverse events of any grade that were reported with an incidence of ≥30% were: dry mouth (27 patients, 47%), constipation (25 patients, 44%), fatigue (22 patients, 39%), muscular weakness (21 patients, 37%), headache 521

J. Cortes et al.

Table 2. Exposure to Rebastinib by Dose Level.

Assigned Dose Level (mg/dose)a 57 QD 114/150 QD 225/228/300 QD 450/456 QD 600 QD 300 BID 1200 QD 100 QD (T)d 100 BID (T)d 150 BID (T)d 200 BID (T)d Any Dose

Number of Subjects Treated at Protocol Dose Level at Cycle Starta As Initial Due to a Protocol Total Dose Dose Changeb 1 2 3 4 3 0 3 0 4 33 4 57

0 2 4 2 4 4 0 4 12 1 0 20

1 4 7 6 7 4 3 4 16 34 4 57

Number of Cycles

Number of Cycles Completedc

3 5 36 20 28 15 22 36 85 117 10 377

3 3 31 18 26 12 20 36 71 80 6 306

BID: twice daily; QD: once daily; T: tablet. aAssigned dose level refers to protocol dose levels taken at the start of a cycle. Changes occurring within a cycle were not counted.

However, for cycles that started with 0 mg, the previous dose taken was summarized if it was different from the start of the previous cycle. bProtocol dose changes included dose level changes resulting from 1) dose-escalation, 2) dose-reduction, 3) dose level assignment in subjects transitioning from QD to BID, and 4) dose level assignment in subjects transitioning from capsule to tablet. A subject could have recieved multiple different dose levels resulting from more than one protocol dose change reason. cA "completed cycle" is defined as consuming study drug on ≥80% of days within a 28-day cycle. dDoses received in tablet form.

(20 patients, 35%), and nausea and blurred vision (19 patients each, 33%) (Table 4). An overview of AEs occurring by dose level is shown in the Online Supplementary Table S1.

A total of 42 patients (74%) experienced one or more AE of ≥Grade 3 intensity: 30 patients (53%) experienced Grade 3 events, 4 patients (7%) experienced AEs of Grade 4, and 8 patients (14%) experienced AEs of Grade 5. The most common Grade 3 and 4 AEs (incidence >5%) were: muscular weakness (13 patients, 23%), hypertension (4 patients, 7.0%), and dyspnea, fatigue, myalgia, and blurred vision (3 patients each, 5%) (Table 4). Twenty six (46%) of the ≥Grade 3 AEs were considered related (Grade 3/4/5: 25 patients [44%]/0 patients/1 patient [2%], respectively). The most common Grade 3 and 4 related AEs (incidence >5%) were muscular weakness (12 patients, 21%), and myalgia and blurred vision (3 patients each, 5%). A total of 34 patients (60%) experienced serious adverse events (SAEs) (as reported by the investigator) during the study. The most common SAEs included muscular weakness (4 patients, 7%), disease progression, and pneumonia (3 patients each, 5%). In 18 patients (32%), these SAEs were considered related to study treatment. The most common study drug related SAE was muscular weakness (4 patients, 7%). Related SAEs occurred at dose levels of 300 mg BID and 1200 mg QD with PIC; and 100 mg, 150 mg, and 200 mg BID with tablets (Online Supplementary Table S2). No treatment-related SAE was reported for patients treated at dose levels below 300 mg BID (PIC) and 100 mg QD (tablet).

Events suggestive of musculoskeletal disorders (e.g., muscular weakness and myopathy), nervous system disorders (e.g., peripheral neuropathy, paresthesia, and periph522

Table 3. Summary of Adverse Events.

Adverse Event Category

Rebastinib (N=57)

Number (%) of Subjects with AEs by Maximum Intensity Grade All AEs 0=No Change from Normal or Reference Range 0% 1=Mild 7 (12.3%) 2=Moderate 8 (14.0%) 3=Severe 30 (52.6%) 4=Life Threatening or Disabling 4 (7.0%) 5=Death Related to the AE 8 (14.0%) Treatment-related AEs 0=No Change from Normal or Reference Range 0% 1=Mild 11 (19.3%) 2=Moderate 12 (21.1%) 3=Severe 25 (43.9%) 4=Life Threatening or Disabling 0% 5=Death Related to the AE 1 (1.8%)a Number (%) of Subjects with AEs by Causality All AEs Definitely Not 2 (3.5%) Unlikely 6 (10.5%) Possibly 22 (38.6%) Probably 16 (28.1%) Definitely 10 (17.5%) Unknown 1 (1.8%) Number (%) of Subjects With AE classified as DLT All AEs 3 (5.3%) Treatment-related AEs 3 (5.3%) AE: adverse event; DLT: dose-limiting toxicity. aOne patient experienced an AE that resulted in death and the relationship to treatment is unknown (Online Supplementary Table S5).

haematologica | 2017; 102(3)

First-in-human study of rebastinib in relapsed CML

eral sensory neuropathy), as well as visual abnormalities (e.g., blurred vision) were identified as AESIs (Online Supplementary Table S3 and Table S4). Musculoskeletal disorders of all grades was reported in 22 patients who experienced 34 events. In 14 of the 22 patients, musculoskeletal

disorders occurred during Cycle 1. Among the 22 patients, 13 experienced Grade 3 events, and no Grade 4 event occurred. Recovery was documented in 11 of the patients within several days of halting rebastinib administration, and 2 patients discontinued study treatment due to mus-

Figure 1. Mean plasma rebastinib concentration-time profile following multiple-dose administration of rebastinib tablets to steady state. Peripheral blood samples were obtained at the indicated times following administration of an oral dose of rebastinib (tablet formulation at doses of 100 mg BID, 150 mg BID, and 200 mg BID) in patients who had received at least 1 week of continuous dosing. Plasma concentrations of rebastinib were determined by LC-MS/MS as described in Methods. IC80 values for inhibition of BCR-ABL1 phosphorylation of CRKL in ex vivo assays using whole blood from CML patients with BCR-ABL1 or BCR-ABL1 T315I are indicated. BID: twice daily; Conc: concentration; pCRKL: phosphorylated CT10 Regulator of Kinase-like.

Table 4. Frequency of patients with adverse events of any grade and grade 3 or 4 adverse events observed in â&#x2030;Ľ15% of patients by causality.

Preferred Term

Dry mouth Constipation Fatigue Muscular weakness Headache Nausea Blurred vision Diarrhea Dizziness Dysgeusia Vomiting Dyspnoea Paresthesia Arthralgia Hypertension Pain in extremity Abdominal pain Cough Myalgia Chest pain Decreased appetite Hypoesthesia

Number of Patients with AEs N=57

Number of Patients with AEs definitely, probably, or possibly related to study drug N=57

Number of Patients with AEs Grade 3 or 4 N=57

Number of Patients with AEs Grade 3 or 4 definitely, probably, or possibly related to study drug N=57

27 (47.4%) 25 (43.9%) 22 (38.6%) 21 (36.8%) 20 (35.1%) 19 (33.3%) 19 (33.3%) 17 (29.8%) 17 (29.8%) 15 (26.3%) 15 (26.3%) 12 (21.1%) 12 (21.1%) 11 (19.3%) 11 (19.3%) 11 (19.3%) 10 (17.5%) 10 (17.5%) 10 (17.5%) 9 (15.8%) 9 (15.8%) 9 (15.8%)

22 (38.6%) 12 (21.1%) 14 (24.6%) 18 (31.6%) 7 (12.3%) 7 (12.3%) 12 (21.1%) 7 (12.3%) 5 (8.8%) 12 (21.1%) 3 (5.3%) 3 (5.3%) 10 (17.5%) 2 (3.5%) 3 (5.3%) 2 (3.5%) 4 (7.0%) 2 (3.5%) 6 (10.5%) 2 (3.5%) 4 (7.0%) 7 (12.3%)

0 (0.0%) 2 (3.5%) 3 (5.3%) 13 (22.8%) 2 (3.5%) 1 (1.8%) 3 (5.3%) 1 (1.8%) 0 (0.0%) 0 (0.0%) 1 (1.8%) 3 (5.3%) 0 (0.0%) 2 (3.5%) 4 (7.0%) 2 (3.5%) 0 (0.0%) 0 (0.0%) 3 (5.3%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

0 (0.0%) 1 (1.8%) 2 (3.5%) 12 (21.1%) 1 (1.8%) 1 (1.8%) 3 (5.3%) 1 (1.8%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 1 (1.8%) 0 (0.0%) 2 (3.5%) 2 (3.5%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 3 (5.3%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

AE: adverse event.

haematologica | 2017; 102(3)

523

J. Cortes et al.

A

B

Figure 2. Pharmacodynamic analyses of pCRKL and pSTAT5 inhibition pre- and post-rebastinib dose. Serial peripheral blood samples were obtained from patients following the first dose of rebastinib (Cycle 1 Day 1) and on Day 8 of continuous twice daily dosing (Cycle 1 Day 8). Mononuclear cells were isolated, lysed, and the indicated proteins and phosphoproteins analyzed by immunoblotting as described in Methods. (A) Patient with relapsed chronic phase CML BCR-ABL1 V299L mutation who demonstrated >75% inhibition of pCRKL 4h following the Day 8 morning dose. (B) Patient with relapsed chronic phase CML and BCR-ABL1 T315I mutation who demonstrated >90% inhibition of pSTAT5 at 1h following the Day 8 morning dose. In this patient, the degree of inhibition of pCRKL was less pronounced (~25%).

PK: pharmacokinetics; pCRKL: phosphorylated CT10 regulator of kinase-like; CRKL: CT10 regulator of kinase-like; eIF4E: eukaryotic translation initiation factor 4E; pSTAT5: phosphorylated signal transducer and activator of transcription 5; STAT5: signal transducer and activator of transcription 5.

culoskeletal disorders. In the majority of cases, these events were reported during the first cycle of therapy.

Nineteen patients (33%) experienced AEs of any grade collectively suggestive of nervous system disorders. Two patients with Grade 3 nervous system disorders continued on decreased rebastinib dosages, and the nervous system disorder resolved. Twenty-four AEs of visual abnormalities were reported in 19 patients (33%). Four out of 5 patients with 2 reported AEs recovered from at least 1 event. Thirteen out of 14 patients with Grade 1 as the highest severity, 2/2 with Grade 2, and 3/3 with Grade 3 reported recovery from visual abnormalities. Two patients with Grade 3 visual abnormalities discontinued from the study due to the AE. In 10 of the 19 patients, visual abnormalities were first reported during Cycle 1 (Grade 1/2/3: 6/1/3 patients, respectively). Treatment interruption was required in 3 patients; 1 patient with Grade 1, Grade 2, and Grade 3 visual abnormalities each and recovery from the event was documented (data not shown).

In addition, 4 CML patients (7%) experienced AEs suggestive of a potential â&#x2030;ĽGrade 3 cardiomyopathy (congestive cardiac failure in 2 patients, viral cardiomyopathy in 1 patient, and cardiomyopathy in 1 patient). The patient who was diagnosed with viral cardiomyopathy on Day 1 had a medical history of allogeneic bone marrow transplant. He was immediately withdrawn from the study due to this event. Grade 3 congestive cardiac failure in another patient occurred on Day 4 and led to death on Day 7.

Since 1998 this patient had received, among other therapies, interferon, imatinib, nilotinib, and dasatinib. The patient had a history of previous pleural effusion on dasatinib, hypertension, coronary artery disease, chronic dyspnea, and tachycardia. The third patient presented on Day 7 with Grade 3 cardiomyopathy; his symptoms included Grade 3 elevated cardiac enzymes and Grade 2 atrial fibrillation. He had a history of hypertension and received imatinib, nilotinib, and dasatinib as prior CML treatment.

This patient recovered after cardiac conversion to a normal sinus rhythm. The fourth patient presented with 524

Grade 3 bronchitis on Day 211, followed by Grade 3 congestive cardiac failure on Day 213 and died due to cardiac failure 2 days later. This patients prior therapies for CML included imatinib and dasatinib. His medical history included chronic obstructive pulmonary disease, hypertension, and bicuspid aortic stenosis with valve replacement. A summary of all deaths of study patients (n=8) and the attributed causality is provided in the Online Supplementary Table S5; none of the deaths were deemed related to rebastinib.

Pharmacokinetics Rebastinib absorption was rapid with both tablet and PIC. Mean peak plasma concentrations occurred within a median time to maximum concentration (Tmax) of 1 to 1.5 hours using tablets (Figure 1) and 2 hours using PIC (Online Supplementary Figure S2). Exposure, including mean maximum concentration (Cmax), area under the curve from 0 to 4 hours (AUC(0-4h)), area under the curve from 0 to 10 hours (AUC(0-10h)), and trough values, increased with dose (Table 5), but in a less than dose-proportional manner. The elimination of rebastinib appeared to be biphasic, with an initial distribution phase from Tmax of 6-8 hours postdose, followed by a terminal elimination with a half-life (T1/2) of 1215 hr (data not shown). Terminal elimination rates appeared to be similar across dose levels. Exposure was approximately 3-fold greater for tablets compared to PIC (Table 5).

Safety Exposure Analyses Safety exposure analyses investigated the relationship between rebastinib exposure, Cmax and AUC(0-4h), and the occurrence of AESIs in the medical categories of muscle weakness, peripheral neuropathy, cardiomyopathy and visual abnormalities (see Online Supplementary Table S3 for AE terms included in each system organ class). The analysis for Cmax included a total of 50 patients, and 47 patients for AUC(0-4h), who had pharmacokinetics data available for Cycle 1 (Day 8 or Day 15). The values for either Cmax or AUC(0-4h) of evaluable patients were divided into low, haematologica | 2017; 102(3)

First-in-human study of rebastinib in relapsed CML Table 5. Summary of pharmacokinetic parameters following multiple doses of oral rebastinib as powder in capsule or tablet taken once or twice daily.

Dose Group & Regimen

N

Cmax ng/mLa (%CV)

Tmax hrb (range)

Trough ng/mLa

AUC(0-4h) hr*ng/mLa

AUC(0-10h) hr*ng/mLa

57 mg QD 114 mg QD 225/228 mg QD

1 1 6

200 178 223

549 348 183c

1 4

12.5 14.1

528 355

NA NA

450/456 mg QD

6

9.21

450

791c

600 mg QD

5

17

566

805c

1200 mg QD

3

13.1

580

966

100 mg BID (T)

9

25.4d

292

554e

150 mg BID (T)

28

40.2

678f

786g

200 mg BID (T)

4

4.0 1.0 2.0 (1.0-2.0) 4.0 1.5 (1.0-4.0) 1.0 (0.5-2.0) 2.0 (1.0-4.0) 2.0 (1.0-2.0) 2.0 (1.0-4.0) 2.0 (0.5-4.3) 1.0 (0.5-6.0)

9.63 7.7 10.3

300 mg QD 300 mg BID

74.7 67 133 (133) 229 181 (70) 186 (100) 218 (60.1) 284 (40.8) 114 (61.9) 308 (71.9) 377 (73.4)

72.0

752

1180

AUC(0-4h): area under the curve from 0 to 4 hours; AUC(0-10h): area under the curve from 0 to 10 hours; BID: twice daily; Cmax: maximum concentration; %CV: % coefficient of variability; N: number of patients; NA: not available; QD: once daily; T: tablet; Tmax: time to maximum concentration; ageometric mean; bmedian; cN=3; dN=8; eN=4; fN=25; gN=5.

middle, and high tertiles. Rebastinib doses were identified that patients received at the time of the onset of a particular AESI, and patients were then analyzed based on the highest toxicity grade for a particular AE category within each Cmax or AUC(0-4h) value.

This analysis showed a trend towards the occurrence of more AESIs or AESIs of higher severity with increasing Cmax and AUC(0-4h) tertiles (Online Supplementary Table S6).

For example, in the AESI analyses by Cmax, 23, 24, and 24 patients were categorized in the low/middle/high Cmax subgroup. A total of 7 patients (30%), 5 patients (22%), and 13 patients (54%) categorized in the low/middle/high Cmax subgroup experienced nervous system disorders.

Whereas Grade 1 and 2 AESIs occurred in all Cmax subgroups, in the medium and high Cmax subgroups, one Grade 3 nervous system disorder each was also observed.

Clinical efficacy Response assessment was performed locally. All treated CML and AML patients were evaluable for response.

Eighteen of the 52 CML patients (35%) responded. Fifty patients were evaluable for hematologic response. The best hematologic responses were 8 patients with complete hematologic response (T315I: 4 patients), 4 patients with partial hematologic response (T315I: 3), and 3 patients with minor hematologic response (T315I: 0), including 15 patients with chronic phase (CP) CML, and 3 patients with acute phase (AP) CML. Of the 37 patients evaluable for cytogenetic response, 8 patients showed a response (2 with complete cytogenetic response [T315I: 1], 2 with partial cytogenetic response [T315I: 1], and 4 with minor cytogenetic response [T315I: 3]). Four of the 40 evaluable haematologica | 2017; 102(3)

patients achieved a molecular response (all had major molecular response; T315I: 2) (Table 6). Two of these occurred between Cycles 17-19 and were maintained for the duration of the study, while the other two were not durable (maintained for 3-6 cycles). All AML patients were evaluable for hematological and morphological responses.

None of the 5 patients with AML showed a response to rebastinib.

Pharmacokinetic/Pharmacodynamic evaluation Screening (predose) whole blood samples were collected from 31 patients. To determine if rebastinib could inhibit BCR-ABL1 or FLT3 in samples from TKI-resistant patients, blood samples were incubated with rebastinib ex vivo and the phosphorylation of CRKL or FLT3 was measured. Of the 28 samples that were evaluable, in all cases the IC50 for inhibition of phosphorylation was less than 200 nM, including in samples from patients with â&#x2030;Ľ1 mutation in BCR-ABL1, such as T315I. To determine if BCR-ABL1 or FLT3 was inhibited in patients dosed with rebastinib, predose and postdose blood samples from Cycle 1 Day 1 and Cycle 1 Day 8 were analyzed from 48 patients. Of the 45 patients with evaluable samples, significant (>50%) inhibition of phosphorylation of CRKL and/or STAT5 was observed for 20 patients (Figure 2). Clinical response did not appear to be correlated closely to inhibition of phosphorylation of CRKL and/or STAT5 as measured in this assay, as only 8 of the 20 patients with significant inhibition of CRKL or STAT5 had a clinical response.

Conversely, 9 patients who had a clinical response exhibited <50% inhibition of phosphorylation of CRKL and STAT5 at the time points tested. Samples were not collected for 1 patient who had a clinical response.

525

J. Cortes et al.

Table 6. Chronic myeloid leukemia subgroup: individual responses.

Pt â&#x20AC;&#x201C;ID

CML

Dose cohort

T315I mutation (Y/N)

Best hematologic response

Best cytogenetic response

Best molecular response

02-001 03-001 02-002 03-002 02-003 03-004 03-006 01-009 02-005 02-007 02-009 03-011 05-006 05-007 07-004 07-005 07-006 01-011

CML-CP CML-CP CML-CP CML-CP CML-AP CML-CP CML-AP CML-CP CML-AP CML-CP CML-CP CML-CP CML-CP CML-CP CML-CP CML-CP CML-CP CML-CP

114/150 mg 225/228 mg 450/456 mg 450/456 mg 600 mg 600 mg 1200 mg 100 mg BID 100 mg BID 150 mg BID 150 mg BID 150 mg BID 150 mg BID 150 mg BID 150 mg BID 150 mg BID 150 mg BID 200 mg BID

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

MiHR CHR PHR CHR PHR CHR CHR CHR CHR MiHR MiHR No response CHR NE CHR PHR PHR No response

No response No response No response NE No response MiCyR MiCyR CCyR NE No response No response MiCyR CCyR MiCyR No response PCyR NE PCyR

No response No response No response MMR MMR No response MMR No response No response No response No response No response MMR No response No response No response NE No response

AP: accelerated phase; BID: twice daily; CCyR: complete cytogenetic response (0% Ph+ metaphases out of a minimum of 20 metaphases examined); CHR: complete hematologic response; CML: chronic myeloid leukemia; CP: chronic phase; MiCyR: minor cytogenetic response (36%-90% Ph+); MiHR: minor hematologic response; MMR: major molecular response; N: no; NE: not evaluable; PCyR: partial cytogenetic response(1%-35% Ph+); PHR: partial hematologic response; Pt-ID: patient identification; QD: once daily; Y: yes.

In addition to BCR-ABL1 and FLT3, rebastinib is a potent TIE2 inhibitor (unpublished data). In selected patients, elevated plasma levels of the TIE2 ligand angiopoietin 2 (ANG2) were observed on Day 22 compared to baseline levels. ANG2 increases were observed in 19/20 of the Day 22 patient samples, with no change observed in 1/20 samples (data not shown). Both Cmax and trough exposures of rebastinib on Day 8 significantly correlated with increased levels of plasma ANG2. Increased plasma ANG2 levels also correlated with trough rebastinib exposures on Day 22 (data not shown).

Discussion This Phase 1 study of the switch-control inhibitor, rebastinib, was conducted in patients with relapsed CML and AML based on its potency against ABL1 and FLT3.

In vitro studies had established that rebastinib is an effective inhibitor of BCR-ABL1 and other kinases.12 The MTD of continuously administered rebastinib was determined to be 150 mg BID, with dysarthria, muscular weakness, and peripheral neuropathy representing the DLTs at a dose of 200 mg BID. Expansion of the 150 mg BID dose cohort consolidated the safety findings of the dose-escalation.

The most common AEs included muscular weakness, peripheral neuropathy, and visual abnormalities. In addition, 4 patients experienced severe cardiac AEs during the study.

Visual abnormalities have been recently described with other TKIs.20,21 Such events observed with rebastinib were mostly of mild severity, and many of the events resolved with no change to rebastinib dosing. Since there were no 526

significant ocular findings in preclinical animal studies, detailed baseline ophthalmologic examinations were not performed at baseline in this study, and a potential contribution of pre-existing conditions or the underlying disease could not be ruled out. Baseline and serial ophthalmologic examinations will be included in forthcoming studies in order to gain insight into the relatedness of the blurred vision seen with rebastinib in CML and AML patients.

The mechanism by which these events are mediated will also need to be investigated.

A number of TKIs have been associated with a low incidence of cardiomyopathy occurring after months of treatment.22 The rebastinib study herein documented 3 cases of cardiomyopathy occurring during the first 2 weeks of treatment. Since these patients had alternative potential causes of cardiomyopathy, and there was no screening of cardiac function at baseline in this study, the relationship seen between cardiomyopathy events and rebastinib exposure is not clear. However, preclinical studies have suggested that ABL inhibition may lead to myocardial toxicity.23 Future rebastinib studies will include baseline and serial echocardiography studies to carefully investigate cardiac function. The pharmacokinetics parameters of rebastinib indicated rapid absorption of the drug, and an increase, though less than dose-proportional, in exposure with dosage. The T1/2 was 12-15 hours; however, due to the biphasic elimination of rebastinib resulting in trough concentrations of generally 10-20% of Cmax, some accumulation should be expected with a BID dosing regimen.

Average plasma levels of 200 ng/mL (360 nM) were achieved transiently from ~1-4 hours postdose at the MTD of 150 mg BID. At this exposure level, the amount of free drug in circulation is expected to be ~1.3 nM (based haematologica | 2017; 102(3)

First-in-human study of rebastinib in relapsed CML

on determined human plasma protein binding of 99.65%).

With regards to potential off-target effects of rebastinib, the only kinases that would be inhibited by ≥50% at concentrations of <1.3 nM include c-ABL1 (IC50 0.7 nM), TIE2 (IC50 0.058 nM), and TRKA and TRKB (IC50 0.17 nM and 0.42 nM, respectively).12 Further monitoring of AESIs will help to understand the off-target effects or collateral "target" effects of rebastinib.

A safety exposure analysis that investigated the relationship of AESIs of cardiac, eye, musculoskeletal, and peripheral neuropathy/nervous systems disorders to rebastinib exposure measured as Cmax or AUC(0-4h) suggested that AESIs of musculoskeletal and peripheral neuropathy/nervous systems disorders may be related to rebastinib exposure. Regarding the severity of cardiac and ocular disorders, an increase in these disorders with an increase in rebastinib exposure cannot be ruled out. The understanding of safety exposure outcomes for cardiac and ocular disorders will improve from longitudinal monitoring in future studies.

Pharmacodynamic data show that the rebastinib plasma concentrations achieved in patients inhibited BCR-ABL1 in patient samples, including mutant alleles of BCR-ABL1, like T315I. Patients with CML who entered this trial were resistant or intolerant to > 1 of other BCR-ABL1 TKIs. In approximately half of the cases, no mutations were present in BCR-ABL1, possibly indicating activation of other pathways for cell survival. Inhibition of BCR-ABL1 in these patients, as monitored by pharmacodynamic analysis, may not necessarily correlate with clinical response. In addition, postdose levels of rebastinib in certain patients may not have inhibited phosphorylation of targets to the level needed to irreversibly commit leukemic progenitors to apoptosis, and thereby achieve a response.24 While rebastinib exposures at 150 mg BID on average were above or near the level required for 80% inhibition of BCR-ABL1 for the entire dosing period, exposures were only above the IC80 for inhibition of T315I mutant BCRABL1 for ~3 hours after each dose (Figure 2). Based on experience with other BCR-ABL1 inhibitors, a higher dose, e.g., 175 mg BID, if it would have been tested and tolerated, would probably not have led to an improved efficacy outcome in the case of BCR-ABL1 T315I resistance mutations.

When this study was designed, the only drug with activity against the BCR-ABL1 T315I mutation was omacetaxine mepesuccinate, which yields cytogenetic responses in ~20% of CP-CML patients.25 Although the overall cytoge-

References 1. Lugo TG, Pendergast AM, Muller AJ, Witte ON. Tyrosine kinase activity and transformation potency of bcr-abl oncogene products. Science. 1990;247(4946):1079 1082.

2. Kantarjian HM, Talpaz M, O’Brien S, et al.

Imatinib mesylate for Philadelphia Chromosome-positive, chronic-phase myeloid leukemia after failure of interferon: follow-up results. Clin Cancer Res. 2002;8(7):2177-2187.

3. Kantarjian H, Giles F, Wunderle L, et al.

Nilotinib in imatinib-resistant CML and

haematologica | 2017; 102(3)

netic response rate of CML patients to rebastinib (21%) was comparable to that of omacetaxine, the response rate was considerably lower than that which has subsequently been reported with ponatinib in a comparable patient population.11 Thus, continued development of rebastinib in CML may not be justified. Pharmacodynamic analysis suggested the inhibition of other rebastinib-targeted kinases in patients receiving rebastinib, such as TIE2 in particular. Compensatory elevations in circulating ligands upon the inhibition of their cognate receptor kinases have been frequently observed clinically. For example, the inhibition of VEGF receptors 1, 2, and 3 by sunitinib leads to increases in the plasma levels of VEGF and PlGF in patients.26 Thus, increased plasma ANG2 levels observed in patients on the trial herein provide evidence of TIE2 inhibition by rebastinib in humans. Future clinical development of rebastinib will focus more on diseases involving TIE2 or TIE2-expressing macrophages and the tumor microenvironment. The ANG/TIE2 kinase signaling pathway is a pivotal signaling axis in the tumor microenvironment, mediating tumor angiogenesis, tumor cell intravasation, extravasation, therapy resistance, and immunosuppression.27-30 TIE2 is expressed in endothelial cells and a subset of highly protumoral macrophages, referred to as TIE2-expressing macrophages (TEMs), both of which contribute to tumor angiogenesis.31 It has been demonstrated that TEMs expand an immunosuppressive population of Treg cells in the tumor microenvironment.29,32 In addition, it was reported that TEMs mediate tumor cell vascular intravasation and extravasation, and ultimately the formation of metastases.33-35 Data from in vitro assays indicate that the testing of a daily dose lower than the MTD defined in the study herein is warranted for TIE2 inhibition in future studies. This hypothesis is being tested in an ongoing clinical trial (clinicaltrials.gov Identifier:02824575).

Acknowledgments The conduct of the study at MD Anderson Cancer Center was supported in part by MD Anderson Cancer Center Support Grant P30 CA016672 (PI: Dr. Ronald DePinho) and P01 CA049639 (PI: Dr. Richard Champlin). Studies at Tufts Medical Center were supported in part by R01 CA090576 (PI: Dr. Richard Van Etten). The authors thank Molly Hood and Cynthia Leary for performing ex vivo analyses on patient-derived plasma samples, Linda Martin of Deciphera Pharmaceuticals LLC for her assistance in the compilation and verification of key data to support this manuscript, and thank Korinna Pilz for medical writing support.

Philadelphia chromosome-positive ALL. N Engl J Med. 2006;354(24):2542 2551.

4. Nakao M, Yokota S, Iwai T, et al. Internal tandem duplication of the flt3 gene found in acute myeloid leukemia. Leukemia.

1996;10(12):1911-1918.

5. Schiller GJ, Tallman MS, Goldberg SL, et al.

Final results of a randomized phase 2 study showing the clinical benefit of quizartinib (AC220) in patients with FLT3-ITD positive relapsed or refractory acute myeloid leukemia. J Clin Oncol. 2014;32:5s, (suppl;abstr 7100).

6. Wander SA, Levis MJ, Fathi AT. The evolving role of FLT3 inhibitors in acute myeloid

leukemia: quizartinib and beyond. Ther Adv Hematol. 2014;5(3):65–77.

7. Röllig, C, Müller-Tidow C, Hüttmann A, et al. Addition of sorafenib versus placebo to standard therapy in patients aged 60 years or younger with newly diagnosed acute myeloid leukemia (SORAML): a multicentre, phase 2, randomised controlled trial.

Lancet Oncol. 2015;16(16):1691-1699.

8. Stone RM, Mandrekar S, Sanford BL, et al.

The multi-kinase inhibitor midostaurin (M) prolongs survival compared with placebo (P) in combination with daunorubicin (D)/cytarabine(C) induction (Ind), highdose C consolidation (consol), and as main-

527

J. Cortes et al.

9.

10.

11.

12.

13.

14.

15.

16.

528

tenance (maint) therapy in newly diagnosed acute myeloid leukemia (AML) patients (pts) age 18-60 with Flt3 mutations (muts): An international prospective randomized (rand) p-controlled double-blind trial (CALGB10603/RATIFY [Alliance].

Blood 2015;126(23):abstract 6.

Gorre ME, Mohammed M, Ellwood K, Hsu N, Paquette R, Rao PN, Sawyers CL. Clinical resistance to STI-571 cancer therapy caused by BCR-ABL gene mutation or amplification. Science. 2001;293(5531):876880.

Shah NP, Nicoll JM, Nagar B, et al. Multiple BCR-ABL kinase domain mutations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib (STI571) in chronic phase and blast crisis chronic myeloid leukemia.

Cancer Cell. 2002;2(2):117 125.

Cortes JE, Kim DW, Pinilla-Ibarz J, et al. A Phase 2 trial of ponatinib in Philadelphia chromosomeâ&#x20AC;&#x201C;positive leukemias. N Engl J Med. 2013;369(19):1783-1796.

Chan WW, Wise SC, Kaufman MD, et al.

Conformational control inhibition of the BCR-ABL1 tyrosine kinase, including the gatekeeper T315I mutant, by the switch control inhibitor DCC-2036. Cancer Cell.

2011;19(4):556-568.

Azam M, Seeliger MA, Gray NS, Kuriyan J, Daley GQ. Activation of tyrosine kinases by mutation of the gatekeeper threonine. Nat Struct Mol Biol. 2008;15(10):1109-1118.

Roumiantsev S, Shah NP, Gorre ME, et al.

Clinical resistance to the kinase inhibitor STI-571 in chronic myeloid leukemia by mutation of Tyr-253 in the Abl kinase domain P-loop. Proc Natl Acad Sci USA. 2002;99(16):10700-10705.

Eide C.A, Adrian LT, Tyner JW, et al. The ABL switch control inhibitor DCC-2036 is active against the chronic myeloid leukemia mutant BCR-ABLT315I and exhibits a narrow resistance profile. Cancer Research. 2011;71(9):3189-3195.

Redaelli S, Mologni L, Rostagno R, et al.

Three novel patient-derived BCR/ABL mutants show different sensitivity to second and third generation tyrosine kinase

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

inhibitors. Am J Hematol. 2012;87(11): E125-128.

National Cancer Institute Cancer Therapy Evaluation Program.

Common Terminology Criteria for Adverse Events v3.0.

Available from: http://ctep.cancer.gov/reporting/ctc.html. th Last accessed: 8 October 2008.

NCCN Practice Guidelines in Oncology.

Available from: http://www.nccn.org/. Last accessed: 8th October 2008.

Cheson BD, Bennett JM, Kopecky KJ, et al.

Revised recommendations of the International Working Group for diagnosis, standardization of response criteria, treatment outcomes, and reporting standards for therapeutic trials in acute myeloid leukemia. J Clin Oncol. 2003;21(24):46424649.

Renouf DJ, Velazquez-Martin JP, Simpson R, Siu LL, Bedard PL. Ocular toxicity of targeted therapies. J Clin Oncol. 2012; 30(26):3277-3286.

Stjepanovic N, Velazquez-Martin JP, Bedard PL. Ocular toxicities of MEK inhibitors and other targeted therapies. Ann Oncol. 2016; 27(6):998-1005.

Mellor HR, Bell AR, Valentin JP, Roberts RR. Cardiotoxicity associated with targeting kinase pathways in cancer. Toxicol Sci. 2011;120(1):14-32.

Kerkala R, Grazette L, Yacobi R, et al.

Cardiotoxicity of the cancer therapeutic agent imatinib mesylate. Nat Med. 2006; 12(8):908-916.

Shah NP, Kasap C, Weier C, et al. Transient potent BCR-ABL inhibition is sufficient to commit chronic myeloid leukemia cells irreversibly to apoptosis. Cancer Cell. 2008; 14(6):485-493.

Cortes JE, Kantarjian HM, Rea D, et al.

Final analysis of the efficacy and safety of omacetaxine mepesuccinate in patients with chronic- or accelerated-phase chronic myeloid leukemia: Results with 24 months of follow-up. Cancer. 2015; 15;121(10): 1637-1644.

DePrimo SE, Bello CL, Smeraglia J, et al.

Circulating protein biomarkers of pharma-

27.

28.

29.

30.

31.

32.

33.

34.

35.

codynamic activity of sunitinib in patients with metastatic renal cell carcinoma: modulation of VEGF and VEGF-related proteins.

J Transl Med. 2007;5:32.

Lewis CE, Ferrara N. Multiple effects of angiopoietin-2 blockade on tumors.

Cancer Cell. 2011;19(4):431-433.

Hanahan D, Coussens LM. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell.

2012 Mar 20;21(3):309-322.

Coffelt SB, Chen YY, Muthana M, et al.

Angiopoietin 2 stimulates TIE2-expressing monocytes to suppress T cell activation and to promote regulatory T cell expansion. J Immunol. 2011;186(7):4183-4190.

Hughes R, Qian B-Z, Rowan C, et al.

Perivascular M2 macrophages stimulate tumor relapse after chemotherapy. Cancer Res. 2015;75(17):3479-3491.

Mazzieri R, Pucci F, Moi D, et al. Targeting the ANG2/TIE2 axis inhibits tumor growth and metastasis by impairing angiogenesis and disabling rebounds of proangiogenic myeloid cells. Cancer Cell. 2011;19(4):512526.

Ibberson M, Bron S, Guex N, et al. TIE2 and VEGFR kinase activities drive immunosuppressive function of TIE2-expressing monocytes in human breast cancer. Clin Cancer Res. 2013;19(13):3439-3449.

Roussos ET, Goswami S, Balsamo M, et al.

Mena invasive (Mena(INV)) and Mena11a isoforms play distinct roles in breast cancer cell cohesion and association with TMEM. Clin Exp Metastasis. 2011;28(6):515-527.

Dovas A, Gligorijevic B, Chen X, Entenberg D, Condeelis J, Cox D. Visualization of actin polymerization in invasive structures of macrophages and carcinoma cells using photoconvertible beta-actin-Dendra2 fusion proteins. PLoS One. 2011;6(2): e16485.

Harney AS, Arwert EN, Entenberg D, Wang Y, Guo P, Qian B-Z, et al. Real-time imaging reveals local, transient vascular permeability, and tumor cell intravasation stimulated

by TIE2hi macrophageâ&#x20AC;&#x201C;derived VEGFA. Cancer Discov. 2015;5(9):932 943.

haematologica | 2017; 102(3)

ARTICLE

Acute Myeloid Leukemia

CEBPA–double-mutated acute myeloid leukemia displays a unique phenotypic profile: a reliable screening method and insight into biological features

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Francesco Mannelli,1,2 Vanessa Ponziani,1,2 Sara Bencini,1,2 Maria Ida Bonetti,1,2 Matteo Benelli,3* Ilaria Cutini,1,2 Giacomo Gianfaldoni,1,2 Barbara Scappini,1,2 Fabiana Pancani,1,2 Matteo Piccini,1,2 Tommaso Rondelli,4 Roberto Caporale,4 Anna Maria Grazia Gelli,4 Benedetta Peruzzi,4 Marco Chiarini,5 Erika Borlenghi,6 Orietta Spinelli,7 Damiano Giupponi,7 Pamela Zanghì,7 Renato Bassan,8 Alessandro Rambaldi,7 Giuseppe Rossi6 and Alberto Bosi1,2 Unità Funzionale di Ematologia, Università degli Studi, AOU Careggi, Firenze; 2Istituto Toscano Tumori, Firenze; 3SOD Diagnostica Genetica, AOU Careggi, Firenze; 4SOD Laboratorio Centrale, Settore Citometria Clinica, AOU Careggi, Firenze; 5Centro di Ricerca Emato-Oncologica AIL (CREA), Spedali Civili, Brescia; 6Divisione di Ematologia, Spedali Civili, Brescia; 7Unità Strutturale Complessa di Ematologia, Ospedali Riuniti, Bergamo and 8Divisione di Ematologia, Ospedale dell'Angelo & Ospedale SS. Giovanni e Paolo, Mestre-Venezia, Italy 1

* Current address: Centro di Biologia Integrativa, Università di Trento, Trento, Italy

Haematologica 2017 Volume 102(3):529-540

ABSTRACT

utations in CCAAT/enhancer binding protein α (CEBPA) occur in 5-10% of cases of acute myeloid leukemia. CEBPA-doublemutated cases usually bear bi-allelic N- and C-terminal mutations and are associated with a favorable clinical outcome. Identification of CEBPA mutants is challenging because of the variety of mutations, intrinsic characteristics of the gene and technical issues. Several screening methods (fragment-length analysis, gene expression array) have been proposed especially for large-scale clinical use; although efficient, they are limited by specific concerns. We investigated the phenotypic profile of blast and maturing bone marrow cell compartments at diagnosis in 251 cases of acute myeloid leukemia. In this cohort, 16 (6.4%) patients had two CEBPA mutations, whereas ten (4.0%) had a single mutation.

First, we highlighted that the CEBPA-double-mutated subset displays recurrent phenotypic abnormalities in all cell compartments. By mutational analysis after cell sorting, we demonstrated that this common phenotypic signature depends on CEBPA-double-mutated multi-lineage involvement. From a multi-dimensional study of phenotypic data, we developed a classifier including ten core and widely available parameters.

The selected markers on blasts (CD34, CD117, CD7, CD15, CD65), neutrophil (SSC, CD64), monocytic (CD14, CD64) and erythroid (CD117) compartments were able to cluster CEBPA-double-mutated cases. In a validation set of 259 AML cases from three independent centers, our classifier showed excellent performance with 100% specificity and 100% sensitivity. We have, therefore, established a reliable screening method, based upon multi-dimensional analysis of widely available phenotypic parameters. This method provides early results and is suitable for largescale detection of CEBPA-double-mutated status, allowing gene sequencing to be focused in selected cases.

M

haematologica | 2017; 102(3)

Correspondence: francesco.mannelli@unifi.it

Received: July 5, 2016.

Accepted: October 28, 2016.

Pre-published: November 10, 2016. doi:10.3324/haematol.2016.151910 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/529 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

529

F. Mannelli et al.

Introduction Mutations in the transcription factor CCAAT/enhancer binding protein Îą (CEBPA) are found in approximately 10% of cases of acute myeloid leukemia (AML).1-3 Most CEBPA-mutant AML exhibit two mutations, which frequently involve a combination of an N-terminal and a Cterminal gene mutation, typically on different alleles.

Recent comprehensive data have shown that CEBPA-double-mutated (CEBPA-dm) cases, rather than single mutants, are associated with a common gene expression signature4 and a relatively favorable outcome.4-7 Based on these features, CEBPA-dm AML has been recognized as a separate entity in the revised World Health Organization classification.8 The identification of a CEBPA-dm genotype provides crucial prognostic information, since these patients often lack other main predictors of relapse risk.

However, there are several technical issues with CEBPA mutational analysis. First, CEBPA sequencing is known to be difficult because of the high GC content of the gene, which frequently correlates with failure of the polymerase chain reaction, and the presence of background or sequencing artifacts. Sequencing the entire gene enables detection of all mutations but is labor-intensive, especially in a routine context, and requires expertise with unusual variants. Several screening methods have, therefore, been developed. Although efficient and sensitive, polymerase chain reaction-based fragment-length analyses can only detect mutations resulting in a net insertion or deletion and not substitution mutations.9,10 Furthermore, they cannot distinguish a common 6-bp duplication polymorphism from an actual insertion or duplication.11 Next-generation sequencing-based CEBPA studies are able to overcome these difficulties but are not widely available yet. Some reports have proposed gene expression arrays as a screening method for CEBPA-dm status, given the unique profile of these malignancies.4,6,12-14 Although these methods have excellent performance, they require further technology and relative expertise for specific application in this context.

As a screening method for genetic abnormalities, the immunophenotype of AML blasts is often able to predict the main underlying genotypes.15 Generally, the association with phenotype is strong when a few, relevant genetic events are responsible for leukemogenesis [e.g., CBFrelated translocations or t(15;17)], whereas it is weaker when genetic heterogeneity is greater (e.g., normal karyotype with several gene mutations). Furthermore, the strength of the correlation with a certain genotype depends on phenotypic aberrations being rare in AML not characterized by that genotype [e.g., cross-lineage CD19 and t(8;21)]. In fact, CEBPA-mutated AML has not been associated with a specific immunophenotype. Rather, it has been described as showing positivity for commonly expressed antigens, such as CD15, CD7, CD34 and HLADR on blasts.16 Although associated with a mutant status, this phenotypic profile was not able to screen effectively for CEBPA-mutated cases, since about 25% of them were missed.16 Furthermore, it was based on the strict application of the European Group for Immunological Characterization of Leukemia (EGIL) threshold for positivity (i.e., more than 20% of cells),17 which is probably inadequate for dissecting a shared phenotypic signature, especially for frequently expressed antigens.

In this study, we extensively investigated the 530

immunophenotype of CEBPA-mutated AML by analyzing all bone marrow cell compartments at diagnosis and by comparing each compartment with its corresponding normal counterpart in order to highlight aberrations. Our aim was to develop a screening method for CEBPA-mutated AML based on the phenotypic profile, which would be straightforward, widely available and fast, in order to focus molecular techniques on a narrow subset of AML patients.

Methods Patients Patients entering the study had a diagnosis of untreated AML, based on World Health Organization criteria.18 and an available immunophenotypic characterization on bone marrow at diagnosis. When eligible for intensive chemotherapy, patients were treated according to two protocols as specified below. Briefly, from 2006 to March 2007 (protocol 1), patients received standard course induction. High-dose cytarabine was used as first consolidation in patients aged <61 years attaining complete remission. On an intent-to-treat basis, patients aged <55 years with a high-risk karyotype, FLT3-ITD or adverse clinical features were assigned to undergo allogeneic stem cell transplantation. Patients with intermediate cytogenetic risk in the absence of FLT3-ITD and adverse clinical features were allocated to allogeneic stem cell transplantation if a related donor was available. Autologous stem cell transplantation was offered to patients aged <61 years with low-risk cytogenetics, intermediate-risk cytogenetics without sibling donor and high-risk disease not eligible for allogeneic transplantation.

From April 2007 to 2013 (protocol 2), patients were enrolled in the Northern Italy Leukemia Group (NILG) AML 02-06 protocol (Eudract code: 2006-003817-42). This protocol included randomization at induction between a standard ICE induction and an experimental, intensified one. Patients randomized to the experimental arm were excluded from the outcome analysis. A more detailed description of treatment protocols is provided in the Online Supplementary Data.

Only intensively treated non-M3 patients were considered for the outcome analysis. The study was approved by the local institutional review board (protocol number: 2013/0024340), and patients were included after giving written informed consent, in accordance with the Declaration of Helsinki.

Karyotype Cytogenetic analysis was performed on bone marrow cells taken at diagnosis and the results are reported according to the International System for Human Cytogenetic Nomenclature.19

Molecular genetics NPM1, FLT3-ITD and CEBPA mutations were searched for using previously described methods.1,21,22 Further details are reported in the Online Supplementary Data.

Flow cytometry Technical details about flow cytometry sample handling, reagents, acquisition and analysis are reported in the Online Supplementary File. Data were analyzed with Infinicyt software (Cytognos SL, Salamanca, Spain). Some major bone marrow cell compartments were identified: (i) blasts; (ii) maturing neutrophils; (iii) monocytes; and (iv) mature erythroid cells. A series of 79 phenotypic parameters were defined (24 for blasts, 30 for the neutrophils, 14 for the monocytic compartment and 11 for erythroid cells). Parameters were expressed as percentage of positive cells for haematologica | 2017; 102(3)

CEBPA-double-mutated AML has a unique phenotype

an antigen and/or as mean fluorescence intensity (MFI; arbitrary relative linear units, scaled from 0 to 104). Bone marrow samples from 21 healthy donors (male 13, female 8; median age 36 years; range, 20-59) were used to define the normal phenotypic profile (mean value ± two standard deviations for each parameter).

CEBPA mutation analysis on sorted cells Cell sorting was performed using a FACSAria flow cytometer (BD) on diagnostic fresh bone marrow samples from six patients with CEBPA-dm AML. Some customized tubes were designed based on the phenotypic profile at diagnosis in order to sort specific cell fractions: (i) blasts; (ii) monocytes; (iii) maturing neutrophils; (iv) erythroid lineage cells; and (v) T-lymphocytes. Purity checks were performed to ensure sorting quality. Dead cells were excluded by analyzing forward scatter (FSC) versus side scatter (SSC) dot plots. Doublets were excluded by a FSC-height versus FSC-area dot plot. CEBPA mutational analysis was carried out on sorted cell fractions to reveal clonal multi-lineage involvement.

Statistical analysis Data were processed using R software (http://cran.rproject.org). Comparisons between groups were performed using the Mann–Whitney U test. P values <0.05 were considered to denote statistically significant differences. Complete remission was defined using established criteria.23 Principal component analysis was used to visualize the similarity of phenotypic profiles, comparing CEBPA-dm cases with other genotypes. We performed Ward hierarchical clustering to reveal recurrent phenotypic aberrations and used Euclidean distance as the distance measure on phenotypic parameters. Consistent with the cluster-

ing strategy, we developed a Euclidean distance-based classifier on a selected group of phenotypic parameters to predict CEBPAdm status. Samples in the validation dataset showing a distance between their normalized phenotypic data and the CEBPA-dm reference vector less than or equal to a classification threshold were considered “highly probable” cases of CEBPA-dm. In order to allow the method to be reproduced, the R script to perform the prediction of CEBPA-dm status is available in the Online Supplementary Data.

Results Characterization of patients according to CEBPA genotype Between 2006 and 2013, 318 consecutive patients were diagnosed with AML at our Institution. Enrollment criteria for the present study were the availability of: (i) a full immunophenotype (i.e., including all required phenotypic parameters) on bone marrow at diagnosis; (ii) karyotype; and (iii) molecular genetics for NPM1, FLT3 and CEBPA. On the basis of these criteria, 67 patients were excluded because of incomplete immunophenotype on bone marrow (n=32), immunophenotype on peripheral blood (n=23), and lack of molecular genetics and unavailability of a diagnostic cryopreserved specimen (n=12). Thus, 251 patients met all criteria and were studied. Their characteristics are summarized in Table 1. In this cohort, 42 CEBPA mutations were identified in 26 patients (10.3%). Sixteen

Table 1. Characteristics of patients according to CEBPA status.

Characteristics Age, median (range), years Diagnosis de novo secondary WBC, x109/L Hb, g/dL Reticulocytes, median % (abs, 1012/L) Platelets, x109/L BM blasts, % Cytogenetics t(15;17) favorable normal karyotype other intermediate adverse lack of growth not available NPM1 mutated wild-type FLT3 ITD D835 PM wild-type

Total (n=251)

CEBPA-wt (n=225, 89.6%)

CEBPA-sm (n=10, 4.0%)

CEBPA-dm (n=16, 6.4%)

P (wt vs. dm)

P (wt vs. sm)

57 (16-81)

57 (16-81)

60.5 (31-69)

48.5 (23-72)

0.0434

0.77

228 (91.9%) 23 (9.1%) 16.9 (0.5-435.0) 9.0 (3.9-14.9) 0.77 (0.029)

203 (91.2%) 22 (9.8%) 16.5 (0.6-415.0) 8.9 (3.9-14.9) 0.57 (0.022)

9 (90.0%) 1 (10.0%) 57.1 (13.1-435.0) 9.5 (7.0-10.8) 0.32 (0.011)

16 (100.0%) 8.1 (1.2-166.0) 10.6 (4.1-13.4) 2.355 (0.098)

-

1.0

0.54 0.004 <0.0001 (0.16)

0.007 0.8219 0.086 (0.09)

43 (3-815) 90 (15-100)

45 (3-815) 90 (15-100)

54 (7-99) 92 (40-100)

24 (10-193) 80 (40-100)

0.025 0.4575

0.6455 0.4388

18 (7.2%) 37 (14.7%) 119 (47.4%) 29 (11.6%) 38 (15.1%) 8 (3.2%) 2 (0.8%)

18 (8.0%) 37 (16.4%) 97 (43.1%) 28 (8.0%) 38 (16.9%) 5 (2.2%) 2 (0.9%)

9 (90.0%) 1 (10.0%) -

13 (81.3%) 1 (6.2%) 2 (12.5%) -

0.0037 1.0 0.31 -

0.006 0.51 -

67 (26.7%) 184 (73.3%)

59 (26.2%) 166 (73.8%)

8 (80.0%) 2 (20.0%)

-

-

0.0008

50 (19.9%) 6 (2.4%) 195 (77.7%)

47 (20.9%) 5 (2.2%) 173 (76.9%)

3 (30.0%) 1 (10.0%) 6 (60.0%)

16 (100.0%)

-

0.28

WBC: white blood cells; Hb: hemoglobin; abs: absolute count; BM: bone marrow; ITD: internal tandem duplication; PM: point mutation. Lack of growth means no metaphases.

haematologica | 2017; 102(3)

531

F. Mannelli et al.

out of the 26 patients (61.5%) had two CEBPA mutations, whereas the remaining ten (38.5%) had a single mutation.

The 16 patients with two CEBPA mutations had both an N-terminal truncation mutation resulting in p30 CEBPA and a C-terminal mutation affecting the bZIP domain of CEBPA. A summary of detected mutations is reported in Table 2. According to the number of mutations in the CEBPA gene, we divided our cases into: patients with double N- and C-terminal CEBPA mutations (CEBPA-dm, n=16), patients with a single mutation (CEBPA-sm, n=10), and wild-type patients without any mutation (CEBPA-wt, n=225). As regards clinical and biological features at diagnosis, CEBPA-dm patients were younger, had higher hemoglobin values and reticulocyte percentages and lower platelet counts compared to CEBPA-wt subjects.

Consistently with published literature,7 a higher incidence of normal karyotype and no mutations of NPM1 and FLT3 genes were observed in CEBPA-dm cases. CEBPA-sm

patients had higher white blood cell counts, as well as higher incidences of normal karyotype and NPM1 mutations with respect to CEBPA-wt cases.

Clinical outcome Two-hundred and two patients out of 251 had non-M3 AML and were intensively treated. In accordance with previous studies,5-7 CEBPA-dm patients, compared to CEBPA-wt patients, showed a trend toward a higher complete remission rate after the first cycle of treatment (87.5% versus 61.0%, respectively; P=0.0549), longer overall survival (median not reached versus 22.3 months, respectively; P=0.00626; Figure 1A) and longer diseasefree survival (median not reached versus 26.8 months, respectively; P=0.0667; Figure 1B). These findings did not change significantly when patients undergoing allogeneic stem cell transplantation were censored at the time of their transplant (Figure 1C,D).

Table 2. Summary of CEBPA mutations in the primary cohort.

N., sm/dm

Mutation 1 – position in CDS N-term, nt-29_518

1, sm 2, dm 3, sm 4, sm 5, sm 6, sm 7, dm

Middle, nt-469_858

C-term, nt-816_1171

c.62_63dupAG c.62_63dupAG c.247delC c.933_934insCGG c.609_610insGCACCTG c.68dupC c.180_186delGTCCATC

8, sm c.196_199dupGCCT 9, sm 10, dm c.97_112delTTTCCCCGGGGCGCGG 11, dm c.146delC 12, dm c.65_103del (-43bp) 13, dm c.247delC 14, dm c.198_201dupCTAC 15, dm c.247delC 16, dm c.291delC 17, dm c.178_188delACGTCCATCGA (-11bp) 18, sm c.209_210insC 19, dm c.338_339insCCGG 20, dm c.62_63dupAG 21, dm c.247delC 22, dm c.318_319dupTG 23, dm C.217_218insC 24, dm c.144_154del11 25, sm c.148G>T 26, sm

c.888G>A

c.1009A>T

Mutation 1position in protein AA consequence p.S21Rfs*160 p.S21Rfs*160 p.Q83Sfs*159 p.Q311_Q312insR P.Q207Afs*322 p.H24Afs*74 p.S61Tfs*157 p.Y67fs*107 p.V296= p.F33Afs*154 p.P49Rfs*159 p.H24Rfs*143 p.Q83Sfs*159 p.I68Lfs*41 p.Q83Sfs*159 p.T98Rfs*159 p.T60Hfs*103 p.A71Gfs*107 p.G114fs*170 p.S21Rfs*160 p.Q83Sfs*159 p.D107Vfs*54 p.F73Sfs*35 p.P49fs*55 p.E50Ter p.T337S

Mutation 2 – position in CDS C-term, nt-816_1171

Mutation 2 – position in protein AA consequence

c.929_930insAAG

p.T310_Q311insR

c.913_950dup

P.305_317dupQRNVETQQ KVLEL

c.992T>C c.937_939dupAAG c.919_954dup (+36bp) c.916_945dup (+30bp) c.929_930delCGins GGCACCACACCTCTCAA c.937_939dupAAG c.928_929insAGTCTA c.901_924dup

p.L331P p.K313dup p. N307_T318dup (+12AA) p.R306_L315dup (+10AA) p.E309_Q311delinsRHHTSQ

c.937_939dupAAG c.937_939dupAAG c.930_931insCAC c.1066_1071delAACTGC c.1053_1054ins129 c.1065_1066insGCC

p.K313dup p.K313dup p.T310_Q311insH p.N356_C357del p.V351_352ins43 p.G355_N356insA

p.K313dup p.E309_T310insKS p.D301_V308dup

AA: amino acid number; CDS: coding DNA sequence; ins: insertion; del: deletion; dup: duplication, nt: nucleotide. Nucleotides numbered from the major translational start codon at nucleotide position mutated. NCBI Reference Sequence is NM_004364.4. The description of sequence variants is according to the nomenclature of the HVGS site.

532

haematologica | 2017; 102(3)

CEBPA-double-mutated AML has a unique phenotype

CEBPA status and immunophenotypic findings We quantified bone marrow cell compartments at diagnosis and found that their distribution varied widely among patients. The blast compartment represented a median of 45.49% (range, 0.14-97.74) of the global cellularity, the monocytic compartment 5.53% (range, 0.0090.32) and the neutrophil and erythroid series accounted for 9.29% (range, 0.03-71.76) and 2.32% (range, 0.055.96), respectively. Phenotypic parameters were evaluated and compared among CEBPA genotypic groups and also to the cell counterpart in a control group, in order to highlight deviations from the normal phenotypic profile (Online Supplementary Data - Online Supplementary Tables S1-S4). CEBPA-dm cases showed some recurrent abnormalities in blasts and also in major maturing cell compartments in the bone marrow. With respect to control CD34+ cells, blasts from CEBPA-dm patients displayed high and homogeneous expression of immature antigens (CD34, CD117, HLA-DR) with asynchronous maturation (concomitant high expression of CD15, CD65, CD64, cyMPO) and aberrant cross-lineage expression of CD7. Beyond being merely defined as CD7+, CEBPA-dm cases showed a peculiar CD7 expression, since the vast majority of blasts expressed this antigen (Figure 2B). Similar findings were observed for antigens of maturation such as CD15 and

CD65 (Figure 2C and Online Supplementary Figure S2). The median level of expression for these antigens was also significantly higher than observed in CEBPA-wt AML (Online Supplementary Table S1). Five out of 16 (31.3%) CEBPA-dm cases displayed cross-lineage CD56 expression on blasts.

CEBPA-sm cases showed more heterogeneous phenotypic patterns (Figure 2 and Online Supplementary Figure S2).

CEBPA-dm AML displayed several recurrent phenotypic abnormalities in the maturing cell compartment as well.

The most frequently observed abnormalities in the neutrophil compartment were low SSC (35.3% of cases; Figure 2D), lower expression of CD65 and higher expression of CD64 compared both to controls and CEBPA-wt and â&#x20AC;&#x201C;sm cases (Online Supplementary Figure S3). Monocytic cells, although not quantitatively expanded compared to controls (mean percentage 2.6% versus 4.6%), were recurrently characterized by high expression of CD64 (Figure 2E) and low expression of CD36 (Online Supplementary Figure S4). The erythroid compartment was significantly more represented in CEBPA-dm cases (7.2%) than in CEBPA-wt (1.6%) and CEBPAâ&#x20AC;&#x201C;sm (0.7%) cases, being similar to control values (8.9%). Furthermore, CEBPA-dm cases shared a significant increase of more immature stages of erythroid series, as revealed by high expression of CD117 (Figure 2F) and CD105, and some antigenic

A

B

C

D

Figure 1. Survival outcomes according to CEBPA gene status. An outcome analysis was carried out for the 202 of 251 patients who were intensively treated. Kaplan Meier curves are stratified on CEBPA status: CEBPA-wild type (blue), single mutants (green), and double mutants (red) with P values representing the comparison versus wild-type patients. (A) Disease-free survival; (B) overall survival; (C) disease-free survival and (D) overall survival after censoring allo-transplanted patients at the date of transplant.

haematologica | 2017; 102(3)

533

F. Mannelli et al.

A

D

B

C

E

F

Figure 2. Phenotypic profile of blasts according to CEBPA status. Box plots illustrate the distribution of values in CEBPA-dm, -sm, -wt and controls for some core parameters: percentages of (A) CD34, (B) CD15 and (C) CD7 in blasts; (D) SSC signal in neutrophil compartment; (E) CD64 MFI in the monocyte compartment; (F) CD117 in the erythroid compartment. Box plots were generated by R software. Boxes represent the interquartile range containing 50% of the cases; the horizontal line marks the median; dots are single cases.

abnormalities (low CD36, Supplementary Figure S5).

low

CD71)

(Online

CEBPA-double-mutated status and multi-lineage involvement As previously reported, CEBPA-dm AML is often characterized as M1-M2 according to the French-AmericanBritish classification.24 Available published data show that about 20-25% of CEBPA-dm cases are associated with multi-lineage dysplasia, as defined by the World Health Organization (i.e. presence of >50% of dysplastic cells in at least 2 cell lineages).18 In our series, five out of 16 (31.3%) CEBPA-dm cases showed multi-lineage dysplasia by morphology. Specifically, erythroid dysplasia was observed in the majority of patients (10 out of 16, 62.5%), which is relatively higher than expected for a de novo, intermediate karyotype category. In this respect, the morphological findings are consistent with phenotypic data: as reported above, maturing cell compartments, and especially the erythroid one, showed aberrant phenotypic patterns that were recurrent in this genotypic subset. We thus investigated whether CEBPA mutations were clonally represented in maturing cell lineages. In order to do this, we performed CEBPA mutational status analysis after separation by fluorescence-activated cell sorting in six out of 16 CEBPA-dm patients from our cohort. Overall, post-sorting acquisition of isolated cell fractions documented a purity of 97Âą1%. We were able to isolate blast, neutrophil, monocytic and erythroid cell compartments from all six 534

patients; T-lymphocytes were employed as a negative control. In addition to blast cells, all sorted myeloid populations showed a CEBPA-dm status, whereas T lymphocytes were CEBPA-wt. The data from one illustrative case are shown in Figure 3.

Multidimensional analysis and classifier definition Although recurrent in CEBPA-dm cases, most phenotypic abnormalities showed a variable degree of overlap with the distribution of values observed in CEBPA-wt and CEBPAâ&#x20AC;&#x201C;sm patients. Consequently, the expression of no single antigen was able to discriminate CEBPA genotype.

We, therefore, processed our data by multidimensional analysis in order to verify the capability of the whole phenotypic profile, including blasts and more mature compartments, to separate the genotypic groups. First we used principal component analysis to compare CEBPA-dm cases to some genotypic subsets one by one (Figure 4).

With this method we observed a clear distinction of CEBPA-dm cases from cases bearing AML1-ETO, CBFBMYH11, and NPM1 mutations and a complex karyotype.

A partial overlap emerged for CEBPA-sm cases, essentially due to one case (Figure 4E) resembling a CEBPA-dm phenotype, which had a normal karyotype and was NPM1wt and FLT3-wt; neither homozygosity nor a second CEBPA mutation was identified after gene re-sequencing on sorted blasts (Online Supplementary Figure S6). No more sample was available for additional analyses (e.g., nextgeneration sequencing). haematologica | 2017; 102(3)

CEBPA-double-mutated AML has a unique phenotype

A

B

C

D

E

Figure 3. CEBPA mutational analysis on sorted cell fractions in one CEBPA-double-mutated patient. Cell compartments are shown on the left, with core phenotypic parameters for (A) blasts, (B) neutrophils, (C) monocytes, (D) erythroid cells, and (E) T-lymphocytes. In the corresponding plots, ungated cells are in gray whereas the relevant cell population is highlighted by color: red for blasts, purple for neutrophils, blue for monocytes, green for erythroid cells and orange for T-lymphocytes. The relative data from CEBPA mutational analysis are reported on the right, together with mutation type.

haematologica | 2017; 102(3)

535

F. Mannelli et al.

We then carried out an unsupervised clustering analysis (Figure 5). This approach was able to collect CEBPA-dm cases into a well-separated cluster. CEBPA-sm cases did not group separately, probably due to the influence on phenotype of other relevant gene mutations (e.g., NPM1).

We also carried out hierarchical clustering within selected subsets, such as the intermediate-risk karyotype category (Online Supplementary Figure S7). Since CEBPA-mutated AML has been associated with EGIL-based positivity for CD7 on blasts,17 we repeated our analysis within CD7+ cases in our cohort (Online Supplementary Figure S8). Our systematic approach provided clustering of CEBPA-dm patients even in these subgroup analyses. Given the average poor prognostic significance of CD7 expression in

A

B

C

D

E

536

F

AML, we studied outcome in CD7+ cases (Online Supplementary Figure S9): CEBPA-dm was confirmed to have a favorable impact in this phenotypic context.

To gain insight into potential influences of additional genetic changes on phenotype, we studied 12 (out of 16) CEBPA-dm cases for mutations of TET2 and GATA2 genes, which are known to be enriched in this subset (Online Supplementary Table S5). The presence of a mutated status did not influence clustering in the whole cohort nor within the CEBPA-dm group (data not shown).

In order to define a suitable classifier we carried out a selection of parameters from the initial group of 79.

Selection criteria were first based on coupled comparisons of single phenotypic parameters among CEBPA-dm versus

Figure 4. Principal component analysis of CEBPA-dm cases versus other genotypes. The multidimensional analysis of the whole phenotypic profile was able to distinguish CEBPA-dm cases from other genotypic groups: AML bearing (A) AML1ETO, (B) CBFB-MYH11, (D) NPM1 mutations, (E) complex karyotype. (C) CEBPA-single mutant cases show a wide distribution in the plot area and a partial overlap essentially due to a case (arrow) resembling a CEBPA-dm phenotypic profile. Bi-plots are generated by the combination of the first two principal components (PC), featured by the highest values of variance. Ellipses graphically represent the area of the 95% confidence interval of the distribution for the principal components. Samples outside the ellipse are outliers. Principal component analysis was carried out by R software.

haematologica | 2017; 102(3)

CEBPA-double-mutated AML has a unique phenotype

CEBPA-wt, CEBPA-sm or controls. We then selected and tested several restricted groups of parameters in principal component analysis and hierarchical clustering. Finally, we chose one set of ten parameters (Table 3) that preserved the ability to separate CEBPA-dm cases in principal component analysis (Online Supplementary Figure S10) and clustering analysis (Online Supplementary Figure S11). The selected markers were: CD34, CD117, CD7, CD15, CD65 on blasts; SSC, CD64 on cells of the neutrophil compartment; CD14, CD64 on the monocytic compartment and CD117 on erythroid cells. Furthermore, we studied the efficacy of the parameter set at clustering in a group of AML samples (n=94), with data also acquired by a FACSCanto II flow cytometer (Online Supplementary Figure S12) in order to prove that the method was not affected by

the instrument type. This classifier was thus tested as a potential screening method for CEBPA-dm genotype in AML.

Validation of the classifier on an independent cohort In order to validate the classifier prospectively, we used a large independent cohort (n=259) of unselected AML cases from three centers (Bergamo, Brescia and Venice).

FCS files, blinded as regards clinical and biological features, were sent electronically to the coordinating center.

The files were then analyzed and parameters tabulated. A group of controls (n=21) from both centers was analyzed in parallel to provide a homogeneous reference frame. The SSC signal of neutrophils was normalized on lymphocyte SSC. Applying our Euclidean distance-based classifier, a

Figure 5. Unsupervised hierarchical clustering according to genotypic groups. Cluster analysis of controls (n=21) and AML cases (n=251) based on the phenotypic parameters of all bone marrow cell compartments at diagnosis. The CEBPA-double-mutated subset clearly grouped in a separate cluster (dark green in the upper bar). CEBPA-single mutated cases displayed a heterogeneous distribution (light green in the upper bar). Columns represent individual bone marrow samples; rows represent the normalized log2 ratios of each parameter analyzed in a given cell compartment divided by the mean value obtained for that parameter in all control samples. The value of each parameter is represented in a color code according to control values: blue represents expression greater than the mean, red represents expression lower than the mean, white when not available; color intensity represents the magnitude of the deviation from the mean. Cluster analysis was carried out using R software.

haematologica | 2017; 102(3)

537

F. Mannelli et al.

score was attributed to each case of the validation cohort (Online Supplementary Table S6). Below a defined threshold, 12 AML cases were considered as “highly probable” CEBPA-dm. Twelve out of the 12 turned out to bear double CEBPA mutations. Of note, no CEBPA-dm cases were missed by the classifier (i.e., there were no false negatives). Ten out of the 12 CEBPA-dm cases had a combination of N- and C-terminal mutations. The remaining two cases showed different mutation patterns: one had one Nterminal mutation and a nonsense mutation (c.569C>A) in the middle of the coding sequence; the other had two biallelic C-terminal mutations confirmed by next-generation sequencing (Online Supplementary File – Online Supplementary Table S7). The validation set included six CEBPA-sm cases, which were not highlighted by the classifier. Considering CEBPA-dm genotype as the target, the sensitivity and specificity of the classifier were both 100%, as were the positive and negative predictive values (Table 3). Our classifier was thus validated as a reliable screening method for CEBPA-dm status on an independent cohort of AML cases.

Discussion The identification of CEBPA-dm status in AML has major clinical importance, allowing relapse risk to be stratified properly for post-remission treatment. However, most molecular screening methods for its detection have a number of technical problems. In our study, we developed an immunophenotype-based screening approach.

Through an extensive phenotypic analysis of a cohort of 251 AML cases, we found that several phenotypic aberrations occurred recurrently on blasts and on maturing cell compartments in the subset of CEBPA-dm cases. Blasts showed features of maturation asynchrony with expression of CD34 and CD117 concomitant with high-intensity CD15, CD65 and MPO. Further, there was cross-lineage expression of CD7 by the whole blast cell population (Figure 1B). This finding is consistent with previous reports correlating the expression of CD7 in AML to loss of wild-type CEBPA due to mutations4,16 or silencing by epigenetic mechanisms.26-28 The neutrophil compartment showed reduced SSC signals and overexpression of CD64, with the latter also being seen in monocytes. The erythroid series was quantitatively expanded in CEBPA-dm cases in comparison to both CEBPA-wt and CEBPA-sm cases, especially at its more immature stages. In fact, the lack of normal CEBPA function has been associated with an imbalance of the transcriptional program of

hematopoietic cells, highlighted by the gene expression profile (upregulation of genes involved in erythroid differentiation, downregulation of HOX gene members),29-31 by microRNA (over-expression of the miR-181 family)31 and long non-coding RNA (induction of UCA1 lncRNA)32 signatures. The functional consequences of CEBPA disruption would thus lead to a block in granulocytic differentiation and a preferential redirection toward the erythroid lineage.31 This is consistent with the frequent observation of erythroid dysplasia in CEBPA-dm patients in a previous study25 and in our cohort. To get insight into these data, we documented a CEBPA-dm status in all sorted myeloid cell compartments in six CEBPA-dm AML cases (Figure 3).

Our findings are a proof-of-principle of the correlation between phenotypic abnormalities and CEBPA-dm status, indicating the multi-lineage involvement and thus common clonal origin of different lineages. Moreover these data account for the observed phenotypic homogeneity, due to “CEBPA-mutated dependent” pathways of maturation.

The multidimensional analysis of the entire phenotypic profile was able to separate CEBPA-dm cases efficiently from all the other genotypes. These results are coherent with reported gene expression profile data4,6,13 and the common phenotypic signature further confirms that CEBPA-dm represents a distinct AML subset. From the initial list of 79 parameters, we built a classifier from a core group of ten parameters (Table 3), strictly required by basic AML diagnostic recommendations.20 We then applied this classifier to an independent validation set of AML cases (n=259) from three other centers. Our classifier performed extremely well (Table 3) in terms of sensitivity and specificity (100%), and no CEBPA-dm cases were missed. This is probably the most important feature such a screening technique should have in order to avoid overtreatment (i.e. allogeneic transplantation) of patients with a favorable outcome with chemotherapy. The concomitant presence of an FLT3-ITD mutation in one patient in the validation dataset did not affect its correct classification as CEBPA-dm. The profile of one CEBPA-sm case in the primary cohort overlapped that of the CEBPA-dm group in principal component analysis. Interestingly, this case had a normal karyotype and no NPM1 or FLT3 mutations, suggesting that in this genetic context, a single mutation might affect the immunophenotype similarly to CEBPA-dm status. It is worth noting that the application of the classifier was not impaired by intrinsic interlaboratory variability or by the use of different instruments, suggesting high reproducibility besides stringent standardization of the method.

Table 3. Parameters of the classifier according to cell compartment and performance in the validation cohort as far as concerns prediction of a CEBPA-double-mutated status.

Cell compartment Blasts Neutrophils Monocytes Erythroid cells Performance

CD34%, CD117 MFI, CD7%, CD15%, CD65% SSC, CD64 MFI CD14%, CD64 MFI CD117% n 259

TP 12

FN 0

TN 247

FP 0

Sensitivity 100%

LL 100

95% CI UL 100

95% CI Specificity 100%

LL 100

UL 100

TP: true positive; FN: false negative; TN: true negative; FP: false positive; CI: confidence interval; LL: lower level; UL: upper level.

538

haematologica | 2017; 102(3)

CEBPA-double-mutated AML has a unique phenotype

Beyond being technically challenging, interpretation of the CEBPA mutation pattern can sometimes be debatable and still crucial in individual cases in terms of prognosis.

The study of functional consequences of CEBPA mutations suggests that the key point of convergence is the exclusive formation of p30/p30 homodimers.33 This scenario is supposed to be shared by bi-allelic N-terminal and C-terminal mutations, as well as by the rarer combinations of two N-terminal mutations or an N-terminal mutation with a frameshift/nonsense mutation in the central part of CEBPA.33 One case from the validation set displayed the latter pattern and one showed an even rarer7 combination of two C-terminal bi-allelic mutations. Of note, both of these cases clustered together with the other CEBPA-dm cases (Online Supplementary File – Online Supplementary Table S7). The phenotypic profile might be useful to suspect bi-allelic mutations occurring on the same gene region, because of the difficult interpretation of Sanger sequencing in such a context. In contrast, it has been reported that about 10% of non-homozygous CEBPA-dm cases carry gene mutations in two different subclones, an event of uncertain significance for leukemogenesis and prognosis.12 Our data suggest that the phenotype-based classifier might pick up a shared phenotypic signature downstream to several mutation patterns, all leading to a peculiar functional CEBPA disruption, independently of mutation type. It could, therefore, enable this “classical” mutation pattern to be distinguished from alternative combinations of gene lesions. We have thus drawn a workflow embedding the classifier in the diagnostic

References 1. Pabst T, Mueller B, Zhang P, et al. Dominantnegative mutations of CEBPA, encoding CCAAT/enhancer binding protein(C/EBP ), in acute myeloid leukemia. Nat Genet. 2001;27(3):263–270.

2. Gombart A, Hofmann WK, Kawano S, et al.

Mutations in the gene encoding the transcription factor CCAAT/enhancer binding protein alpha in myelodysplastic syndromes and acute myeloid leukemias. Blood.

2002;99(4):1332–1340.

3. Preudhomme C, Sagot C, Boissel N, et al.

Favorable prognostic significance of CEBPA mutations in patients with de novo acute myeloid leukemia: a study from the Acute Leukemia French Association (ALFA). Blood.

2002;100(8):2717–2723.

4. Wouters B, Löwenberg B, ErpelinckVerschueren C, van Putten W, Valk P, Delwel R. Double CEBPA mutations, but not single CEBPA mutations, define a subgroup of acute myeloid leukemia with a distinctive gene expression profile that is uniquely associated with a favorable outcome. Blood.

2009;113(13):3088–3091.

5. Schlenk RF, Döhner K, Krauter J, et al.

Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia.

N Engl J Med. 2008;358(18):1909–1918.

6. Dufour A, Schneider F, Metzeler K, et al.

Acute myeloid leukemia with biallelic CEBPA gene mutations and normal karyotype represents a distinct genetic entity associated with a favorable clinical outcome.

J Clin Oncol. 2010;28(4):570–577.

7. Green C, Koo K, Hills R, Burnett A, Linch D, Gale R. Prognostic significance of CEBPA

haematologica | 2017; 102(3)

8.

9.

10.

11.

12.

13.

14.

work-up of AML (Online Supplementary Figure S13). This would provide insight into CEBPA-related leukemogenesis and obviously translate into quickly available prognostic information.

Being based on phenotypic data, our approach provides very early results and this goes beyond the mere speeding up of a focused molecular study. Although it is well-recognized that main genetic prognostic factors drive only the post-complete remission phase, knowledge about them since the outset is often meaningful for the clinical management of patients with AML. In conclusion, we established a reliable and straightforward screening method, based simply on the multidimensional analysis of widely available phenotypic parameters, suitable for large-scale detection of CEBPAdm status and potentially able to overcome technical issues related to molecular methods. Our approach provides very early results, allowing entire CEBPA sequencing to be performed in only selected cases. The method has high specificity and sensitivity, as demonstrated in an independent AML cohort. This is of major clinical significance, since CEBPA-dm patients show a favorable prognosis, and knowledge about the CEBPA genotype status permits the use of proportional treatment modalities.

Acknowledgments The authors would like to thank the Istituto Toscano Tumori, Ente Cassa di Risparmio di Firenze (2009-15520) and Regione Toscana (Bando Salute 2009 – research n 46) for funding this study.

mutations in a large cohort of younger adult patients with acute myeloid leukemia: impact of double CEBPA mutations and the interaction with FLT3 and NPM1 mutations.

J Clin Oncol. 2010;28(16):2739–2747.

Arber D, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood.

2016;127(20): 2391–2405.

Benthaus T, Schneider F, Mellert G, et al.

Rapid and sensitive screening for CEBPA mutations in acute myeloid leukaemia. Br J Haematol. 2008;143(2):230–239.

Fuster O, Barragán E, Bolufer P, et al.

Fragment length analysis screening for detection of CEBPA mutations in intermediate-risk karyotype acute myeloid leukemia.

Ann Hematol. 2011;91(1):1–7.

Wouters BJ, Louwers I, Valk PJ, Löwenberg B, Delwel R. A recurrent in-frame insertion in a CEBPA transactivation domain is a polymorphism rather than a mutation that does not affect gene expression profiling-based clustering of AML. Blood. 2007;109(1): 389– 390.

Behdad A, Weigelin H, Elenitoba-Johnson K, Betz B. A clinical grade sequencing-based assay for CEBPA mutation testing report of a large series of myeloid neoplasms. J Mol Diagn. 2015;17(1):76–84.

Taskesen E, Bullinger L, Corbacioglu A, et al.

Prognostic impact, concurrent genetic mutations, and gene expression features of AML with CEBPA mutations in a cohort of 1182 cytogenetically normal AML patients: further evidence for CEBPA double mutant AML as a distinctive disease entity. Blood.

2011;117(8):2469–2475.

Van Vliet MH, Burgmer P, de Quartel L, et al.

15.

16.

17.

18.

19.

20.

21.

22.

Detection of CEBPA double mutants in acute myeloid leukemia using a custom gene expression array. Gen Test Mol Biomakers.

2013;17(5):395–400.

Hrusák O, Porwit-MacDonald A. Antigen expression patterns reflecting genotype of acute leukemias. Leukemia. 2002;16(7): 1233–1258.

Lin LI, Chen CY, Lin DT, et al.

Characterization of CEBPA mutations in acute myeloid leukemia: most patients with CEBPA mutations have biallelic mutations and show a distinct immunophenotype of the leukemic cells. Clin Cancer Res. 2005;11(4):1372–1379.

Bene MC, Castoldi G, Knapp W, et al.

Proposals for the immunological classification of acute leukemias. European Group for the Immunological Characterization of Leukemias (EGIL). Leukemia. 1995;9(10): 1783–1786.

Vardiman J, Harris N, Brunning R. The World Health Organization (WHO) classification of the myeloid neoplasms. Blood.

2002;100(7):2292–2302.

Mitelman FP. An International System for Human Cytogenetic Nomenclature. 1995 S. Karger, Basel.

Döhner H, Estey E, Amadori S, et al.

Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood.

2010;115(3):453–474.

Noguera NI, Ammatuna E, Zangrilli D, et al.

Simultaneous detection of NPM1 and FLT3ITD mutations by capillary electrophoresis in acute myeloid leukemia. Leukemia.

2005;19(8):1479–1482.

Falini B, Martelli M, Bolli N, et al.

539

F. Mannelli et al.

Immunohistochemistry predicts nucleophosmin (NPM) mutations in acute myeloid leukemia. Blood. 2006;108(6): 1999–2005.

23. Cheson B, Bennett J, Kopecky K, et al.

Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol.

2003;21(24):4642–4649.

24. Fröhling S, Schlenk R, Stolze I, et al. CEBPA mutations in younger adults with acute myeloid leukemia and normal cytogenetics: prognostic relevance and analysis of cooperating mutations. J Clin Oncol. 2004;22 (4):624–633.

25. Bacher U, Schnittger S, Macijewski K, et al.

Multilineage dysplasia does not influence prognosis in CEBPA-mutated AML, supporting the WHO proposal to classify these patients as a unique entity. Blood. 2012;119

540

(20):4719–4722.

26. Röhrs S, Scherr M, Romani J, Zaborski M, Drexler H, Quentmeier H. CD7 in acute myeloid leukemia: correlation with loss of wild-type CEBPA, consequence of epigenetic regulation. J Hematol Oncol. 2010; 3:15.

27. Wouters BJ, Jordà MA, Keeshan K, et al.

Distinct gene expression profiles of acute myeloid/T-lymphoid leukemia with silenced CEBPA and mutations in NOTCH1.

Blood. 2007;110(10):3706–3714.

28. Fasan A, Alpermann T, Haferlach C, et al.

Frequency and prognostic impact of CEBPA proximal, distal and core promoter methylation in normal karyotype AML: a study on 623 cases. PLoS One. 2013;8(2):e54365.

29. Heath V, Suh HC, Holman M, et al.

C/EBPalpha deficiency results in hyperproliferation of hematopoietic progenitor cells and disrupts macrophage development in vitro and in vivo. Blood. 2004;104(6): 1639– 1647.

30. Zhang P, Iwasaki-Arai J, Iwasaki H, et al.

Enhancement of hematopoietic stem cell repopulating capacity and self-renewal in the absence of the transcription factor C/EBP alpha. Immunity. 2004;21(6): 853–863.

31. Marcucci G, Maharry K, Radmacher MD, et al. Prognostic significance of, and gene and microRNA expression signature associated with, CEBPA mutations in cytogenetically normal acute myeloid leukemia with highrisk molecular features: a Cancer and Leukemia Group B study. J Clin Oncol.

2008;26(31):5078-5087.

32. Hughes JM, Legnini I, Salvatori B, et al.

C/EBPalpha-p30 protein induces expression of the oncogenic long non-coding RNA UCA1 in acute myeloid leukemia.

Oncotarget. 2015;6(21):18534-18544.

33. Ohlsson E, Schuster MB, Hasemann M, Porse BT. The multifaceted functions of C/EBP in normal and malignant hematopoiesis. Leukemia. 2015;30(4):767775.

haematologica | 2017; 102(3)

ARTICLE

Acute Lymphoblastic Leukemia

Tumor suppressors BTG1 and IKZF1 cooperate during mouse leukemia development and increase relapse risk in B-cell precursor acute lymphoblastic leukemia patients Blanca Scheijen,1* Judith M. Boer,2* René Marke,1* Esther Tijchon,1 Dorette van Ingen Schenau,1 Esmé Waanders,3,4 Liesbeth van Emst,1 Laurens T. van der Meer,1 Rob Pieters,4 Gabriele Escherich,5 Martin A. Horstmann,5 Edwin Sonneveld,6 Nicola Venn,7 Rosemary Sutton,7 Luciano Dalla-Pozza,8 Roland P. Kuiper,3,4 Peter M. Hoogerbrugge,4 Monique L. den Boer2 and Frank N. van Leeuwen1

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(3):541-551

Laboratory of Pediatric Oncology, Radboud university medical center, Nijmegen, the Netherlands; 2Department of Pediatric Oncology, Erasmus MC-Sophia Children's Hospital, Erasmus University Medical Center, Rotterdam, the Netherlands; 3Department of Human Genetics, Radboud university medical center, Nijmegen, the Netherlands; 4 Princess Máxima Center for Pediatric Oncology, Utrecht, the Netherlands; 5Research Institute Children's Cancer Center and Clinic of Pediatric Hematology and Oncology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 6Dutch Childhood Oncology Group, The Hague, the Netherlands; 7Australian and New Zealand Children’s Oncology Group, Children’s Cancer Institute Australia, Lowy Cancer Research Centre, University of New South Wales, Sydney, Australia and 8Oncology Unit, The Children’s Hospital at Westmead, Sydney, Australia 1

*BS, JMB and RM contributed equally to this work

ABSTRACT

D

eletions and mutations affecting lymphoid transcription factor IKZF1 (IKAROS) are associated with an increased relapse risk and poor outcome in B-cell precursor acute lymphoblastic leukemia. However, additional genetic events may either enhance or negate the effects of IKZF1 deletions on prognosis. In a large discovery cohort of 533 childhood B-cell precursor acute lymphoblastic leukemia patients, we observed that single-copy losses of BTG1 were significantly enriched in IKZF1-deleted B-cell precursor acute lymphoblastic leukemia (P=0.007). While BTG1 deletions alone had no impact on prognosis, the combined presence of BTG1 and IKZF1 deletions was associated with a significantly lower 5-year event-free survival (P=0.0003) and a higher 5-year cumulative incidence of relapse (P=0.005), when compared with IKZF1-deleted cases without BTG1 aberrations. In contrast, other copy number losses commonly observed in B-cell precursor acute lymphoblastic leukemia, such as CDKN2A/B, PAX5, EBF1 or RB1, did not affect the outcome of IKZF1-deleted acute lymphoblastic leukemia patients. To establish whether the combined loss of IKZF1 and BTG1 function cooperate in leukemogenesis, Btg1-deficient mice were crossed onto an Ikzf1 heterozygous background. We observed that loss of Btg1 increased the tumor incidence of Ikzf1+/- mice in a dose-dependent manner. Moreover, murine B cells deficient for Btg1 and Ikzf1+/- displayed increased resistance to glucocorticoids, but not to other chemotherapeutic drugs. Together, our results identify BTG1 as a tumor suppressor in leukemia that, when deleted, strongly enhances the risk of relapse in IKZF1-deleted B-cell precursor acute lymphoblastic leukemia, and augments the glucocorticoid resistance phenotype mediated by the loss of IKZF1 function. haematologica | 2017; 102(3)

Correspondence: blanca.scheijen@radboudumc.nl

Received: July 19, 2016.

Accepted: December 14, 2016.

Pre-published: December 15, 2016. doi:10.3324/haematol.2016.153023 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/541 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

541

B. Scheijen et al.

Introduction Acute lymphoblastic leukemia (ALL) is the most common form of cancer in children and is characterized by recurrent genetic aberrations and chromosomal abnormalities, which represent distinct genetic subtypes that are used for risk stratification.1 In the past, we and others have demonstrated that genomic alterations affecting the lymphoid transcription factor gene IKZF1 represent a strong prognostic factor associated with relapse in childhood B-cell precursor acute lymphoblastic leukemia (BCPALL).2-5 Recently, we established that loss of IKZF1 affects glucocorticoid (GC)-mediated gene regulation and confers GC resistance in BCP-ALL.6 Deletion of IKZF1 is a hallmark of BCR-ABL1-positive BCP-ALL,7 and within this high-risk cytogenetic subtype, IKZF1 loss is associated with an even worse outcome.8 IKZF1 gene lesions are also frequently present in high-risk leukemias with a Philadelphia- or BCR-ABL1-like expression signature,2,5,9 which often carry activated tyrosine kinases (i.e., ABL1, PDGFRB, JAK2, FLT3) or mutations targeting the RAS pathway.10-12 Mouse studies have confirmed that two of these genomic alterations, namely BCR-ABL1 and mutant RAS, cooperate with loss of IKZF1 during leukemic transformation.13,14 On the other hand, ERG gene deletions constitute a specific subtype of BCP-ALL with favorable outcome, despite the frequent co-occurrence of IKZF1 deletions.15,16 Thus, specific genetic interactions may modulate the tumor suppressive functions of IKZF1 during leukemia development at initial diagnosis and at relapse after chemotherapy treatment.

Besides IKZF1, many other recurrent genetic aberrations have been observed in BCP-ALL, which include deletions affecting BTG1, CDKN2A/B, EBF1, ETV6, PAX5, RAG1, RB1 and TCF3.17-19 For some of these co-occurring genetic lesions, synergistic effects have been reported on leukemia development. For example, loss of Rag1 was shown to accelerate the onset of B-cell lymphoblastic leukemia in Cdkn2a/p19Arf-deficient mice,20 while increasing the incidence of T-cell lymphomas in Tcf3–/– mice.21 Furthermore, combined heterozygosity for Ebf1 and Pax5 results in a strong increase in the frequency of pro-B-cell leukemia in mice.22 Herein we describe the impact of deletions affecting the transcriptional coregulator B-cell translocation gene 1 (BTG1), in co-occurrence with IKZF1 loss on both leukemogenesis as well as outcome. We demonstrate that BTG1 deletions are enriched in IKZF1-deleted pediatric BCP-ALL cases, and correlate with increased relapse risk in this patient group. Using mouse knockout models, we further demonstrate that loss of Btg1 cooperates with Ikzf1+/- in the onset of ALL. Finally, our data indicate that both BTG1 and IKZF1 are important determinants of the GC therapy response, and that combined loss of these tumor suppressors enhances GC resistance.

enrolled in the Australian and New Zealand Children’s Haematology and Oncology Group (ANZCHOG) ALL8 protocol.

In accordance with the Declaration of Helsinki, written informed consent was obtained from parents or legal guardians, and institutional review boards approved the use of excess diagnostic material for research purposes. Details on the patient cohorts, treatment regimens and outcomes, were described previously.5,23,24

Statistics patient data Cumulative incidence of relapse (CIR) was estimated using a competing risks model, equality of CIRs was tested with Gray's test. Relapse and non-response to induction chemotherapy were considered as events, with death and secondary malignancy as competing events. Event-free survival (EFS) was calculated with non-response, relapse, secondary malignancy and death considered as events. EFS probabilities were estimated using the KaplanMeier method, and survival data between groups were compared using univariate and multivariate Cox regression analyses. The proportions of patients with IKZF1 and other B-cell development gene deletions as well as other categorical variables were compared using the Fisher’s exact test. All P-values are two-sided, and a significance level of 0.05 or less was considered to be significant.

Analyses were performed in R 3.0.1 (2013-05-16), using the packages cmprsk version 2.2-7, mstate version 0.2.7, and survival version 2.37-4.25-27

Mice Btg1 and Ikzf1 (IkNeo) knockout lines have been described previously,28,29 and were intercrossed on a C57Bl/6J background. Mice were maintained under specific pathogen-free conditions at our Central Animal Laboratory facility. Genotyping of the offspring was performed by polymerase chain reaction (PCR) (primer sequences are listed in the Online Supplementary Table S1). Animal experiments were approved by the Animal Experimental Committee of the Radboud university medical center and were performed in accordance with institutional and national guidelines.

Functional characterization of murine lymphocytes Detailed information on functional characterization of normal and leukemic lymphocytes by flow cytometry, immunohistochemistry, immunoglobulin (IG)/T-cell receptor (TR) PCR and cell viability assays can be found in the Online Supplementary Methods section. To analyze the glucocorticoid response in B lymphocytes, mononuclear splenocytes were isolated from wild-type and the different knockout mice, and stimulated in vitro with 5 mg/mL lipopolysaccharide (LPS) for 48 hours. The obtained activated B lymphocytes (≥ 80% B220+) were isolated by ficoll gradient and cultured for another 48 hours in the absence or presence of the synthetic glucocorticoids prednisolone or dexamethasone.

Thereafter, relative cell viability was assessed by MTS assay and AnnexinV/7-AAD staining.

Results BTG1 deletions are enriched in IKZF1-deleted pediatric BCP-ALL

Methods Clinical samples The discovery cohort comprised of 533 pediatric patients with newly diagnosed BCP-ALL from three consecutive Dutch Childhood Oncology Group trials (DCOG ALL-8, ALL-9 and ALL10) and two German Cooperative ALL trials (COALL 06-97 and 07-03). The validation cohort consisted of 515 pediatric patients 542

Gene deletion of the tumor suppressor IKZF1, creating either dominant-negative IKZF1 isoforms or haploinsufficiency, is an important predictor of poor outcome in BCPALL,2-5 but to what extent other additional common single gene deletions, such as CDKN2A/B, PAX5, BTG1, ETV6, EBF1 or RB1 impact the prognostic value of IKZF1 has not been clearly established. To address this question, we studied a previously described childhood BCP-ALL cohort haematologica | 2017; 102(3)

Combined BTG1 and IKZF1 deletions affect outcome

Table 1. Co-occurrence of IKZF1 deletion with other common gene deletions in BCP-ALL.

BTG1 no deletion BTG1 deletion PAX5 no deletion PAX5 deletion CDKN2A/B no deletion CDKN2A/B deletion EBF1 no deletion EBF1 deletion RB1 no deletion RB1 deletion ETV6 no deletion ETV6 deletion

IKZF1 no deletion n

IKZF1 deletion %

n

%

397 31 330 98 298 130 405 23 400 28 296 132

0.93 0.07 0.77 0.23 0.70 0.30 0.95 0.05 0.93 0.07 0.69 0.31

88 17 55 50 53 52 96 9 91 14 80 25

0.84 0.16 0.52 0.48 0.50 0.50 0.91 0.09 0.87 0.13 0.76 0.24

of 533 cases enrolled in consecutive DCOG and COALL trials.5 The representation of the different BCP-ALL subtypes was similar between the DCOG and COALL cohorts, and comparable to that described in the literature (Online Supplementary Table S2). We identified 105 BCPALL patient samples containing an IKZF1 deletion. Within the IKZF1-deleted group, we observed a significant enrichment for BTG1 deletions (P=0.007), where 17 of the 105 IKZF1-deleted cases (16%) harbored BTG1 deletions as compared to 31 of the 428 IKZF1 wild-type cases (7%) (Table 1). These focal BTG1 deletions mainly covered the second exon of BTG1 and downstream adjacent sequences, as described previously.30 Similarly, deletions affecting PAX5 (P<0.0001), CDKN2A/B (P=0.0003) and RB1 (P=0.026) were present at higher frequencies in IKZF1-deleted cases, whilst this was not observed for EBF1 or ETV6 deletions (Table 1). To the contrary, in the BTG1-deleted group we observed significant enrichment for IKZF1 (P=0.007), EBF1 (P=0.0011), RB1 (P=0.042), and ETV6 (P=0.0046) deletions (Online Supplementary Table S3), which is in agreement with our previous findings.30 Previously, it has been shown that both IKZF1 and BTG1 deletions are strongly enriched in the cytogenetic BCR-ABL1 subtype.7,31 Consistent with this notion, we observed that of 24 BCR-ABL1-positive cases, 7 (29%) harbored deletions in both IKZF1 and BTG1 (Table 2), which represents 47% of IKZF1-deleted and 100% of the BTG1-deleted cases found in the BCR-ABL1-positive group (Table 2).

Combined deletions of BTG1 and IKZF1 predict inferior outcome in BCP-ALL We next compared clinical characteristics of BTG1;IKZF1 double-deleted cases, sole deletion of IKZF1 and sole deletion of BTG1 cases, and cases without IKZF1 or BTG1 deletion in our complete childhood BCP-ALL cohort. The characteristics of these four groups were similar to the total cohort with respect to gender and treatment protocol (Table 2). The double-deleted and sole deletion of IKZF1 groups contained more patients over 10 years of age and increased white blood cell counts, and hence more National Cancer Institute (NCI)-Rome criteria high-risk cases32 (Table 2). We compared the EFS and CIR haematologica | 2017; 102(3)

Fisher’s Test P

Odds ratio

0.0071

2.5

1.40E-06

3.1

3.40E-04

2.2

0.25

1.6

0.026

2.2

0.19

0.7

between cases with both IKZF1 and BTG1 deletions (BTG1-del;IKZF1-del, n=17) with the cases of sole deletion of IKZF1 (BTG1-wt;IKZF1-del, n=88) (Figure 1A,B).

The 5-year CIR in IKZF1- plus BTG1-deleted cases was 53% ± 13% compared with 28% ± 5% in the cases of sole deletion of IKZF1 (P=0.005; Table 3A). Similarly, the 5year EFS was lower in double-deleted cases compared with the cases of sole deletion of IKZF1 (HR 3.5, P=0.0003; Table 3A). In contrast, the cases of sole deletion of BTG1 (BTG1-del;IKZF1-wt, n=31) showed a similar outcome (5-year CIR:10% ± 6%) to the reference cases without IKZF1 or BTG1 deletions (n=397; 5-year CIR:13% ± 2%). The synergistic effect of loss of BTG1 and IKZF1 on outcome remained after correction for subtype in the Cox model (Table 3A), and after leaving out the BCR-ABL1-positive cases (Figure 1C,D; Table 3B). For the BCR-ABL1-positive cases, BTG1 deletions did not further impact the poor treatment outcome as observed for the cases of sole deletion of IKZF1 (Figure 1E,F; Table 3C).

As deletions of PAX5, CDKN2A/B and RB1 were similarly enriched in IKZF1-deleted BCP-ALL, we examined the impact of these deletions on the outcome of IKZF1-deleted cases. In contrast to BTG1;IKZF1 double-deleted patients, the outcome of patients with co-occurring PAX5, CDKN2A/B or RB1 deletions did not differ from cases of sole deletion of IKZF1 (Figure 2; Online Supplementary Table S4). Similarly, co-occurrences of IKZF1 deletions with either EBF1 or ETV6 deletions did not affect outcome compared with the sole deletion of IKZF1. To validate our findings, we analyzed the Australian and New Zealand ANZCHOG ALL8 cohort (n=515)23,24 to assess the prognostic value of BTG1;IKZF1 double-deletions. In this cohort, 6 out of 11 BTG1-del;IKZF1-del patients developed a relapse (Online Supplementary Figure S1). The 5-year CIR in the BTG1;IKZF1 double-deleted group was 61% ± 19% versus 35% ± 6% in the group with the sole deletion of IKZF1 (P=0.19; Table 3D). Hence, the same trend was observed in this independent validation cohort, albeit statistically non-significant. Together, these data indicate that BTG1 deletions in an unselected leukemia population have no prognostic value, but BTG1 copy number losses specifically exacerbate the effects of IKZF1 deletion on inferior outcome in BCP-ALL. 543

B. Scheijen et al.

Table 2. Pediatric BCP-ALL patient characteristics in the BTG1 and IKZF1 sole and BTG1;IKZF1 double-deleted groups.

Patients' characteristics

Sex Female Male Age (years) < 10 ≥ 10 WBC (cells/nl) < 50 ≥ 50 NCI-Rome SR HR Protocol DCOG COALL Subtype ETV6-RUNX1 HeH B-other BCR-ABL1-like TCF3r BCR-ABL1 MLLr

BTG1-del; IKZF1-del; n=17 (3%) n %

BTG1-wt IKZF1-del n=88 (20%) n %

BTG1-del; IKZF1-wt n=31 (6%) n %

BTG1-wt; IKZF1-wt n=397 (74%) n %

5 11

29% 71%

44 44

50% 50%

11 20

35% 65%

183 214

11 6

65% 35%

60 28

68% 32%

26 5

84% 16%

11 6

65% 35%

55 33

62.5% 37.5%

23 8

8 9

47% 53%

39 49

44% 56%

14 3

82% 18%

63 25

2 1 4 3 0 7 0

12% 6% 24% 18% 0% 41% 0%

2 16 23 38 0 8 1

Total n=533 n

%

46% 54%

243 290

46% 54%

330 67

83% 17%

427 106

80% 20%

74% 26%

299 96

76% 24%

388 143

73% 27%

19 12

61% 39%

245 151

62% 38%

311 221

58% 42%

72% 28%

25 6

81% 19%

311 86

78% 22%

413 120

77% 23%

2% 18% 26% 43% 0% 9% 1%

24 2 5 0 0 0 0

77% 7% 16% 0% 0% 0% 0%

126 107 75 51 19 9 10

32% 27% 19% 13% 5% 2% 3%

154 126 107 92 19 24 11

29% 24% 20% 17% 4% 5% 2%

Fisher’s P

0.3

0.0054

0.066

0.016

0.54

<0.0001 0.0055 0.41 <0.0001 0.093 <0.0001 0.89

WBC count for 2 cases missing, NCI Risk for 1 case missing, both in the BTG1-wt ;IKZF1-wt group. del: deletion; WBC: white blood cell; NCI-Rome HR: high-risk defined by age at diagnosis ≥ 10 years and/or WBC ≥ 50 cells/nl; HeH: high hyperdiploid (51-65 chromosomes); P-value: Fisher's exact test across the four deletions groups; wt: wild-type; DCOG: Dutch Childhood Oncology Group; COALL: German Cooperative ALL; SR: standard-risk.

Leukemia predisposition in Btg1 knockout mice Although deletions affecting the BTG1 gene are a frequent event in BCP-ALL, a direct role for BTG1 in leukemia development has not been reported. Therefore, we first examined the tumor suppressive function of BTG1 using a constitutive Btg1 knockout line harboring a Neo-cassette in the first coding exon.28 We previously reported that these Btg1 knockout mice display defective B-cell development with a 25% reduction in the amount of progenitor B cells within the bone marrow compartment, mainly affecting the pre-B and immature B-cell stage.33 Furthermore, Btg1 is required for optimal proliferative expansion of early progenitor B cells in methylcellulose in response to interleukin-7. At the same time, there was no obvious defect in the development of myeloid and T-lymphoid cells in these Btg1-deficient animals.33 In the study herein, mice that carried either one or two copies of the Btg1 knockout allele were followed over a period of 18 months, along with control littermates. About 6% (n=3/49) of the wild-type C57BL6/J mice developed B-cell lymphomas between the age of 14 and 18 months (Table 4), which is consistent with previous observations.34 Within the same time period 6% of the Btg1+/- (n=2/34) and 18% of the Btg1-/- (n=6/33) mice developed T-cell leukemia exclusively (Table 4), characterized by enlarged primary lymphoid organs, such as the spleen and lymph nodes, and focal infiltration of leukemic T cells into peripheral organs, such as the lungs and liver. These Btg1-deficient T-cell leukemias expressed the T-cell surface marker CD3, and displayed clonal T-cell receptor (TR) rearrangements (Figure 3B). In addition, these CD3+ T-cell 544

leukemias not only showed increased expression of the Tcell activation marker CD44, but also large numbers of B220+ cells within the infiltrated areas of tissues, such as the liver or lungs, and affected lymph nodes (Online Supplementary Figure S2). There was no evidence for clonal immunoglobulin gene rearrangements in these Btg1-/T-cell leukemias (Online Supplementary Figure S2), suggesting the presence of a substantial number of non-malignant B lymphocytes in proximity to these leukemic T cells.

These data show that, although somatic BTG1 deletions predominantly occur in BCP-ALL, Btg1-deficiency in the mouse germline predisposes exclusively to T-cell malignancies. This predilection for T-lineage leukemias is also observed in other knockout mouse models targeting genes commonly deleted in BCP-ALL, such as Ikzf1 mutant mice.35,36

Loss of Btg1 increases leukemia incidence in Ikzf1+/- mice To investigate cooperation between BTG1 and IKZF1 during leukemogenesis, we intercrossed Btg1-deficient mice with haplodeficient Ikzf1 mice using the IkNeo mouse line,29 which harbors a Neo-floxed knock-in allele combined with a Pax5-IRES-GFP complementary DNA (cDNA) at the first coding exon of Ikzf1, thereby creating an Ikzf1 null allele.29 These mice are only viable as a heterozygous knockout line (Ikzf1+/-). First, we analyzed the phenotype of young animals (age 6-12 weeks) to assess the effect of the Ikzf1+/- allele on Btg1-deficiency in B- and T-lymphoid development. Ikzf1+/-mice, like Btg1-/- mice, displayed a moderate reduction in the fraction of B220+ cells in bone marrow (BM) and the spleen (Online haematologica | 2017; 102(3)

Combined BTG1 and IKZF1 deletions affect outcome Table 3. CIR and EFS analysis of BTG1;IKZF1 double-deleted pediatric BCP-ALL cases compared with cases of sole deleted IKZF1.

A BCP-ALL1 (n=533) Deletion

Total

Relapse

Death

Univariate 5-yr CIR (SE)

Gray P

EFS HR (95% CI)

Cox P

BTG1;IKZF1

17

10

2

53% (13%)

0.005

3.5 (1.8-6.8)

0.0003

2.5 (1.0-5.9)

IKZF1 BTG1 none

88 31 397

25 3 52

5 0 12

28% (5%) 10% (6%) 13% (2%)

Death

5-yr CIR (SE)

Univariate Gray P

EFS HR (95% CI)

Cox P

EFS HR (95% CI)

0.011

4.4 (1.9-10.3)

0.0007

6.7 (2.7-16)

EFS HR (95% CI)

Cox P

Multivariate EFS HR (95% CI)

Cox P

0.8 (0.3-2.5)

0.7

ND

ND

EFS HR (95% CI)

Cox P

Multivariate EFS HR (95% CI)

Cox P

01.83 (0.8-4.2)

0.16

ND

ND

B BCP-ALL1 without BCR-ABL1 (n=509) Deletion Total Relapse BTG1;IKZF1

10

6

1

50% (17%)

IKZF1 BTG1 none

80 31 388

21 3 52

2 0 10

26% (5%) 10% (6%) 13% (2%)

Death

5-yr CIR (SE)

Univariate Gray P 0.7

C BCP-ALL1 BCR-ABL1 only (n=24) Deletion Total Relapse BTG1;IKZF1

7

4

1

57% (22%)

IKZF1 BTG1 none

8 0 9

4 0 0

3 0 2

50% (21%)

Death

5-yr CIR (SE)

Univariate Gray P 0.19

D BCP-ALL2 validation cohort (n=515) Deletion Total Relapse

Multivariate EFS HR (95% CI)

Cox P 0.043

Multivariate Cox P 0.00003

0% (0%)

BTG1;IKZF1*

11

6

0

61% (19%)

IKZF1 BTG1*

69 45

23 13

4 0

35% (6%) 27% (7%)

none**

390

55

4

13% (2%)

Discovery cohort: pediatric patients from consecutive DCOG/COALL trials. 2Validation cohort: pediatric patients from ANZCHOG ALL8 trial. * Report of 1 secondary malignancy in these groups; ** Report of 6 independent secondary malignancies in this group. CIR: cumulative incidence of relapse; SE: standard error; EFS: event-free survival; HR: hazard ratio; CI: confidence interval; multivariate: corrected for BCP-ALL subtype and stratified for study cohort (DCOG, COALL); none: neither BTG1 nor IKZF1 deletion: BCP-ALL: B-cell precursor acute lymphoblastic leukemia; ND: not determined.

1

Supplementary Figure S3). This correlated with a partial block at the pre-pro-B-cell stage (Hardy fraction A) and the pre-B-cell stage (Hardy fraction D) in Ikzf1+/- mice (Online Supplementary Figure S3). Btg1-/-;Ikzf1+/- mice showed an even stronger reduction in B220+ cells, with additive effects at both B220+CD43+ and B220+CD43â&#x20AC;&#x201C; differentiation stages in BM (Hardy fractions A to E) (Online Supplementary Figure S3). In contrast, Btg1-/-;Ikzf1+/- mice, similar to Btg1-/- and Ikzf1+/- single knockout animals, showed no major defects in postnatal thymic T-cell development (Online Supplementary Figure S4).

Next, we followed Ikzf1+/-, Btg1-/-;Ikzf1+/- and Btg1-/;Ikzf1+/- mice for a period of 18 months. In our cohort, 14% (n=5/36) of the Ikzf1+/- mice developed T-cell leukemia between 3 and 18 months of age, with a median age of 12 months (Figure 3A; Table 4). Similar to Btg1-deficient leukemia, the Ikzf1+/- T-cell leukemias showed infiltration haematologica | 2017; 102(3)

of CD3+ leukemic blasts into distant organs and clonal TR rearrangements (Figure 3). Interestingly, we observed an increased leukemia incidence upon combined loss of Btg1 and Ikzf1 (Figure 3A), which is consistent with our hypothesis that BTG1 deletions cooperate with IKZF1 aberration to induce human BCP-ALL. We found that 26% of the Btg1+/-;Ikzf1+/- mice (n=10/35) and 40% of the Btg1-/;Ikzf1+/- (n=12/30) animals developed T-cell leukemia, while leukemias in Btg1-/-;Ikzf1+/- appeared with a slightly shorter latency (9.4 months) relative to Ikzf1+/- mice (12.4 months) (P=0.011) (Table 4). Tumors in the Btg1-/-;Ikzf1+/compound mice were characterized by significantly higher leukocyte counts in peripheral blood compared to single knockout animals (Figure 3B,C), and strong infiltration of leukemic cells into the liver and lungs (Figure 3B,D), as well as clonal TR rearrangements (Figure 3E). Flow cytometric analysis of the different T-cell leukemias revealed 545

B. Scheijen et al.

that most of the Btg1-/-;Ikzf1+/- leukemias were CD4+CD8+ double positive T cells with ongoing differentiation towards CD4 or CD8 single positive stage (Figure 3F, Online Supplementary Table S5). Moreover, isolated leukemic T cells, derived from all the different genetic backgrounds included in our studies (Ikzf1+/-, Btg1-/- and

Btg1-/-;Ikzf1+/-), could be serially transplanted into syngeneic C57BL6/J mice giving rise to similar (oligo)clonal T-cell leukemias (Online Supplementary Figure S5). Taken together, our data demonstrate that loss of Btg1 cooperates with haplodeficiency for Ikzf1 during mouse leukemia development in a dose-dependent manner.

A

B

C

D

E

F

Figure 1. Cumulative incidence of relapse (CIR) and event-free survival (EFS) curves for pediatric BCP-ALL cases with or without IKZF1 and BTG1 deletions. (A) CIR and (B) EFS curves for total BCP-ALL cohort (n=533), IKZF1 plus BTG1 deletion, (C) CIR and (D) EFS curves for BCP-ALL without BCR-ABL1-positive cases (n=509), IKZF1 plus BTG1 deletion, (E) CIR (F) EFS curves for BCR-ABL1-positive cases (n=24), IKZF1 plus BTG1 deletion. Colors: black, IKZF1 and BTG1 wild-type; green, IKZF1 wild-type, BTG1-deleted; red, IKZF1-deleted, BTG1 wild-type; blue, both BTG1- and IKZF1-deleted. For CIR graphs (A,C,E) the Gray P-value and for the EFS graphs (B,D,F) the Cox P-value is indicated comparing BTG1-del;IKZF1-del with BTG1-wt;IKZF1-del. wt: wild-type; del: deletion; n: total number; BCP-ALL: B-cell precursor acute lymphoblastic leukemia.

546

haematologica | 2017; 102(3)

Combined BTG1 and IKZF1 deletions affect outcome

A

B

C

D

E

Figure 2. Cumulative incidence of relapse (CIR) curves for pediatric BCP-ALL cases with or without IKZF1 deletions in combination with other single common gene deletions. (A) IKZF1 plus PAX5, (B) IKZF1 plus CDKN2A/B, (C) IKZF1 plus RB1, (D) IKZF1 plus EBF1, (E) IKZF1 plus ETV6 (n=533). Colors: black, none of the indicated deletions; green, IKZF1 wild-type, indicated gene deleted; red, IKZF1-deleted, indicated gene wild-type; blue, both IKZF1 and indicated gene deleted. wt: wild-type; del: deletion; n: total number; BCP-ALL: B-cell precursor acute lymphoblastic leukemia.

Table 4. Characteristics of lymphoid tumors derived from single and intercrossed Btg1 and Ikzf1 knockout lines.

Genotype

Wild-type Ikzf1+/Btg1+/Btg1+/-;Ikzf1+/Btg1–/– Btg1–/–;Ikzf1+/-

Tumor incidence (% mice) 3/49 5/36 2/34 10/35 6/33 12/30

(6%) (14%) (6%) (29%) (18%) (40%)

Mean age (months)

Tumor phenotype

16.4 12.4 15.8 13.6 17.4 9.4

B-cell lymphoma T-cell leukemia T-cell leukemia T-cell leukemia T-cell leukemia T-cell leukemia

Ikzf1

P-valuea Btg1+/-

Btg1–/–

0.271 0.134 0.703 0.011

0.271 0.013 0.143 0.001

0.703 0.143 0.231 0.026

+/-

Chi-square analysis comparing incidence of T-cell leukemia to indicated genotype at 18 months.

a

haematologica | 2017; 102(3)

547

B. Scheijen et al.

BTG1 modifies glucocorticoid resistance mediated by loss of IKZF1 While these experiments confirm the genetic interaction between BTG1 deletions and IKZF1 aberrations during leukemogenesis, they do not explain the poor outcome observed in patients showing a combined loss of BTG1 and IKZF1. Recently, we established that inferior outcome related to IKZF1 deletions in BCP-ALL is correlated with

A

B

C

D

E

F

an attenuated in vivo day 8 prednisolone response and increased GC resistance in IKZF1-deleted primary leukemic cells, as determined by in vitro MTT assays.6 These results could be recapitulated using primary splenic B cells isolated from Ikzf1+/- mice, which revealed that nonleukemic Ikzf1+/- B cells are also less sensitive towards GCinduced apoptosis.6 Based on our previous findings that BTG1 regulates glucocorticoid receptor activation,37 we

Figure 3. Leukemia incidence and phenotype of Btg1 knockout mice intercrossed with haplodeficient Ikzf1 animals. (A) Kaplan-Meier survival curve indicates the leukemia event-free survival in wild-type, Ikzf1+/-, Btg1+/-, Btg1+/-;Ikzf1+/-, Btg1-/-, and Btg1-/-;Ikzf1+/- mice over a time period of 17 months. Leukemia incidence is significantly increased in Btg1-/-;Ikzf1+/- mice as compared to Ikzf1+/- mice (P=0.011). (B) Peripheral blood smear stained with Giemsa and immunohistochemistry for CD3 on the lung and liver tissues of diseased Ikzf1+/-, Btg1-/-, Btg1+/-;Ikzf1+/-, and Btg1-/-;Ikzf1+/- mice. (C) Quantification of blast counts in peripheral blood smear stained with Giemsa of diseased wild-type, Ikzf1+/-, Btg1-/-, Btg1+/-;Ikzf1+/- and Btg1-/-;Ikzf1+/- mice (n=3). The percentage of leukemic blasts is indicated. ***P<0.001. (D) Quantification of T-cell (CD3) infiltration into the liver and lungs of diseased wild-type, Ikzf1+/-, Btg1-/-, Btg1+/-;Ikzf1+/- and Btg1-/-;Ikzf1+/- mice (n=3) using FIJI software.

Data are represented as the positively stained area divided by the total area measured, with standard errors of the mean P-values (two-sided t-test). *P<0.05, **P<0.01, and ***P<0.001. (E) T-cell receptor beta 2 (Trb2) gene rearrangement analysis by PCR on control tissue, B-cell lymphoma derived from wild-type mice (#3863), and T-cell leukemias derived from Ikzf1+/- (#13135, #350), Btg1-/- (#160), Btg1+/-;Ikzf1+/- (#5103 and #5102) and Btg1-/-;Ikzf1+/- mice (#9073, #7169, #7173).

(F) Flow cytometry analyzing CD4 and CD8 expression on T-cell leukemia samples of two Ikzf1+/-, Btg1-/-, Btg1+/-;Ikzf1+/- and Btg1-/-;Ikzf1+/- mice. WT: wild-type; M: 1 kb DNA ladder marker; Ctrl: control DNA.

548

haematologica | 2017; 102(3)

Combined BTG1 and IKZF1 deletions affect outcome

investigated whether loss of BTG1 would impact the GC response in primary murine B cells. To this end, B cells isolated from WT, Btg1-/-, Ikzf1-/- and Btg1-/-;Ikzf1+/- mice, and obtained after lipopolysaccharide activation, were stimulated for 48 hours with increasing concentrations of prednisolone or dexamethasone and subjected to MTS assays to assess relative cell survival. While Btg1-deficiency alone had no effect on GC-induced apoptosis, Ikzf1-haplodeficient B cells showed enhanced cell survival as compared to WT (Figure 4A), similar to our previous findings.6 Importantly, Btg1-/-;Ikzf1+/- B cells showed an even stronger resistance to GC-induced apoptosis when compared to Ikzf1+/- B cells (P<0.001). These findings were confirmed

by AnnexinV staining, demonstrating a significantly smaller apoptotic fraction in Btg1-/-;Ikzf1+/- B cells relative to Ikzf1+/- B cells (P<0.001) (Figure 4B). Analyses of primary Btg1-/-, Ikzf1+/- and Btg1-/-;Ikzf1+/- thymocytes revealed no differential sensitivity to GC-induced apoptosis as compared to WT (Online Supplementary Figure S6A). Next, we assessed whether loss of Btg1 and Ikzf1 would promote resistance in B cells to other chemotherapeutic drugs commonly used in the treatment of BCP-ALL patients, including 6-mercaptopurine, doxorubicin, vincristine and

asparaginase. However, Btg1-/-;Ikzf1+/- and Btg1-/-;Ikzf1+/- B cells showed similar cell survival in comparison to control cells (Online Supplementary Figure S6B).

Together, these data argue that loss of tumor suppressor BTG1 enhances GC resistance in the context of IKZF1 deletions, which may explain the inferior treatment outcome observed in patients showing combined loss of BTG1 and IKZF1.

Discussion BCP-ALL is a heterogeneous disease characterized by recurrent deletions enriched in specific genetic subtypes.1 For instance, it is known that deletions affecting the transcriptional co-regulator BTG1 are unevenly distributed among cytogenetic subgroups, as we and others have shown that BTG1 deletions are strongly enriched in ETV6RUNX1-positive leukemia as well as BCR-ABL1-positive ALL.30,31,38 The presence of these lesions in such distinct BCP-ALL subgroups may relate to the fact that deletions of IKZF1 and BTG1 appear to be the result of illegitimate RAG recombination,30 as is the case for several of the other

A

B

Figure 4. Glucocorticoid resistance of B cells isolated from Btg1 knockout mice intercrossed with haplodeficient Ikzf1 animals. (A) Splenic B cells isolated from wild-type, Ikzf1+/-, Btg1-/- and Btg1-/-;Ikzf1+/- mice were activated by LPS for 48 hours and subsequently treated in vitro for 48 hours with increasing concentrations of prednisolone (PRED, left panel) or dexamethasone (DEX, right panel) and analyzed by MTS assay (n=6). All values were normalized to untreated (UT) B cells. Error bars represent Âą standard error of the mean (SEM). P-values were calculated based on the differences of the best-fit curve using two-way ANOVA. (B) AnnexinV/7AAD staining of WT, Ikzf1+/-, Btg1-/- and Btg1-/-;Ikzf1+/- B-lymphocytes after 2 mM prednisolone or 5 mM dexamethasone treatment for 48 hours (n=4). The fraction of AnnexinV-positive cells was determined. Data represent means, and error bars represent SEM. P-values (two-sided t-test) are indicated. ***P<0.001.

haematologica | 2017; 102(3)

549

B. Scheijen et al.

commonly deleted genes, such as EBF1 and PAX5. In addition to these earlier observations, we find a specific cooccurrence of BTG1 and IKZF1 gene deletions across cytogenetic subtypes, suggesting that the combined loss of BTG1 and IKZF1 may actively contribute to leukemogenesis. Previous studies, mostly carried out in mouse models, revealed that Ikzf1 and Arf alterations in BCR-ABL1-positive ALL synergize to promote the development of leukemia by conferring stem cell-like properties.39 This leads us to hypothesize that the preponderance of BTG1;IKZF1 double-deletions in this particular subgroup may have similar consequences, although this remains to be assessed in well-established BCR-ABL1-positive mouse models. As BTG1 and IKZF1 deletions also (co)occur in lymphoid blast crises of chronic myeloid leukemia (CML),7,31 it will be interesting to study if and how the combined loss of BTG1 and IKZF1 drive the progression of this disease. Of all the common copy number losses analyzed in our study, including CDKN2A/B, PAX5, EBF1, and RB1, only loss of BTG1 appears to worsen the outcome of IKZF1-deleted ALL. Our data are consistent with the findings of Moorman et al. showing that specific combinations of different deletions impact the outcome in BCP-ALL.40 A number of different knockout mouse models have provided insight into the role of commonly deleted transcription factors in early hematopoiesis and spontaneous tumor incidence. It is evident that several of these transcriptional regulators play an important role as lymphoid specification factors and are essential for normal lymphopoiesis.41-44 However, in the mouse, loss of these early B-cell transcription factors affected in BCP-ALL, such as E2A45 and IKZF1,46 gives rise to T-cell malignancies. E2adeficient tumors are characterized by a strong increase in c-Myc expression,45 an oncogene known to promote the development of T-cell lymphomas.47 Similarly, while IKZF1 deletions predominantly occur in human BCP-ALL, heterozygous Ikzf1 knockout and dominant-negative Ikzf1 mice develop T-cell malignancies exclusively, which has been attributed to activation of the Notch pathway.36,48 In our studies we observed a lower incidence of T-cell leukemia in mice as compared to what has been reported for some other genetically engineered Ikzf1 mouse models, where expression of dominant-negative isoforms or hypomorphic knockout alleles of Ikzf1 yielded a higher susceptibility to T-cell malignancies.35,36,48 Similar to mice heterozygous knockout for Ikzf1, Btg1 knockout mice develop T-cell leukemia, while BTG1 deletions are almost exclusively found in human BCP-ALL.30 However, consis-

References 1. Hunger SP, Mullighan CG. Redefining ALL classification: toward detecting high-risk ALL and implementing precision medicine.

Blood. 2015;125(26): 3977-3987.

2. Mullighan CG, Su X, Zhang J, et al.

Deletion of IKZF1 and prognosis in acute lymphoblastic leukemia. N Engl J Med. 2009;360(5):470-480.

3. Kuiper RP, Waanders E, van der Velden VH, et al. IKZF1 deletions predict relapse in uniformly treated pediatric precursor B-ALL. Leukemia. 2010;24(7):1258-1264.

550

tent with our finding that monoallelic BTG1 deletions are enriched in human BCP-ALL cases with IKZF1 aberrations, Btg1+/-;Ikzf1+/- mice are more prone to develop leukemia relative to Btg1+/- single knockout mice (P=0.013).

In addition, we observed a significant acceleration in the onset of disease in Btg1-/-;Ikzf1+/- mice as compared to Btg1-/- (P=0.026) or Ikzf1+/- mice (P=0.011), indicating that loss of these tumor suppressor genes cooperates during leukemogenesis. Genomic DNA analyses further indicate that both the wild-type Btg1 allele and Ikzf1 allele are maintained in the Btg1+/-;Ikzf1+/- leukemias (data not shown), arguing that Btg1 and Ikzf1 dosage contribute to leukemia development. These data confirm that BTG1 acts as a tumor suppressor gene that cooperates with IKZF1 loss during leukemia development.

Another important finding in this study is that BTG1 deletions define a high-risk group within the IKZF1-deleted subtype. Our finding that loss of BTG1 specifically enhances GC resistance mediated by Ikzf1-haplodeficiency implies that the prognostic value of BTG1 and IKZF1 deletions could be dependent on the upfront treatment and dose of synthetic glucocorticoids used. However, this remains to be established in future studies. The relation between BTG1 deletions and inferior outcome was recently confirmed with the analyses of a relapsed BCP-ALL cohort, showing that BTG1 and NR3C1 deletions were associated with a higher risk of disease progression.49 Collectively, our data demonstrate that BTG1 is a prognostic factor and regulator of the GC response, particularly in the context of IKZF1-deletions.

Acknowledgments The authors would like to thank Marieke von Lindern and Meinrad Busslinger for providing the Btg1 and Ikzf1 knockout mice, respectively. We would like to thank Arian van der Veer for performing the MLPA experiments.

Funding This work was supported by grants of the Stichting Kinderen Kankervrij (KiKa 2009-55; KiKa 2010-77; KiKa 2014-132) and Stichting KOC Nijmegen. Work on the validation cohort was supported by NHMRC Australian APP1057746. MLdB has been supported by grants from the Dutch Cancer Society KWF (AMC 2008-4265), the Paediatric Oncology Foundation Rotterdam, the European Union's Seventh Framework Program (FP7/2007-2013 ENCCA grant HEALTH-F2-2011-261474); RPK and PH have been supported by grants from the Dutch Cancer Society KWF (KUN 2009-4298); EW is supported by a fellowship from the Dutch Cancer Society KWF.

4. Waanders E, van der Velden VH, van der Schoot CE, et al. Integrated use of minimal residual disease classification and IKZF1 alteration status accurately predicts 79% of relapses in pediatric acute lymphoblastic leukemia. Leukemia. 2011;25(2):254-258.

5. van der Veer A, Waanders E, Pieters R, et al.

Independent prognostic value of BCRABL1-like signature and IKZF1 deletion, but not high CRLF2 expression, in children with B-cell precursor ALL. Blood. 2013;122 (15):2622-2629.

6. Marke R, Havinga J, Cloos J, et al. Tumor suppressor IKZF1 mediates glucocorticoid

resistance in B-cell precursor acute lymphoblastic leukemia. Leukemia. 2016;30(7): 1599-1603.

7. Mullighan CG, Miller CB, Radtke I, et al.

BCR-ABL1 lymphoblastic leukaemia is characterized by the deletion of Ikaros.

Nature. 2008;453(7191):110-114.

8. van der Veer A, Zaliova M, Mottadelli F, et al. IKZF1 status as a prognostic feature in BCR-ABL1-positive childhood ALL. Blood.

2014;123(11):1691-1698.

9. Den Boer ML, van Slegtenhorst M, De Menezes RX, et al. A subtype of childhood acute lymphoblastic leukaemia with poor

haematologica | 2017; 102(3)

Combined BTG1 and IKZF1 deletions affect outcome

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

treatment outcome: a genome-wide classification study. Lancet Oncol. 2009; 10(2):125-134.

Zhang J, Mullighan CG, Harvey RC, et al.

Key pathways are frequently mutated in high-risk childhood acute lymphoblastic leukemia: a report from the Children's Oncology Group. Blood. 2011;118(11): 3080-3087.

Roberts KG, Morin RD, Zhang J, et al.

Genetic alterations activating kinase and cytokine receptor signaling in high-risk acute lymphoblastic leukemia. Cancer Cell.

2012;22(2):153-166.

Roberts KG, Pei D, Campana D, et al.

Outcomes of children with BCR-ABL1-like acute lymphoblastic leukemia treated with risk-directed therapy based on the levels of minimal residual disease. J Clin Oncol.

2014;32(27):3012-3020.

Dail M, Li Q, McDaniel A, et al. Mutant Ikzf1, KrasG12D, and Notch1 cooperate in T lineage leukemogenesis and modulate responses to targeted agents. Proc Natl Acad Sci U S A. 2010;107(11):5106-5111.

Virely C, Moulin S, Cobaleda C, et al.

Haploinsufficiency of the IKZF1 (IKAROS) tumor suppressor gene cooperates with BCR-ABL in a transgenic model of acute lymphoblastic leukemia.

Leukemia.

2010;24(6):1200-1204.

Clappier E, Auclerc MF, Rapion J, et al. An intragenic ERG deletion is a marker of an oncogenic subtype of B-cell precursor acute lymphoblastic leukemia with a favorable outcome despite frequent IKZF1 deletions.

Leukemia. 2014;28(1):70-77.

Zaliova M, Zimmermannova O, Dorge P, et al. ERG deletion is associated with CD2 and attenuates the negative impact of IKZF1 deletion in childhood acute lymphoblastic leukemia. Leukemia. 2014;28(1): 182-185.

Kuiper RP, Schoenmakers EF, van Reijmersdal SV, et al. High-resolution genomic profiling of childhood ALL reveals novel recurrent genetic lesions affecting pathways involved in lymphocyte differentiation and cell cycle progression.

Leukemia. 2007;21(6):1258-1266.

Mullighan CG, Goorha S, Radtke I, et al.

Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia.

Nature. 2007 12;446(7137):758-764.

Tijchon E, Havinga J, van Leeuwen FN, Scheijen B. B-lineage transcription factors and cooperating gene lesions required for leukemia development.

Leukemia.

2013;27(3):541-552.

Hauer J, Mullighan C, Morillon E, et al. Loss of p19Arf in a Rag1(-/-) B-cell precursor population initiates acute B-lymphoblastic leukemia. Blood. 2011;118(3): 544-553.

Engel I, Murre C. Disruption of pre-TCR expression accelerates lymphomagenesis in E2A-deficient mice. Proc Natl Acad Sci U S A. 2002;99(17):11322-11327.

haematologica | 2017; 102(3)

22. Prasad MA, Ungerback J, Ahsberg J, et al.

Ebf1 heterozygosity results in increased DNA damage in pro-B cells and their synergistic transformation by Pax5 haploinsufficiency. Blood. 2015;125(26):4052-4059.

23. Marshall GM, Dalla Pozza L, Sutton R, et al. High-risk childhood acute lymphoblastic leukemia in first remission treated with novel intensive chemotherapy and allogeneic transplantation.

Leukemia.

2013;27(7):1497-1503.

24. Sutton R, Shaw PJ, Venn NC, et al.

Persistent MRD before and after allogeneic BMT predicts relapse in children with acute lymphoblastic leukaemia. Br J Haematol.

2015 ;168(3):395-404.

25. Gray RJ. cmprsk: Subdistribution Analysis of Competing Risks. R package version 2.26 2013 [Available from: http://CRAN.Rproject.org/package=cmprsk].Last accessed: 12th January 2017.

26. de Wreede LC, Fiocco M. Putter, H. mstate: An R Package for the Analysis of Competing Risks and Multi-State Models.

Stat Med. 2011;38(7):1-30.

27. Therneau T. A Package for Survival Analysis in S. R package version 2.36-12.

2012.

28. Farioli-Vecchioli S, Micheli L, Saraulli D, et al. Btg1 is Required to Maintain the Pool of Stem and Progenitor Cells of the Dentate Gyrus and Subventricular Zone. Front Neurosci. 2012;6:124.

29. Souabni A, Cobaleda C, Schebesta M, Busslinger M. Pax5 promotes B lymphopoiesis and blocks T cell development by repressing Notch1. Immunity. 2002; 17(6):781-793.

30. Waanders E, Scheijen B, van der Meer LT, et al. The origin and nature of tightly clustered BTG1 deletions in precursor B-cell acute lymphoblastic leukemia support a model of multiclonal evolution. PLoS Genet. 2012;8(2):e1002533.

31. Xie J, Wang Q, Wang Q, et al. High frequency of BTG1 deletions in patients with BCR-ABL1-positive acute leukemia. Cancer Genet. 2014;207(5):226-230.

32. Smith M, Arthur D, Camitta B, et al.

Uniform approach to risk classification and treatment assignment for children with acute lymphoblastic leukemia. J Clin Oncol. 1996;14(1):18-24.

33. Tijchon E, van Emst L, Yuniati L, et al.

Tumor suppressors BTG1 and BTG2 regulate early mouse B-cell development.

Haematologica. 2016;101(7):e272-276.

34. Haines DC, Chattopadhyay S, Ward JM. Pathology of aging B6;129 mice. Toxicol Pathol. 2001;29(6):653-61.

35. Winandy S, Wu P, Georgopoulos K. A dominant mutation in the Ikaros gene leads to rapid development of leukemia and lymphoma. Cell. 1995;83(2):289-99.

36. Mantha S, Ward M, McCafferty J, t al.

Activating Notch1 mutations are an early

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

event in T-cell malignancy of Ikaros point mutant Plastic/+ mice. Leuk Res. 2007;31(3):321-327. van Galen JC, Kuiper RP, van Emst L, et al.

BTG1 regulates glucocorticoid receptor autoinduction in acute lymphoblastic leukemia. Blood. 2010;115(23):4810-4819.

Tsuzuki S, Karnan S, Horibe K, et al.

Genetic abnormalities involved in t(12;21) TEL-AML1 acute lymphoblastic leukemia: analysis by means of array-based comparative genomic hybridization. Cancer Sci. 2007;98(5):698-706.

Churchman ML, Low J, Qu C, et al.

Efficacy of Retinoids in IKZF1-Mutated BCR-ABL1 Acute Lymphoblastic Leukemia. Cancer Cell. 2015;28(3):343356.

Moorman AV, Enshaei A, Schwab C, t al. A novel integrated cytogenetic and genomic classification refines risk stratification in pediatric acute lymphoblastic leukemia.

Blood. 2014;124(9):1434-1444.

Georgopoulos K, Bigby M, Wang JH, t al.

The Ikaros gene is required for the development of all lymphoid lineages. Cell.

1994;79(1):143-156.

Wang JH, Nichogiannopoulou A, Wu L, et al. Selective defects in the development of the fetal and adult lymphoid system in mice with an Ikaros null mutation.

Immunity. 1996;5(6):537-549.

Engel I, Johns C, Bain G, Rivera RR, Murre C. Early thymocyte development is regulated by modulation of E2A protein activity.

The Journal of experimental medicine.

2001;194(6):733-745.

Kee BL, Bain G, Murre C. IL-7Ralpha and E47: independent pathways required for development of multipotent lymphoid progenitors. The EMBO journal. 2002;21(12):103-13.

Bain G, Engel I, Robanus Maandag EC, et al. E2A deficiency leads to abnormalities in alphabeta T-cell development and to rapid development of T-cell lymphomas. Mol Cell Biol. 1997;17(8):4782-4791.

Kastner P, Chan S. Role of Ikaros in T-cell acute lymphoblastic leukemia. World J Biol Chem. 2011;2(6):108-114.

Stewart M, Cameron E, Campbell M, et al.

Conditional expression and oncogenicity of c-myc linked to a CD2 gene dominant control region. Int J Cancer. 1993;53(6):10231030.

Dumortier A, Jeannet R, Kirstetter P, et al.

Notch activation is an early and critical event during T-Cell leukemogenesis in Ikaros-deficient mice. Mol Cell Biol.

2006;26(1):209-220.

Irving JA, Enshaei A, Parker CA, et al.

Integration of genetic and clinical risk factors improves prognostication in relapsed childhood B-cell precursor acute lymphoblastic leukemia. Blood. 2016; 128 (7):911-922.

551

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Lymphoblastic Leukemia

Ferrata Storti Foundation

Population pharmacokinetics of intravenous Erwinia asparaginase in pediatric acute lymphoblastic leukemia patients

Sebastiaan D.T. Sassen,1 Ron A.A. Mathôt,2 Rob Pieters,3 Robin Q.H. Kloos,1 Valérie de Haas,4 Gertjan J.L. Kaspers,3,5 Cor van den Bos,6 Wim J.E. Tissing,7 Maroeska te Loo,8 Marc B. Bierings,9 Wouter J.W. Kollen,10 Christian M. Zwaan1 and Inge M. van der Sluis1 CMZ and IMvdS contributed equally to this work.

Department of Pediatric Oncology, Erasmus MC-Sophia Children’s Hospital, Rotterdam; Department of Hospital Pharmacy, Academic Medical Center, University of Amsterdam; Princess Máxima Center for Pediatric Oncology, Utrecht; 4Dutch Childhood Oncology Group, The Hague; 5Department of Pediatric Oncology, VU University Medical Center, Amsterdam; 6Department of Pediatric Oncology, Emma Children’s Hospital, Academic Medical Center, Amsterdam; 7Department of Pediatric Oncology, Beatrix Children’s Hospital, University Medical Center, Groningen; 8Department of Pediatric HematoOncology, Radboud University Nijmegen Medical Center; 9Pediatric Blood and Marrow Transplantation Program, University Medical Center Utrecht/Wilhelmina Children’s Hospital and 10Department of Pediatric Immunology, Hemato-Oncology and Stem Cell Transplantation, Leiden University Medical Center, the Netherlands 1 2 3

Haematologica 2017 Volume 102(3):552-561

ABSTRACT

Correspondence: i.vandersluis@erasmusmc.nl

Received: May 17, 2016.

Accepted: October 27, 2016.

Pre-published: November 10, 2016. doi:10.3324/haematol.2016.149195 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/552 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

552

E

rwinia asparaginase is an important component in the treatment of pediatric acute lymphoblastic leukemia. A large variability in serum concentrations has been observed after intravenous Erwinia asparaginase. Currently, Dutch Childhood Oncology Group protocols dose alterations are based on trough concentrations to ensure adequate asparaginase activity (≥100 IU/L). The aim of this study was to describe the population pharmacokinetics of intravenous Erwinia asparaginase to quantify and gather insight into inter-individual and inter-occasion variability. The starting dose was evaluated on the basis of the derived population pharmacokinetic parameters. In a multicenter prospective observational study, a total of 714 blood samples were collected from 51 children (age 1-17 years) with acute lymphoblastic leukemia. The starting dose was 20,000 IU/m2 three times a week and adjusted according to trough levels from week three onwards. A population pharmacokinetic model was developed using NONMEM®. A 2-compartment linear model with allometric scaling best described the data. Inter-individual and interoccasion variability of clearance were 33% and 13%, respectively.

Clearance in the first month of treatment was 14% higher (P<0.01).

Monte Carlo simulations with our pharmacokinetic model demonstrated that patients with a low weight might require higher doses to achieve similar concentrations compared to patients with high weight. The current starting dose of 20,000 IU/m2 might result in inadequate concentrations, especially for smaller, lower weight patients, hence dose adjustments based on individual clearance are recommended. The protocols were approved by the institutional review boards. (Registered at NTR 3379 Dutch Trial Register; www.trialregister.nl).

haematologica | 2017; 102(3)

Pop PK of i.v. Erwinia asparaginase in pediatric ALL patients

Introduction Asparaginase is an enzyme that catalyzes the hydrolysis of asparagine to aspartic acid and ammonia. Leukemic cells rely on exogenous supplies of asparagine for their protein synthesis. Hence the depletion of asparagine results in cell death.1 Asparaginase has become an important component in the treatment of pediatric acute lymphoblastic leukemia (ALL), and therefore every effort should be made to expose the patient to the protocol-prescribed cumulative dose.2-6 Erwinia asparaginase is derived from Erwinia chrysanthemi, whereas the other forms of asparaginase [native Escherichia coli (E. coli) and polyethylene glycol (PEG)-asparaginase] are derived from E. coli.

Currently the E. coli derivatives are the first choice in treatment naïve patients, and the use of Erwinia asparaginase is indicated in patients who develop hypersensitivity to the E. coli-derived asparaginase or in the case of silent inactivation of E. coli asparaginase.4,7,8 Little is known about the pharmacokinetics (PK) of Erwinia asparaginase, especially after intravenous administration. Currently, in the Dutch Childhood Oncology Group (DCOG) ALL-11 protocol (and the preceding ALL10), the interval and/or dose alterations of Erwinia asparaginase are based on therapeutic drug monitoring (TDM), with the aim of keeping the trough asparaginase activity above the 100 IU/L threshold, which leads to complete asparagine depletion.7,9-13 A concentration of 100 IU/L is considered safe concerning asparagine depletion; however, complete depletion has been observed at lower concentrations.14-16 Currently, the consensus for the target threshold is 100 IU/L.17 No evidence-based guidelines for increasing or decreasing the dose are available and doseadaptations are based on empirical knowledge.

The PK data of different asparaginase forms are not transferrable. Intramuscular native E. coli asparaginase has an elimination half-life of 1.3 days,18 and intravenous recombinant and native E. coli asparaginase 17.3-19.0 hours (h),19,20 and follow linear elimination, whereas E. coli PEG-asparaginase has time-dependent pharmacokinetics and a half-life with an observed range of 2.4-11.8 days.9,18,21,22 Erwinia asparaginase has the shortest half-life with means of 12.6-22.1 h after intramuscular administration,7,9,23,24 and 6.4-7.6 h after intravenous administration.23,25 The aim of the present study was to describe the population PK of intravenously administered Erwinia asparaginase in pediatric ALL patients, to quantify the random parameters inter-individual (IIV), inter-occasion (IOV) and residual variability, and to determine the association between patients’ characteristics and the PK parameters.

Quantification of random parameters is important for proper TDM, as IIV can be compensated by TDM, whereas IOV cannot. In addition, the current starting dose was evaluated considering a trough concentration above 100 IU/L.

2004-April 1, 2012) or ALL-11 (since April 1, 2012 and ongoing) protocols with Erwinia asparaginase (Erwinase®; EUSA Pharma, Europe) after the development of an allergy to E. coli-derived asparaginase or silent inactivation of E. coli-derived asparaginase.

The starting dose of Erwinia asparaginase was 20,000 IU/m2 intravenously (i.v.) over 1 h thrice weekly (Mon/Wed/Fri). The dose was fixed for the first two weeks. Subsequently, if the 72 h concentration was more than 100 IU/L, the dose interval was adjusted to twice weekly. Additionally in ALL-11, the dose was increased based on clinical expertise if the 72 h concentration was less than 100 IU/L. When this was not sufficient, the dose interval was set to every other day. The TDM samples were collected between May 1, 2009 and February 5, 2015. These are trough samples prior to their next dose, predominantly 48 or 72 h after the last Erwinia asparaginase administration. For the purpose of this study, additional peak concentrations were collected between 1 and 4 h in a subset of patients. Erwinia asparaginase activity concentrations in serum were analyzed as previously described by us, with a lower limit of quantification (LLoQ) of less than 5 IU/L.26 The protocols were approved by a central Institutional Review Board.

Pharmacokinetic analysis A detailed description of the pharmacokinetic analysis methods is provided in the Online Supplementary Appendix. Time profiles of log transformed Erwinia asparaginase concentrations were analyzed using the non-linear effects modeling approach implemented in non-linear mixed effects modeling (NONMEM). The data were initially fitted to a 1-compartment linear model without an absorption compartment. Subsequently more complex models were evaluated; improvement of the fit was evaluated by the precision of the estimated PK parameters, change in the objective function values (OFV), goodness-of-fit plots (GOF), and visual predictive checks (VPC). A 3.84-point decrease in OFV for one degree of freedom was considered a significant improvement (P<0.05).

The data were obtained in a pediatric population, hence PK parameters were allometrically scaled to adequately describe the parameters across a wide range of body weights. For allometric scaling, standard fixed exponent values of 0.75 for the flow dependent physiological process parameters and 1 for volumerelated parameters were used.27-29 Inter-patient and inter-occasion variability in clearances and volumes of distribution were characterized with exponential models. An additive error model was used to describe the residual error in plasma concentrations.

After the finalization of the structural model, covariate models were built by a stepwise forward inclusion procedure. The covariate with the greatest reduction in OFV was added to the base model. This was iterated over all the covariates until no statistically significant decrease in OFV occurred. For the internal validation of the model, non-parametric bootstrap procedures (n=1000) were performed and VPCs were obtained. The final model including covariates was used to perform Monte Carlo simulations (n=5000) for different doses and patient weight.

Results Patients and samples

Methods Patients’ characteristics and treatment The study was designed as a prospective multicenter study in 7 pediatric oncology centers in the Netherlands. Children aged 1-18 years with acute lymphoblastic leukemia (ALL) were eligible for the study when treated according to the DCOG ALL-10 (Nov 1, haematologica | 2017; 102(3)

During the study period, 53 patients were switched from E. coli-derived asparaginase to Erwinia asparaginase due to allergic reactions or silent inactivation. Data from 51 of these 53 patients were included in the PK analysis.

Two patients were excluded due to Erwinia asparaginase concentrations below the LLoQ. One of these patients had anti-Erwinia asparaginase antibodies neutralizing the drug and prohibiting sufficient exposure, for the other patient 553

S.D.T. Sassen et al.

the reason for unmeasurable concentrations is not known.

Both patients discontinued treatment with Erwinia asparaginase. A summary of patients’ characteristics is shown in Table 1. A total of 741 samples were available, with a median of 11 samples per patient (range 2-43 samples). Twenty-seven samples (3.6%) were excluded from the analysis due to: missing sampling data (n=20), no measurable asparaginase (n=4), or unrealistic high concentrations (n=3) due to sampling artefacts. The four unmeasurable trough concentrations were all from the same patient; however, this patient did have measurable Erwinia asparaginase concentrations within 24 h after administration.

Samples were collected for two weeks up to 12 months after the start of Erwinia asparaginase treatment. Samples were predominantly trough concentrations taken around 48 h (52.2%) and 72 h (36.8%) after Erwinia asparaginase administration. Figure 1 shows the combined Erwinia asparaginase concentrations versus the time after dose for all patients throughout therapy, demonstrating a large variability. The concentrations can be stratified for the first two weeks of treatment (no dose adjustments), and after two weeks with potential adjustments (TDM) in Erwinia asparaginase dose frequency (both ALL-10 and ALL-11 protocol) and dose (ALL-11 protocol only). The median asparaginase trough concentrations two days after administration (range 42-50 h) for, respectively, the first two weeks and during TDM, were 166.2 IU/L [interquartile range (IQR) 103.4-270.1] and 191.0 IU/L (IQR 115.0296.5); 75.4% and 82.6% of the patients had asparaginase trough concentrations of 100 IU/L or more; 90.16% and 91.30% had trough concentrations of 50 IU/L or more.

Three days after administration (65-80 h), the median trough concentrations for, respectively, the first two weeks and during TDM, were 48.4 IU/L (IQR 29.9-104.7) and 83.7 IU/L (IQR 49.5-98.7); 26.4% and 39.5% of the patients had trough concentrations of 100 IU/L or more; 50% and 74.3% had trough concentrations of 50 IU/L or more.

The summary of the number of samples and time points can be found in Table 2. A total of 311 samples (43.6%) were collected in the first month of Erwinia asparaginase treatment. The number of samples during the following months ranged from 86 (in month 2) to 1 (in month 12).

Especially trough concentrations taken at 72 h frequently dropped below the desired therapeutic target threshold of 100 IU/L. Eleven patients (21.6%) were switched from thrice to twice weekly interval after asparaginase 72 h trough concentrations of more than 100 IU/L during their treatment. A total of 117 samples (15.5%) were drawn during a twice-weekly interval.

Pharmacokinetic model Both 1- and 2-compartment models were evaluated.

The estimated PK parameters were normalized to a weight of 70 kg using the ¾ allometric model. The 2-compartment model provided a better fit to the data than a 1compartment model based on the OFV and the goodnessof-fit plots. The OFV significantly decreased 127.2 points from 366.5 to 239.3 (P<0.001). A slight decrease in additive error was observed from 0.70 IU/L (1-compartment model) to 0.64 IU/L (2-compartment model). Addition of a third compartment did not improve the model. A Michaelis-Menten elimination model did not improve the model either.

554

The estimated IIV on CL was 36%, whereas this parameter could not be estimated for distribution of the central compartment (Vc), intercompartmental clearance (Q) and distribution of the peripheral compartment (Vp).

Complete removal of the a priori incorporated allometric scaling from the model, based on body weight (standardized for 70 kg) and fixed exponents, resulted in a worse model with a 8.3 points increase in OFV and an increase of the IIV for CL from 33% to 40%. Samples were collected throughout therapy on different occasions. Addition of IOV resulted in an improvement of the population model with an estimated value of 14%. By incorporation of the IOV, the OFV decreased with 44.9 points, additive error decreased from 0.64 to 0.57, and the IIV for CL decreased from 36% to 33%. The shrinkage was 6% for IIV on CL, 32% for IOV on CL and 9% for residual variability. This was considered the structural model and was used for the stepwise forward inclusion of covariates.

All covariates were tested one at a time for improvement of the structural model. The clearance in the first month was 14% higher in comparison to the subsequent months with a decrease of 17.0 points in OFV (P<0.001).

Dose as a covariate on CL also improved the model (P<0.05). However, during TDM, the dose is adjusted according to the patients’ asparaginase concentrations and Table 1. Patients’ characteristics.

Patients characteristics Total patients, n Age, y Median Range Sex, n, % Male Female Weight, kg Median Range BSA m2 Median Range

51 6 1.9-17.7 32 (62.7%) 19 (37.3%) 24.5 11.7-99.0 0.92 0.53-2.22

y: years; kg: kilograms; BSA: body surface area.

Table 2. Samples for pharmokinetic analysis.

Samples Total samples Evaluable Not evaluable Pre-TDM During TDM Samples per patient Median Range Sample time* Within 5 h 5-40 h After 48 h After 72 h 80-120 h

N (%) 741 714 (96.4%) 27 (3.6%) 225 (31.5%) 489 (68.5%) 11 2-43 17 (2.4%) 9 (1.3%) 373 (52.2%) 263 (36.8%) 52 (7.3%)

TDM: therapeutic drug monitoring; h: hours.*Sample time is the time after last Erwinia asparaginase administration in hours.

haematologica | 2017; 102(3)

Pop PK of i.v. Erwinia asparaginase in pediatric ALL patients

therefore indirectly for their clearance, which explains the correlation between dose and clearance. Similarly, an association between dose interval and clearance was detected.

With dose interval as covariate, the OFV significantly decreased 6.11 points (P<0.05) and patients with twice weekly dosing (n=11) were associated with a 24% lower CL in comparison with patients on thrice weekly dosing (n=40). Due to TDM, dose and dose interval were not incorporated in the model. The other covariates [age, weight, height, BSA, sex, treatment protocol (ALL-10 or ALL-11) and treatment center] did not result in a significant improvement of the base model.

The final model was a 2-compartment model with fixed exponents allometric scaling, a correction factor for increased clearance in the first month, and inter-individual and inter-occasion variability on clearance. The parameter estimates of the final model were: CL 0.44 L/h/70kg, Vc (central compartment) 3.2 L/70kg, Q (intercompartmental clearance) 0.15 L/h/70kg and Vp (peripheral compartment) 2.9 L/70kg. The calculated half-lives (t1/2) for the 2-compartment model are T1/2,ι 3.5 h and T1/2,β 19.6 h, which represent respectively the distribution phase and elimination phase. The inter-individual variability of clearance was 33%. There was an inter-occasion variability of 13% based on monthly intervals (Table 3). The basic goodnessof-fit plots (Figure 2) show good model performance.

Individual predictions versus observation are well distributed around the unity line. The weighted residuals are within a good range (2 to -2) and evenly distributed. The population predictions show a deviation from the unity line for the lower concentrations where limited samples were available.

Model validation The non-parametric bootstrap procedure was per-

formed to test the robustness of the model. A total 998 of the 1000 runs were successful. The results are shown in Table 3. The estimates of the final model are in accordance with the results from the 1000 bootstrap replicates. The plot of the prediction corrected visual predictive check (pcVPC) shows the median and 90% interval of the observed asparaginase concentrations (Figure 3). The model adequately predicts the time course of the asparaginase plasma concentration during the first three time frames (0-36 h, 36-60 h and 60-84 h). However, an underprediction of median concentrations and 5th and 95th percentiles was observed in the 84-118 h timeframe.

Evaluation of the starting dose To investigate the appropriateness of the starting dose, Monte Carlo simulations with varying weights and doses were performed utilizing the developed population PK model. Based on the simulations, patients with a lower body weight appeared to require higher weight-normalized starting doses to achieve sufficient Erwinia asparaginase concentrations after 48 h. Patients weighing 100 kg require 500 IU/kg compared to doses exceeding 1000 IU/kg for patients with a body weight below 20 kg in order to have Erwinia asparaginase concentrations of 100 IU/L or more in 75% of the patients at 48 h after administration (Figures 4 and 5). With the current starting dose of 20,000 IU/m2, approximately 75% of the patients with a body weight of more than 50 kg would have concentrations of 100 IU/L or more after 48 h. For patients with a body weight between 30-50 kg, the suggested starting dose would be 25,000 IU/m2, and for patients with a body weight of 10-30 kg doses of 25,000-37,000 IU/m2 to achieve Erwinia asparaginase concentrations of 100 IU/L or more after 48 h in at least 75% of the patients after the first dose. To achieve this in 90% of the patients, starting

Figure 1. Asparaginase concentration (IU/L) versus time after dose in hours (h) for all patients and all occasions on a semi-log plot. This shows the large inter-patient variability in plasma concentrations as collected throughout treatment. Triangles (black) show concentrations prior to possible dose adjustments (first 2 weeks) and circles (blue) show observed concentrations with possible dose adjustments [therapeutic drug monitoring (TDM) after week two].

haematologica | 2017; 102(3)

555

S.D.T. Sassen et al.

Figure 2. Goodness-of-fit plots final model. (Upper left) Predicted population concentrations versus observed concentrations of the final model. (Upper right) Predicted individual concentrations versus observed concentrations of the final model. (Lower left) Individual weighted residuals versus individual predictions, (lower right) conditional weighted residuals versus time after dose. h: time in hours; IWRES: individual weighted residual predictions.

dose for all patients would have to be above 33,000 IU/m2.

The dose in IU/kg was converted to IU/m2 utilizing corresponding BSA to weight for pediatric oncology patients.30 Figures 5 and 6 show required starting dose (in IU/kg and IU/m2) versus weight to achieve 48 h-trough concentration of 100 IU/L or more (Figure 5) and of 50 IU/L or more (Figure 6) in a percentage of patients ranging from 10% to 90%. The simulations with the final PK model are in accordance with the observations made in the patient data. Starting doses of 36 patients were evaluable (Erwinia asparaginase dose of 20,000 IU/m2 and an available 48 h sample after administration of the first dose). With respect to weight, 12 of 31 patients (39%) weighing less than 50 kg and and 1 of 5 patients (20%) weighing 50 kg or more had 48 h trough concentrations below 100 IU/L. Although these numbers are small, it shows the same trend in weight-concentration relationship as the Monte Carlo simulations.

Discussion The PK of Erwinia asparaginase was best described with a 2-compartment model with linear elimination and, therefore, more similar to native asparaginase than PEGylated asparaginase which has time-dependent elimination, probably due to the polyethylene glycol.21 There appears to be a negative correlation between weight and weight-normalized clearance, where the patients with a lower weight require higher weight-normalized doses based on Monte Carlo simulations with the final PK model. The same trend was observed in the actual patient data; however, only a small number of patients had a body 556

weight of more than 50 kg. In addition, clearance in the first month was significantly higher.

Asparaginase is an important component in the treatment of pediatric ALL, where it contributes 10%-20% to the total treatment outcome.6,11,12,19,31,32 However, treatment with asparaginase at suboptimal dose schedules leads to an inferior outcome.11-13,32 Asparaginase is available in different molecular forms (eg. PEGylated or native). The PK properties of these different forms are, however, not similar, and therefore can not be used interchangeably.9,18,33 Erwinia asparaginase has a shorter half-life in comparison with the other asparaginase forms; this results in lower concentrations and exposure when administered at equal dose schedules and, consequently, worse outcome.11-13 In North-America, intramuscular injection of Erwinia asparaginase was the only Food and Drug Administration (FDA)-approved method of administration, but this has recently been extended with i.v. administration.34 In Europe, the predominant method of Erwinia asparaginase administration is i.v. administration. Several PK studies of intramuscular administration have been published.8,9,23,24 This is important in studying the PK because, in addition to the asparaginase molecule or pharmaceutical form, the route of administration may also influence the PK. The PK of intramuscular-administered Erwinia asparaginase differs from i.v. administration due to the presence of a ratelimiting step in the absorption phase.23,24,34 Bypassing the absorption from the muscle will result in faster elimination and probably more predictable concentrations, as variability in absorption rate is eliminated. The calculated terminal half-life was 19.6 h, which is similar to the terminal half-life of the i.v.-administered native E. coli asparaginase of 19.0 h.22 Previously published studies with i.v. haematologica | 2017; 102(3)

Pop PK of i.v. Erwinia asparaginase in pediatric ALL patients

Figure 3. Visual predictive check. Prediction corrected (Pred Corr) visual prediction plot of observed log asparaginase concentrations versus time after last Erwinia asparaginase dose in hours (h) of the final model. Red solid line shows the median observed concentrations and the surrounding opaque red area the simulation based 95% interval for the median. Red dashed lines indicate the observed 5% and 95% percentiles; surrounding opaque blue areas show the simulated 95% confidence intervals for the corresponding predicted percentiles.

Table 3. Population parameter estimates and non-parametric bootstrap.

Estimate

CL (L/h/70kg) Vc (L/70kg) Q (L/h/70kg) Vp (L/70kg) CL month 1 diff IIV CL (%) IOV (%) Residual error (IU/L)

0.44 3.2 0.15 2.9 1.1 33 13 0.57

Parameter estimates RSE (%) 95%CI 95%CI (lower) CI (upper) 11% 22% 44% 35% 3% 20 18 6.4%

0.35 1.8 0.02 0.9 1.06 20 8 0.50

0.53 4.6 0.28 5.0 1.22 45.2 17.2 0.65

Shrink (%) 5.5 31.9 8.5

Median

Bootstrap 95%CI (lower)

95%CI (upper)

0.42 3.3 0.13 2.7 1.12 30 11 0.57

0.32 2.2 0.01 0.7 1.05 18 6 0.50

0.54 6.9 0.33 5.3 1.20 42 17 0.63

RSE: residual standard error; CI: confidence interval; CL: population estimate for clearance; Vc: population estimate of apparent volume of distribution in central compartment; Q: population estimate for intercompartmental clearance; Vp: population estimate of apparent volume of distribution in peripheral compartment; CL month 1 diff: population estimate difference in clearance of first month; IIV: inter-individual variability; IOV: inter-occasion variability.

Erwinia asparaginase showed a half-life of 6.4 h and 7.5 h.23,25 However, our study uses a 2-compartment model which has a fast elimination half-life of 3.5 h during the distribution phase and a slower elimination half-life of 19.6 h for the terminal elimination phase. This presents itself in a concentration time curve with a steep decline in the first phase followed by a slower decline in the second; this can also be observed in Figures 1 and 3.

During TDM, a large variability in Erwinia asparaginase concentrations was also observed after i.v.-administered Erwinia asparaginase (Figure 1). Hence the population PK of Erwinia asparaginase was studied to evaluate the elimination of Erwinia asparaginase from the body in a quantitative manner, and to explain and quantify the variability in order to improve individual dosing to achieve sufficient haematologica | 2017; 102(3)

asparaginase concentrations prior to their next dose. PKbased TDM dosing can compensate for IIV, but not for IOV. In this study, the IIV was 33% and the IOV was 13%, which is favorable for PK-based dosing.

The development of the PK model was successful despite the limited number of peak concentrations. The parameters were estimated with good precision; the shrinkage of IIV and the residual error was small. The bootstrap estimates were also in accordance with the model estimates. The VPC showed the model predictions to be in line with the observations, except for the last time frame (84-118 h), which showed a slight under-prediction (Figure 3). However, samples in this timeframe were patients who were switched to twice-weekly Erwinia asparaginase administration due to high asparaginase con557

S.D.T. Sassen et al.

A

B

Figure 4. Simulated Erwinia asparaginase concentrations for a 10 kg and a 100 kg patient. Median and the 25% and 75% percentiles of the patients who achieve asparaginase concentrations (y-axis) after 48 hours for different Erwinia asparaginase doses (xaxis). (A) Concentrations for a 10 kg patient and (B) a 100 kg patient. Red dashed line is the target trough concentration of 100 IU/L.

centrations, hence presumably characterized by a lower clearance. Calculations based on the total population would, therefore, over-predict the clearance in this group, resulting in the under-prediction of the Erwinia asparaginase concentrations. When the predictions are corrected for the dose interval, the under-prediction disappears and is in accordance with the observations (Online Supplementary Figure S1).

Monte Carlo simulations of patients with different body weights and starting doses of Erwinia asparaginase were performed aiming at trough concentrations of 100 IU/L or more. Based on the simulations, the current starting dose of 20,000 IU/m2 seems rather low, especially for children with a body weight less than 50 kg. If this starting dose is used, close monitoring of the patient is required to ensure sufficient Erwinia asparaginase concentrations. The PK model uses the standard allometric scaling based on weight, which is the gold standard for allometric scaling in population PK. Unlike weight-based scaling, it is unclear how scaling based on BSA should be implemented. The implementation of BSA in the PK model might depend on the chosen method of BSA calculation (e.g. Mosteller, Dubois & Dubois, Haycock), as these methods use different internal (exponential) correction factors.35-37 However, to our knowledge, this has still not been studied.

558

Therefore, weight-based allometric scaling was used for the development of the PK model. Monte Carlo simulations were also expressed on a per weight basis.

Additional simulations were performed with scaling based on BSA resulting in similar results. For clinical convenience, the dose was converted to IU/m2 using the corresponding BSA to weight in pediatric oncology patients for the representation of Figures 5B and 6B.30 With the registered dose of 25,000 IU/m2, approximately 75% of the patients have Erwinia asparaginase concentrations above 100 IU/L 48 h after the first dose and 90%-100% of the patients above 50 IU/L. When increasing the dose to achieve sufficient trough concentrations, one has to keep in mind that the peak concentrations (Cmax) and exposure (AUC) increase as well.

This might lead to side effects which include hypersensitivity or infusion reactions, pancreatitis, liver abnormalities, central neurotoxicities, glucose intolerance or coagulation abnormalities,24,38,39 although we showed no significant correlation between asparaginase activity concentrations and pancreatitis, thrombosis or central neurotoxicities.39 High concentrations of asparaginase were associated with high triglyceride and high cholesterol concentrations, and were more pronounced in children aged ten years or older. This might be explained by the lower weight norhaematologica | 2017; 102(3)

Pop PK of i.v. Erwinia asparaginase in pediatric ALL patients

A

B

Figure 5. Erwinia asparaginase starting dose versus patient weight to achieve 100 IU/L or more after 48 hours (h). (A) Required starting dose in IU/kg and (B) IU/m2 versus weight of patients in kilograms (kg) to achieve 100 IU/L or more after 48 h. Median (solid line), 25% and 75% percentiles (dashed line) and 10% and 90% percentiles (dotted line) of the patients with asparaginase concentrations of 100 IU/L or more with different weight (x-axis) and different starting doses (y-axis).

malized clearance in older children.39 Information concerning specific toxic concentrations was not available therefore maximum concentrations (Cmax) and exposure (area under the curve) were not evaluated. In addition, due to the increased clearance in the first month, the plasma concentrations will be lower compared to the following months. Hence increasing the dose might not be necessary.

A potential limitation of this study was that samples were collected during the TDM procedure and predominantly withdrawn at 48 and 72 h after administration.

Additional peak samples were collected during the first hours after administration. Patients were at home in the period between the peak (first hours after dose) and trough concentrations (prior to the next dose), therefore Erwinia asparaginase concentrations within this timeframe were not available. However, with Erwinia asparaginase dosing, the aim is to achieve sufficient trough concentrations to assure complete asparagine depletion prior to the next dose (which is after 48 or 72 h), and the dose will be adjusted according to those time points. Two patients were excluded from the analysis due to non-measurable asparaginase levels. One patient had antibodies, which could explain the lack of asparaginase; the other patient might have had a very fast asparaginase clearance.

Excluding these patients could result in a lower variability and estimated clearance. haematologica | 2017; 102(3)

With the registered Erwinia asparaginase dose of 25,000 IU/m2, approximately 25% of the patients will not have 48 h trough concentrations above 100 IU/L after their first dose. This will be more pronounced in those patients with a low body weight. However, the PK analysis showed an increased clearance in the first month, therefore Erwinia asparaginase concentrations will increase after the first month. Monitoring the plasma concentrations and adjusting the dose for the individual patients presented with concentrations below target threshold is recommended.

Asparagine is completely depleted with Erwinia asparaginase concentrations of 100 IU/L or more, although some studies show complete depletion at lower concentrations.7,9,13,15,16,40 Increasing the dose for the group as a whole might lead to unnecessarily high concentrations in the majority of the patients with concentrations already over 100 IU/L, hence resulting in possible (long-term) side effects and unnecessary costs. Therefore, individual dose adjustments are recommended.

The optimal treatment would be dose adjustments based on patientsâ&#x20AC;&#x2122; individual PK parameters. With the PK model developed in this study, the individualized dose requirements can be calculated via post-hoc Bayesian analysis. This might reduce possible under-exposure that could potentially result in relapse, as well as reduce high concentrations. A prospective randomized controlled trial could compare conventional dosing versus individualized PK559

S.D.T. Sassen et al.

A

B

Figure 6. Erwinia asparaginase starting dose versus patient body surface area to achieve 50 IU/L or more after 48 hours (h). (A) Required starting dose in IU/kg and (B) IU/m2 versus weight of patients in kilograms (kg) to achieve 50 IU/L or more after 48 h. The median (solid line), 25% and 75% percentiles (dashed line) and 10% and 90% percentiles (dotted line) of the patients with asparaginase concentrations of 50 IU/L or more with different weight (x-axis) and different starting doses (y-axis).

based Erwinia asparaginase dosing to evaluate whether individual Erwinia asparaginase concentrations would improve and whether this affects treatment outcome.

However, due to the limited number of Erwinia asparaginase-treated patients in the Netherlands, this is not possible and should be performed on an international level. A rational approach should be adopted for dose management to ensure adequate trough concentrations.

References 1. Dubbers A, Wurthwein G, Muller HJ, et al.

Asparagine synthetase activity in paediatric acute leukaemias: AML-M5 subtype shows lowest activity. Br J Haematol. 2000; 109(2):427-429.

2. Capizzi RL, Bertino JR, Skeel RT, et al.

Clinical, Biochemical, Pharmacological, and Immunological Studies. Ann Intern Med. 1971;74(6):893-901.

3. Jaffe N, Traggis D, Das L, et al. L-asparaginase in the treatment of neoplastic diseases in children. Cancer Res. 1971;31(7):942949.

4. Pieters R, Hunger SP, Boos J, et al. Lasparaginase treatment in acute lymphoblastic leukemia. Cancer. 2011;117: 238-249.

5. Pession A. Long-Term Results of a Randomized Trial on Extended Use of High Dose L-Asparaginase for Standard Risk Childhood Acute Lymphoblastic Leukemia.

J Clin Oncol. 2005;23(28):7161-7167.

560

Acknowledgments We thank the patients and their parents, participating centers and (research) nurses for their help. We thank the pediatric oncology/hematology laboratory of the Erasmus MC – Sophia Children’s Hospital in Rotterdam, the Netherlands, for the analysis of the samples.

6. Silverman LB, Gelber RD, Dalton VK, et al.

Improved outcome for children with acute lymphoblastic leukemia: results of DanaFarber Consortium Protocol 91-01. Blood.

2001;97(5):1211-1218.

7. Salzer WL, Asselin B, Supko JG, et al.

Erwinia asparaginase achieves therapeutic activity after pegaspargase allergy: A report from the Children’s Oncology Group.

Blood. 2013;122(4):507-514.

8. Salzer W, Seibel N, Smith M. Erwinia asparaginase in pediatric acute lymphoblastic leukemia. Expert Opin Biol Ther. 2012; 12(10):1407-1414.

9. Avramis VI, Spence SA. Clinical pharmacology of asparaginases in the United States: asparaginase population pharmacokinetic and pharmacodynamic (PK-PD) models (NONMEM) in adult and pediatric ALL patients. J Pediatr Hematol Oncol. 2007; 29(4):239-247.

10. Tong WH, Pieters R, Kaspers GJL, et al. A prospective study on drug monitoring of PEGasparaginase and Erwinia asparaginase

and asparaginase antibodies in pediatric acute lymphoblastic leukemia. Blood.

2014;123(13):2026-2033.

11. Duval M, Suciu S, Ferster A, et al.

Comparison of Escherichia coli-asparaginase with Erwinia-asparaginase in the treatment of childhood lymphoid malignancies: results of a randomized European Organisation for Research and Treatment of Cancer-Children’s Leukemia Group phase 3 trial. Blood. 2002;99(8):2734-2739.

12. Moghrabi A, Levy DE, Asselin B, et al.

Results of the Dana-Farber Cancer Institute ALL Consortium Protocol 95-01 for children with acute lymphoblastic leukemia.

Blood. 2007;109(3):896-904.

13. Kwok CS, Kham SK, Ariffin H, Lin HP, Quah TC, Yeoh AE. Minimal residual disease (MRD) measurement as a tool to compare the efficacy of chemotherapeutic drug regimens using Escherichia Coli-asparaginase or Erwinia-asparaginase in childhood acute lymphoblastic leukemia (ALL).

Pediatr Blood Cancer. 2006;47(3):299-304.

haematologica | 2017; 102(3)

Pop PK of i.v. Erwinia asparaginase in pediatric ALL patients 14. Rizzari C, Citterio M, Zucchetti M, et al. A pharmacological study on pegylated asparaginase used in front-line treatment of children with acute lymphoblastic leukemia. Haematologica. 2006;91(1):2431.

15. Ahlke E, Nowak-Göttl U, SchulzeWesthoff P, et al. Dose reduction of asparaginase under pharmacokinetic and pharmacodynamic control during induction therapy in children with acute lymphoblastic leukaemia. Br J Haematol. 1997; 96(4):675-681.

16. Rizzari C, Zucchetti M, Conter V, et al. Lasparagine depletion and L-asparaginase activity in children with acute lymphoblastic leukemia receiving i.m. or i.v. Erwinia C. or E. coli L-asparaginase as first exposure.

Ann Oncol. 2000;11(2):189-193.

17. van der Sluis IM, Vrooman LM, Pieters R, et al. Consensus expert recommendations for identification and management of asparaginase hypersensitivity and silent inactivation. Haematologica. 2016;101(3): 279-285.

18. Asselin BL, Whitin JC, Coppola DJ, Rupp IP, Sallan SE, Cohen HJ. Comparative pharmacokinetic studies of three asparaginase preparations. J Clin Oncol. 1993;11(9): 1780-1786.

19. Pieters R, Appel I, Kuehnel H-J, et al.

Pharmacokinetics, pharmacodynamics, efficacy, and safety of a new recombinant asparaginase preparation in children with previously untreated acute lymphoblastic leukemia: a randomized phase 2 clinical trial. Blood. 2008;112(13):4832-4838.

20. Borghorst S, Pieters R, Kuehnel H-J, Boos J, Hempel G. Population pharmacokinetics of native Escherichia coli asparaginase. Pediatr Hematol Oncol. Taylor & Francis; 2012; 29(2):154-165.

21. Hempel G, Müller HJ, Lanvers-Kaminsky C, Würthwein G, Hoppe A, Boos J. A population pharmacokinetic model for pegylated-asparaginase in children. Br J Haematol.

2010;148(1):119-125.

22. Borghorst S, Pieters R, Kuehnel H-J, Boos J, Hempel G. Population pharmacokinetics of native Escherichia coli asparaginase. Pediatr

haematologica | 2017; 102(3)

23.

24.

25.

26.

27.

28.

29.

30.

31.

Hematol Oncol. Taylor & Francis; 2012; 29(2):154-165.

Albertsen BK, Jakobsen P, Schrøder H, Schmiegelow K, Carlsen NT. Pharmacokinetics of Erwinia asparaginase after intravenous and intramuscular administration. Cancer Chemother Pharmacol.

2001;48(1):77-82.

Albertsen BK, Schrøder H, Ingerslev J, et al.

Comparison of intramuscular therapy with Erwinia asparaginase and asparaginase Medac: Pharmacokinetics, pharmacodynamics, formation of antibodies and influence on the coagulation system. Br J Haematol. 2001;115(4):983-990.

Vrooman LM, Kirov II, Dreyer ZE, et al.

Activity and Toxicity of Intravenous Erwinia Asparaginase Following Allergy to E. coli-Derived Asparaginase in Children and Adolescents With Acute Lymphoblastic Leukemia. Pediatr Blood Cancer. 2016;63(2):228-233.

Tong WH, Pieters R, Hop WCJ, LanversKaminsky C, Boos J, van der Sluis IM. No evidence of increased asparagine levels in the bone marrow of patients with acute lymphoblastic leukemia during asparaginase therapy. Pediatr Blood Cancer. 2013; 60(2):258-261.

Anderson BJ, Holford NHG. Tips and traps analyzing pediatric PK data. Paediatr Anaesth. 2011;21(3):222-237.

Holford NH. A size standard for pharmacokinetics. Clin Pharmacokinet. 1996;30(5): 329-332.

Wang C, Peeters MYM, Allegaert K, et al. A bodyweight-dependent allometric exponent for scaling clearance across the human life-span. Pharm Res. 2012;29(6):15701581.

Sharkey I, Boddy AV, Wallace H, et al. Body surface area estimation in children using weight alone: application in paediatric oncology. Br J Cancer. 2001;85(1):23-28.

Abshire TC, Pollock BH, Billett AL, Bradley P, Buchanan GR. Weekly polyethylene glycol conjugated L-asparaginase compared with biweekly dosing produces superior induction remission rates in childhood relapsed acute lymphoblastic leukemia: a

32.

33.

34.

35.

36.

37.

38.

39.

40.

Pediatric Oncology Group Study. Blood.

2000;96(5):1709-1715.

Paolucci G, Vecchi V, Favre C, et al.

Treatment of childhood acute lymphoblastic leukemia. Long-term results of the AIEOP-ALL 87 study. Haematologica.

2001;86(5):478-484.

Boos J, Werber G, Ahlke E, et al.

Monitoring of asparaginase activity and asparagine levels in children on different asparaginase preparations. Eur J Cancer.

1996;32A(9):1544-1550.

Vrooman LM, Kirov II, Dreyer ZE, et al.

Activity and toxicity of intravenous Erwinia asparaginase following allergy to E. coli-derived asparaginase in children and adolescents with acute lymphoblastic leukemia. Pediatr Blood Cancer. 2016; 63(2):228-233.

Mosteller RD. Simplified calculation of body-surface area. N Engl J Med. 1987; 317(17):1098.

Haycock GB, Schwartz GJ, Wisotsky DH. Geometric method for measuring body surface area: a height-weight formula validated in infants, children, and adults. J Pediatr.

1978;93(1):62-66.

Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition.

1989;5(5):303-311.

Erwinaze® (asparaginase Erwinia Chrysanthemi). [Package insert] Palo Alto, CA: Jazz Pharmaceuticals, Inc; [Internet].

2014 [cited 29 June 2016]. Available from: http://www.accessdata.fda.gov/drugsatfda_docs/label/2014/125359s085lbl.pdf.

Last accessed: 29 June 2016.

Tong WH, Pieters R, de Groot-Kruseman HA, et al. The toxicity of very prolonged courses of PEGasparaginase or Erwinia asparaginase in relation to asparaginase activity, with a special focus on dyslipidemia. Haematologica. 2014;99(11):17161721.

Rizzari C, Citterio M, Zucchetti M, et al. A pharmacological study on pegylated asparaginase used in front-line treatment of children with acute lymphoblastic leukemia.

Haematologica. 2006;91(1):24-31.

561

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

T cells in chronic lymphocytic leukemia display dysregulated expression of immune checkpoints and activation markers Marzia Palma,1,2 Giusy Gentilcore,1 Kia Heimersson,1 Fariba Mozaffari,1 Barbro Näsman-Glaser,1 Emma Young,3 Richard Rosenquist,3 Lotta Hansson,1,2 Anders Österborg1,2 and Håkan Mellstedt1 AÖ and HM contributed equally to this work

Haematologica 2017 Volume 102(3):562-572

Immune and Gene Therapy Laboratory, Department of Oncology & Pathology, Cancer Centre Karolinska, Karolinska Institutet, Stockholm; 2Department of Hematology, Karolinska University Hospital, Stockholm and 3Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala University, Sweden 1

ABSTRACT

C

Correspondence: marzia.palma@karolinska.se

Received: June 14, 2016.

Accepted: November 17, 2016.

Pre-published: December 7, 2016 doi:10.3324/haematol.2016.151100 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/562

hronic lymphocytic leukemia is characterized by impaired immune functions largely due to profound T-cell defects. T-cell functions also depend on co-signaling receptors, inhibitory or stimulatory, known as immune checkpoints, including cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) and programmed death-1 (PD-1).

Here we analyzed the T-cell phenotype focusing on immune checkpoints and activation markers in chronic lymphocytic leukemia patients (n=80) with different clinical characteristics and compared them to healthy controls. In general, patients had higher absolute numbers of CD3+ cells and the CD8+ subset was particularly expanded in previously treated patients. Progressive patients had higher numbers of CD4+ and CD8+ cells expressing PD-1 compared to healthy controls, which was more pronounced in previously treated patients (P=0.0003 and P=0.001, respectively). A significant increase in antigen-experienced T cells was observed in patients within both the CD4+ and CD8+ subsets, with a significantly higher PD-1 expression. Higher numbers of CD4+ and CD8+ cells with intracellular CTLA-4 were observed in patients, as well as high numbers of proliferating (Ki67+) and activated (CD69+) CD4+ and CD8+ cells, more pronounced in patients with active disease. The numbers of Th1, Th2, Th17 and regulatory T cells were substantially increased in patients compared to controls (P<0.05), albeit decreasing to low levels in pre-treated patients. In conclusion, chronic lymphocytic leukemia T cells display increased expression of immune checkpoints, abnormal subset distribution, and a higher proportion of proliferating cells compared to healthy T cells. Disease activity and previous treatment shape the T-cell profile of chronic lymphocytic leukemia patients in different ways.

©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

562

Introduction Chronic lymphocytic leukemia (CLL) patients have dysregulated immune functions resulting in impaired antitumor immunity and increased risk for infections.

Profound defects in T-cell functions have been described as an imbalance of T-cell subsets,1 defective immune synapse formation with antigen presenting cells,2 impaired cytotoxic effector function3 and high frequency of regulatory T cells (Tregs).4-6 A number of co-signaling receptors, known as immune checkpoints, participate in the regulation of T-cell-driven immune responses. CD28 is constitutively expressed on CD4+ and CD8+ T cells providing the primary co-stimulatory signal for T-cell activation.7 Upon T-cell stimulation, a number of cell surface molecules belonging to the haematologica | 2017; 102(3)

Immune checkpoints and activation markers in CLL

tumor necrosis factor (TNF)-receptor family are up-regulated and deliver co-stimulatory signals. CD137 is such a molecule. It binds to its ligand (CD137L), a member of the TNF family, expressed on macrophages, activated B cells and dendritic cells (DCs) enhancing T-cell proliferation and cytolytic effector functions.8 The CD28 homolog cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) is expressed on activated T cells and has an inhibitory role in regulating T-cell activation.9 Another negative co-stimulatory molecule is programmed death-1 (PD-1), expressed on activated CD4+ and CD8+ T cells, natural killer (NK) T cells, B cells, activated monocytes and DCs. Upon binding to the ligands PD-L1 and PD-L2, PD-1 inhibits T-cell functions reducing T-cell receptor signaling and target cell lysis.10 Except for cells of the macrophage lineage, PD-L1 has low expression in normal tissues; on the other hand, it is highly expressed on various tumors and can be further enhanced by tumor environmental factors.11,12 The PD-1/PD-L1 pathway is considered a central regulator of T-cell exhaustion, a condition characterized by deteriorated Tcell effector function due to chronic antigen stimulation.13 Tumor cells as well as pathogens exploit these inhibitory signals to hamper immune eradication.

The aim of the present study was to evaluate the expression of the immune checkpoints CD137, CTLA-4 and PD-1 in CLL patients at different phases of the disease as well as the distribution of various CD4+ and CD8+ T-cell subsets. We aimed to provide a comprehensive analysis of T-cell phenotype in CLL and to discriminate between alterations that are due to the disease itself and disease activity and those related to treatment.

We show that diverse T-cell alterations are related to distinct clinical situations, i.e. disease activity and previous treatment. The PD-1 receptor expression was markedly increased in active disease, especially in heavily treated patients.

Methods Patients Peripheral blood samples from 80 CLL patients were collected at the Department of Hematology, Karolinska University Hospital, Stockholm, Sweden, as well as from 9 age- and sex-matched controls. Patients were grouped according to disease activity: non-progressive versus progressive (i.e. fulfilling criteria for active disease14). Characteristics of the patients and controls are shown in Table 1. Ninety percent of the patients were cytomegalovirus (CMV)-positive, in line with the prevalence in the Swedish population aged over 60 years.15 The research project was approved by the regional ethics committee (www.epn.se) and conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from study participants.

Flow cytometric analysis of lymphocyte subsets on whole blood Cells were washed after lysis of red blood cell, resuspended in Cell Staining Buffer (CSB) (BioLegend, San Diego, CA, USA) and stained with CD19-AF488, CD16+56-PE, CD4-PerCp, CD3-PECy7, CD8-APC and CD45-AF700 (Bio-Legend). After incubation and washing, cells were resuspended in CSB and analyzed by FACSCanto II flow cytometer and the FACSDiva v.6.1.3 (BD Biosciences, San Diego, CA, USA) or FlowJo v.8.8.2 (TreeStar, Ashland, OR, USA) softwares. haematologica | 2017; 102(3)

Isolation of peripheral blood mononuclear cells and cell culture conditions Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood by density gradient centrifugation on a FicollHypaque gradient (GE Healthcare, Uppsala, Sweden) and washed twice with Dulbecco’s Phosphate-Buffered Saline 0.9% (DPBS) (Gibco, Life Technologies, Carlsbad, CA, USA). Cells were freshly used or stored in liquid nitrogen until use. After thawing, PBMC were analyzed immediately unless used for stimulation experiments.

3x106 PBMC were cultured for 72 hours in humidified air with 5% CO2 at 37°C in RPMI 1640 medium (GIBCO, Life Technologies, Carlsbad, CA, USA) supplemented with heat-inactivated autologous serum for fresh samples and pooled normal human AB+ serum for frozen samples in the presence of phytohemagglutinin (10 mg/mL) (PHA-M, Sigma Aldrig, St. Louis, MO, USA). PBMC cultured in medium alone were used as controls.

Flow cytometric analysis of PBMC Peripheral blood mononuclear cells were washed with CSB (BioLegend). The following antibodies were used: CD19-AF488 and -PE-Cy7, CD16/CD56-PE, CD4-PerCp, -FITC and -AF700, CD3-PE-Cy7, -AF700 and -PerCP, CD8-APC and -AF700, CD45AF700, CD5-PE and -PerCP, CD45RO-FITC, HLA-DR-PerCp, CD25-APC, CD45RA-AF488, PD-1 (CD279)-PE, CD69-AF488, CTLA-4 (CD152)-PE, Ki-67-AF647 (Bio-Legend), CCR6 (CD196)PE, CCR4 (CD194)-PE, CD127-PE-Cy7, CXCR3 (CD183)-APC, CCR7 (CD197)-AF647, PD-L1 (CD274)-PE (BD-Biosciences) and the appropriate isotype controls. Further details are provided in the Online Supplementary Appendix.

Sequence analysis of IGHV–IGHD–IGHJ rearrangements IGHV-IGHD-IGHJ rearrangements were determined through PCR amplification, Sanger sequencing and subsequent sequence interpretation following established international guidelines and using the IMGT® databases and the IMGT/V-QUEST tool (http://www.imgt.org), as previously reported.16 IGHV gene mutational status was defined as either mutated or unmutated based on the clinically relevant 98% cut-off value for identity to the closest germline gene.17,18 Subset #2 cases (IGHV3-21/IGLV3-21 usage, mixed IGHV mutation status) are listed together with IGHVunmutated cases since this entity is a recognised adverse-prognostic group.19

Statistical analyses Statistical analyses were performed using the GraphPad Prism software 6.0 (GraphPad Software, La Jolla, CA, USA). All tests were two-sided, and P<0.05 was considered statistically significant. Further details are provided in the Online Supplementary Appendix.

Results CLL patients had higher absolute numbers of T cells and the number of CD8+ T cells was related to treatment Chronic lymphocytic leukemia patients had higher numbers of CD3+ cells compared to controls (Online Supplementary Table S1); the difference was statistically significant for all the patient subgroups. No difference was observed for CD4+ cells, while CD8+ cells were higher in pre-treated progressive patients compared to controls as well as non-progressive (P=0.02 and P=0.001, respectively), irrespective of type of treatment (alemtuzumab or not; fludarabine/cyclophosphamide or not) and IGHV mutational status (data not shown).

563

M. Palma et al.

Table 1. Characteristics of patients and controls at time of testing.

Controls (n=9) Sex Age

Male, n Female, n Years (median, range)

6 3 65 (45-79)

Patients (n=80) Disease phasea Sex Age Modified Rai stagea

IGHV status

Previous treatment

Chemotherapy-refractory

Male, n Female, n Years, median (range) Low risk, n Intermediate risk, n High risk, n Mutated, n Unmutated/subset #2, n Undefined, n No, n 17p deletion/TP53 mutation Yes, n 17p deletion/TP53 mutation Previous therapy Chlorambucil FC or FCR BR Alemtuzumab Other ≥ 2 previous treatment lines No, n Yes, n

Non-progressive (n=39) 21 18 67 (49-82) 28 11 0 22* 10 3 39 0/4 0 0/0

Chemotherapy-refractory and/or 17p deletion/TP53 mutation

Progressive (n=41) 29 12 69 (47-85) 2 11 28 17** 24 0 22 2/22 19 7/19

Total (n=80) 48 30 69 (47-85) 29 21 28 43 34 3 60 18

5 10 2 7 3 10 13/19 6/19 11/19

a As defined by Hallek et al.14 *Of which 3 borderline IGHV mutated (97%-98% identity). **Of which 2 borderline IGHV mutated. FC: fludarabine + cyclophosphamide; FCR: fludarabine + cyclophosphamide + rituximab; BR: bendamustine + rituximab.

PD-1 expression was increased in T cells from pre-treated progressive CLL patients Compared to controls, CLL patients had higher numbers of PD-1-expressing CD4+ T cells, which related to disease activity and previous treatment. No difference was observed between non-progressive patients and controls (median 258 vs.169/mL; P=0.1), while progressive patients had higher numbers compared to controls (median 315 and 521/mL for untreated and treated patients; P=0.01 and P=0.0003, respectively). This was observed regardless of IGHV mutational status. Pre-treated patients with progressive disease had higher numbers of PD-1+CD4+ cells as compared to non-progressive (P=0.008) (Figure 1A). There was a moderate positive correlation between PD-1+ CD4+ T cells and total lymphocyte count (r=0.36, P=0.001) (Figure 1C).

No difference was seen with regard to CD8+ T cells comparing patients and controls, with the exception of previously treated progressive patients, who had higher numbers of PD-1+ CD8+ T cells compared to controls and non-progressive patients (median 389 vs. 121 vs. 143/mL; P=0.001 and P=0.007, respectively) (Figure 1B).

Subgrouping based on the IGHV mutational status showed that this was only observed in the unmutated group. A moderate positive correlation was also observed 564

between PD-1+CD8+ T cells and total lymphocyte count (r=0.43, P<0.0001) (Figure 1D). No expression of PD-L1 on CLL cells was noted (data not shown).

Progressive CLL patients had an increase in PD-1+ antigen-experienced T cells A subset of patients (n=33) was analyzed for the distribution of CD4+ and CD8+ memory T cells. By CD45RA and CCR7 staining, T-cell subpopulations were identified as naïve (CD45RA+/CCR7+), central memory (CD45RA– /CCR7+), effector memory (CD45RA–/CCR7–) and effector (CD45RA+/CCR7–). CLL patients had higher absolute numbers of CD4+ effector memory cells compared to controls irrespective of disease activity and previous treatment. The frequency of central memory T cells was also higher in CLL patients compared to controls, but only in those untreated. Naïve T cells were dramatically reduced in pre-treated patients compared to controls and untreated. No difference was seen for effector T cells (Figure 2A).

Similarly, higher absolute numbers of CD8+ effector memory and effector cells were found in progressive patients compared to controls. This was observed in both untreated and pre-treated patients, but in this latter group, this held true only for patients who had received alemtuzumab, who had higher numbers of CD8+ effector haematologica | 2017; 102(3)

Immune checkpoints and activation markers in CLL

A

B

C

D P=0.001

P<0.0001

Figure 1. PD-1 expression in T cells from chronic lymphocytic leukemia (CLL) patients and controls. Absolute numbers of (A) PD-1+CD4+ and (B) PD-1+CD8+ T cells from progressive (P) and non-progressive (NP) CLL patients compared to controls. Box plots display cumulative data with line at median. *P<0.05, **P<0.005, ***P<0.0005. Correlation of (C) PD-1+CD4+ and (D) PD-1+CD8+ T cells with total lymphocyte count at the time of testing (n=80).

memory (P=0.001) and effector (P=0.007) cells compared to controls. Untreated patients with progressive disease had higher numbers compared to non-progressive. CLL patients and controls had comparable numbers of naĂŻve T cells, which were significantly reduced in pre-treated compared to untreated patients. No difference was noted with regard to the numbers of central memory T cells (Figure 2B).

PD-1 expression within the CD4+ population was higher among memory T-cell subsets in patients as compared to controls, with the exception of previously treated patients in which the numbers of PD-1+CD4+ naĂŻve, central memory and effector cells were similar to controls (Figure 2C).

CLL patients irrespective of disease phase and previous treatment had higher numbers of CD8+ effector memory and effector cells expressing PD-1 compared to controls.

The numbers of PD-1+ CD8+ naĂŻve cells were low in CLL patients, but higher in untreated patients compared to controls (P=0.03 and P=0.01 for non-progressive and progressive untreated, respectively) (Figure 2D).

CTLA-4 was only detected intracellularly in CLL T cells No expression of surface CTLA-4 was seen in either CD4+ or CD8+ cells from CLL patients and controls.

Intracellular CTLA-4 was, however, expressed in a higher number of CD4+ T cells in CLL patients as compared to haematologica | 2017; 102(3)

controls (median 329/mL for non-progressive patients, 717/mL for progressive untreated and 317/mL for progressive pre-treated vs. 136/mL for controls; P<0.005) (Figure 3A). Numbers of CD4+ cells with intracellular CTLA-4 expression were higher both in patients treated with alemtuzumab (P=0.001) and cyclophosphamide/fludarabine (P=0.0007) compared to controls. A positive correlation was observed between the numbers of CD4+ T cells with intracellular CTLA-4 and total lymphocyte count (r=0.50, P=0.003). CLL patients also had higher numbers of CD8+ cells with intracellular CTLA-4 compared to controls (median 23/mL for non-progressive patients, 79/mL for progressive untreated and 59/mL for progressive pre-treated vs. 7.5/mL for controls; P<0.0001). Both untreated and previously treated patients with progressive disease had higher numbers of CD8+ cells with intracellular CTLA-4 as compared to non-progressive (P<0.05) (Figure 3B), which correlated positively with the total lymphocyte count (r=0.38, P=0.03).

T cells from progressive CLL patients displayed an activated phenotype but no expression of the co-stimulatory molecule CD137 Lower numbers of CD69+CD4+ cells were noted in nonprogressive compared to progressive CLL patients irrespective of previous treatment (median 30/mL in untreated 565

M. Palma et al.

A

B

C

D

controls

non-progressive CLL

untreated progressive CLL

pre-treated progressive CLL

Figure 2. Comparative analysis of T-cell memory subsets in chronic lymphocytic leukemia (CLL) patients compared to controls. Absolute numbers of (A) CD4+, (B) CD8+, (C) CD4+PD-1+ and (D) CD8+PD-1+ naĂŻve, central memory (CM), effector memory (EM) and effector (EMRA) cells in healthy controls (n=9) were compared to non-progressive (n=13), untreated progressive (n=8) and pre-treated progressive (n=12) CLL patients. Box plots display cumulative data with line at median. Only significant statistical values are reported. *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001.

and 22/mL in pre-treated compared to 5/mL in non-progressive; P=0.002) (Figure 4A). Non-progressive patients had lower numbers of CD69+CD8+ cells compared to controls (median 6/mL in non-progressive and 27/mL in controls; P=0.008), untreated progressive (median 16.5/mL; P=0.009) and previously treated progressive (median 40/m; P=0.0003) patients (Figure 4B). A moderate positive correlation was observed between the total lymphocyte count and the numbers of CD69+CD4+ T cells (r=0.39, P=0.0004) and CD69+CD8+ T cells (r=0.34, P=0.002). No expression of CD137 was observed on T cells from CLL patients and controls (data not shown).

Expression of immune checkpoints and activation markers could be induced on CLL T cells It is known that T-cell stimulation leads to upregulation of immune checkpoints and activation markers on the cell surface.8-10,20,21 We, therefore, stimulated T cells from CLL patients and controls for 72 hours with PHA. This method was chosen rather than others of unspecific T-cell stimulation to more closely reflect physiological conditions and avoid interference with the flow-cytometry staining.

PD-1 expression increased markedly on both CD4+ and CD8+ cells, and similarly in CLL patients and controls (Online Supplementary Figure S1A). Surface CTLA-4, which was virtually absent at baseline both in CLL patients and 566

controls, was induced on CD4+ cells both from CLL patients and controls (median % CTLA-4+CD4+ cells after PHA stimulation was 2.9 in non-progressive, 14.9 in progressive patients and 4.8 in controls) (Online Supplementary Figure S1B). CD69 expression also increased in both CD4+ and CD8+ cells and to a similar degree in CLL patients and controls (Online Supplementary Figure S1C), while CD137 expression increased to a higher extent in T cells from progressive patients compared to controls (P=0.03 and 0.01 for the CD4+ and CD8+ cells, respectively) (Online Supplementary Figure S1D). We also studied expression of CD103, a marker for alloantigen-induced CD8+ Tregs20 and found that the percentage of CD103+CD8+ T cells increased more in controls than in CLL patients (median increase 4.8% in controls vs. 0.7% in non-progressive and 0% in progressive patients; P=0.01 and P=0.004, respectively) (Online Supplementary Figure S1E).

Proliferating T cells were significantly higher in CLL patients compared to controls and correlated with disease activity The percentage of proliferating (Ki67+) circulating tumor cells (CD5+CD19+) in CLL patients was low (<1%) irrespective of disease activity and previous treatment (data not shown). However, CLL patients had higher absolute numbers of proliferating CD8+ T cells compared to conhaematologica | 2017; 102(3)

Immune checkpoints and activation markers in CLL

Table 2. Summary of the different T-cell subpopulations and T cells expressing immune checkpoints or activation / proliferation markers as compared between the different studied subject groups. (A) CD4+ T cells. (B) CD8+ T cells.

A Disease non-progressive CLL vs. healthy controls

Effect of disease phase progressive untreated vs. non-progressive

Treatment progressive pre-treated vs. progressive untreated

Disease non-progressive CLL vs. healthy controls

Effect of disease phase progressive untreated vs. non-progressive

Treatment progressive pre-treated vs. progressive untreated

↔ ↔ ↔ ↑ (**) ↑ (**) ↔ ↑(***) ↔ (****) ↑ (*) ↔ ↑ (*) ↔

CD4+ PD1+ Tnaive TCM TEM TEMRA i.c. CTLA-4+ CD69+ Th1 Th2 Th17 Tregs Ki67+

B

CD8+ PD1+ Tnaive TCM TEM TEMRA i.c. CTLA-4+ CD69+ Ki67+

↔ ↔ ↔ ↔ ↔ ↔ ↑ (****) ↓ (**) ↑(*)

↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑(**) ↔ ↔ ↔ ↔ ↔

↔ ↔ ↔ ↔ ↑(**) ↑(*) ↑ (*) ↑ (**) ↑(*)

↔ ↔ ↓ (***) ↓ (*) ↔ ↔ ↔ ↔ ↓ (***) ↓ (**) ↓ (**) ↔ ↔

↔ ↔ ↓ (**) ↔ ↔ ↔ ↔ ↔ ↔

Statistically significant differences are symbolized as follows: ↔ : no difference; ↑ higher; ↓ : lower. i.c.: intracellular. *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001.

trols irrespective of disease activity and previous treatment (median 10/mL for non-progressive, 36/mL for progressive untreated, 33/mL for progressive treated vs. 5/mL for controls; P=0.02, P=0.0002 and P<0.0001, respectively). Higher numbers of proliferating CD4+ T cells were observed also in progressive patients compared to controls, irrespective of previous treatment (median 75/mL for progressive untreated and 41/mL for progressive treated vs.

10/mL for controls; P=0.006 and P=0.001, respectively) (Figure 4C and D).

Distribution of functional CD4+ T-helper subpopulations in relation to disease activity T-helper subpopulations were defined by CCR6 and CXCR3 expression as Th1 (CCR6–/CXCR3+), Th2 (CCR6–/CXCR3–) and Th17 (CCR6+/ CXCR3–) cells. Tregs were defined by the expression of CD4, CD25 and CCR4 and CD127low.22,23 Both non-progressive and progressive untreated CLL patients had higher numbers of Th1 cells compared to conhaematologica | 2017; 102(3)

trols (median 406 and 1064 vs. 139/mL; P<0.0001 and P=0.001, respectively). However, progressive treated patients had lower Th1 numbers (median 62/mL) compared to both controls (P=0.009) and the other patient groups (P<0.0001 and P=0.0001 compared to non-progressive and progressive untreated, respectively). The number of Th1 cells was lower only in patients treated with cyclophosphamide/fludarabine compared to controls (P=0.01).

Higher Th2 numbers were observed in non-progressive patients compared to controls (median 833 and 599/mL; P=0.03), but progressive pre-treated patients had lower numbers of Th2 cells compared to untreated (P<0.005). The numbers of Th17 cells were higher in progressive untreated patients compared to controls (median 196 and 109/mL; P=0.04) but progressive pre-treated patients had lower Th17 numbers compared to controls (P=0.03) and untreated patients (P=0.002 and P=0.003, for non-progressive and progressive untreated, respectively) (Figure 5A).

No difference was observed in the percentage of Tregs comparing CLL patients and controls (median 4.8% for 567

M. Palma et al.

A

B

Figure 3. Intracellular CTLA-4 expression in T cells from chronic lymphocytic leukemia (CLL) patients and controls. Absolute numbers of (A) CD4+ and (B) CD8+ T cells with intracellular (i.c.) CTLA-4 expression in CLL patients and controls. Box plots display cumulative data with line at median. Only significant statistical values are reported. NP: non-progressive; P: progressive. *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001.

non-progressive, 4.2% for progressive CLL patients and 4.2% for controls, respectively; P=0.5) (Online Supplementary Table S2). However, the absolute number of Tregs was higher in untreated CLL patients compared to controls (median 72/mL for non-progressive and 78/mL for progressive untreated vs. 37/mL for controls; P=0.04 and P=0.002, respectively), while no difference was seen for progressive pre-treated patients (median 54/mL) (Figure 5B). Nevertheless, Tregs were higher in patients pre-treated with cyclophosphamide/fludarabine as well compared to controls (P=0.04). Low numbers of CD8+ cells expressing CD103 were observed in CLL patients, though higher in non-progressive (n=27) and progressive untreated (n=14) patients compared to progressive previously treated patients (n=10) (median 3/mL vs. 0.2/mL; P=0.006 and P=0.002, respectively).

Discussion In the present study, we analyzed the T-cell phenotype focusing on immune checkpoints and activation markers in CLL patients with different clinical characteristics. Since the total T-cell numbers may vary considerably between CLL patients and healthy individuals, between patients in different phases of the disease, and depending on previous treatments, we chose to compare absolute cell numbers.

Percentage numbers are reported in Online Supplementary Table S2.

Increased T-lymphocyte counts, as well as expansion of CD8+ and CD4+ T cells, have been described in CLL,24-26 with a relatively higher increase in CD8+ cells resulting in a low CD4/CD8 ratio compared to controls.27,28 We found that CLL patients irrespective of disease phase and previous treatment had significantly higher numbers of CD3+ cells compared to controls. There was no significant difference in the distribution of the CD4+ and CD8+ subsets within the CD3+ population between untreated patients and controls. Nevertheless, pre-treated patients had signif568

icantly higher numbers of CD8+ cells.

Several studies have investigated the expression of PD1 and CTLA-4 in CLL patients, but the results are contradictory. An increase in PD-1+CD8+ T cells in CLL patients, particularly within the effector memory subset, was noted

by Riches et al.,3 while Tonino et al.29 found that PD-1 expression was decreased. Brusa et al.30 found significantly higher PD-1 expression in CD4+ and CD8+ T cells from CLL patients, but could not identify any association of significance between PD-1 expression and disease stage, treatment requirements or unfavorable molecular or cytogenetic markers. Novak et al.31 recently reported higher numbers of PD-1-expressing T cells within both the CD4+ and CD8+ subsets in CLL patients but no significant difference between patients in different phases of the disease.

Finally, an association between the PD-1/PD-L1 axis and T-cell dysfunction in progressive disease has been reported.32-34 A comprehensive summary of the relative changes we observed in absolute numbers of T-cell subpopulations and T cells expressing immune checkpoints or activation/proliferation markers in different subgroups of CLL patients compared to healthy controls is provided in Figure 6. In contrast to a previous report,31 we observed that the absolute numbers of CD4+ cells expressing PD-1 were significantly increased only in CLL patients with progressive disease compared to controls. The difference was more marked for pre-treated patients. Within the CD8+ subset, only pre-treated patients had significantly higher numbers of PD-1+-expressing cells compared to controls. This observation may indicate that T cells in progressive patients display features of exhaustion, which seemed to be accentuated after treatment. Whether this may relate to the treatment per se or to the fact that previously treated patients have more advanced disease cannot be fully elucidated at present.

It is known that the distribution of memory T-cell subsets is altered in CLL patients. The expression of PD-1 on CD4+ effector memory cells is considered to be a marker haematologica | 2017; 102(3)

Immune checkpoints and activation markers in CLL

A

B

C

D

Figure 4. CD69 and intracellular Ki67 expression in T cells from chronic lymphocytic leukemia (CLL) patients and controls. Absolute numbers of (A) CD69+CD4+, (B) CD69+CD8+, (C) Ki67+CD4+, (D) Ki67+CD8+ T cells from progressive (P) and non-progressive (NP) CLL patients and healthy controls. Box plots display cumulative data with line at median. Only significant statistical values are reported. *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001.

of chronic activation.35,36 We noted that CLL patients had higher absolute numbers of CD4+ effector memory cells expressing PD-1 compared to controls irrespective of disease phase and previous treatment. CD4+ central memory cells also displayed high PD-1 expression. This subset was expanded in CLL patients, but only in those untreated.

Moreover, naĂŻve CD4+ cells expressing PD-1 were significantly higher in untreated CLL patients compared to controls. Effector CD4+ cells were not expanded but showed a high PD-1 expression. Collectively, these data may indicate a persistent (chronic) antigen exposure in CLL patients inducing T-cell exhaustion in all the CD4+ subsets, preferentially those antigen-experienced (CD45ROâ&#x20AC;&#x201C;), i.e. central memory and effector memory cells.

High numbers of effector memory cells were observed in the CD8+ subset in all the patients, and significantly higher PD-1 expression was observed in progressive patients. High numbers of PD-1+ effector cells were also observed in all the patient subgroups. T cells from CLL patients display higher expression of inhibitory receptors, including PD-1, irrespective of CMV status.3,32 Therefore, it is unlikely that the observed T-cell subset distribution is due to chronic stimulation by the CMV antigen, but is most likely due to stimulation by tumor antigens.

Contrary to previous reports,30,33,37 we found no expression of PD-L1 on CLL cells. The reason for this is unclear, haematologica | 2017; 102(3)

but might be due to the fact that we analyzed PD-L1 expression on freshly isolated cells, while other studies have analyzed the expression on purified CD19+ cells.

One previous study reported that surface expression of CTLA-4 was decreased in CLL patients compared to controls,38 while another study noted no difference3 and three other studies39-41 showed an increase. However, a significantly higher intracellular CTLA-4 expression in CLL T cells was found in all the previously published studies.38,40,41 Moreover, Motta et al.40 and Scrivener et al.38 could not see any significant enhancement of CTLA-4 expression after in vitro stimulation, while Frydecka et al.39 found that CTLA-4 expression increased over time in T cells but with a different kinetics to controls.

We noted no expression of surface CTLA-4 in either CLL patients or controls, but this could be induced in CD4+ cells by in vitro stimulation, in particular in non-progressive patients. However, intracellular CTLA-4 expression was high in both CD4+ and CD8+ cells of CLL patients compared to controls. A hallmark of CTLA-4 is the trafficking to and from the plasma membrane following TCR stimulation.9,42 CTLA-4 is engaged in the primary phase of T-cell activation, which might explain why chronically activated, exhausted T cells lack surface expression.

CD137 is poorly expressed or not at all in the resting T-cell state but up-regulated upon activation.8 In line with 569

M. Palma et al.

this, we observed no expression of CD137 on freshly isolated CLL T cells, but expression could be induced in both CD4+ and CD8+ cells by in vitro stimulation, in particular in progressive patients.

Chronic lymphocytic leukemia patients had higher numbers of Th1, Th2 and Th17 cells compared to controls. No significant difference between non-progressive and progressive patients was observed. This is in contrast to previous data based on cytokine production, showing increased secretion of IL-4 in CLL, suggested to be due to a Th2 polarization during disease progression.25,43,44 We observed that previously treated progressive patients had significantly lower numbers of all three subsets.

Consistent with previous data,4,5 we found that absolute numbers of Tregs were higher in untreated CLL patients compared to controls, independent of disease phase, but lower in previously treated patients.

Finally, we confirmed that both CD4+ and CD8+ T cells in progressive CLL patients display an activated pheno-

type (CD69+), as also shown previously.45 Moreover CLL patients had significantly higher numbers of proliferating CD4+ and CD8+ T cells, which was more evident at disease progression.

Taken together, our results suggest that disease activity and previous treatment have a different impact on T-cell profile in CLL. The disease per se implies a number of changes in T cells (Table 2). At disease progression the most remarkable alteration occurring in the CD4+ subset is an increase in CD69+ cells, while in the CD8+ subset more extensive changes take place. In addition to higher numbers of CD69+ cells, within the CD8+ subset, higher numbers of proliferating (Ki67+), effector memory and effector cells were noted. However, PD-1 and CTLA-4 expression in progressive disease were so high that it is reasonable to assume that these cells have heavily impaired immune functions, as also suggested by previously published data.30,32 CLL treatment also seemed to dramatically affect T cells, in particular the CD4+ subset, in which a decrease

A A

B

B

controls

non-progressive CLL

untreated progressive CLL

pre-treated progressive CLL

Figure 5. Comparative analysis of functional CD4+ T-helper subpopulations (Th1/Th2/Th17) and regulatory T cells (Tregs) in chronic lymphocytic leukemia (CLL) patients and controls. Absolute numbers of (A) Th1, Th2, Th17 and (B) Tregs cells in non-progressive (NP) (n=13), untreated progressive (P) (n=8), and pre-treated progressive (n=12) CLL patients compared with healthy controls (n=9). Box plots display cumulative data with line at median. Only significant statistical values are reported. *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001.

570

Figure 6. Relative change in absolute numbers of different T-cell subpopulations and T cells expressing immune checkpoints or activation / proliferation markers in chronic lymphocytic leukemia (CLL) patients compared to healthy controls. Relative change in the (A) CD4+ subset and (B) CD8+ subset calculated as median value patients/median value controls in non-progressive (green bars), untreated progressive (yellow bars) and pre-treated progressive (black bars) CLL patients.

haematologica | 2017; 102(3)

Immune checkpoints and activation markers in CLL

of all T-helper subsets (Th1, Th2, Th17) was observed. A decrease in naĂŻve T cells in both the CD4+ and the CD8+ subsets was also related to therapy. We tried to define more specifically the impact of different treatment regimens on T-cell phenotype by further subgrouping the patients into those who had received alemtuzumab and those who had received fludarabine/cyclophosphamide, since these drugs have a known effect on T cells.46,47 The number of Th1 cells was significantly lower while Tregs were higher in patients treated with cyclophosphamide/fludarabine compared to controls; intracellular CTLA-4 expression seemed to be affected by both pretreatment with both alemtuzumab and cyclophosphamide.

Different treatments did not seem to have a different impact on the expression of immune checkpoints and activation markers. Overall, the IGHV mutational status seemed to have a minor impact. Unfortunately, we do not have cytogenetic data for all the patients, since in Sweden analysis by interphase fluorescence in situ hybridization is routinely performed only in patients requiring therapy.

Therapeutic interference with T-cell exhaustion by tar-

References 1. Mellstedt H, Choudhury A. T and B cells in B-chronic lymphocytic leukaemia: Faust, Mephistopheles and the pact with the Devil. Cancer Immunol Immunother.

2006;55(2):210-220.

2. Ramsay AG, Johson AJ, Lee AM, et al.

Chronic lymphocytic leukemia T cells show impaired immunological synapse formation that can be reversed with an immunomodulating drug. J Clin Invest.

2008;118(7):2427-2437.

3. Riches JC, Davies JK, McClanahan F, et al.

T cells from CLL patients exhibit features of T-cell exhaustion but retain capacity for cytokine production. Blood. 2013; 121(9):1612-1621.

4. D'Arena G, Laurenti L, Minervini MM, et al. Regulatory T-cell number is increased in chronic lymphocytic leukemia patients and correlates with progressive disease. Leuk Res. 2011;35(3):363-368.

5. Dasgupta, A, Mahapatra M, Saxena R. A study for proposal of use of regulatory T cells as a prognostic marker and establishing an optimal threshold level for their expression in chronic lymphocytic leukemia.

Leuk Lymphoma.

2015;56(6):1831-1838.

6. Giannopoulos K, Schmitt M, Kowal M, et al. Characterization of regulatory T cells in patients with B-cell chronic lymphocytic leukemia. Oncol Rep. 2008;20(3):677-682.

7. Bocko D, Kosmaczewska A, Ciszak L, et al.

CD28 costimulatory molecule--expression, structure and function. Arch Immunol Ther Exp (Warsz). 2002;50(3):169-177.

8. Melero L, Hirschhorn-Cymerman D, Morales-Kastresana A, et al. Agonist antibodies to TNFR molecules that costimulate T and NK cells. Clin Cancer Res. 2013;19(5):1044-1053.

9. Walker LS, Sansom DM. The emerging role of CTLA4 as a cell-extrinsic regulator of T cell responses. Nat Rev Immunol.

2011;11(12):852-863.

haematologica | 2017; 102(3)

geting co-stimulatory and inhibitory pathways may be beneficial to increase anti-tumor T-cell responses in CLL patients. In particular, immune checkpoint blockade with anti-PD1 mAb might be successful also in heavily pretreated chemo-refractory patients. Even though PD-1 blockade alone might not be enough to reanimate exhausted T cells in CLL,48 a combined approach either with targeted drugs or immunotherapies directed against different receptors might be a rewarding approach in this patient subgroup.

Acknowledgments The authors thank the patients and donors who consented to the use of their cell samples for this study. The authors thank Ms Leila Relander for excellent secretarial help.

Funding This work was supported by grants from The Swedish Cancer Society, the Swedish Research Council, The Cancer Society in Stockholm, King Gustav V Jubilee Fund, The Cancer and Allergy Foundation, The Karolinska Institutet Foundations, The Stockholm County Council and AFA Insurance.

10. Keir ME, Butte MJ, Freeman GJ, et al. PD-1 and its ligands in tolerance and immunity.

Annu Rev Immunol. 2008;26:677-704.

11. Dong H, Strome SE, Salomao DR, et al.

Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion. Nat Med. 2002;8(8):793800.

12. Curiel TJ, Wei S, Dong H, et al. Blockade of B7-H1 improves myeloid dendritic cellmediated antitumor immunity. Nat Med. 2003;9(5):562-567.

13. Wherry EJ. T cell exhaustion. Nat Immunol. 2011;12(6):492-499.

14. Hallek M, Cheson BD, Catovsky D, et al.

Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008; 111(12):5446-5456.

15. Strindhall J, Skog M, Ernerudh J, et al. The inverted CD4/CD8 ratio and associated parameters in 66-year-old individuals: the Swedish HEXA immune study. Age. 2013;35(3):985-981.

16. Murray F, Darzentas N, Hadzidimitriou A, et al. Stereotyped patterns of somatic hypermutation in subsets of patients with chronic lymphocytic leukemia: implications for the role of antigen selection in leukemogenesis. Blood. 2008;111(3):15241533.

17. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood. 1999;94(6):18401847.

18. Hamblin TJ, Davis Z, Gardiner A, et al.

Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood. 1999; 94(6):1848-1854.

19. Baliakas P, Agathangelidis A, Hadzidimitriou A, et al. Not all IGHV3-21 chronic lymphocytic leukemias are equal:

20.

21.

22.

23.

24.

25.

26.

27.

28.

prognostic considerations. Blood. 2015; 125(5):856-859.

Uss E, Rowshani AT, Hooibrink B, et al.

CD103 is a marker for alloantigen-induced regulatory CD8+ T cells. J Immunol. 2006; 177(5):2775-2783.

Yamashita I, Nagata T, Tada T, et al. CD69 cell surface expression identifies developing thymocytes which audition for T cell antigen receptor-mediated positive selection.

Int Immunol. 1993;5(9):1139-1150.

Liu W, Putnam AL, Xu-Yu Z, et al. CD127 expression inversely correlates with FoxP3 and suppressive function of human CD4+ T reg cells. J Exp Med. 2006;203(7):17011711.

Seddiki N, Santer-Nana B, Martinson J, et al. Expression of interleukin (IL)-2 and IL-7 receptors discriminates between human regulatory and activated T cells. J Exp Med. 2006;203(7):1693-1700.

Serrano D, Monteiro J, Allen SL, et al.

Clonal expansion within the CD4+CD57+ and CD8+CD57+ T cell subsets in chronic lymphocytic leukemia. J Immunol. 1997; 158(3):1482-1489.

Porakishvili N, Roschupkina T, Kalber T, et al. Expansion of CD4+ T cells with a cytotoxic phenotype in patients with B-chronic lymphocytic leukaemia (B-CLL). Clin Exp Immunol. 2001;126(1):29-36.

Goolsby CL, Kuchnio M, Finn WG, et al.

Expansions of clonal and oligoclonal T cells in B-cell chronic lymphocytic leukemia are primarily restricted to the CD3(+)CD8(+) T-cell population. Cytometry. 2000; 42(3):188-195.

Christopoulos P, Pfeifer D, BartholomĂŠ K, et al. Definition and characterization of the systemic T-cell dysregulation in untreated indolent B-cell lymphoma and very early CLL. Blood. 2011;117(14):3836-3846.

Kimby E, Mellstedt H, Nilsson B, et al. T lymphocyte subpopulations in chronic lymphocytic leukemia of B cell type in relation to immunoglobulin isotype(s) on the leukemic clone and to clinical features. Eur

571

M. Palma et al.

J Haematol. 1987;38(3):261-267.

29. Tonino SH, van de Berg PJ, Yong SL, et al.

Expansion of effector T cells associated with decreased PD-1 expression in patients with indolent B cell lymphomas and chronic lymphocytic leukemia. Leuk Lymphoma.

2012;53(9):1785-1794.

30. Brusa D, Serra S, Coscia M, et al. The PD1/PD-L1 axis contributes to T-cell dysfunction in chronic lymphocytic leukemia.

Haematologica. 2013;98(6):953-963.

31. Novak M, Prochazka V, Turcsanyi P, et al.

Numbers of CD8+PD-1+ and CD4+PD-1+ Cells in Peripheral Blood of Patients with Chronic Lymphocytic Leukemia Are Independent of Binet Stage and Are Significantly Higher Compared to Healthy Volunteers. Acta Haematol. 2015; 134(4):208-214.

32. Gassner FJ, Zaborsky N, Neureiter D, et al.

Chemotherapy-induced augmentation of T cells expressing inhibitory receptors is reversed by treatment with lenalidomide in chronic lymphocytic leukemia.

Haematologica. 2014;99(5):67-69.

33. Ramsay AG, Clear AJ, Fatah R, et al. Multiple inhibitory ligands induce impaired T-cell immunologic synapse function in chronic lymphocytic leukemia that can be blocked with lenalidomide: establishing a reversible immune evasion mechanism in human cancer. Blood. 2012;120(7):1412-1421.

34. Nunes C, Wong R, Mason M, et al.

Expansion of a CD8(+)PD-1(+) replicative senescence phenotype in early stage CLL patients is associated with inverted CD4:CD8 ratios and disease progression.

Clin Cancer Res. 2012;18(3):678-687.

572

35. Li M, Sun XH, Zhu XJ, et al. HBcAg induces PD-1 upregulation on CD4+T cells through activation of JNK, ERK and PI3K/AKT pathways in chronic hepatitis-B-infected patients. Lab Invest. 2012;92(2):295-304.

36. Adekambi T, Ibegbu CC, Kalokhe AS, et al.

Distinct effector memory CD4+ T cell signatures in latent Mycobacterium tuberculosis infection, BCG vaccination and clinically resolved tuberculosis. PLoS One. 2012; 7(4):e36046.

37. Jitschin R, Braun M, BĂźttner M, et al. CLLcells induce IDOhi CD14+HLA-DRlo myeloid-derived suppressor cells that inhibit T-cell responses and promote TRegs. Blood. 2014;124(5):750-760.

38. Scrivener S, Kaminski ER, Demaine A, et al.

Analysis of the expression of critical activation/interaction markers on peripheral blood T cells in B-cell chronic lymphocytic leukaemia: evidence of immune dysregulation. Br J Haematol. 2001;112(4):959-964.

39. Frydecka I, Kosmaczewska A, Bocko D, et al. Alterations of the expression of T-cellrelated costimulatory CD28 and downregulatory CD152 (CTLA-4) molecules in patients with B-cell chronic lymphocytic leukaemia. Br J Cancer. 2004;90:2042- 2048.

40. Motta M, rassenti L, Shelvin BJ, et al.

Increased expression of CD152 (CTLA-4) by normal T lymphocytes in untreated patients with B-cell chronic lymphocytic leukemia. Leukemia. 2005;19(10):17881793.

41. Rossmann ED, Jeddi-Tehrani M, Osterborg A, et al. T-cell signaling and costimulatory molecules in B-chronic lymphocytic leukemia (B-CLL): an increased abnormal

42.

43.

44.

45.

46.

47.

48.

expression by advancing stage. Leukemia.

2003;17(11):2252-2254.

Schneider H, Rudd CE. Rudd, Diverse mechanisms regulate the surface expression of immunotherapeutic target ctla-4.

Front Immunol. 2014;5:619.

Podhorecka M, Dmoszynska A, Rolinski J, et al. T type 1/type 2 subsets balance in Bcell chronic lymphocytic leukemia--the three-color flow cytometry analysis. Leuk Res. 2002;26(7):657-660.

Horna P, Sotomayor EM. Cellular and molecular mechanisms of tumor-induced T-cell tolerance. Curr Cancer Drug Targets.

2007;7(1):41-53.

Del Poeta G, Del Principe MI, Zucchetto A, et al. CD69 is independently prognostic in chronic lymphocytic leukemia: a comprehensive clinical and biological profiling study. Haematologica. 2012;97(2):279-287.

Gassner FJ, Weiss L, Geisberger R, et al.

Fludarabine modulates composition and function of the T cell pool in patients with chronic lymphocytic leukaemia. Cancer Immunol Immunother. 2011;60(1):75-85.

Lundin J, Porwit-MacDonald A, Rossmann ED, et al. Cellular immune reconstitution after subcutaneous alemtuzumab (antiCD52 monoclonal antibody, CAMPATH1H) treatment as first-line therapy for B-cell chronic lymphocytic leukaemia. Leukemia.

2004;18(3):484-490.

Ding W, Dong H, Call TG, et al. PD-1 Blockade with Pembrolizumab (MK-3475) in Relapsed/Refractory CLL Including Richter Transformation: An Early Efficacy Report from a Phase 2 Trial (MC1485).

Blood. 2015;23:834.

haematologica | 2017; 102(3)

ARTICLE

Non-Hodgkin Lymphoma

The small FOXP1 isoform predominantly expressed in activated B cell-like diffuse large B-cell lymphoma and full-length FOXP1 exert similar oncogenic and transcriptional activity in human B cells

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Martine van Keimpema,1 Leonie J. Grüneberg,1 Esther J.M. Schilder-Tol,1 Monique E.C.M. Oud,1 Esther A. Beuling,1 Paul J. Hensbergen,2 Johann de Jong,3 Steven T. Pals1,* and Marcel Spaargaren1*

Department of Pathology, Lymphoma and Myeloma Center Amsterdam (LYMMCARE), Academic Medical Center; 2Center for Proteomics and Metabolomics, Leiden University Medical Center and 3Division of Molecular Carcinogenesis, Netherlands Cancer institute, Amsterdam, The Netherlands 1

Haematologica 2017 Volume 102(3):573-583

*STP and MS contributed equally to this work

ABSTRACT

T

he forkhead transcription factor FOXP1 is generally regarded as an oncogene in activated B cell-like diffuse large B-cell lymphoma. Previous studies have suggested that a small isoform of FOXP1 rather than full-length FOXP1, may possess this oncogenic activity. Corroborating those studies, we herein show that activated B cell-like diffuse large B-cell lymphoma cell lines and primary activated B cell-like diffuse large B-cell lymphoma cells predominantly express a small FOXP1 isoform, and that the 5’-end of the Foxp1 gene is a common insertion site in murine lymphomas in leukemia virus- and transposon-mediated insertional mutagenesis screens. By combined mass spectrometry, (quantative) reverse transcription polymerase chain reaction/sequencing, and small interfering ribonucleic acid-mediated gene silencing, we determined that the small FOXP1 isoform predominantly expressed in activated B cell-like diffuse large B-cell lymphoma lacks the N-terminal 100 amino acids of full-length FOXP1. Aberrant overexpression of this FOXP1 isoform (ΔN100) in primary human B cells revealed its oncogenic capacity; it repressed apoptosis and plasma cell differentiation. However, no difference in potency was found between this small FOXP1 isoform and full-length FOXP1. Furthermore, overexpression of full-length FOXP1 or this small FOXP1 isoform in primary B cells and diffuse large B-cell lymphoma cell lines resulted in similar gene regulation. Taken together, our data indicate that this small FOXP1 isoform and full-length FOXP1 have comparable oncogenic and transcriptional activity in human B cells, suggesting that aberrant expression or overexpression of FOXP1, irrespective of the specific isoform, contributes to lymphomagenesis. These novel insights further enhance the value of FOXP1 for the diagnostics, prognostics, and treatment of diffuse large B-cell lymphoma patients.

Introduction The forkhead transcription factor FOXP1 plays an important role in a wide variety of biological processes, including T- and B-cell development and function.1-5 Furthermore, FOXP1 has been recognized as a potential oncogene in hepatocellular carcinoma, pancreatic cancer, and various types of B-cell non-Hodgkin lymphomas.1-4 In hepatocellular carcinoma, diffuse large B-cell lymphoma (DLBCL), and mucosa-associated lymphoid tissue (MALT) lymphoma, overexpression of FOXP1, by chromosomal translocations, copy number alterations, or other means, haematologica | 2017; 102(3)

Correspondence: marcel.spaargaren@amc.uva.nl

Received: September 14, 2016.

Accepted: November 24, 2016.

Pre-published: December 1, 2016. doi:10.3324/haematol.2016.156455 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/573 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

573

M. van Keimpema et al.

is associated with poor prognosis and transformation to aggressive lymphoma.3,5,6 Rare but recurrent chromosomal translocations affecting FOXP1 have been found in activated B-cell (ABC)-DLBCL and MALT lymphoma.

The majority of these translocations involve FOXP1 and the immunoglobulin heavy chain (IgH) enhancer (t(3;14)(p13;q32)).7-10 These FOXP1-IgH rearrangements mostly affect the 5′ untranslated region of FOXP1 and result in overexpression of full-length FOXP1.11 NonIG/FOXP1 rearrangements have also been described, and these often target FOXP1 downstream of its first coding exon, resulting in increased expression of N-terminally truncated FOXP1 isoforms.12 In addition, expression levels of FOXP1 can be used as a discriminator between the ABC and germinal center (GC) subtypes of DLBCL, which are biologically distinct disease entities. ABCDLBCL combines high FOXP1 expression with an unfavorable prognosis, supporting an oncogenic role of FOXP1.13,14 Paradoxically, FOXP1 is located on a chromosomal region that is associated with a loss of heterozygosity and deletions in a number of solid tumors.1,15 In line with this, FOXP1 transcriptional activity is inhibited in a large number of epithelial malignancies by either a decrease in FOXP1 messenger ribonucleic acid (mRNA), a decrease in FOXP1 protein levels, or by aberrant cytoplasmic localization of FOXP1.16 Moreover, high FOXP1 expression is associated with favorable prognosis in breast cancer, lung cancer, epithelial ovarian carcinoma and peripheral T–cell lymphoma.17-22 A possible explanation for the above-mentioned apparently contradictory role of FOXP1 as either an oncogene or a tumor suppressor gene was presented with the identification of smaller FOXP1 isoforms (encoding proteins with N-terminal deletions), that are preferentially expressed in ABC-DLBCL.23,24 It was proposed that these smaller FOXP1 isoforms might have oncogenic potential in B-cell non-Hodgkin lymphomas, whereas the fulllength protein might function as a tumor suppressor.23,25 The hypothesis that loss of the FOXP1 N-terminus might be linked to malignancy is further supported by a study in which Foxp1 was identified as the second most frequent viral integration sites that results in avian nephroblastoma.26 These insertions clustered within the second coding exon of Foxp1, but did not affect mRNA expression levels,26 suggesting that they might result in expression of an N-terminally truncated Foxp1 protein. Moreover, in contrast to FOXP1-IGH translocations, non-IG/FOXP1 rearrangements, which cause increased expression of Nterminally truncated FOXP1 isoforms, are found as secondary genetic hits acquired during the clinical course of various B-cell neoplasms, suggesting that these smaller isoforms might be involved in disease progression.12 Thus, several lines of evidence suggest that the smaller FOXP1 isoforms, rather than full-length FOXP1 (FOXP1FL), might act as oncogenes in B-cell malignancies.

However, functional studies with these smaller FOXP1 isoforms in B cells, including a direct comparison with the actions of FOXP1-FL, are lacking. This became even more relevant as recent studies by our own and other laboratories have shown that high FOXP1 expression can contribute to B-cell lymphomagenesis by promoting B-cell survival,27-29 inhibiting plasma cell differentiation,30,31 potentiating Wnt/β-catenin signaling,32 and suppressing major histocompatibility complex (MHC) class II expression.28,33 Therefore, we herein determined the identity of the small 574

FOXP1 isoform (FOXP1-iso) predominantly expressed in ABC-DLBCL and studied its oncogenic potential and transcriptional activity, in direct comparison to FOXP1-FL in DLBCL cell lines and primary human B cells.

Methods Constructs LZRS-FOXP1-IRES and LZRS-BCL6-IRES-GFP were generated as previously described.27,30 A FOXP1-iso construct that starts translation from the second coding ATG (in exon 6; M101), encoding a 100 AA N-terminally deleted FOXP1, was PCR-cloned and subcloned into LZRS.

B-cell cultures, cell-lines, retroviral transduction and siRNA-mediated knockdown With informed consent and approval of the Academic Medical Center review board, human B cells were isolated, cultured, and retrovirally transduced, followed by 3 day recovery, after which cells were passaged and cultured with or (for plasma cell differentiation experiments, or as indicated) without CD40L-L cells, essentially as described.30 For microarray analysis, after transduction cells were cultured without cytokines for three days.

DLBCL cell lines OCI-Ly1, OCI-Ly3, OCI-Ly7 and OCI-Ly10 were obtained and cultured as described.34 HBL1, TMD8 and RIVA were kindly provided by Dr. G. Lenz (Munster, Germany).

Retroviral transduction and small interfering ribonucleic acid (siRNA)-mediated knockdown were performed as described.27

Immunoblotting 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting with rabbit polyclonal anti-FOXP1 ab16645 (Abcam) was performed as described.30

Quantitative reverse transcription polymerase chain reaction (RT-PCR) Quantitative PCR was conducted exactly as described.27

Mass spectrometry FOXP1 protein was immunoprecipitated from OCI-Ly10 cells using anti-FOXP1 antibody (JC12) and analyzed by SDS-PAGE and silver stained. Protein bands corresponding to FOXP1 full-length and the smaller isoform were trypsin-digested as described35 and analyzed by nano-LC-ESI-ion trap MS/MS and by MALDI-ToF MS. For details, see the Online Supplementary Information. FOXP1 peptides identified by both methods in the FOXP1 full-length and isoform protein bands, respectively, were combined to get the overall picture of exons covered by these peptides.

Flow cytometry, proliferation and cell cycle analysis, and Caspase-Glo 3/7 assay For identification of plasma cells, cells were stained with antihuman CD38-APC and CD20-PE and analyzed on a FACSCanto.

Flow cytometry of green fluorescent protein (GFP) and/or yellow fluorescent protein (YFP) fluorescence, proliferation and cell-cycle analysis by eFluor 670 and propidium iodide, and the Caspase-Glo 3/7 assay (Promega, Madison, WI, USA), was performed as described.27

Enzyme-linked immunosorbent assay (ELISA) and enzyme-linked immunospot (ELISPOT) assay Human immunoglobulin G (IgG), immunoglobulin M (IgM), haematologica | 2017; 102(3)

Identification and oncogenic activity of FOXP1-iso

immunoglobulin A (IgA) and IgG isotypes ELISA, and IgG and IgM ELISPOTs were performed as described.30

Microarray analysis Microarray analysis was performed essentially as described.27,36 Data were analyzed and heatmaps were generated using the microarray analysis and visualization platform R2.

patients (7/11) and ABC-DLBCL cell lines (4/6) predominantly express the smaller FOXP1 protein species, whereas all GC-DLBCL patients, GC-DLBCL cell lines, and peripheral blood B cells from healthy donors predominantly express FOXP1-FL protein (Figure 1A). These results corroborate a previous study by Brown et al.,23 suggesting that a small isoform of FOXP1 rather than FOXP1FL, might have oncogenic potential in ABC-DLBCL.

Results

The 5’-end of the Foxp1 gene is a frequent target of insertional mutagenesis in murine lymphoma models

A small FOXP1 isoform is highly expressed in ABC-DLBCL

Additional support for an oncogenic role of FOXP1, including FOXP1-iso, is provided by insertional mutagenesis screens in murine lymphoma models, using either murine leukemia virus (MuLV) or sleeping beauty (SB) transposon.41,42 In these datasets Foxp1 ranked number 40 and number 13 of the most frequently targeted genes, respectively42 (and unpublished results). Herein, we show that in both models, most insertions are present at the 5’end of Foxp1 (26/26 and 47/51, respectively), and several insertions are found between the start of the first coding exon (mouse Foxp1 exon 4; corresponding to human

FOXP1 is a discriminator (between the ABC and GC subtype), prognosticator, and putative oncogene in DLBCL.5,13,14,37-40 To study the expression of FOXP1 in DLBCL, we analyzed a panel of DLBCL patient biopsies (n=18) and the ABC-DLBCL cell lines OCI-Ly3, OCI-Ly10, U2932, RIVA, HBL1, and TMD8, and the GC-DLBCL cell lines OCI-Ly-1, OCI-Ly7, and OCI-Ly18, by immunoblotting. We observed that the majority of non-GC-DLBCL

GC-DLBCL patients

non GC-DLBCL patients

A

GC-DLBCL

ABC-DLBCL

B

Figure 1. A small FOXP1 isoform is highly expressed in ABC-DLBCL patients, and the 5’end of the Foxp1 gene is a frequent target of insertional mutagenesis in mouse lymphoma models. (A) Western blot analysis for FOXP1 protein expression in tissue samples of non-GC- and GC-DLBCL patients, DLBCL cell lines (i.e., the ABC-DLBCL cell lines OCI-Ly3, OCI-Ly10, U2932, RIVA, HBL1, and TMD8 and the GC-DLBCL cell lines OCI-Ly-1, OCI-Ly7, and OCI-Ly18), and human peripheral blood B cells. NBC: naïve B cell (CD19+CD27–); MBC: memory B cell (CD19+CD27+). Expression of FOXP1 (iso + FL) was relative to β-actin. For comparison, in the blots with DLBCL patient biopsies the FOXP1 levels of OCI-Ly3 were normalized to an arbitrary level of 10 units; please note that shorter exposures were used for accurate quantification of FOXP1 expression in OCI-Ly3. (B) locations of insertions found in the Foxp1 gene in MuLV and sleeping beauty (SB) transposon insertional mutagenesis screens in murine lymphoma models. DLBCL: diffuse large B-cell lymphoma; GC: germinal center; ABC: activated B-cell; iso: isoform; FL: full-length; PBMC: peripheral blood mononuclear cell; MuLV: murine leukemia virus.

haematologica | 2017; 102(3)

575

M. van Keimpema et al.

FOXP1 exon 6) and the third coding exon (7/26 and 9/51, respectively) (Figure 1B). This is of specific interest because truncations in this region might lead to expression of an N-terminally truncated Foxp1 isoform due to the presence of an in frame ATG site in the third coding exon (Figure 1B).

Taken together, the predominant expression of FOXP1iso in ABC-DLBCL, and the clustering of insertions at the 5’ end of the Foxp1 gene, in particular between the first and third coding exon in mouse lymphoma models, indicate that deregulated expression of FOXP1 or (N-terminally) truncated FOXP1 is oncogenic in lymphomas.

The major FOXP1 isoform expressed in ABC-DLBCL lacks the N-terminal 100 amino acids To identify the FOXP1-iso protein specie(s), we isolated FOXP1-FL and the major FOXP1-iso protein from the ABC-DLBCL cell line OCI-Ly10 and performed mass spectrometric analysis of trypsin-digested protein (Figure 2A).

Whereas protein fragments encoded by exon 9, 10, and 13 to 19 were found in mass spectrometric analysis of both FOXP1-FL and FOXP1-iso, three fragments were exclusively detected in the full-length protein: one is encoded by exons 13 and 14 and the two others by exons 6 and 7 (Figure 2A). Intriguingly, the absence of exon 6, the first coding exon, would result in translation of a protein that starts at the first available in-frame start codon, being an ATG site in exon 8; the encoded protein would lack the Nterminal 100 amino acids (AAs), which would be in line with the observed difference in molecular weight between FOXP1-iso and FOXP1-FL. Next, to investigate the potential presence of internal deletions in FOXP1-iso transcripts, we conducted RT-PCR analysis of the GC-DLBCL cell line OCI-Ly1 and the ABCDLBCL cell line OCI-Ly3, the latter expressing high levels of FOXP1-iso (Figure 1A). Employing primer pairs spanning multiple exons (Online Supplementary Figure S1A), we did not detect any shorter products (Online Supplementary Figure S1B). Indeed, sequencing of these RT-PCR products, as well as additional RT-PCR products covering single exons 3 to 21, confirmed the lack of deletions, insertions or mutations in FOXP1 encoding exons in these cell lines.

These data demonstrate that the major isoform does not contain any “gaps” of 1 or 2 consecutive internal exons in the region of exons 4 to 19. Notably, any protein product of a transcript that lacks more than 2 exons would be too small in size to represent the major FOXP1 isoform. Thus, our combined spectrometry and PCR/sequencing data indicate that the FOXP1-iso transcript lacks mRNA encoded by exons preceding exon 8.

The previous non-quantitative PCR setup may not reveal the absence of multiple 5’ or 3’ exon-encoded sequences in the FOXP1-iso transcript: primers against these regions will bind the transcript encoding FOXP1-FL. Therefore, by quantitative RT-PCR, we compared expression of several 5’ versus 3’ exon-encoded sequences in cell lines which predominantly express either FOXP1-iso (OCI-Ly3 and OCI-Ly10), or FOXP1-FL (OCI-Ly1 and OCI-Ly7) (Figure 2B), and in ABC- and GC-DLBCL patient biopsies (Figure 2C). Interestingly, whereas ABC-DLBCL cell lines and patient biopsies show relatively high expression of the FOXP1 C-terminal region (exons 18-20), they show relatively low expression of regions upstream of exon 7 (Figure 2B,C). In addition, we also performed qRTPCRs with primers targeting alternative exons 4a and 7c, 576

previously described to be present in transcripts encoding small isoforms of FOXP1 (Figure 2D).23 Although mRNA containing these alternative exons could be detected, expression levels were very low and did not correlate with expression of FOXP1-iso. Together with the lack of internal exon deletions, these results strongly suggest that exon 6 and preceding exons are absent in the FOXP1-iso transcript.

Finally, to confirm the absence of exon 6 in the transcript encoding FOXP1-iso, we specifically targeted FOXP1 transcripts containing exon 6 with an exon 6 specific siRNA pool (Figure 2E). Whereas nucleofection of DLBCL cell lines with an siRNA that targets exon 18 silenced expression of both FOXP1-FL and FOXP1-iso, the exon 6 siRNA exclusively reduced expression of FOXP1FL, demonstrating that the FOXP1-iso transcript indeed does not contain exon 6 (Figure 2E).

Combined, our mass spectrometry, RT-PCR, and knockdown data firmly demonstrated the absence of any internal deletions in FOXP1-iso, that FOXP1-iso is encoded by an mRNA that lacks exon 6, and that translation starts from exon 8 resulting in a protein product that lacks the 100 N-terminal AAs (Figure 2F).

FOXP1-iso and FOXP1-FL display similar effects on expansion and survival of primary human B cells We have previously shown that overexpression of FOXP1-FL in primary human B cells directly represses a panel of pro-apoptotic genes, and promotes the expansion of these cells by inhibiting caspase-dependent apoptosis, without affecting B-cell proliferation.27 These data indicate that high FOXP1 expression might contribute to lymphomagenesis by promoting B-cell survival. To investigate the effect of FOXP1-iso on B-cell survival and proliferation, we generated a retroviral expression vector encoding a FOXP1 protein lacking the first 100 AAs (“FOXP1-iso”), followed by IRES-YFP (FOXP1-iso-IRES-YFP). To compare the effects of FOXP1-FL versus FOXP1-iso overexpression, primary human B cells were retrovirally transduced with FOXP1-FL-IRES-YFP, FOXP1-iso-IRES-YFP or controlIRES-YFP (as a negative control) (Figure 3A), and cultured on L cells expressing CD40 ligand (CD40L-L cells), in the presence of interleukin (IL)-21 and IL-2. Consistent with our previous study, the percentage of YFP positive cells in FOXP1-FL transduced cultures rapidly increased with time (Figure 3B). Interestingly, a similar increase in the percentage of YFP positive cells was observed in cultures transduced with FOXP1-iso, indicating that FOXP1–FL and FOXP1-iso are equally potent in promoting selective outgrowth of primary human B cells (Figure 3B).

To assess whether FOXP1-iso also promotes B-cell survival, viable cells of the YFP positive and negative fractions of FOXP1-FL, FOXP1-iso or empty vector transduced cells were sorted, and the percentage of live cells was monitored by flow cytometry during 1 week of subsequent culturing. A clear, and similar, survival benefit was observed in cells transduced with either FOXP1-FL or FOXP1-iso (Figure 3C). Moreover, caspase 3/7 activity was significantly and similarly reduced in FOXP1-FL and FOXP1-iso transduced cells (Figure 3D). Furthermore, expression of a panel of 7 pro-apoptotic genes (HRK, BIK, RASSF6, AIM2, EAF2, TP53INP1 and TP63), previously shown to be repressed by FOXP1-FL, was repressed to a similar extent

by FOXP1-iso overexpression (Figure 3E). Together, these results indicate that FOXP1-iso promotes the same surhaematologica | 2017; 102(3)

Identification and oncogenic activity of FOXP1-iso

A

B

C

E

D

F

Figure 2. Identification of the FOXP1-iso transcript and protein. (A) FOXP1 protein was immunoprecipitated from OCI-Ly10 cells and FOXP1-FL and FOXP1-iso proteins were isolated from a silverstained gel (left). The proteins were trypsin-digested and fragments were subsequently analyzed by mass spectrometry. Fragments found in both proteins are indicated in purple, fragments exclusively found in the FOXP1-FL protein are indicated in green. Different color intensities are used to discriminate individual fragments (right). Alternating exons are indicated by regular and bold font. Amino acids overlapping exon boundaries are indicated in red. Methionine-101 is indicated by yellow shading. (B,C) qRT-PCR analysis for the levels of 5â&#x20AC;&#x2122;FOXP1 exons in cell lines with high levels of FOXP1-FL (OCI-Ly1 and OCI-Ly7) or of FOXP1-iso (OCI-Ly3 and OCI-Ly10) (B), and in ABC-DLBCL and GC-DLBCL patient biopsies. Numbers correspond to numbers of biopsies in Figure 1A (C). Expression levels were normalized for HPRT expression levels and to expression levels in OCI-Ly3 (B) or in ABC-DLBCL patient 1 (C). (D) qRT-PCR analysis for the levels of mRNA transcripts containing alternative exons 4a and 7c. Upper panel: diagram illustrating the exon structure of alternative spliced isoforms that encode for FOXP1-FL (1), or that have been previously described by Brown et al. to encode for the most prominently expressed isoforms (2,3).23 Lower panel: qRT-PCR analysis for the levels of expression of transcript 1, 2, and 3 in cell lines with high levels of FOXP1-FL (OCI-ly1 and OCI-ly7) or of FOXP1-iso (OCI-ly3 and OCI-ly10). Please note the difference in y-axis values. (E) DLBCL cell lines were transfected with siRNA against FOXP1 exon 6, siRNA against FOXP1 exon 18, control siRNA, or were treated in the same way without addition of siRNA (-). Two days after nucleofection with siRNA against FOXP1, cell lysates were harvested and immunoblotted for FOXP1. β-tubulin was used as a loading control. *indicates a non-specific background band. (F) Diagram illustrating the exon structure of FOXP1-FL and FOXP1-iso. FL: fulllength; iso: isoform; sictrl: si control; siFOXP1: small interfering RNA directed against FOXP1; ab: antibody.

haematologica | 2017; 102(3)

577

M. van Keimpema et al.

A

B

C

D

E

F

Figure 3. FOXP1-iso exerts similar effects on B-cell survival as FOXP1-FL. Memory B cells were sorted from human peripheral blood and transduced with either FOXP1-IRES-YFP, FOXP1-iso-IRES-YFP, BCL6-IRES-GFP (F) or ctrl-IRES-YFP, and cultured with CD40L-L cells, IL-21 and IL-2 (A-D,F) or without CD40L-L cells as of day 3 after transduction (E). (A) Representative example of FOXP1 and FOXP1-iso overexpression in primary YFP+ B cells. 6 days after transduction YFP+ cells were sorted and analyzed by immunoblotting. β-tubulin was used as loading control. (B) FOXP1-IRES-YFP, FOXP1-iso-IRES-YFP, and ctrl-IRES-YFP transduced B cells were continuously cultured with IL-21 and IL-2 and CD40L-L cells. The percentage of transduced cells in each culture was followed over time by FACS analysis and normalized to the percentage of transduced cells at day 3 after transduction. Mean ± SD of three independent experiments are shown. Significant differences were observed at day 15 and 20 after transduction between FOXP1 transduced cells vs. ctrl transduced cells (*P<0.05) and between FOXP1-iso transduced and control transduced cells (**P<0.01). (C) A total of 6-7 days after transduction, live YFP positive and negative fractions of FOXP1 and control vector single transduced cultures were sorted. After a recovery period of 4-5 days, the percentage of cells in the FSC/SCC live gate was determined by flow cytometry at three consecutive time points. The data were normalized to the percentage of living cells measured at the first time point. Mean ± SEM of three independent experiments are shown. Significant differences were observed at day 6 after transduction between FOXP1 transduced cells vs. ctrl transduced cells (**P<0.01) and between FOXP1-iso transduced and control transduced cells (*P<0.05). (D) Six days after transduction, the YFP positive fractions of FOXP1-FL, FOXP1-iso, and control vector-transduced cultures were sorted.

After a recovery period of 5-7 days, caspase 3/7 activity was determined by the caspase glo3/7 assay. Values were corrected for the number of living cells as determined by FACS analysis. Mean ± SD of five independent experiments are shown (t-test **P<0.01, *** P<0.001). (E) Six days after transduction YFP positive cells were sorted. Gene expression levels of FOXP1 repressed pro-apoptotic genes were analyzed by quantitative RT-PCR. Expression levels were normalized to expression levels in control transduced cells. Mean ± SEM of at least three independent experiments (except for TP63) are shown (one sample t-test, **P<0.01, ***P<0.001).

(F) 7 days after transduction, cells were labeled with eFluor 670 and cultured for four days, after which the eFluor 670 intensity was determined by flow cytometry.

Representative graphs of two independent experiments are shown. FL: full-length; iso: isoform; ctrl: control.

578

haematologica | 2017; 102(3)

Identification and oncogenic activity of FOXP1-iso

vival pathways with a similar potency as FOXP1-FL in primary human B cells.

Whereas FOXP1-FL promotes selective outgrowth and inhibits apoptosis of primary human B cells, it does not affect proliferation in these cells.27 To investigate whether FOXP1-iso would have the additional capacity to affect proliferation of primary human B cells, we analyzed proliferation in FOXP1-iso, FOXP1-FL, BCL6 (as a positive control), or empty vector (as a negative control) transduced cultures, by labeling the cells with the proliferation dye eFluor 670. After four days of culturing, eFluor 670

A

dilution was not increased in FOXP1-FL nor in FOXP1-isotransduced cells (Figure 3F). In accordance, cell cycle analysis indicated no major changes in cell cycle distribution upon FOXP1-FL or FOXP1-iso overexpression (Online Supplementary Figure S2). BCL6 overexpression, on the other hand, did result in increased eFluor 670 dilution and a higher percentage of cells in synthesis (S) phase (Figure 3F and Online Supplementary Figure S2).

Combined, these results show that FOXP1-iso overexpression has a very similar effect on B-cell outgrowth as that observed with FOXP1-FL, both qualitatively and

Figure 4. FOXP1-iso exerts similar effects on plasma cell differentiation as FOXP1-FL. CD19+CD27+ MBCs were sorted from human peripheral blood and transduced with FOXP1-FL-IRES-YFP, FOXP1-iso-IRES-YFP, or control empty vector and cultured under conditions that promote PC differentiation. Six days after transduction YFP positive cells were sorted (A,C,D). (A) Gene expression levels of PRDM1, IRF4, XBP1, FOXP1, BCL6 and PAX5 were analyzed in sorted cells by qRT-PCR. Expression levels in FOXP1-FL and FOXP1-iso- and controltransduced cells were normalized to β-actin and then to expression levels in control transduced cells. Mean ± SEM of five independent experiments are shown. (**P<0.01; ***P<0.001). (B) Six days after transduction, YFP positive cells were analyzed for CD20 and CD38 surface expression by flow cytometry.

Representative dot plots of one out of 10 independent experiments are shown (left panel). Percentages of CD38+ cells in FOXP1-FL and FOXP1-iso-transduced cultures were normalized to control cultures. Mean ± SD values of 10 independent experiments are shown (right panel) (one sample t-test, ***P<0.001). (C) Equal numbers of sorted cells (50000) were cultured with IL21 and IL-2 for an additional 24 hours. Thereafter, the supernatants were collected, and IgG protein levels were analyzed by ELISA. Levels were normalized to levels in control transduced cells. Mean ± SD of three independent experiments are shown (one sample t-test, *P<0.05, **P<0.01). (D) Equal numbers of sorted cells were plated onto membranes in serial dilutions and cultured with IL-21 and IL-2 for an additional 18 hours, after which numbers of IgM or IgG secreting cells were determined by ELISPOT. Spot numbers were normalized to numbers in stimulated, control transduced cells. Mean ± SD of four independent experiments are shown. (one sample t-test, *P<0.01). FL: fulllength; iso: isoform; IgG: immunoglobulin G; ctrl: control.

B

C

D

haematologica | 2017; 102(3)

579

M. van Keimpema et al.

quantitatively. Like FOXP1-FL, overexpression of FOXP1iso represses a panel of 7 pro-apoptotic genes, promotes survival, and prevents caspase-dependent apoptosis, but does not affect B-cell proliferation.

FOXP1-iso and FOXP1-FL display similar effects on plasma cell differentiation of primary human memory B cells We have previously demonstrated that overexpression of FOXP1-FL in primary human B cells also represses plasma cell differentiation by inhibiting the expression of master regulators of plasma cell differentiation.30 These results indicate that the oncogenic potential of FOXP1 in B cells not only lies in its ability to promote cell survival but also in its ability to block terminal differentiation.

Therefore, sorted CD27+ memory B cells from human peripheral blood were transduced with FOXP1-FL, FOXP1-iso or control empty vector, and cultured under conditions that drive their differentiation towards cells that have acquired a plasma cell phenotype (CD20â&#x20AC;&#x201C;CD38+), display strongly increased expression of the master regulators of plasma cell differentiation PRDM1, IRF4 and XBP1, and secrete immunoglobulins.30 Overexpression of FOXP1-FL or FOXP1-iso resulted in a strong, similar repression of PRDM1, IRF4, and XBP1 in these cells, as determined by qRT-PCRs analysis (Figure 4A). Furthermore, flow cytometry analysis indicated an equal reduction in the formation of CD20â&#x20AC;&#x201C;CD38+ plasma cells upon overexpression of either FOXP1-FL or FOXP1iso (Figure 4B). Moreover, the secretion of IgGs, as determined by ELISA (Figure 4C), and the formation of IgG secreting plasma cells, as determined by ELISPOT (Figure 4D), were reduced to a similar extent upon overexpression of either FOXP1 isoform. These results indicate that FOXP1-FL and FOXP1-iso are equally potent in repressing plasma cell differentiation of primary human B cells.

FOXP1-iso and FOXP1-FL show similar gene expression regulation in DLBCL cell-lines Thus far, our results revealed that the previously described effects of aberrant overexpression of FOXP1FL in primary human B cells, i.e., promotion of B-cell survival and inhibition of plasma cell differentiation, are similarly and equally executed by FOXP1-iso. To obtain a broader insight into potential differences in the functions of FOXP1-iso and FOXP1-FL, we performed gene expression microarray analysis of the GC-DLBCL OCILy1 and OCI-Ly7 (which predominantly express FOXP1FL) and the ABC-DLBCL cell-line OCI-Ly10 (which predominantly expresses FOXP1-iso) (Figure 1A), retrovirally transduced with FOXP1-IRES-YFP, FOXP1-iso-IRESYFP or control empty vector (Figure 5). Upon overexpression of either FOXP1-FL or FOXP1-iso the expression of 79 genes was changed by at least 1.5-fold in at least two cell lines. Chromatin immunoprecipitation sequencing (ChIP-Seq) analysis established that the majority of these genes were also directly bound within 20 kb of the transcription start site by FOXP1 in at least 2 DLBCL cell lines (Figure 5). Among the directly bound and repressed genes are AICDA, encoding activation-induced cytidine deaminase (AID), and several genes characteristic for GC-DLBCL, i.e., MYBL1, MME and LPP,43-45 which, like FOXP1, are also important prognosticators for DLBCL patient survival (high expression of those three genes and low expression of FOXP1 being correlated with better 580

Figure 5. FOXP1-iso and FOXP1-FL regulate the same genes in DLBCL cell lines. DLBCL cell lines were retrovirally transduced with FOXP1-FL-IRES-YFP, FOXP1-iso-IRES-YFP, or control-IRES-YFP, and YFP positive cells were sorted.

Gene expression microarray analysis was performed on these samples. All genes of which the expression was changed by either FOXP1-FL or FOXP1-iso overexpression by at least 1.5-fold in at least two cell lines are shown. Data are represented as z-scores calculated within samples of each cell line. Genes that were identified by ChIP-seq analysis to be bound by FOXP1 within 20kb of their TSS in at least 2 DLBCL cell lines are indicated by an asterisk. The gray squares indicate expression beneath the threshold value (= no expression). FL: fulllength; iso: isoform: CTRL: control.

haematologica | 2017; 102(3)

Identification and oncogenic activity of FOXP1-iso

prognosis),45 indicating an important role for FOXP1 in defining the discriminatory gene signatures of ABC vs.

GC-DLBCL. Importantly, however, all FOXP1-FL-regulated genes were also regulated by FOXP1-iso overexpression and vice versa, and always in the same direction.

Thus, no functional differences in gene regulation were observed between FOPX1-FL and FOXP1-iso in DLBCL cell lines.

Discussion FOXP1 has long been recognized as a putative oncogene in ABC-DLBCL, but molecular mechanisms underlying its oncogenic potential, i.e., promotion of B-cell survival, inhibition of plasma cell differentiation, potentiation of Wnt signaling, and suppression of MHC class II expression have only recently been described.27-33 Previous studies have suggested that smaller FOXP1 isoforms, which are preferentially expressed in ABC-DLBCLs, rather than FOXP1-FL protein, might exert oncogenic roles.1,23 It has been proposed that these smaller isoforms may either act as a dominant negative over FOXP-FL or may have a deregulated and/or distinct function.1,23,26 However, until now, functional studies, directly addressing the activity of these smaller FOXP1 isoforms and comparing the actions of these isoforms with full-length FOXP1 in B cells, have not been performed. Herein, we established that the major isoform of FOXP1 expressed in ABC-DLBCL, FOXP1-iso, lacks the N-terminal 100 AAs, and that aberrant overexpression of this FOXP1-iso (Î&#x201D;N100) in primary human B cells results in a similar and equally strong promotion of Bcell survival and inhibition of plasma cell differentiation as overexpression of FOXP1-FL. Moreover, gene expression microarray analysis in DLBCL cell lines and qRT-PCR analysis in primary human B cells did not reveal any differences in regulation of gene expression by these different protein species.

Foxp1 has previously been identified as the most frequent integration site in a piggyBac transposon screen in a pancreatic tumorigenesis model, and the second most frequent viral integration site in an avian nephroblastoma model.4,26 The insertions in the latter model were clustered around the second coding exon of Foxp1 (corresponding to human FOXP1 exon 7), which may lead to expression of an N-terminally truncated Foxp1 isoform due to the presence of an in-frame ATG site in the third coding exon (Figure 1B). Herein, we show that Foxp1 insertions in insertional mutagenesis screens in murine lymphoma models, using either MuLV or SB transposon, cluster at the 5â&#x20AC;&#x2122; end of Foxp1, upstream of the third coding exon.

Therefore, these insertions might also lead to expression of an N-terminally truncated Foxp1 isoform. Together these insertional mutagenesis screens support the hypothesis that specifically N-terminally truncated Foxp1 isoforms possess oncogenic function.

The exact identity of the major FOXP1-iso protein expressed in ABC-DLBCL has not been completely clarified before. Using semi-quantitative RT-PCR analysis, Brown et al. proposed that smaller isoforms were encoded by two alternative FOXP1 splice variants, one of which might start translation in exon 8 (isoform 2), resulting in a protein that lacks the N-terminal 100 AAs, whereas the other starts translation in exon 6, but does not contain exons 8 and 9, resulting in an internal N-terminal deletion haematologica | 2017; 102(3)

(isoform 3).23 By mass spectrometry, (q)RT-PCR, sequencing, and siRNA-mediated knockdown studies, we established that the major FOXP1-iso expressed in ABC-DLBCL cell lines is encoded by a transcript that lacks exon 6, but does contain exons 7-21, resulting in expression of a protein that lacks the first 100 N-terminal AAs. During the preparation of this manuscript, essentially similar results were reported by Banham's group.46 Since we found evidence for only one transcript encoding for the major isoform, the detection of two FOXP1-iso protein species by western blotting suggests that these two protein species may have undergone different post-translational protein modifications.

A previous study by Wang et al. showed that the murine Foxp1D isoform, which lacks the N-terminus of the protein, exerts stronger repression of an SV40 promoter-driven luciferase construct in transfected HEK293T cells, compared to Foxp1-FL.47 However, we could not find any differences in gene repression strength between FOXP1-FL and FOXP1iso upon overexpression in primary human B cells (Figures 3E and 4A), suggesting that at least these human isoforms do not differ in their strength of transcriptional regulation.

Moreover, the functional consequences of overexpression of these two protein species, at least in the responses studied by us, i.e., survival, proliferation and plasma cell differentiation, were equally affected by both. Notably, a recent study by Walker et al. showed that FOXP1-FL and N-terminal truncated FOXP1 proteins show similar effects on Wnt reporter activity in HEK293T cells, indicating that these two isoforms might also equally potentiate Wnt signaling in DLBCL.32 The differences with the study by Wang et al. might be explained by dissimilarities in species (murine versus human FOXP1), cell types (HEK293T versus DLBCL cell lines) or the assay system being used (artificial versus endogenous promoters). Together, these results suggests that, at least in B cells and DLBCL cell lines, the effect of FOXP1-FL and FOXP1-iso on gene expression and cellular responses seem to be similar.

Our results imply that expression of FOXP1-iso in B-cell non-Hodgkin lymphomas versus expression of FOXP1-FL in epithelial tissues cannot explain its paradoxical actions as an oncogene in the former and as a tumor suppressor in the latter tissues. These opposing functions of FOXP1 in different cell types are therefore more likely to be due to tissue-specific expression of transcriptional co-regulators or regulation of transcriptional targets, an intriguing and important aspect that warrants future studies. By ChIPseq analysis of FOXP1 in DLBCL cell lines, we previously established that FOXP1 binding sites in these cells are enriched for binding sites and consensus motifs for other transcription factors that regulate ABC-DLBCL pathogenesis.27 By analogy to its family member FOXP3,48,49 these results suggest that FOXP1 might preferentially bind sites co-occupied by other transcription factors and regulate gene expression through physical or functional interaction with those factors.27 Therefore, the transcriptional outcome of high FOXP1 expression in different cell types might be dictated by expression levels of other transcriptional co-regulators in these cells.

Combined, our results indicate that FOXP1-FL and FOXP1-iso do not differ in terms of oncogenic activity and gene regulation in B cells and DLBCLs, and that total levels of FOXP1 expression, rather than the relative expression of specific isoforms, contributes to the oncogenic potential of FOXP1 in B-cell lymphomas. The specific mechanisms 581

M. van Keimpema et al.

underlying transcriptional regulation of FOXP1-FL versus FOXP1-iso expression remain to be established.

Interestingly, however, a previous study showed that activation of the B-cell receptor in primary B cells resulted in specific induction of FOXP1-iso expression.23 Since ABCDLBCLs are characterized by chronic active B-cell-receptor signaling,50 this might explain the relatively high expression of FOXP1-iso in these lymphomas. Importantly, however, since the total levels of FOXP1 are also higher in ABC- as compared to GC-DLBCL,13,14,38-40 deregulated expression of FOXP1, irrespective of the isoform, may contribute to the worse prognosis of ABC-DLBCL patients,5,38-40 thereby reflecting its role in lymphomagenesis. Taken together, these novel insights into the function of FOXP1 isoforms in

References 1. Koon HB, Ippolito GC, Banham AH, Tucker PW. FOXP1: a potential therapeutic target in cancer. Expert Opin Ther Targets.

2007;11(7):955-965.

2. Katoh M, Igarashi M, Fukuda H, Nakagama H, Katoh M. Cancer genetics and genomics of human FOX family genes. Cancer Lett.

2013;328(2):198-206.

3. Zhang Y, Zhang S, Wang X, et al. Prognostic significance of FOXP1 as an oncogene in hepatocellular carcinoma. J Clin Pathol.

2012;65(6):528-533.

4. Rad R, Rad L, Wang W, et al. A conditional piggyBac transposition system for genetic screening in mice identifies oncogenic networks in pancreatic cancer. Nat Genet.

2015;47:47-56.

5 Banham AH, Connors JM, Brown PJ, et al.

Expression of the FOXP1 transcription factor is strongly associated with inferior survival in patients with diffuse large B-cell lymphoma. Clin Cancer Res. 2005;11(3): 1065-1072.

6. Sagaert X, de Paepe P, Libbrecht L, et al.

Forkhead box protein P1 expression in mucosa-associated lymphoid tissue lymphomas predicts poor prognosis and transformation to diffuse large B-cell lymphoma.

J Clin Oncol. 2006;24(16):2490-2497.

7. Streubel B, Vinatzer U, Lamprecht A, Raderer M, Chott A. T(3;14)(p14.1;q32) involving IGH and FOXP1 is a novel recurrent chromosomal aberration in MALT lymphoma. Leukemia. 2005;19(4):652-658.

8. Fenton JA, Schuuring E, Barrans SL, et al. t(3;14)(p14;q32) Results in aberrant expression of FOXP1 in a case of diffuse large Bcell lymphoma. Genes Chromosom Cancer. 2006;45(2):164-168.

9. Wlodarska I, Veyt E, de Paepe P, et al.

FOXP1, a gene highly expressed in a subset of diffuse large B-cell lymphoma, is recurrently targeted by genomic aberrations.

Leukemia. 2005;19(8):1299-1305.

10. Haralambieva E, Adam P, Ventura R, et al.

Genetic rearrangement of FOXP1 is predominantly detected in a subset of diffuse large B-cell lymphomas with extranodal presentation. Leukemia. 2006;20(7):13001303.

11. Goatly A, Bacon CM, Nakamura S, et al.

FOXP1 abnormalities in lymphoma: translocation breakpoint mapping reveals insights into deregulated transcriptional

582

controlling the transcriptional program, survival and differentiation of B cells advance our understanding of the role of FOXP1 in lymphomagenesis and further enhance the value of FOXP1 for diagnostics, prognostics, and treatment of DLBCL patients.

Acknowledgments The authors thank Berend Hooibrink and Toni van Capel for FACS sorting, and Jan Koster and Richard Volckman for providing help with microarray analysis.

Funding This study was supported by a grant from the Dutch Cancer Society.

control. Mod Pathol. 2008;21(7):902-911.

12. Rouhigharabaei L, Finalet Ferreiro J, Tousseyn T et al. Non-IG aberrations of FOXP1 in B-Cell malignancies lead to an aberrant expression of N-truncated isoforms of FOXP1. PLoS ONE. 2014;9(1): e85851.

13. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275-282.

14. Staudt LM, Dave S. The biology of human lymphoid malignancies revealed by gene expression profiling. Adv Immunol.

2005;87:163-208.

15. Krohn A, Seidel A, Burkhardt L, et al.

Recurrent deletion of 3p13 targets multiple tumour suppressor genes and defines a distinct subgroup of aggressive ERG fusionpositive prostate cancers. J Pathol.

2013;231(1):130-141.

16. Banham AH, Beasley N, Campo E, et al.

The FOXP1 winged helix transcription factor is a novel candidate tumor suppressor gene on chromosome 3p. Cancer Res. 2001;61(24):8820-8829.

17. Fox SB, Brown P, Han C, et al. Expression of the forkhead transcription Factor FOXP1 is associated with estrogen receptor alpha and improved survival in primary human breast carcinomas.

Clin Cancer Res. 2004;10(10):3521-3527.

18. Shigekawa T, Ijichi N, Ikeda K, et al.

FOXP1, an estrogen-inducible transcription factor, modulates cell proliferation in breast cancer cells and 5-year recurrence-free survival of patients with tamoxifen-treated breast cancer. Horm Cancer. 2011;2(5):286297.

19. Rayoo M, Yan M, Takano EA et al.

Expression of the forkhead box transcription factor FOXP1 is associated with oestrogen receptor alpha, oestrogen receptor beta and improved survival in familial breast cancers. J Clin Pathol. 2009;62(10): 896-902.

20. Yamada S, Sato F, Xia H, et al. Forkhead box P1 overexpression and its clinicopathologic significance in peripheral T-cell lymphoma, not otherwise specified. Human Pathol.

2012;43(8):1322-1327.

21. Zhu L, Hu Z, Liu J, Gao J, Lin B. Gene expression profile analysis identifies metastasis and chemoresistance-associated genes in epithelial ovarian carcinoma cells. Med Oncol. 2014;32(1):1-13.

22. Feng J, Zhang X, Zhu H, Wang X, Ni S, Huang J. High expression of FoxP1 is associated with improved survival in patients with non-small cell lung cancer. Am J Clin Pathol. 2012;138(2):230-235.

23. Brown PJ, Ashe SL, Leich E, et al.

Potentially oncogenic B-cell activationinduced smaller isoforms of FOXP1 are highly expressed in the activated B cell-like subtype of DLBCL. Blood. 2008;111(5): 2816-2824.

24. Courts C, Brunn A, Montesinos-Rongen M, et al. Preferential Expression of Truncated Isoforms of FOXP1 in Primary Central Nervous System Lymphoma. J Neuropathol Exp Neurol. 2009;68(9):727-741.

25. Green MR, Gandhi MK, Courtney MJ, Marlton P, Griffiths L. Relative abundance of full-length and truncated FOXP1 isoforms is associated with differential NFkappaB activity in Follicular Lymphoma.

Leuk Res. 2009;33(12):1699-1702.

26. Pajer P, Pecenka V, Kralova J, et al.

Identification of potential human oncogenes by mapping the common viral integration sites in avian nephroblastoma.

Cancer Res. 2006;66(1):78-86.

27. van Keimpema M, Gruneberg LJ, Mokry M, et al. FOXP1 directly represses transcription of pro-apoptotic genes and cooperates with NF-kappaB to promote survival of human B-cells. Blood. 2014;124(23):3431-3440.

28. Dekker JD, Park D, Shaffer AL, et al.

Subtype-specific addiction of the activated B-cell subset of diffuse large B-cell lymphoma to FOXP1. Proc Natl Acad Sci. 2016;113(5):E577-586.

29. Flori M, Schmid CA, Sumrall ET, et al. The hematopoietic oncoprotein FOXP1 promotes tumor cell survival in diffuse large Bcell lymphoma by repressing S1PR2 signaling. Blood. 2016;127(11):1438-1448.

30. van Keimpema M, GrĂźneberg LJ, Mokry M, et al. The forkhead transcription factor FOXP1 represses human plasma cell differentiation. Blood. 2015;126(18):2098-2109.

31. Tsai DY, Hung KH, Lin IY et al. Uncovering MicroRNA regulatory hubs that modulate plasma Cell differentiation. Sci Rep. 2015;11(5):17957.

32. Walker MP, Stopford CM, Cederlund M, et al. FOXP1 potentiates Wnt/beta-catenin signaling in diffuse large B cell lymphoma.

Sci Signal. 2015;8(362):ra12.

33. Brown PJ, Wong KK, Felce SL et al. FOXP1 suppresses immune response signatures and MHC class II expression in activated B-

haematologica | 2017; 102(3)

Identification and oncogenic activity of FOXP1-iso

34.

35.

36.

37.

38.

cell-like diffuse large B-cell lymphomas.

Leukemia. 2015;30(3):605-616.

Tjin EPM, Groen RWJ, Vogelzang I, et al.

Functional analysis of HGF/MET signaling and aberrant HGF-activator expression in diffuse large B-cell lymphoma. Blood.

2006;107(2):760-768.

Zauner G, Hoffmann M, Rapp E, et al.

Glycoproteomic analysis of human fibrogen reveals novel regions of O-glycosylation. J Proteome Res. 2012;11(12):58045814.

Kocemba KA, van Andel H, de HaanKramer A, et al. The hypoxia target adrenomedullin is aberrantly expressed in multiple myeloma and promotes angiogenesis. Leukemia. 2013; 27(8):1729-1737.

Hu CR, Wang JH, Wang R, Sun Q, Chen LB. Both FOXP1 and p65 expression are adverse risk factors in diffuse large B-cell lymphoma: A retrospective study in China.

Acta Histochem. 2013;115(2):137-143.

Hoeller S, Schneider A, Haralambieva E, Dirnhofer S, Tzankov A. FOXP1 protein overexpression is associated with inferior outcome in nodal diffuse large B-cell lymphomas with non-germinal centre phenotype, independent of gains and structural aberrations at 3p14.1. Histopathology.

2010;57(1):73-80.

haematologica | 2017; 102(3)

39. Barrans SL, Fenton JA, Banham A, Owen RG, Jack AS. Strong expression of FOXP1 identifies a distinct subset of diffuse large Bcell lymphoma (DLBCL) patients with poor outcome. Blood. 2004;104(9):2933-2935.

40. Yu B, Zhou X, Li B, Xiao X, Yan S, Shi D. FOXP1 expression and its clinicopathologic significance in nodal and extranodal diffuse large B-cell lymphoma. Ann Hematol.

2011;90(6):701-708.

41. Uren AG, Kool J, Matentzoglu K, et al.

Large-scale mutagenesis in p19ARF- and p53-deficient mice identifies cancer genes and their collaborative networks. Cell.

2008;133(4):727-741.

42. de Jong J, de Ridder J, van der Weyden L, et al. Computational identification of insertional mutagenesis targets for cancer gene discovery. Nucleic Acids Res. 2011;39(15): e105.

43. Alizadeh AA, Eisen MB, Davis RE et al.

Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-511.

44. Wright G, Tan B, Rosenwald A, Hurt EH, Wiestner A, Staudt LM. A gene expressionbased method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma.

Proc Natl Acad Sci USA. 2003;100 (17):9991-9996.

45. Jais JP, Haioun C, Molina TJ, et al. The expression of 16 genes related to the cell of origin and immune response predicts survival in elderly patients with diffuse large B-cell lymphoma treated with CHOP and rituximab. Leukemia. 2008;22(10):19171924.

46. Brown PJ, Gascoyne DM, Lyne L, et al. Nterminally truncated FOXP1 protein expression and alternate internal FOXP1 promoter usage in normal and malignant B cells. Haematologica. 2016;101(7):861-71.

47. Wang B, Lin D, Li C, Tucker P. Multiple domains define the expression and regulatory properties of Foxp1 forkhead transcriptional repressors. J Biol Chem. 2003; 278(27):24259-24268.

48. Rudra D, deRoos P, Chaudhry A et al.

Transcription factor Foxp3 and its protein partners form a complex regulatory network. Nat Immunol. 2012;13(10):10101019.

49. Samstein RM, Arvey A, Josefowicz SZ, et al. Foxp3 exploits a pre-existent enhancer landscape for regulatory T cell lineage specification. Cell. 2012;151(1):153-166.

50. Davis RE, Ngo VN, Lenz G, et al. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature. 2010;463 (7277):88-92.

583

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Non-Hodgkin Leukemia

Ferrata Storti Foundation

Factors related to the relative survival of patients with diffuse large B-cell lymphoma in a population-based study in France: does socio-economic status have a role?

Sandra Le Guyader-Peyrou,1,2 Sébastien Orazio,1,2 Olivier Dejardin,3 Marc Maynadié,4 Xavier Troussard5,6 and Alain Monnereau1,2

Haematologica 2017 Volume 102(3):584-592

Registre des Hémopathies Malignes de la Gironde, Institut Bergonie, Bordeaux; 2 University of Bordeaux, Inserm, Bordeaux Population Health Research Center, Team EPICENE, UMR 1219, F-33000; 3University Hospital of Caen, U1086 INSERM UCBN «Cancers & Préventions»; 4Registre des Hémopathies Malignes de Côte d’Or, EA4184, Université de Bourgogne, Dijon; 5Registre des Hémopathies Malignes de Basse Normandie, Caen and 6Laboratoire d’Hématologie, CHU de Caen, France 1

ABSTRACT

T

Correspondence: s.leguyaderpeyrou@bordeaux.unicancer.fr

Received: July 19, 2016.

Accepted: November 24, 2016.

Pre-published: December 1, 2016. doi:10.3324/haematol.2016.152918 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/584 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

584

he survival of patients with diffuse large B-cell lymphoma has increased during the last decade as a result of addition of antiCD20 to anthracycline-based chemotherapy. Although the trend is encouraging, there are persistent differences in survival within and between the USA and European countries suggesting that non-biological factors play a role. Our aim was to investigate the influence of such factors on relative survival of patients with diffuse large B-cell lymphoma.

We conducted a retrospective, multicenter, registry-based study in France on 1165 incident cases of diffuse large B-cell lymphoma between 2002 and 2008. Relative survival analyses were performed and missing data were controlled with the multiple imputation method. In a multivariate analysis, adjusted for age, sex and International Prognostic Index, we confirmed that time period was associated with a better 5-year relative survival. The registry area, the medical specialty of the care department (onco-hematology versus other), the time to travel to the nearest teaching hospital, the place of treatment (teaching versus not-teaching hospital borderline significance), a comorbidity burden and marital status were independently associated with the 5-year relative survival. Adjusted for first-course treatment, inclusion in a clinical trial and treatment discussion in a multidisciplinary meeting were strongly associated with a better survival outcome. In contrast, socio-economic status (determined using the European Deprivation Index) was not associated with outcome.

Despite therapeutic advances, various non-biological factors affected the relative survival of patients with diffuse large B-cell lymphoma. The notion of lymphoma-specific expertise seems to be essential to achieve optimal care management and reopens the debate regarding centralization of these patients’ care in hematology/oncology departments.

Introduction Non-Hodgkin lymphoma (NHL) is the most frequent hematologic malignancy in the world and comprises a heterogeneous group of more than 40 different subtypes.1 Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of NHL, accounting for up to 25-30% of all cases globally, with an age-adjusted incidence rate of 5.0 cases per 100,000 person-years in both sexes worldwide.2-4 haematologica | 2017; 102(3)

Socio-economic factors and survival in DLBCL patients

Although DLBCL is curable in many cases, it remains an aggressive disease and fatal if left untreated or treated improperly. DLBCL usually affects adults over 60 years old, although it occurs in patients of all ages, including children, and needs many courses of curative treatments (polychemotherapy associated or not with immunotherapy followed by radiotherapy for localized disease). Recent data from the USA5 show a significant reduction in DLBCL mortality, reflecting a better survival. Positive trends in DLBCL survival were also observed in population-based studies in France and Europe beginning in the early 2000s.6-8 However, if these trends in DLBCL survival are due to clinical advances in the treatment of the disease (i.e., the introduction of rituximab), they may not be equally distributed in the population. Indeed, persistent differences in DLBCL survival are observed within and between countries (USA and European countries) suggesting the role of variations in access to/quality of care and availability of new drugs.

Moreover, a growing body of literature suggests a persistent relationship between non-biological factors such as socio-economic status (SES) and health status that may influence survival of patients with various common cancers. Individual characteristics (e.g., age, sex, marital status),9,10 contextual data such as a high Deprivation Index (living in a poorer district),11,12 living in a rural area,13,14 living far away from the referral center,15,16 being treated in a community hospital17,18 and low hospital volume19 have been associated with poorer outcome.

However, only a few studies assessing the impact of non-biological factors on NHL survival have been reported and most of them focused on the influence of SES or place of residence on NHL survival20-22 or, more recently, specifically DLBCL survival.23-25 These latest studies took into account, in their analyses, the introduction of rituximab in DLBCL treatment in 2002.

The aim of this study was to investigate the influence of socio-economic determinants, care management and place of care on relative survival of DLBCL patients during the early rituximab era.

Methods Data source Our study concerns all DLBCL cases diagnosed between 01/01/2002 and 12/31/2008 and collected in three populationbased registries of hematologic malignancies in France (BasseNormandie, Côte d’Or and Gironde). The cases were classified according to the International Classification of Diseases for Oncology 3rd edition using morphology codes: 9678/3, 9679/3, 9680/3, and 9684/3.26,27 All pathology reports were reviewed to ascertain the diagnosis of DLBCL. The study was approved by the French national consultative committee.

roadmap database (Multinet TéléAtlas©), and expressed as travel time in minutes.

Vital status was determined from the date of diagnosis to the death or until 30th June, 2013 using the Repertoire National d’Identification des Personnes Physiques (RNIPP). Loss to follow-up was <2%.

Aggregate data of the study population Residential address at diagnosis was geocoded and allocated to an Ilôts Regroupés pour l’Information Statistique (IRIS) the smallest geographical area for which census data are available. We used the French ecological European Deprivation Index (EDI),28 which attributes a social deprivation score to each IRIS,29 as a proxy measure of individual SES at the time of diagnosis. Residential address at diagnosis was defined according to the rural or urban commuting area code. Care provision was determined by the density of general practitioners per IRIS.

Statistical analysis

Patients’ characteristics were compared with the t test or χ2 test, as appropriate. The percentage of missing values was also provided.

We first estimated net survival using the Pohar Perme unbiased method for descriptive analyses.30 To evaluate the impact of prognostic factors on relative survival, we used the Esteve approach.31 Data were missing for a few variables in our dataset, with more than 10% of values missing for three variables. In this context, a complete-case method multivariate analysis would have dropped 35% of the subjects from the dataset. We, therefore, used a multiple imputations by chained equations (MICE) method to estimate supplementary variability of the Esteve model parameters due to the missing data.32-34 Finally, we fitted two multivariate analyses: in model A, the date of origin was the date of diagnosis; in model B, the date of origin was the date of the first treatment and care management variables were added into the model.

The stability of the results regarding SES was also tested by conducting the same analyses using the Townsend score.35 Finally, we described the proportion of patients treated with anthracyclinebased chemotherapy (ABC) and immunotherapy over time by study region. To comfort assumptions regarding type of hospital and medical specialty, we conducted two sensitive analyses on survival with different criteria to characterize the level of lymphoma-specific expertise. One analysis involved the case volume as the total number of DLBCL treated by each center during the period of the study with three cutoffs determined by fitting segmented regression. The second analysis was based on combinations of two variables “type of hospital” and “medical specialties” (4 modalities).

As we were using ecological data, we used a hierarchical model.

However, since the number of patients per IRIS was 1.7 on average, hierarchical models did not bring any supplementary information to residual variability (data not shown).

The statistical analyses were performed with Stata® 14 and R® (version 3.2.2).

Individual data of the study population We collected socio-demographic details, medical data and information about care management. Place of care was classified as the reference center, being either a “teaching hospital” (university or specialized oncology hospital) or “not-teaching hospital” (private clinic or community hospital). First medical contact (general practitioner or specialist) and medical specialty (hematology/oncology versus other specialties) for care management were also noted.

Distances between the place of residence and the nearest reference care center were calculated with ArcGis10© combined with a haematologica | 2017; 102(3)

Results Patients’ characteristics A total of 1,312 DLBCL were diagnosed between 2002 and 2008 in the three regions (Figure 1). After applying exclusion criteria, the population analyzed was composed of 1,165 subjects. Table 1 shows the distribution of patients’ characteristics according to the period of diagno585

S. Le Guyader-Peyrou et al.

Figure 1. Flowchart of the survival analysis of patients with diffuse large B-cell lymphoma (2002-2008).

DLBCL: diffuse large B-cell lymphoma; AIDS: acquired immunodeficiency syndrome.

sis (2002-2005 versus 2006-2008). Eighty-seven percent of patients (1017/1165) were treated in oncology/hematology departments. Fifty-nine percent of DLBCL patients (690/1165) were referred to teaching hospitals for their management.

The median of age at diagnosis was 72 years (range, 9 99) and the sex ratio (male/female) was 1.1 (610/556).

Clinical outcome and factors associated with relative survival of patients with diffuse large B-cell lymphoma The 5-year net survival for the entire cohort was 59%.

Figure 2 shows the plots of net survival probability according to groups of patients divided by period of diagnosis, marital status, medical specialty, and registry area.

The final models are shown in Tables 2 and 3. Adjusted for age, sex and International Prognostic Index (IPI) score, living alone [adjusted excess hazard ratio (adjusted EHR): 1.41; 95% confidence interval (95% CI): 1.01-1.99], having mild or severe comorbidity (adjusted EHR: 1.64; 95% CI: 1.22-2.22) and having been diagnosed and treated in Basse-Normandie county (adjusted EHR: 1.49; 95% CI: 1.19-1.85) were independently associated with unfavorable relative survival in model A (Table 2). Conversely, having been diagnosed during the later study period (2006-2008 compared to 2002-2005) (adjusted EHR: 0.71; 95% CI: 0.58-0.88) and treated in an oncohematology department (adjusted EHR: 0.34; 95% CI: 0.27-0.43) were factors independently associated with a favorable relative survival in model A. The results were comparable in the final multivariate model B, with treatment information incorporated, as reported in Table 3, except for the period of diagnosis that was no longer associated with relative survival. First-course treatment with a combination of ABC and rituximab was strongly associated with a better survival outcome (adjusted EHR: 0.32; 95% CI: 0.25-0.41), together with inclusion in a clinical trial (adjusted EHR: 0.44; 95% CI: 0.28-0.69), or having treatment discussed in 586

a multidisciplinary meeting (adjusted EHR: 0.69; 95% CI: 0.50-0.96). Patients treated in a teaching hospital had a better outcome (adjusted EHR: 0.82; 95% CI: 0.66-1.01) although the association was of borderline statistical significance (P=0.05) in model A, and no longer statistically significant after introduction of treatment information (model B, Table 3). SES variables (French EDI, urban/rural area also used as means of considering issues of accessibility or medical density) were not associated with relative survival of DLBCL patients in either model A or B. Using other cutoff points of EDI, splitting in quintiles or as a continuous variable did not change the results. We observed comparable results with the introduction of the Townsend index (a common deprivation measure used in the UK) in the place of the EDI. We refined our adjustment on prognostic factors by using each item of information contained in the IPI score (i.e., Eastern Cooperative Oncology Group Performance Status, age, Ann Arbor stage, serum lactate dehydrogenase level, number of extranodal sites of disease) rather than the score itself and found comparable results. Adjusting for age as a continuous variable also gave similar results. Our results were stable when each center was excluded in turn from the analyses, except for the association with marital status that became non-significant after exclusion of the Gironde center.

Our sensitive analysis on lymphoma-specific expertise showed that, adjusted for age, sex, IPI, living alone, comorbidity, registry area and period, patients treated in “high-volume centers” had a better relative survival (adjusted EHR: 0.42; 95% CI: 0.31-0.57) than those treated in “low-volume centers” with a trend between the four modalities. Moreover, a multivariate survival analysis including a variable mixing the medical specialties and types of hospital suggested that specialty was more important than type of hospital: patients treated in hematology/oncology departments had a better survival than haematologica | 2017; 102(3)

Socio-economic factors and survival in DLBCL patients

patients treated in other specialty departments whether or not these were in teaching hospitals (adjusted EHR: 0.28; 95% CI 0.20-0.39 for treatment in a hematology/oncology

department in a not-teaching hospital, and adjusted EHR: 0.24; 95% CI 0.18-0.33 for treatment in a hematology/ oncology department in a teaching hospital) versus adjust-

Table 1. Socio-demographics and clinical characteristics of diffuse large B-cell lymphoma patients at diagnosis, divided by study period.

Characteristic Age group 9-57 years 58-72 years 73-80 years 81-99 years Male sex Marital status: alone (single, widower or divorced) Unknown Moderate or severe comorbidity (ACE-27) Unknown Poor Performance Status (ECOG ≥ 2) Unknown Presence of B symptoms Unknown Ann Arbor stage at diagnosis III or IV (disseminated) Unknown Involvement of extranodal sites >1 Unknown Elevated lactate dehydrogenase International Prognostic Index score Low (0-1) Intermediate (2-3) High (4-5) Unknown Registry area Gironde Côte d'Or Basse-Normandie High GP density GP as first medical contact Unknown Place of care Private or community hospital Teaching hospital Medical specialty Other specialties Onco/hematology Travel time to nearest reference center ≤15 minutes 16-44 minutes 45-118 minutes Unknown Rural area Socioeconomic status (EDI national score) More deprived (last two quintiles) Unknown Inclusion in clinical trial (yes)* Unknown Multidisciplinary meeting (<90 days)* Unknown ABC chemotherapy & immunotherapy at 1st line * Unknown

2002-2005 N=634 %

2006-2008 N=531 %

P-value†

Total N=1165

% 0.18

170 170 161 133 337 181 96 174 22 210 30 196 130 306 2 87 25 288

26.8 26.8 25.4 22.0 53.1 28.5 15.1 27.4 3.5 33.1 4.7 30.9 20.5 48.3 0.3 13.7 3.9 45.4

121 143 130 137 273 165 99 140 8 167 32 184 99 308 1 113 20 241

22.8 26.9 24.5 25.8 51.4 31.1 18.6 26.4 1.5 31.4 6.0 34.6 18.6 58.0 0.2 21.3 3.8 45.4

291 313 291 270 610 346 195 314 30 377 62 380 229 614 3 200 45 529

25.0 26.9 25.0 23.2 52.4 29.7 16.7 26.9 2.6 32.4 5.3 32.6 19.6 52.7 0.3 17.2 3.9 45.4

182 210 234 8

28.7 33.1 36.9 1.3

137 145 237 12

25.8 27.3 44.6 2.3

319 355 471 20

27.4 30.5 40.4 1.7

289 101 244 305 400 63

45.6 15.9 38.5 48.1 63.1 9.9

243 88 200 257 386 41

45.8 16.6 37.6 48.4 72.7 7.7

532 189 444 293 786 104

45.7 16.2 38.1 48.2 67.5 8.8

262 372

41.3 58.7

213 318

40.1 59.9

475 690

40.8 59.2

85 549

13.4 86.6

63 468

11.9 88.1

148 1017

12.7 87.3

221 212 199 2 171

34.9 33.4 31.4 0.3 27.0

167 176 188 0 160

31.4 33.1 35.4 0 30.1

388 388 387 2 331

33.3 33.3 33.2 0.2 28.4

301 3 68 10 107 155 366 1

47.5 0.5 10.7 1.6 16.9 24.4 57.7 0.2

248 0 99 3 232 101 395 1

46.7 0 18.6 0.6 43.7 19.0 74.4 0.2

549 3 167 13 339 256 761 2

47.1 0.3 14.3 1.1 29.1 21.9 65.3 0.2

0.55 0.09 0.15 0.56 0.37 0.004 0.003 0.97 0.01

0.94

0.49 <10-3 0.67

0.43

0.25

0.23 0.27

<10-3 <10-3 <10-3

ACE27: Adult Comorbidity Evaluation; ECOG: Eastern Cooperative Oncology Group; LDH: lactate dehydrogenase; GP: general practitioner. EDI: European Deprivation Index; ABC: anthracycline-based. *Information collected only for patients treated (n=1111). †P-value calculated without unknown data.

haematologica | 2017; 102(3)

587

S. Le Guyader-Peyrou et al.

ed EHR: 0.60; 95% CI: 0.39-0.92 for patients treated in teaching hospitals in other specialty departments (Table 4).

Regarding the use of rituximab, and more specifically ABC with immunotherapy, as first-line treatment, we observed that the implementation of this regimen was less frequent in Basse-Normandie than in the other two regions: 78.9% of DLBCL patients in Basse-Normandie were given ABC with immunotherapy during the study period 2006-2008, compared with 92.2% of the patients in the two other regions in the same time period. This difference is comparable to that during the first study period (2002-2005) although the proportions were lower (i.e., 60.3% versus 70.2%).

Discussion To our knowledge, this study assessing the impact of social disparities, care management and place of care on DLBCL patientsâ&#x20AC;&#x2122; survival is the first performed in France to date. After adjusting for age, sex and IPI, our results suggest that SES (measured by the EDI) is not associated with relative survival of DLBCL patients. However, we observed a better survival during the later study period.

This positive trend in survival is likely to be explained by

the addition of immunotherapy to ABC regimens in frontline therapy (official agreement in France in 2002): study period was no longer associated with DLBCL survival when treatment variables were entered into the model.

Moreover, the area in which a patient is diagnosed, his or her medical care team (onco-hematology versus other) and to a lesser extent the type of treatment center (teaching versus not-teaching hospital) are independently associated with better 5-year relative survival. Lastly, a higher comorbidity burden and being single are independently associated with poorer survival.

The result on SES and DLBCL survival is consistent with a French study on mortality which did not find any statistically significant relative indices of inequality related to education for NHL mortality, although the study was conducted in the pre-rituximab era.36,37 Only one study published in 2014 evaluated the role of SES in specific DLBCL survival.23 In contrast to our results, the authors reported lower DLBCL survival in patients with a lower SES, with the association being more pronounced in the modern treatment era after the introduction of rituximab and in younger patients. As the authors stated, inadequate insurance coverage with additional financial burden due to modern treatments may be associated with increased DLBCL mortality. Differences in health care systems could, therefore, explain our different results, as the

A

C

B

D

Figure 2. Unadjusted net survival in the 5 years after diagnosis for patients with diffuse large B-cell lymphoma (2002-2008). Patients are divided by: (A) period of diagnosis; (B) registry area; (C) treatment department (oncohematology vs. other medical specialities); and (D) marital status.

588

haematologica | 2017; 102(3)

Socio-economic factors and survival in DLBCL patients

French universal health care system may be better able to obviate barriers to access necessary NHL care, as demonstrated in Canada by Darmawilarta et al.38 and more recently, in Germany for patients with acute myeloid leukemia.39 A Scandinavian study also found a relationship between low SES (specifically, educational level) and poor relative survival outcome of NHL patients overall, without any subtype analysis.40 In this study, performed in the prerituximab era, the authors observed a difference in survival mainly due to differences in excess mortality rates within the first year after diagnosis of NHL. They also observed a relationship between comorbidity and poor survival but did not analyze variables simultaneously, although the latter could partly explain the association between education and DLBCL survival.

Our result suggesting that unmarried people have lower DLBCL relative survival than married people, independently of other prognostic factors, is in agreement with findings of other recent studies.41 Kravdal et al. also reported a 15% excess of all-cause mortality in never-married men or women and divorced male cancer patients, compared to married people of the same sex.42 These results highlight the potentially significant impact that social sup-

Table 2. Factors related to the relative survival of diffuse large B-cell lymphoma patients in the 5 years following diagnosis in a multivariate Esteve model with MICE; model A (no information on treatment) (n=1165).

Characteristics Age group 9-57 years 58-72 years 73-80 years 81-99 years Sex Male Female Marital status Married Alone (single, widower or divorced) Comorbidity (ACE-27) None Mild Moderate & severe Risk group (IPI) Low Intermediate High Registry area Gironde Côte d'Or Basse-Normandie Year of diagnosis 2002-2005 2006-2008 Medical specialties Other specialties Onco-hematology Place of treatment Private or community hospital Teaching hospital

Multivariate Esteve model with MICE EHR 95% CI P-value P<10-3 1 1.06 1.77 2.68

[0.72-1.55] [1.22-2.57] [1.84-3.92] 0.001

1 0.72

[0.56-0.93] 0.002

1 1.41

[1.01-1.99] P<10-3

1 1.49 1.64

[1.12-1.98] [1.22-2.22] P<10-3

1 3.42 5.77

[2.18-5.38] [3.69-9.03] P<10-3

1 0.84 1.49

[0.60-1.18] [1.19-1.85] 0.005

1 0.71

[0.58-0.88] -3

P<10 1 0.34

[0.27-0.43] 0.055

1 0.82

[0.66-1.01]

MICE: multiple imputation by chained equation; EHR: excess hazard ratio; ACE27: Adult Comorbidity Evaluation; IPI: International Prognostic Index.

haematologica | 2017; 102(3)

port can have on cancer detection, treatment, and survival.

However, our observation of a persistent "registry area effect" on DLBCL survival could be interpreted as a residual indirect effect of patients’ SES on survival since patients living in the Basse-Normandie area were more deprived, more likely to live in rural areas, and further from the regional hospital compared to patients living in the other regions investigated. These patients were older (median age: 73.5 years), more frequently had a Performance Status ≥ 2, and more frequently had disseminated disease at diagnosis (i.e. Ann Arbor stage III/IV) than those in other registry areas. On the other hand, the fact that patients from this area were less frequently treated with rituximab or ABC with rituximab (whatever the study period), and less frequently included in clinical trials Table 3. Factors related to the relative survival of diffuse large B-cell lymphoma patients in the 5 years following diagnosis in a multivariate Esteve model with MICE; model B (no information on treatment) (n=1111).

Characteristics

Multivariate Esteve model with MICE EHR 95% CI P-value

Age group 9-57 years 1 58-72 years 1.18 73-80 years 1.61 81-99 years 2.09 Sex Male 1 Female 0.74 Marital status Married 1 Alone (single. widower or divorced) 1.52 Comorbidity (ACE-27) None 1 Mild 1.51 Moderate & severe 1.67 Risk group (IPI) Low 1 Intermediate 2.78 High 6.47 Registry area Gironde 1 Côte d’Or 0.95 Basse-Normandie 1.32 Time to travel to nearest reference center ≤15 minutes 1 16-44 minutes 1.42 45-118 minutes 1.11 Medical specialties Other specialties 1 Onco-hematology 0.45 Inclusion in trial No 1 Yes 0.44 Multidisciplinary meeting (<90 days) No 1 Yes 0.69 ABC chemotherapy & immunotherapy No 1 Yes 0.32

P<10-3 [0.78-1.78] [1.08-2.41] [1.39-3.16] 0.004 [0.57-0.96] 0.001 [1.10-2.08] 0.002 [1.10-2.07] [1.21-2.31] P<10-3 [1.72-4.47] [4.04-10.38] 0.047 [0.66-1.36] [1.03-1.69] 0.016 [1.07-1.87] [0.82-1.49] P<10-3 [0.34-0.61] P<10-3 [0.28-0.69] 0.001 [0.50-0.96] P<10-3 [0.25-0.41]

Note: model B applied only to patients treated (n=1111), patients who died prematurely without treatment are excluded from the analysis. MICE: multiple imputation by chained equation; EHR: excess hazard ratio; ACE27: Adult Comorbidity Evaluation; IPI: International Prognostic Index; ABC: anthracycline-based chemotherapy.

589

S. Le Guyader-Peyrou et al.

compared to patients in the other areas also suggests a delay in the implementation of these treatments in this specific region, as shown by Flowers et al.43 Thus, lower survival in Basse-Normadie than in other areas could be interpreted as a consequence of a poorer care provision in hematology. This interpretation is in agreement with the lower number of hematologists per inhabitant in Basse-Normadie and may have contributed to the creation in 2016 of a hematology Institute in Caen to reinforce care provision in this region.

Another explanation of DLBCL survival disparities would be the existence of discrepancies in other non-measured risk factors linked to poorer survival such as cigarette smoking or other risk factors with a higher prevalence in Basse-Normandie than in the other areas.44 Beside the benefit of new treatments, other improvements in patients’ management, such as better work-up, better prognostication, better treatment decisions that could be aggregated in concepts such as "included in a clinical trial" or " treatment discussed by a multidisciplinary team”, could also have led to better survival. Prior reports suggest that both factors are related to better survival outcomes.45 In our study, we found a borderline association with teaching/not-teaching hospital. It has been demonstrated, in a large panel of cancer survival studies, that patients treated in teaching hospitals have a better survival than patients treated in community hospitals. The effect of centralizing operative treatment might improve survival but these results always concerned solid tumors treated first by surgery and were correlated with increased hospital procedure volume.19 Regarding hematologic malignancies, few studies have explored the role of the care provider in a universal health care system setting in which each patient can choose his/her place of care management. The type of care center has been linked to survival, with a benefit being shown for patients with acute myeloid leukemia and DLBCL being treated in high volume hospitals.37,46 More recently, Lamy et al.47 suggested that even in a universal health care system, disparities in the management of DLBCL patients still exist depending on the type of care center, even after adjustment for differences between patients.

The first explanation of such a small difference that we found in survival between patients treated in teaching or not-teaching hospitals could be related to training and the hematology/oncology network: all hematologists/oncologists are trained in teaching hospitals and they continue to maintain strong relationships and collaborate together after their training when they work in private clinics through multidisciplinary discussion of treatment decisions and clinical trial participation. In France, there is a single cooperative group, named the Lymphoma Study Association (LYSA), which has the mission of bringing together professionals specializing in the field of lymphoma in both public and private hospitals in order to promote basic and clinical research, improve prevention, management, and treatment of lymphoma patients, and share their knowledge on lymphoma. This collaborative system works well in hematology/oncology and created a strong network all over the country. However, this type of cooperation could be reinforced between hematology/ oncology and other departments (mostly internal medicine) that also treat NHL patients in teaching or private hospitals.

590

More broadly speaking, this result could be organizational and may also be an effect of the last two French cancer plans (2003-2013) that implemented several actions aimed to reinforce the quality of care throughout the country. Several measures were taken to ensure all patients access to medical expertise whatever his/her place of treatment (better access to clinical trials and innovative treatment, panel review of tumor samples by expert pathologists and systematic discussion of medical records by clinical experts thereby promoting multidisciplinary management). Each NHL patient’s treatment must be discussed in a multidisciplinary meeting with the strict application of latest guidelines regardless of whether the patient is being treated in a teaching or not-teaching hospital and all patients must have better access to targeted treatment and clinical trials.

The major strengths of our study are related to its population-based design that limits selection bias making the results generalizable to a larger DLBCL patient population.

In our study, the 5-year net survival was 58%, which is very close to the rate observed at the national level with data from all cancer registries in a recent population-based study (57%).48 Secondly, we simultaneously analyzed a large set of variables in relation to DLBCL survival, adjusting for major known prognostic factors (age, sex and IPI).

Moreover, a senior registrar performed a systematic centralized review of all the pathology reports with particular attention to cases of ‘not otherwise specified’ NHL and transformation from follicular lymphoma.

This study had some limitations: we did not have access to information regarding individual risk factors, biomarkers that could be used for prognostication or official thresholds of hospitals’ expertise as for solid tumors. The last lack may explain why we did not find a clear relationship between the type of the hospital and survival outcomes. However, we did bring to light the importance of being treated in a specialized hematology or oncology department rather than in another type of department (internal medicine, polyvalent medicine, pneumology, etc.). This issue regarding the importance of medical specialty as a factor influencing survival of cancer patients has already been shown by other authors.49 As we mentioned above, we observed differences in the characteristics of patients treated in teaching hospitals and

Table 4. Effect of volume of cases and diffuse large B-cell lymphoma expertise on relative survival in the 5 years following diagnosis in a multivariate Esteve model with MICE.

Characteristics

Multivariate Esteve model with MICE EHR* 95% CI P-value

Volume of cases Low volume centers 1 Low intermediate volume centers 0.53 High intermediate volume centers 0.51 High volume centers 0.39 Hospital status & medical specialties† Not-teaching/other specialties 1 Teaching/other specialties 0.60 Not-teaching/onco-hematology 0.28 Teaching/onco-hematology 0.24

<10-3 [0.35-0.79] [0.36-0.73] [0.29-0.54] <10-3 [0.39-0.92] [0.20-0.39] [0.18-0.33]

MICE: multiple imputation by chained equation; EHR: excess hazard ratio; *adjusted for age, sex, marital status, comorbidity, risk group (IPI), period and registry area. †Not teaching: private and community hospitals – other specialties: not hematology or oncology department.

haematologica | 2017; 102(3)

Socio-economic factors and survival in DLBCL patients

those treated in not-teaching hospitals and between those treated in onco-hematology departments and those treated in other departments (e.g., internal medicine).

However, these characteristics (age, sex, Performance Status, comorbidity, etc.) were taken into account in the multivariate survival models.

This study suggests that type of specialty care is more important than type of hospital. All our findings, together with the need to treat patients (at least the elderly) near their place of residence and the observation of a recent increase in the number of survivors with NHL (positive trends in survival), reinforce the authors’ opinion that the implementation of hematology networks (combining routine, guidelines and clinical research) is a better solution than centralizing the care of NHL patients in teaching hospitals. The organization of different levels of expertise over the territory (whatever the hospital status) is helpful to prevent delays and give access to the cutting edge of hematology instead of triggering the bottleneck that would be created in teaching hospitals if all NHL patients were to be referred to such structures.

In the near future, we may expect an increase in elderly NHL cases and hematology/oncology departments will not be able to take care all these new patients.

Alternatively, cooperation could be reinforced between hematology/oncology departments and other departments taking care of NHL patients, in all types of hospital.

The onco-geriatric units that have been created all over France during the last 5 years are a step in that direction.

Since information was missing for a limited but not negligible number of cases regarding marital status, B symptoms, multidisciplinary teamwork and other health care behaviors, we used the MICE method, based on a Monte Carlo Markov chain algorithm under a missing random data hypothesis. This method is used to manage incomplete observations, avoiding biased estimates and improving their precision.50

References 6.

1. Ferlay J, Soerjomataram I, Ervik M, et al.

GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11. Lyon, France: International Agency for Research on Cancer; 2013. Available from: http://globocan.iarc.fr, accessed on October 10th, 2016.

2. Stewart B, Wild CP (eds.), International Agency for Research on Cancer, WHO. (2014) World Cancer Report 2014 [Online].

Available at: http://www.thehealthwell.info/node/72584 5 , accessed on 10th October 2016.

3. Sant M, Allemani C, Tereanu C, et al.

Incidence of hematologic malignancies in Europe by morphologic subtype: results of the HAEMACARE project.

Blood.

2010;116(19):3724-3734.

4. Morton LM, Wang SS, Devesa SS, Hartge P, Weisenburger DD, Linet MS. Lymphoma incidence patterns by WHO subtype in the United States, 1992-2001. Blood. 2006;107 (1):265-276.

5. Howlader N, Morton LM, Feuer EJ, Besson C, Engels EA. Contributions of subtypes of non-Hodgkin lymphoma to mortality

haematologica | 2017; 102(3)

7.

8.

9.

10.

11.

The social welfare system in France is intended to give free access to all types of hospital and innovative treatments. Our results showing no association between an aggregate SES index and survival outcome do not call into question this organization. However, heterogeneity in care management and later introduction of the use of innovative drugs in some regions could explain survival differences between registry areas, as could other nonmeasured risk factors that should be prospectively collected in new cohorts, together with genetic factors.

The notion of lymphoma-specific expertise seems to be essential for the best DLBCL management and raises the question of centralization of NHL patients’ care in hematology/oncology departments. However, the expected number of new NHL cases for the next 20 years (due to demographic variations in the elderly and improved NHL survival) argues more for the implementation of collaborative networks with close communication between hematology/oncology departments and other medical departments rather than centralization of NHL care.

Acknowledgments The authors would like to thank all those who contributed to recording cancer data in the registries, in particular the laboratories and departments of anatomy, cytology, and pathology; the medical informatics departments of the public and private hospitals and general practitioners and specialists. The authors also express their gratitude to JM. Poncet, Dr. A. Collignon, S. Boulanger, H. Rachou and A. Viale for collecting data; L. Launay of INSERM U1086 for geocoding addresses and applying the EDI and Dr. R. Nookala of the Institut Bergonié for the medical writing service.

Funding This work was supported by the French “Institut National du Cancer” (INCa).

Trends. Cancer Epidemiol Biomarkers Prev.

2016;25(1):174-179.

Sant M, Minicozzi P, Mounier M, et al.

Survival for haematological malignancies in Europe between 1997 and 2008 by region and age: results of EUROCARE-5, a population-based study. Lancet Oncol. 2014;15(9): 931-942.

Mounier M, Bossard N, Belot A, et al. Trends in excess mortality in follicular lymphoma at a population level. Eur J Haematol.

2015;94(2):120-129.

Monnereau A, Troussard X, Belot A, et al.

Unbiased estimates of long-term net survival of hematological malignancy patients detailed by major subtypes in France. Int J Cancer. 2013;132(10):2378-2387.

Faggiano F, Partanen T, Kogevinas M, Boffetta P. Socioeconomic differences in cancer incidence and mortality. IARC Sci Publ.

1997;138:65-176.

Borate UM, Mineishi S, Costa LJ. Non biological factors affecting survival in younger patients with acute myeloid leukemia.

Cancer. 2015;121(21):3877-3884.

Coleman MP, Rachet B, Woods LM, et al.

Trends and socioeconomic inequalities in cancer survival in England and Wales to

2001. Br J Cancer. 2004;90(7):1367-1373.

12. Eggleston KS, Coker AL, Williams M, Tortolero-Luna G, Martin JB, Tortolero SR. Cervical cancer survival by socioeconomic status, race/ethnicity, and place of residence in Texas, 1995-2001. J Women Health.

2006;15(8):941-945.

13. Meilleur A, Subramanian SV, Plascak J, Fisher JL, Paskett ED, Lamont EB. Rural residence and cancer outcomes in the US: issues and challenges.

Cancer Epidemiol Biomarkers Prev. 2013;22(10):1657-1667.

14. Parikh-Patel A, Bates JH, and Campleman S. Colorectal cancer stage at diagnosis by socioeconomic and urban/rural status in California, 1988-2000. Cancer. 2006;107(5 Suppl):1189-1195.

15. Dejardin O, Jones AP, Rachet B, et al. The influence of geographical access to health care and material deprivation on colorectal cancer survival: evidence from France and England. Health Place. 2014;30:36-44.

16. Ambroggi M, Biasini C, Del Giovane C, Fornari F, Cavanna L. Distance as a barrier to cancer diagnosis and treatment: review of the literature. Oncologist. 2015;20(12): 13781385.

17. Birkmeyer NJ, Goodney PP, Stukel TA,

591

S. Le Guyader-Peyrou et al.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

592

Hillner BE, Birkmeyer JD. Do cancer centers designated by the National Cancer Institute have better surgical outcomes? Cancer.

2005;103(3):435-441.

Chaudhry R, Goel V, Sawka C. Breast cancer survival by teaching status of the initial treating hospital. CMAJ. 2001;164(2):183188.

Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital procedure volume on outcomes following surgery for colon cancer. JAMA. 2000;284 (23):3028-3035.

Keegan TH, McLure LA, Foran JM, Clarke CA. Improvements in survival after follicular lymphoma by race/ethnicity and socioeconomic status: a population-based study. J Clin Oncol. 2009;27(18):3044-3051.

Keegan TH, Clarke CA, Chang ET, Shema SJ, Glaser SL. Disparities in survival after Hodgkin lymphoma: a population-based study. Cancer Causes Control. 2009;20(10): 1881-1892.

Loberiza FR, Cannon AJ, Weisenburger DD, et al. Survival disparities in patients with lymphoma according to place of residence and treatment provider: a population-based study. J Clin Oncol. 2009;27(32):5376-5382.

Tao L, Foran JM, Clarke CA, Gomez SL, Keegan TH. Socioeconomic disparities in mortality after diffuse large B-cell lymphoma in the modern treatment era. Blood.

2014;123(23):3553-3562.

Lee B, Goktepe O, Hay H, et al. Effect of place of residence and treatment on survival outcomes in patients with diffuse large Bcell lymphoma in British Columbia.

Oncologist. 2014;19(3):283-290.

Flowers CR, Nastoupil LJ. Socioeconomic disparities in lymphoma. Blood. 2014;123 (23):3530-3531.

Morton LM, Turner JJ, Cerhan JR, et al.

Proposed classification of lymphoid neoplasms for epidemiologic research from the Pathology Working Group of the International Lymphoma Epidemiology Consortium (InterLymph). Blood. 2007;110 (2):695-708.

Fritz A, Percy C, Jack A, et al. International classification of diseases for oncology, 3rd edition. World Health Organisation, 2000.

Pornet C, Delpierre C, Dejardin O, et al.

Construction of an adaptable European transnational ecological deprivation index:

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

the French version. J Epidemiol Community Health. 2012;66(11):982-989.

Woods LM, Rachet B, Coleman MP. Choice of geographic unit influences socioeconomic inequalities in breast cancer survival. Br J Cancer. 2005;92(7):1279-1282.

Pohar Perme M, Stare J, Esteve J. On estimation in relative survival. Biometrics.

2012;68(1):113-120.

Estève J, Benhamou E, Croasdale M, Raymond L. Relative survival and the estimation of net survival. Stat Med.1990;9(5): 529-538.

Giorgi R, Belot A, Gaudart J, Launoy G, French Network of Cancer Registries FRANCIM. The performance of multiple imputation for missing covariate data within the context of regression relative survival analysis. Stat Med. 2008;27(30):6310-6331.

White IR, Royston P, Wood A. Tutorial in biostatistics: multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30(4):377-399.

Nur U, Shack LG, Rachet B, Carpenter JR, Coleman MP. Modelling relative survival in the presence of incomplete data: a tutorial.

Int J Epidemiology. 2010;39(1):118-128.

Townsend P. Deprivation. J Soc Policy.

1987;16:125-146.

Menvielle G, Chastang JF, Luce D, Leclerc A, Groupe EDISC. Changing social disparities and mortality in France (1968-1996): cause of death analysis by educational level. Rev Epidemiol Sante Publique. 2007;55(2):97-105.

Borel C, Lamy S, Compaci G, et al. A longitudinal study of non-medical determinants of adherence to R-CHOP therapy for diffuse large B-cell lymphoma: implication for survival. BMC Cancer.2015;15:288-299.

Darmawikarta D, Pole JD, Greenberg M. The association between socioeconomic status and survival among children with Hodgkin and non-Hodgkin lymphomas in a universal health care system. Pediatr Blood Cancer. 2013; 60(7):1171-1177.

Erdmann F, Kaatsch P, Zeeb H, Roman E, Lightfoot T, Schüz J. Survival from childhood acute lymphoblastic leukaemia in West Germany: does socio-demographic background matter? Eur J Cancer. 2014; 50(7):1345-1353.

Roswall N, Olsen A, Christensen J, Rugbjerg K, Mellemkjaer L. Social inequality and incidence of and survival from Hodgkin lym-

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

phoma, non-Hodgkin lymphoma and leukaemia in a population-based study in Denmark, 1994-2003. Eur J Cancer.

2008;44(14):2058-2073.

Aizer A, Chen MH, McCarthy EP, et al.

Marital status and survival in patients with cancer. J Clin Oncol. 2013;31(31):3869-3876.

Kravdal O. The impact of marital status on cancer survival. Soc Sci Med. 2001;52(3): 357-368.

Flowers CR, Fedewa SA, Chen AY, et al.

Disparities in the early adoption of chemoimmunotherapy for diffuse large Bcell lymphoma in the United States. Cancer Epidemiol Biomarkers Prev. 2012;21(9): 1520-1530.

Battaglioli T, Gorini G, Costantini AS, et al.

Cigarette smoking and alcohol consumption as determinants of survival in nonHodgkin's lymphoma: a population-based study. Ann Oncol. 2006;17(8):1283-1289.

Pillay B, Wootten AC, Crowe H, et al. The impact of multidisciplinary team meetings on patient assessment, management and outcomes in oncology settings: a systematic review of the literature. Cancer Treat Rev. 2016;42:56-72.

Giri S, Pathak R, Aryal MR, Karmacharya P, Bhatt VR, Martin MG. Impact of hospital volume on outcomes of patients undergoing chemotherapy for acute myeloid leukemia: a matched cohort study. Blood. 2015;125 (21):3359-3360.

Lamy S, Bettiol C, Grosclaude P, et al. The care center influences the management of lymphoma patients in a universal health care system: an observational cohort study.

BMC Health Serv Res. 2016;16(a):336-346.

Monnereau A, Urhy Z, Bossard N, et al.

Survie des personnes atteintes de cancer en France métropolitaine, 1989-2013. Partie 2 – Hémopathies malignes. Saint-Maurice: Institut de veille sanitaire; 2015. Available at: http://www.invs.sante.fr or http://www.ecancer.fr. Accessed on 10th October 2016.

Engelen MJ, Kos HE, Willemse PH, et al.

Surgery by consultant gynecologic oncologists improves survival in patients with ovarian carcinoma. Cancer. 2006;106(3):589-598.

Falcaro M, Nur U, Rachet B, Carpenter JR. Estimating excess hazard ratios and net survival when covariate data are missing: strategies for multiple imputation.

Epidemiology. 2015;26(3):421-428.

haematologica | 2017; 102(3)

ARTICLE

Plasma Cell Disorders

Evaluation of the Revised International Staging System in an independent cohort of unselected patients with multiple myeloma

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Efstathios Kastritis, Evangelos Terpos, Maria Roussou, Maria Gavriatopoulou, Magdalini Migkou, Evangelos Eleutherakis-Papaiakovou, Despoina Fotiou, Dimitrios Ziogas, Ioannis Panagiotidis, Eftychia Kafantari, Stavroula Giannouli, Athanasios Zomas, Konstantinos Konstantopoulos and Meletios A. Dimopoulos Department of Clinical Therapeutics, National and Kapodistrian University of Athens, School of Medicine, Greece

Haematologica 2017 Volume 102(3):593-599

ABSTRACT

T

he Revised International Staging System (R-ISS) was recently introduced in order to improve risk stratification over that provided by the widely used standard International Staging System. In addition to the parameters of the standard system, the R-ISS incorporates the presence of chromosomal abnormalities detected by interphase fluorescence in situ hybridization [t(4;14), t(14;16) and del17p] and elevated serum lactate dehydrogenase. The R-ISS was formulated on the basis of a large dataset of selected patients who had participated in clinical trials and has not been validated in an independent cohort of unselected patients. Thus, we evaluated the R-ISS in 475 consecutive, unselected patients, treated in a single center. Our patients were older and more often had severe renal dysfunction than those in the original publication on the R-ISS. As regards distribution by group, 18% had R-ISS-1, 64.5% R-ISS-2 and 18% R-ISS-3.

According to R-ISS group, the 5-year survival rate was 77%, 53% and 19% for R-ISS-1, -2 and -3, respectively (P<0.001). The R-ISS could identify three groups with distinct outcomes among patients treated with or without autologous stem cell transplantation, among those treated with either bortezomib-based or immunomodulatory drug-based primary therapy and in patients ≤65, 66-75 or >75 years. However, in patients with severe renal dysfunction the distinction between groups was less clear. In conclusion, our data in consecutive, unselected patients, with differences in the characteristics and treatment approaches compared to the original International Myeloma Working Group cohort, verified that R-ISS is a robust tool for risk stratification of newly diagnosed patients with symptomatic myeloma.

Introduction Multiple myeloma is a heterogeneous disease and the development of a risk stratification tool has been challenging. Both disease and host characteristics are crucial and should be incorporated in a staging system. The Durie-Salmon staging system (which was based on the levels of M protein, the number of lytic bone lesions, hemoglobin values, serum calcium levels and creatinine) was used for several years.1 In 2003, the simpler but robust International Staging System (ISS) was introduced; this system is based on β2-microblobulin and serum albumin levels, and since its introduction it has been the standard for risk stratification of patients with multiple myeloma.2 Several data have shown that chromosomal abnormalities detected by interphase fluorescence in situ hybridization (iFISH) [mainly t(4;14), t(14;16) and del17p] are strong prognostic factors, reflecting the inherent haematologica | 2017; 102(3)

Correspondence: mdimop@med.uoa.gr

Received: February 29, 2016.

Accepted: October 21, 2016.

Pre-published: October 27, 2016. doi:10.3324/haematol.2016.145078 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/593

©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

593

E. Kastritis et al.

genetic characteristics of the disease.3-6 In addition, elevated serum lactate dehydrogenase (LDH) has been consistently associated with poor prognosis.7-9 In order to improve the prognostic performance of the ISS, the International Myeloma Working Group (IMWG) revised the current ISS by adding high-risk cytogenetics [t(4;14), t(14;16) and del17p by iFISH] and elevated serum LDH and thus the Revised-ISS (R-ISS) was proposed as the new system.10 The formulation of the R-ISS was based on a very large number of patients (3,060 patients) from several independent large prospective trials, who were carefully monitored and reviewed. However, clinical trials exclude patients with severe renal dysfunction or with poor performance status.

Thus, the aim of the current analysis was to validate the new R-ISS in an independent cohort of unselected, consecutive patients with symptomatic myeloma, treated with contemporary regimens and followed rigorously in a single center.

Methods Consecutive patients with symptomatic myeloma who were treated in our center and who had available ISS stage, cytogenetics [by iFISH for del17p, t(4;14) and t(14;16)], and serum LDH were included in this analysis. Between 2007 and 2014, 475 of the 625 (76%) consecutive patients who started therapy in our center fulfilled the above criteria. Approval for the analysis and publication of the data was obtained from the Scientific/Ethics Committee of “Alexandra” hospital.

R-ISS-1 includes patients with ISS-1 (serum β2-microblobulin level <3.5 mg/L and serum albumin level ≥3.5 g/dL), no high-risk cytogenetic abnormalities in iFISH [such as del(17p) and/or t(4;14) and/or t(14;16)] and normal LDH levels (below the upper limit of normal). R-ISS-3 includes patients with ISS-3 (serum β2microglobulin level >5.5 mg/L) and either high-risk cytogenetic abnormalities in iFISH or elevated LDH levels. R-ISS-2 includes all the other possible combinations.10 Renal function was evaluated by the estimated glomerular filtration rate (eGFR) which was calculated using the Modification of Diet in Renal Diseases formula.

Interphase fluorescence in situ hybridization studies Plasma cells were separated using anti-CD138–coated magnetic MicroBeads (Miltenyi Biotech, San Diego, CA, USA) following the manufacturer’s instructions. At least 100 nuclei were analyzed using a fluorescent light microscope. Patients were considered positive for del17p if ≥20% of the nuclei were positive and for t(4;14) or t(14;16) if ≥10% were positive.11

Statistical analysis Comparisons for categorical variables among different groups were made with the chi-square test, using the Fisher exact test when appropriate. Overall survival was measured from the date of treatment initiation until the date of death or date of last follow up. Progression-free survival was calculated from the date of initiation of therapy until the first date of confirmed progression or death from any cause. Time to event curves were plotted with the method of Kaplan and Meier, and comparisons among groups were made using the log-rank test. For multivariate analysis, factors associated with time to event were introduced into a Cox proportional hazards model. IBM SPSS v20 software (SPSS Inc., Chicago, IL, USA) was used for the statistical analysis.

594

Results Characteristics of the patients and comparison to International Myeloma Working Group cohort The analysis included 475 patients with available data on cytogenetics [t(4;14), t(14;16) and del17p by iFISH], serum LDH and ISS. There was no difference in the distribution per ISS stage or elevated LDH between those with or without available cytogenetics, but patients for whom cytogenetic data were available were younger (43% versus 31% were ≤65 years, P=0.012) and less often received primary treatment with conventional chemotherapy (2% versus 10%), while the frequencies of primary therapy with bortezomib (50% versus 44%), lenalidomide (28% versus 21.5%) and thalidomide (20% versus 24%) were not significantly different.

The median age of the patients was 67 years (range, 2791 years); 53% of them were >65 years, while 25% were

Table 1. Characteristics of the patients in the analysis.

N=475 Age median range, years 67(27-91) Age ≤65 years, n(%) 222 (47%) Age 66-75 years, n(%) 133 (28%) Age >75 years, n(%) 119 (25%) Males / females 265 (53%)/ 210 (47%) Median (range) ECOG Performance Status, n.

1 (0-4) ECOG PS ≥2 44% ECOG PS ≥2 in patients ≤65 years 33% ECOG PS ≥2 in patients 65-75 years 53% ECOG PS ≥2 in patients ≥76 years 54% ISS-1 115 (24%) ISS-2 163 (34%) ISS-3 197 (42%) High risk cytogenetics [del17p or t(4;14) or t(14;16)] 112 (23.5%) Increased LDH (>250 IU/L) 70 (15%) R-ISS-1 86 (18%) R-ISS-2 306 (64%) R-ISS-3 83 (18%) Durie-Salmon stage IA 31 (6.5%) Durie-Salmon stage IB 252 (53%) Durie-Salmon stage IIA 93 (19,5)% Durie-Salmon stage IIB 3 (<1%) Durie-Salmon stage IIIA 39 (8%) Durie-Salmon stage IIIB 57 (12%) Primary therapy Chemotherapy 41 (9%) Thalidomide 92 (19%) Lenalidomide 110 (23%) Bortezomib 233 (49%) ASCT 170 (36%) Serum creatinine ≥ 2 mg/dL 98 (21%) Median (range) eGFR in mL/min/1.73 m2 67 (<5 - >150) eGFR < 30 mL/min/1.73 m2 96 (20%) Hemoglobin < 10 g/dL 234 (49%) Platelet count <130x109/L 56 (12%) Calcium > 11 mg/dL 85 (18%) ECOG: Eastern Cooperative Oncology Group; PS: Performance Status; ISS: International Staging System; LDH: lactate dehydrogenase; R-ISS: revised International Staging System; ASCT: autologous stem cell transplanation; eGFR: estimated glomerular filtration rate.

haematologica | 2017; 102(3)

Evaluation of the R-ISS

>75 years of age. Compared to the IMWG cohort, our patients were older since in the IMWG cohort only 32% were >65 years. The median eGFR was 67 mL/min/1.73 m2 and 20% of the patients had an eGFR <30 mL/min/1.73 m2 (Table 1). In comparison, most of the studies that were included in the dataset for the formulation of the R-ISS had excluded patients with low eGFR and all of them had excluded patients on dialysis. Only 8.6% of our patients did not receive novel agents (thalidomide, lenalidomide or bortezomib) as primary therapy; 42% received immunomodulatory drugs (19% thalidomide-based, 23% lenalidomide-based) and 49% bortezomib-based primary therapy, while 36% underwent autologous stem cell transplantation (ASCT) as part of their frontline therapy (Table 1). Online Supplementary Table S1 presents the detailed distribution per regimen. In the IMWG cohort, 65% of the patients had received ASCT, 6% had received first-line therapy with conventional chemotherapy, 44% had received proteasome inhibitors and 66% had received immunomodulatory drugs. Online Supplementary Table S2 summarizes the differences and similarities between the IMWG cohort and our series of patients.

According to standard ISS, 24% of the patients in our cohort were rated as ISS-1, 34% as ISS-2 and 42% as ISS3 (Table 1). The distribution according to Durie-Salmon staging system is also presented in Table 1. High-risk cytogenetics [t(4;14), del17p or t(14;16)] were present in 23.5% of the patients and elevated LDH was present in 15%. In the IMWG cohort 38% of the patients were rated as ISS1, 38% as ISS-2 and 24% as ISS-3. Thus, the patients in our cohort more often had ISS-3 and less often ISS-1 disease, probably reflecting the unselected nature of our population, which also included patients with severe renal impairment. In contrast, high-risk cytogenetics and elevated LDH were not different among the two cohorts (24% and 13% in the IMWG cohort, respectively) (Online Supplementary Table S2).

Revised International Staging System distribution According to the R-ISS, 86 (18%) patients were rated as

A

R-ISS-1, 83 (18%) rated as R-ISS-3 and 306 (64%) were rated as R-ISS-2. The distribution within the R-ISS in the original IMWG cohort was 28% for R-ISS-1, 62% for RISS-2 and 10% for R-ISS-3. The higher percentage of patients with R-ISS-3 in our cohort was due to the higher proportion of patients with ISS-3 compared to that in the IMWG cohort, since the frequency of high-risk cytogenetics and elevated LDH were similar between the two cohorts. The R-ISS distribution in those â&#x2030;¤65 years was 21%, 60% and 19% for R-ISS-1, -2 and -3, respectively; among patients 66-75 years it was 19%, 63% and 18%, and among those >75 years it was 11%, 74% and 15%, respectively (P=0.128). The differences in the distribution of stages of ISS and R-ISS between the IMWG cohort and our series of patients are presented in Online Supplementary Table S2.

Outcomes by Revised International Staging System group The median follow-up of the entire cohort was 40 months; 57% of the patients have progressed or died and 63% remain alive. The median progression-free survival was 27 months and estimated median overall survival was 63 months. The median progression-free survival for patients rated as R-ISS-1, R-ISS-2 and R-ISS-3 was 34, 28 and 17 months, respectively (P<0.001) (Figure 1A).

According to the R-ISS, the probability of overall survival at 3 years was 83%, 69% and 45% and that at 5 years was 77%, 53% and 19% for patients rated as R-ISS-1, R-ISS-2 and R-ISS-3, respectively (Figure 1B; P<0.001 and Table 3).

We then evaluated outcomes according to the R-ISS in

Table 2. Distribution of 475 patients with symptomatic myeloma between R-ISS and ISS stages.

R-ISS-1 R-ISS-2 R-ISS-3 Total ISS

ISS-1

ISS-2

ISS-3

Total R-ISS

86 29 0 115

0 163 0 163

0 114 83 197

86 306 83 475

B

Figure 1. Survival outcomes in 475 patients according to R-ISS group. (A) Progression-free survival (PFS) and (B) overall survival.

haematologica | 2017; 102(3)

595

E. Kastritis et al. patients ≤65, 66-75 and >75 years. The 5-year overall survival rate was 50%, 32% and 21% for patients ≤65 years, 66-75 and >75 years, respectively. In patients ≤65 years, the 5-year overall survival rate was 84%, 71% and 29% for patients in R-ISS-1, R-ISS-2 and R-ISS-3, respectively (P<0.001); for patients 66-75 years, it was 73%, 43% and 18% (P=0.001), while in patients >75 years, the 5-year overall survival rate was 59%, 33% and 0%, respectively (P=0.122). Thus, the R-ISS identified a group of patients >75 years old with a favorable prognosis (Figure 2).

In patients who were not treated with high-dose melphalan and ASCT, the 5-year overall survival rate according to R-ISS stage was 64%, 41% and 13% for RISS-1, RISS-2 and RISS-3 patients, respectively (Figure 3A, P<0.001), while for patients treated with high-dose mel-

phalan and ASCT, the corresponding figures were 93%, 77% and 32%, respectively (Figure 3B; P<0.001).

Regarding the type of primary therapy, the 5-year probability of overall survival for patients treated with bortezomib-based upfront therapy was 95%, 69% and 18% for those in the R-ISS-1, RISS-2 and RISS-3 groups, respectively (Figure 3C; P<0.001) and the corresponding figures for

A

Table 3. Univariate analysis for factors associated with survival using Cox regression.

N=475 Survival (months) R-ISS-1 R-ISS-2 R-ISS-3 ISS-1 ISS-2 ISS-3 Durie-Salmon stage IA Durie-Salmon stage IB Durie-Salmon stage IIA Durie-Salmon stage IIB Durie-Salmon stage IIIA Durie-Salmon stage IIIB Age ≤65 years (A) Age 66-75 years (B) Age >75 years (C) Males Females High risk cytogenetics [del17p or t(4;14) or t(14;16)] absent High risk cytogenetics [del17p or t(4;14) or t(14;16)] present Normal LDH (<250 IU/L) Increased LDH (≥250 IU/L) Primary therapy Chemotherapy only Thalidomide Lenalidomide Bortezomib ASCT No ASCT eGFR ≥30 mL/min/1.73 m2 eGFR <30 mL/min/1.73 m2 Hemoglobin ≥10 g/dL Hemoglobin <10 g/dL Platelet count ≥130x109/L Platelet count <130x109/L Calcium <11 mg/dL Calcium ≥11 mg/dL

126 66 29 126 86 86 75 53 86 42 60 30 109 51 35 66 62 86

P-value HR (95% CI) <0.001

II vs. I: 1.9 (1.14-3.3) III vs. I: 4.2 (2.4-7.5) <0.001 II vs. I: 1.75 (1.1-2.8) III vs. I: 3.5 (2.2-5.5) 0.002

<0.001 B vs A:1,9 (1,3-2,7) C vs A: 3.2 (2.2-4.7) 0.647

1.6 (1.2-2.2)

0.005

1.8 (1.2-2.6)

0.002

44 70 41 93 53 45.5 93 109.5 45.5 86 40 93 46 70 43 70 44

C 0.047

0.35 (0.25-0.5)

<0.001

2 (1.45-2.8)

<0.001

1.8 (1.3-2.4)

<0.001

1.8 (1.2-2.7)

0.005

1.6 (1.1-2.3)

0.013

Note: P-values for interactions between age and ASCT, between age and type of therapy and between ASCT and type of primary therapy were highly significant. HR: hazard ratio, 95%CI: 95% confidence interval for the HR; R-ISS: revised International Staging System; ISS: International Staging System; LDH: lactate dehydrogenase; ASCT: autologous stem cell transplanation; GFR: estimated glomerular filtration rate.

596

B

IIA vs. IA: 1.28 (0.66-2.5) IIIA vs. IA: 1.67(0.82-3.4) IB vs. IA: 3 (0.66-13.8) IIB vs. IA: 2.03 (0.9-4.4) IIIB vs. IA: 2.9 (1.4-5.9)

Figure 2. Survival of patients of different age groups according to R-ISS stage.

(A) Patients ≤65 years, (B) patients 66-75 years, and (C) patients ≥76 years.

haematologica | 2017; 102(3)

Evaluation of the R-ISS

those treated with immunomodulatory drugs were 68%, 41% and 23%, respectively (Figure 3D; P=0.002).

We also evaluated the R-ISS in patients with different degrees of renal dysfunction. In patients with eGFR â&#x2030;Ľ30 mL/min/1.73 m2, the 5-year overall survival of patients with R-ISS-1, RISS-2 and RISS-3 disease was 76%, 56% and 28% respectively (Figure 4A; P<0.001). In patients with eGFR <30 mL/min/1.73 m2, no patients had R-ISS-1 disease and the median overall survival for R-ISS-2 versus R-ISS-3 patients was 42 versus 32 months (P=0.354), while the probability of 5-year overall survival was 40% versus 8%, respectively (Figure 4B); however, the number of patients in each subgroup does not provide enough statistical power to confirm whether the differences in overall survival were statistically significant. Online Supplementary Figures S2-S5 show the survival curves for each subgroup for the R-ISS versus the ISS. Because there was a strong interaction between therapy and R-ISS distribution, as well as between R-ISS stage and age, we performed a multivariate analysis, which showed that R-ISS stage was an independent prognostic factor

associated with survival with a hazard ratio of 1.68 for RISS-2 versus R-ISS-1 and 3.8 for R-ISS-3 versus R-ISS-1 (Table 4).

Discussion Validation of a prognostic system is an important step for acceptance of the system as a prognostic tool and its incorporation in everyday practice. Our data indicate that the R-ISS, which combines the ISS together with characteristics of the myeloma clone (cytogenetic abnormalities) and elements of the aggressiveness of the plasma cells (reflected by serum LDH), provides significant prognostic information.

Furthermore, the R-ISS retains its prognostic significance even in a population of patients with significant differences from the original IMWG cohort (detailed in Online Supplementary Table S2). It is important that the performance of a risk assessment tool is not restricted to a specific patient population so that it can be used more

A

B

C

D

Figure 3. Survival according to the R-ISS for patients who did or did not receive ASCT and depending on first-line treatment regimen. (A) Patients treated with firstline regimens based on bortezomib (B) Patients treated with first-line regimens based on immunomodulatory drugs (IMID). (C) Patients who did not receive highdose melphalan autologous stem cell transplantation (HDM-ASCT). (D) Patients who received HDM-ASCT.

haematologica | 2017; 102(3)

597

E. Kastritis et al.

A

B

Figure 4. Overall survival according to R-ISS stage in patients categorized by renal fuction. (A) Patients with eGFR â&#x2030;Ľ30 mL/min/1.73 m2. (B) Patients with eGFR <30 mL/min/1.73 m2

extensively. Indeed, our patients were older, less often received ASCT and more often had severe renal dysfunction than the patients in the IMWG cohort.10 Nonetheless, the R-ISS identified subgroups with very different outcomes among those treated or not with ASCT and among elderly or younger patients. Among patients with severe renal dysfunction (20% of patients in out cohort) only patients with R-ISS-2 or RISS-3 were identified. This is expected because such patients almost invariably have ISS-2 or more often ISS-3 disease due to elevated β2microglobulin, related both to disease burden and renal dysfunction. The median overall survival in patients with severe renal impairment (i.e. those with an eGFR <30 mL/min/1.73 m2) was not significantly different; however, a careful inspection of the curves revealed that there is a clear separation later in the course, resulting in different 5year survival rates.

In the IMWG cohort, FISH studies were performed in different laboratories with different cutoffs for positivity of del17p or t(4;14).10 In contrast, in our patients the cutoffs were defined from the outset for del17p, t(4;14) and t(14;16), according to recommendations from the European Myeloma Network (EMN).11 The rates of highrisk cytogenetics were not, however, very different between our cohort and the IMWG one. The differences in the age composition between the two populations, the fact that the frequency of certain cytogenetic abnormalities may be lower among elderly patients,12,13 and the use of different cutoffs for positivity may have affected the rates of presence of cytogenetic abnormalities. It is important to note that all our patients had serum LDH measured in the same laboratory, thus, reducing inter-laboratory variability.

As in the original IMWG cohort, most of our patients received therapy with contemporary regimens. Thus, we must postulate that the R-ISS is applicable mostly in patients treated with such regimens, since only very few patients received chemotherapy alone as part of their primary therapy. Importantly, the R-ISS remains robust among patients treated with proteasome inhibitors (bortezomib) or immunomodulatory drugs (thalidomide or 598

Table 4. Multivariate analysis of factors associated with overall survival in univariate analysis, with R-ISS in the model.

HR R-ISS-1 (reference) R-ISS-2 R-ISS-3 No HDM-ASCT Conventional chemo (reference) Thalidomide-based Lenalidomide-based Bortezomib-based Age â&#x2030;¤ 65 years (reference) Age 66-75 years Age > 75 years Hemoglobin <10 g/dL Calcium > 11 mg/dL Platelets < 130x109/L eGFR < 30 mL/min/1.73 m2

1 1.68 3.83 1.72 1 1.13 1.17 0.93 1 1.17 1.87 1.245 1.136 1.651 1.322

95% CI

P-value

1.03-2.9 2.1-7.1 1.1-2.9

0.044 <0.001 0.043

0.62-2.03 0.63-2.1 0.52-1.67

0.690 0.616 0.805

0.71-1.94 1.1-3.2 0.890-1.742 0.757-1.706 1.074-2.538 0.889-1.966

0.53 0.02 0.201 0.538 0.022 0.168

HR: hazard ratio, 95% CI: 95% confidence interval for HR; R-ISS: revised International Staging System; HDM-ASCT: high-dose melphalan-autologous stem cell transplantation; chemo: chemotherapy; eGFR: estimated glomerular filtration rate.

lenalidomide) and also among patients treated with or without high-dose melphalan. Whether the R-ISS performs similarly in patients treated with monoclonal antibodies or other novel agents still needs to be evaluated.

It is interesting that the median survival of our cohort of myeloma patients is projected to exceed 5 years (median projected overall survival, 63 months). This is important because our patients were an unselected population, were quite elderly, only a minority had received ASCT and 20% presented with severe renal dysfunction. Such data are indicative of the progress that has been achieved over the past decade in the treatment of this disease, mostly due to the introduction of new therapies and to improvements in the use of the currently available treatment options. It is also important that a subset of patients with myeloma (mostly those with R-ISS-1) have a very high probability of a long disease course, with an expected survival well beyond a decade. In contrast, the poor outcome of haematologica | 2017; 102(3)

Evaluation of the R-ISS

patients at high risk, especially of those with R-ISS-3, emphatically demonstrates the need for new therapies and innovative approaches for the treatment of myeloma.

The R-ISS is based on three laboratory variables obtained from a blood sample and the presence or absence of three cytogenetic abnormalities evaluated by iFISH of a bone marrow aspiration sample. This system should be adopted in everyday clinical practice because it provides significant prognostic information. However, there are no prospective data supporting different treatment strategies for patients belonging to different risk groups at diagnosis.

Given the poor prognosis of patients with R-ISS-3 disease, it is reasonable to consider a more intensive treatment strategy and exploration of innovative treatments and drugs for such patients, who should be strongly encouraged to participate in clinical trials.

References 1. 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.

2. Greipp PR, San Miguel J, Durie BG, et al.

International staging system for multiple myeloma. J Clin Oncol. 2005;23(15):34123420.

3. Fonseca R, Barlogie B, Bataille R, et al.

Genetics and cytogenetics of multiple myeloma: a workshop report. Cancer Res. 2004;64(4):1546-1558.

4. Munshi NC, Anderson KC, Bergsagel PL, et al. Consensus recommendations for risk stratification in multiple myeloma: report of the International Myeloma Workshop Consensus Panel 2. Blood. 2011;117(18): 4696-4700.

haematologica | 2017; 102(3)

In conclusion, our series of consecutive, unselected patients with symptomatic myeloma, with significant differences in their characteristics and treatment approaches compared to the original IMWG cohort, verified that R-ISS provides significant prognostic information and that it allows the identification of three different groups of patients with clearly different outcomes. Thus, the R-ISS should be used as a standard for the risk stratification of patients with myeloma, for example in the stratification of patients in clinical trials. External validation is crucial and it would be useful to further validate the R-ISS in other cohorts of patients. The potential prognostic role of the RISS in patients with relapsed disease may also be evaluated, since this risk stratification tool is so far applicable to newly diagnosed patients and no data exist about its performance beyond first-line therapy.

5. Corre J, Avet-Loiseau H. The impact of genomics on the management of myeloma.

J Natl Compr Canc Netw. 2011;9(10):12001206.

6. Avet-Loiseau H. Role of genetics in prognostication in myeloma. Best Pract Res Clin Haematol. 2007;20(4):625-635.

7. Gkotzamanidou M, Kastritis E, Gavriatopoulou MR, et al. Increased serum lactate dehydrongenase should be included among the variables that define very-high-risk multiple myeloma. Clin Lymphoma Myeloma Leuk. 2011;11(5): 409-413.

8. Terpos E, Katodritou E, Roussou M, et al.

High serum lactate dehydrogenase adds prognostic value to the international myeloma staging system even in the era of novel agents. Eur J Haematol. 2010;85(2): 114-119.

9. Dimopoulos MA, Barlogie B, Smith TL, Alexanian R. High serum lactate dehydrogenase level as a marker for drug resistance

10.

11.

12.

13.

and short survival in multiple myeloma.

Ann Intern Med. 1991;115(12):931-935.

Palumbo A, Avet-Loiseau H, Oliva S, et al.

Revised international staging system for multiple myeloma: a report from International Myeloma Working Group. J Clin Oncol. 2015;33(26):2863-2869.

Ross FM, Avet-Loiseau H, Ameye G, et al.

Report from the European Myeloma Network on interphase FISH in multiple myeloma and related disorders.

Haematologica. 2012;97(8):1272-1277.

Ross FM, Ibrahim AH, Vilain-Holmes A, et al. Age has a profound effect on the incidence and significance of chromosome abnormalities in myeloma. Leukemia.

2005;19(9):1634-1642.

Dimopoulos MA, Terpos E, Gavriatopoulou M, et al. Myeloma in the octogenarians: disease characteristics and clinical outcomes in the era of modern antimyeloma therapy. Blood. 2014;124(21): 4738-4738.

599

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Cell Therapy & Immunotherapy

Ferrata Storti Foundation

Effect of high-dose plerixafor on CD34+ cell mobilization in healthy stem cell donors: results of a randomized crossover trial

Jeremy Pantin,1* Enkhtsetseg Purev,2* Xin Tian,3 Lisa Cook,2 Theresa Donohue-Jerussi,2 Elena Cho,2 Robert Reger,2 Matthew Hsieh,2 Hanh Khuu,4 Gary Calandra,5 Nancy L. Geller3 and Richard W. Childs2

Hematology and Oncology, Department of Medicine, Augusta University, GA; Hematology Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD; 3Office of Biostatistics Research, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD; 4Department of Transfusion Medicine, Clinical Research Center, National Institutes of Health, Bethesda, MD and 5 Sanofi Oncology, Cambridge, MA, USA 1 2

Haematologica 2017 Volume 102(3):600-609

*JP and EP contributed equally to this work.

ABSTRACT

H

Correspondence: childsr@nih.gov

Received: April 7, 2016.

Accepted: October 10, 2016.

Pre-published: October 20, 2016. doi:10.3324/haematol.2016.147132 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/3/600 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright.

All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode.

Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

600

ematopoietic stem cells can be mobilized from healthy donors using single-agent plerixafor without granulocyte colony-stimulating factor and, following allogeneic transplantation, can result in sustained donor-derived hematopoiesis. However, when a single dose of plerixafor is administered at a conventional 240 mg/kg dose, approximately one-third of donors will fail to mobilize the minimally acceptable dose of CD34+ cells needed for allogeneic transplantation. We conducted an open-label, randomized trial to assess the safety and activity of highdose (480 mg/kg) plerixafor in CD34+ cell mobilization in healthy donors.

Subjects were randomly assigned to receive either a high dose or a conventional dose (240 mg/kg) of plerixafor, given as a single subcutaneous injection, in a two-sequence, two-period, crossover design. Each treatment period was separated by a 2-week minimum washout period. The primary endpoint was the peak CD34+ count in the blood, with secondary endpoints of CD34+ cell area under the curve (AUC), CD34+ count at 24 hours, and time to peak CD34+ following the administration of plerixafor. We randomized 23 subjects to the two treatment sequences and 20 subjects received both doses of plerixafor. Peak CD34+ count in the blood was significantly increased (mean 32.2 versus 27.8 cells/µL, P=0.0009) and CD34+ cell AUC over 24 hours was significantly increased (mean 553 versus 446 h cells/mL, P<0.0001) following the administration of the 480 mg/kg dose of plerixafor compared with the 240 mg/kg dose.

Remarkably, of seven subjects who mobilized poorly (peak CD34+ ≤20 cells/mL) after the 240 mg/kg dose of plerixafor, six achieved higher peak CD34+ cell numbers and all achieved higher CD34+ AUC over 24 hours after the 480 mg/kg dose. No grade 3 or worse drug-related adverse events were observed. This study establishes that high-dose plerixafor can be safely administered in healthy donors and mobilizes greater numbers of CD34+ cells than conventional-dose plerixafor, which may improve CD34+ graft yields and reduce the number of apheresis procedures needed to collect sufficient stem cells for allogeneic transplantation. (ClinicalTrials.gov, identifier: NCT00322127) Introduction Granulocyte colony-stimulating factor (G-CSF) is the most commonly used cytokine to mobilize and collect peripheral blood progenitor cells from healthy donors for allogeneic transplantation. Although hematopoietic progenitors in G-CSFhaematologica | 2017; 102(3)

Stem cell mobilization using high-dose plerixafor

mobilized grafts differ from bone-marrow harvested grafts in several respects, the long-term engraftment potential of G-CSF-mobilized peripheral blood progenitor cell grafts is well documented.1 Reductions in bone marrow stromal cell derived factor-1 and up-regulation in CXCR4 constitute one of the multiple mechanisms through which G-CSF mobilizes CD34+ cells into the circulation.2 Although GCSF successfully mobilizes peripheral blood progenitor cells in most healthy donors,3 this agent often leads to inadequate mobilization of autologous peripheral blood progenitor cells in patients with hematologic malignancies who have received prior chemotherapy.4-7 Plerixafor is a bicyclam compound that inhibits the binding of stromal cell derived factor-1 to its cognate receptor CXCR4. Following administration of plerixafor, CD34+ cells are rapidly released into the circulation and can be collected by apheresis.8-11 The Food and Drug Administration has approved plerixafor at the 240 mg/kg dose for use in conjunction with G-CSF to mobilize autografts for transplantation in patients with non-Hodgkin lymphoma and multiple myeloma.12 Investigators have recently shown that allogeneic grafts can also be mobilized from healthy donors and non-human primates using single-agent plerixafor without G-CSF and, following transplantation, can result in sustained donor-derived hematopoiesis.10,13-15 However, when a single dose of plerixafor is administered at a conventional 240 mg/kg dose, approximately one-third of donors will fail to mobilize the minimally acceptable dose of CD34+ cells needed for allogeneic transplantation.14,16 Healthy volunteers tolerate plerixafor without experiencing many of the adverse events associated with G-CSF administration.10,16,17 Following a 240 mg/kg subcutaneous (sc) plerixafor injection, white blood cell counts peak at 6â&#x20AC;&#x201C; 10 h and return to baseline levels by approximately 24 h.10,16 Additionally, plerixafor has a short half-life10,18 and following administration for 3 consecutive days, a consistent and reversible increase in peripheral blood CD34+ cells can be achieved,19 with a rapid pharmacodynamic washout effect.

Initial phase I studies suggested optimal CD34+ cell mobilization was achieved following the sc administration of plerixafor 240 mg/kg. We performed a subsequent dose escalation study and observed that a single dose of plerixafor up to a dose of 480 mg/kg sc was safe in humans, and had a higher plerixafor Cmax and CD34+ cell area under the curve (AUC). However, the small number of subjects in that trial precluded determination of whether the efficacy

of CD34+ cell mobilization was improved with the higher 480 mg/kg dose of plerixafor.16 To test whether high-dose plerixafor is safe and more efficient at mobilizing stem cells for allogeneic transplantation compared to conventionaldose plerixafor, we conducted a randomized, crossover trial comparing CD34+ mobilization in healthy subjects mobilized with a single dose of sc plerixafor given at either a high dose (480 mg/kg) or a conventional dose (240 mg/kg).

Methods Study design and participants This was a phase II, open-label, single-center, investigator-initiated, randomized, crossover trial, conducted between October 12, 2010, and June 8, 2012, in healthy stem cell donors at the National Institutes of Health Clinical Center in Bethesda (MD, USA). Study subjects were randomly assigned in a 1:1 ratio to receive an initial dose of either 240 mg/kg plerixafor (followed by crossover to a 480 mg/kg dose of plerixafor) or an initial dose of 480 mg/kg plerixafor (followed by crossover to 240 mg/kg plerixafor) as shown in Figure 1. Each treatment period was separated by a minimum washout period of 2 weeks. Eligibility criteria were age between 18 and 50 years and normal renal, hepatic, and hematologic function. Exclusion criteria were cerebrovascular disease, cardiac disease, a Framingham 10-year coronary risk of >10%, positive human immunodeficiency virus status, a history of hepatitis B or C infection, a history of cancer in the preceding 5 years, a history of autoimmune disease, pregnant and lactating women, or a history of stroke or transient ischemic attack.

We selected a crossover trial design to evaluate the difference in activity endpoints between high-dose and conventional-dose plerixafor based on within-subject comparisons because CD34+ cell mobilization varies considerably between stem cell donors.16 The short half-life and rapid washout of plerixafor enable the use of this randomized crossover design with a minimum of a 2-week washout period to avoid a carryover effect and allow for an estimate of the direct treatment effect of high-dose plerixafor on CD34+ cell mobilization.

This trial conformed with the principles of the Declaration of Helsinki and was approved by the National Heart, Lung, and Blood Institute Institutional Review Board. All participants provided written informed consent.

Procedures Healthy stem cell donors were screened and assessed for eligi-

Figure 1. Study design.

haematologica | 2017; 102(3)

601

J. Pantin et al.

bility. Eligible study participants were admitted to the National Institutes of Health Clinical Center for dosing and phlebotomy and underwent continuous cardiac telemetry monitoring for 24 h after each injection of plerixafor. Complete blood counts and serum chemistry panels were assessed in all subjects within 24 h prior to each dose of plerixafor and 1 week after each administration. Electrocardiography was done prior to dosing and before discharge. Complete blood counts and CD34+ cell counts were enumerated immediately before (time 0) and 2, 4, 6, 8, 10, 12, 14, 18 and 24 h after each sc injection of plerixafor. Blood samples were evaluated for CD34+ cell content using the International Society of Hematotherapy and Graft Engineering guidelines.20 The absolute number of circulating CD34+ cells/mL was determined by standard flow cytometry techniques. Fluorochromes used for cell surface staining were CD45 APC-Cy7 and CD34 APC (IgG1 class antibody, BD Biosciences, San Jose, CA, USA) and cells were identified using the BD FACSCanto instrument (BD Biosciences).

FACSDiva software (BD Biosciences) was used for the analysis.

To assess colony-forming units (CFU), peripheral blood samples obtained immediately before and 6 h after dosing with plerixafor were compared, as described previously.16

mined to have 96% power to detect a difference of 10 cells/mL (SD of the difference 12) in peak CD34+ counts using a two-sided paired t-test at the 0.05 significance level. Efficacy was analyzed on an intention-to-treat basis and all randomized participants were included. The analysis of activity endpoints based on paired differences was also assessed in all participants who received both doses of study drug. All study subjects were included in the safety analysis.

Repeated measures analysis of variance was used for the primary efficacy analysis to compare circulating CD34+ cell numbers and AUC between the two plerixafor dose levels. The secondary analysis of paired data used paired t-tests or Wilcoxon signed rank tests for non-normally distributed data. The period effect and carryover effect were tested using unpaired t-tests or Mann-Whitney U tests to examine whether the order of administration of the two plerixafor doses affected the overall activities and differences in CD34+ cell mobilization, respectively. The incidences of adverse events between the two dose groups were compared by the Fisher exact test. R statistical software (version 3.2.2) was used for all statistical analyses.

Outcomes

Results

The primary endpoint was the peak CD34+ cell count in the blood within 24 h following plerixafor injection. Secondary endpoints were CD34+ cell AUC over 24 h, CD34+ count at 24 h, time to peak CD34+ cell count, and CFU quantified from peripheral blood samples. CD34+ cell AUC was calculated using the trapezoidal rule. The safety endpoint was the tolerability and toxicity, including incidence and severity of adverse events, evaluated after each injection of plerixafor up until 1 week following its administration. Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 3. Treatment-related serious adverse events were reviewed by an independent medical monitor.

Statistical analysis This randomized crossover study was designed to determine whether the 480 mg/kg dose of plerixafor would mobilize higher peak CD34+ cell numbers in the blood compared to the conventional 240 mg/kg dose. A sample size of 20 subjects was deter-

Table 1. Baseline characteristics.

Age (years) Sex Female Male Weight (kg) Height (cm) Body mass index (kg/m2) Body surface area (m2) Race Caucasian African American Asian CD34+ cells/mL at 0 h

Randomized to plerixafor 240 µg/kg/ 480 µg/kg (n=11)

Randomized to plerixafor 480 µg/kg/ 240 µg/kg (n=12)

Total (n= 23)

30 (19-42)

33 (20-48)

31 (19-48)

4 (36%) 7 (64%) 78 (63- 111) 177 (162-190) 25 (20-35) 2.0 (1.7-2.4)

7 (58%) 5 (42%) 68 (48-108) 171 (154-184) 24 (18-32) 1.8 (1.5-2.3)

11 (48%) 12 (52%) 73 (48-111) 172 (154-190) 25 (18-35) 1.9 (1.5-2.4)

6 (55%) 5 (45%) 0 (0%) 1 (0-4)

9 (75%) 2 (17%) 1 ( 8%) 1 (0-4)

15 (65%) 7 (30%) 1 (4%) 1 (0-4)

Data are median (range) or n (%).

602

Subjects’ characteristics We screened 28 healthy stem cell donors and enrolled 23 into the trial, comprising 11 women and 12 men, with a median age of 31 years (range, 19-48) and a median weight of 73 kg (range, 48-111). Table 1 summarizes the subjects’ baseline characteristics which were comparable between the two randomization groups. Three subjects discontinued the study after receiving the initial dose of plerixafor.

One subject failed to return after receiving the 240 mg/kg dose, and two subjects were withdrawn from the study after receiving the 480 mg/kg dose; neither withdrawal was attributed to plerixafor administration. Therefore, 20 subjects received both doses of plerixafor and completed the study.

CD34+ cell mobilization Table 2 shows the results of primary and secondary outcomes in all 23 randomized subjects. Compared to the 240 µg/kg dose, the 480 mg/kg dose of plerixafor resulted in: higher peak circulating CD34+ cell numbers [mean 32.2 versus 27.8 cells/mL; mean difference 4.6 cells/mL (95% CI: 2.3−6.9), P=0.0009], higher circulating CD34+ cell numbers

Table 2. Treatment effects on primary and secondary outcomes.

Peak CD34+ (cells/mL) CD34+ cell AUC 0-24 h (h cells/mL) Time to CD34+ peak (h) CD34+ at 24 h (cells/mL )

Plerixafor 240 mg/kg mean (SD) (n=21)

Plerixafor 480 mg/kg mean (SD) (n=22)

Treatment difference mean (95% CI)

P value

27.8 (11.7)

32.2 (11.8)

4.6 (2.3-6.9)

0.0009

446 (207)

553 (223)

113 (79-148)

<0.0001

8.4 (1.9)

10.5 (2.9)

2.1 (0.6-3.5)

0.012

10.7 (7.1)

17.8 (11.0)

7.3 (4.7-9.9)

<0.0001

Treatment difference was estimated from the repeated measures analysis of variance.

haematologica | 2017; 102(3)

Stem cell mobilization using high-dose plerixafor

at 24 h [mean 17.8 versus 10.7 cells/mL; mean difference 7.3 cells/mL (95% CI: 4.7−9.9), P<0.0001], and a greater circulating CD34+ AUC over 24 h [mean 553 versus 446 h cells/mL; mean difference 113 h cells/mL (95% CI: 79−148), P<0.0001]. In addition, the time to achieve this higher peak in the peripheral blood CD34+ count with the 480 mg/kg dose was longer by an average of 2.1 h; CD34+ cells peaked at a mean of 10.5 h (range 6-18, median 10) after the 480 mg/kg dose compared with a mean of 8.4 h (range 6-12, median 8) after the 240 mg/kg dose (P=0.012). Figure 2 and Online Supplementary Tables S1 and S2 show the analysis of paired data from 20 individual subjects who received both doses of plerixafor, with each line connecting the same subject at the two dose levels. In most subjects, all of these measures were greater following the administration of the 480 mg/kg dose compared to the 240 mg/kg dose. The peak circulating CD34+ counts were higher in 16 (same in one and lower in three) out of 20 subjects following the administration of plerixafor at the 480 mg/kg dose compared to the 240 mg/kg dose. Other exceptions included one subject who had a higher CD34+ AUC, two subjects who had a higher CD34+ cell number at 24 h, and three subjects who had a longer time to peak in circulating CD34+ cell numbers

A

B

C

D

with the 240 mg/kg of plerixafor. Of note, no evidence of a period effect (P=0.76) or a carryover effect (P=0.59) was detected on the primary endpoint of peak CD34+ count.

These tests were not statistically significant (all P>0.1) for other secondary endpoints. Thus the observed differences in CD34+ cell mobilization represented a direct treatment effect and were not related to the order of administration of the two plerixafor doses.

An analysis of the mean number of CD34+ cells mobilized into the circulation over time (24 h) for both dose cohorts demonstrated the CD34+ cell AUC was greater with the 480 mg/kg plerixafor dose (Figure 3). This, combined with the observation that CD34+ cells peaked 2 h later with the higher plerixafor dose, would affect the timing and duration of apheresis procedures to optimize the CD34+ cell yield collected after administering the 480 mg/kg plerixafor dose. As shown in Figure 3A where higher sustained CD34+ circulating numbers were observed following high-dose plerixafor, a large-volume apheresis spanning ≥4 h initiated 8 h after administration of plerixafor 480 mg/kg would be predicted to provide a greater collection yield of CD34+ cells compared to a volume-matched apheresis collected over the same amount of time, starting

Figure 2. CD34+ cells in the peripheral blood in 20 subjects who received both doses of plerixafor.

Individual subject plots with each line connecting the same subject at both dose levels representing the peak CD34+ peripheral blood count, CD34+ AUC over 24 h, CD34+ count 24 h after plerixafor administration, and time to peak CD34+ count.

haematologica | 2017; 102(3)

603

J. Pantin et al. at any time point, following the 240 mg/kg dose of plerixafor. When we examined paired differences between peripheral blood CD34+ counts following administration of both plerixafor doses over time in individual subjects (Figure 3B), we observed that after 2 h there was a steady increase in the difference between CD34+ counts in subjects who received both plerixafor doses at each sampled timepoint, with these differences being statistically significant over the 8 to 24 h time-frame (shaded regions between curves in Figure 3A). Seven (35%) of the 20 subjects had peak CD34+ counts ≤20 cells/mL after the 240 µg/kg dose of plerixafor and, using conventional criteria, were classified as poor mobilizers (Online Supplementary Figure S1).21 Highdose plerixafor resulted in significantly better CD34+ cell mobilization compared to plerixafor 240 mg/kg in both the poor mobilizer and the good mobilizer subgroups (Figure 3C,D and Figure 4). Remarkably, as shown in the subgroup analysis in Figure 4, six of the seven (86%) poor mobilizers achieved higher peak CD34+ cell numbers, corresponding to a relative increase of 28% (95% CI: 11–46%, P=0.014), and all seven poor mobilizers achieved higher CD34+ AUC over 24 h, corresponding to a relative increase of 36% (95% CI: 18–53%, P=0.005) after the 480 mg/kg dose compared with the 240 mg/kg dose. This increase also represented a relative

improvement of 27% (95% CI: 9–44) and 30% (95% CI: 13–46) in the CD34+ cell AUC over an optimal 4-h and 6-h collection time, respectively, after the 480 mg/kg dose.

Furthermore, there was a trend towards peak CD34+ cell counts and CD34+ AUC increasing more in poor mobilizers than in good mobilizers after high-dose versus conventional-dose plerixafor. In addition to the greater increase in CD34+ counts, there was a significant increase in circulating total white blood cells, lymphocytes, monocytes, and granulocytes, over time following administration of the 480 mg/kg dose of plerixafor compared with the 240 mg/kg dose (Figure 5).

Colony-forming units The comparison of blood erythroid (E) or granulocytemacrophage (GM) CFU colonies is shown in Online Supplementary Figure S2. Compared to baseline CFU counts immediately before dosing, the mean number of CFU-E colonies was slightly increased following each dose of plerixafor. In contrast, there was a significant increase in CFUGM colonies following each dose of plerixafor. The increases in either CFU-E or CFU-GM colonies were comparable based on samples collected 6 h following the high dose and conventional dose of plerixafor.

A

B

C

D

Figure 3. CD34+ cell counts. (A) Mean CD34+ cell counts in the blood over time with one standard error of the mean (SEM) in all subjects who received both doses of plerixafor. The shaded regions indicate when the mean CD34+ counts were significantly different between the two dose cohorts. (B) Mean paired difference in the CD34+ count following administration of the 480 mg/kg and 240 mg/kg dose with 95% CI. Mean CD34+ cell counts with 1 SEM for (C) poor mobilizers and (D) good mobilizers, defined as those with peak CD34+ counts ≤ 20 cells/mL and > 20 cells/mL after the 240 mg/kg dose of plerixafor, respectively.

604

haematologica | 2017; 102(3)

Stem cell mobilization using high-dose plerixafor

Safety

The high dose of plerixafor (480 mg/kg) was well-tolerated. Among all 23 study participants, only one grade 3 serious adverse event occurred in one subject (night terrors and palpitations), which led to overnight hospitalization for observation 3 days after the plerixafor had been administered; this event was attributed to the subject working in a stressful work environment. No treatment-related serious adverse event or adverse events of CTCAE ≥grade 3 occurred. There were eight grade 2 adverse events in six subjects who received the 240 mg/kg dose, and 19 grade 2 adverse events in ten subjects who received the 480 mg/kg dose. The percentages of subjects who had grade 2 adverse events were not statistically significantly different between the two doses (P= 0.29). Table 3 shows all adverse events that were considered at least possibly related to plerixafor.

Commonly reported adverse events consisted of local selflimiting erythema or induration at the site of injection, tachycardia, and the sensation of gastrointestinal bloating, flatulence, nausea, and diarrhea. Most adverse events were mild (grade 1), occurred within 30 min to 4 h after plerixafor administration, and resolved without intervention.

The overall incidences of all adverse events and of adverse events in each grade or category were not significantly different between subjects receiving the high dose or conventional dose of plerixafor.

Discussion This study establishes that high-dose plerixafor can be

safely administered to healthy donors and mobilizes greater numbers of CD34+ cells than conventional-dose plerixafor.

The ability to collect a large quantity of peripheral blood progenitor cells following a single injection of plerixafor without the need for G-CSF makes it an attractive agent for mobilizing donors for allogeneic hematopoietic stem cell transplantation.14,16 Despite its ease of use, we and others have observed that about one-third of healthy donors will fail to mobilize ≥20 CD34+ cells/mL after a single 240 mg/kg dose of plerixafor. These donors, conventionally classified as poor mobilizers, are less likely to have a single apheresis that yields at least 2 x 106 CD34+ cells/kg14,22 and would require one or more additional injections of plerixafor or additional apheresis procedures to collect an adequate number of CD34+ cells. Determining the most efficacious dose of plerixafor in healthy donors would improve allograft CD34+ cell yields and reduce the number of apheresis procedures. Initial phase I studies suggested that peak CD34+ cell mobilization occurred at the highest administered dose of 240 mg/kg plerixafor.10 In a recent dose escalation study, we reported that a single 480 mg/kg dose of plerixafor is safe and effective in mobilizing CD34+ cells.16 However, that study enrolled only a limited number of subjects and did not randomize the order of dosing, thus its design was inadequate to determine whether the efficacy of CD34+ cell mobilization was improved with high-dose plerixafor.

Therefore, to evaluate whether high-dose plerixafor is more efficient in mobilizing CD34+ stem cells compared to conventional-dose plerixafor, we conducted the current

Figure 4. Subgroup analyses of relative differences in CD34+ cell mobilization. “All completed” included all subjects who received both doses of plerixafor. AUC 4h (or AUC 6h) was calculated from 6-10 h (or 6-12 h) for the 240 mg/kg dose of plerixafor, and from 8-12 h (or 8-14 h) for the 480 mg/kg dose of plerixafor, respectively, as the optimal collection time would be expected to start 2 h prior to CD34+ cells peaking in the circulation to achieve the maximum CD34+ AUC time window.

haematologica | 2017; 102(3)

605

J. Pantin et al.

study that utilized a randomized crossover trial design with over 95% power to compare CD34+ mobilization in subjects mobilized with a single dose of plerixafor of either 480 mg/kg or 240 mg/kg. We observed that peak circulating CD34+ cell numbers, circulating CD34+ cell numbers at 24 h, and the CD34+ cell AUC over 24 h were, on average, all significantly higher following the administration of 480 mg/kg compared to 240 mg/kg of plerixafor. Remarkably, the peak circulating CD34+ counts were higher in 16 of 20 of subjects following the administration of plerixafor 480 mg/kg compared to plerixafor 240 mg/kg. Perhaps most importantly, improvements in CD34+ cell mobilization with high-dose plerixafor were most pronounced in those who were poor mobilizers with conventional-dose plerixafor. In this context, all seven subjects who mobilized poorly with plerixafor 240 mg/kg achieved a significantly higher CD34+ AUC after the 480 mg/kg dose (relative increase 36%) and six of seven (86%) had higher peak CD34+ cell numbers (relative increase 28%) following high-dose plerixafor (Figure 4 and Online Supplementary Figure S1). These increases represented relative improvements of 27% and 30% in AUC over an optimal 4-h and 6h collection time, respectively, after the 480 mg/kg dose.

This difference would be expected to make a significant clinical impact as those who would mobilize poorly and subsequently fail to yield enough CD34+ cells at the 240 mg/kg dose would be more likely to mobilize an adequate

number of CD34+ cells at the higher 480 mg/kg dose of plerixafor. We noted high-dose plerixafor was well-tolerated in the 22 subjects who received this agent, and was not associated with treatment-related serious adverse events or adverse events of grade 3 or higher. Although the absolute number of grade 2 adverse events was slightly higher in subjects who received the 480 mg/kg dose than in those who received the 240 mg/kg dose, there was no significant difference in the severity and overall incidence of adverse events between subjects receiving high-dose and conventional-dose plerixafor. Most adverse events were mild, transient, and resolved without treatment.

A limitation of our study is that we compared CD34+ mobilization in the peripheral blood without performing an apheresis procedure. However, numerous reports have shown that the number of circulating CD34+ cells in peripheral blood correlates strongly with CD34+ graft yield and predicts the number of CD34+ cells that will be collected during apheresis.21,23,24 Achieving higher peak circulating CD34+ cell numbers and a higher circulating CD34+ AUC are clinically important outcomes that would be expected to improve the yield of hematopoietic progenitor cells collected during apheresis procedures. The target for collection for unmanipulated allogeneic peripheral blood progenitor cells is frequently 5x106 CD34+ cells/kg recipient body weight with a minimally acceptable number of 2x106 CD34+ cells/kg.23,25 Most collection centers do not have real-

Figure 5. Mean circulating white blood cells, absolute lymphocytes, absolute granulocytes, and absolute monocytes over time with one standard error of mean for both dose cohorts. The shaded regions indicate when the mean circulating cell counts were significantly different between the two dose cohorts.

606

haematologica | 2017; 102(3)

Stem cell mobilization using high-dose plerixafor

time information on the number of CD34+ cells mobilized in individual subjects at the time the apheresis collection is initiated. In this context, apheresis volumes are fixed, and in subjects who mobilize poorly, often fail to collect an adequate number of CD34+ cells needed for transplantation. Utilization of higher-dose plerixafor would likely increase the probability that the target dose of CD34+ progenitor cells can be successfully collected following a single apheresis procedure. It is important to consider that the

efficiency with which CD34+ cells are collected directly affects the yield of CD34+ cells in apheresis collections. At our center, CD34+ cell efficiencies typically average between 30-40%, which is consistent with reports from other apheresis centers.24 Based on a validated formula24 and the within-subject comparison data in this study (Online Supplementary Figure S3A), for a 24 liter apheresis processing volume, a recipient with a 70 kg body weight, and a 30% collection efficiency, 7/20 (35%) poor mobiliz-

Table 3. Treatment-related adverse events in all 23 subjects.

240 µg/kg (n = 21 subjects) Adverse event Allergy Injection site reaction Hives Blood/bone marrow Absolute neutrophil count decreased Hemoglobin, low White blood cells decreased Cardiac Premature ventricular contraction Sinus bradycardia Sinus tachycardia Arrhythmia Vasovagal episode Hypotension Pulses fuller Constitutional Altered sleep/insomnia Fatigue Night sweats Gastrointestinal Abdominal distention/bloating Pain/cramping Diarrhea Flatulence (gas) Heart burn Nausea Vomiting Other Metabolic/laboratory Albumin, decreased Alanine aminotransferase, increased Bilirubin, increased Lactate dehydrogenase, increased Total protein, decreased Electrolyte abnormalities Neurology/pain/vision Dizziness Facial paresthesia Headache Tingling Blurred vision Vivid dreams Pulmonary Cough Chest wall tightness Hiccups Dyspnea Upper respiratory infection

480 µg/kg (n = 22 subjects)

Grade 1

Grade 2

Grade 1

Grade 2

16 (76%) −

− −

19 (86%) −

− 1 (5%)

1 (5%) 4 (19%) 1 (5%)

− − −

− 1 (5%) −

− − −

3 (14%) 2 (10%) 10 (48%) − − − 1 (5%)

− − − − − − −

6 (27%) 1 (5%) 13 (59%) 2 (9%) − − −

− − − − 1 (5%) 1 (5%) −

1 (5%) 3 (14%) −

− − −

1 (5%) 2 (9%) 1 (5%)

− − 1 (5%)

3 (14%) 1 (5%) 3 (14%) 6 (29%) − 2 (10%) − 1 (5%)

1 (5%) − 1 (5%) 1 (5%) 1 (5%) − − −

2 (9%) 3 (14%) 4 (18%) 6 (27%) − 4 (18%) − −

2 (9%) − − 2 (9%) 2 (9%) − 1 (5%) −

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

− − − − − 1 (5%)

2 (9%) − 2 (9%) 1 (5%) − 5 (23%)

− − − − − 2 (9%)

1 (5%) 3 (14%) 1 (5%) − − −

− 1 (5%) 1 (5%) − − −

1 (5%) 4 (18%) 1 (5%) 1 (5%) 2 (9%) 3 (14%)

− 3 (14%) 2 (9%) − − −

1 (5%) 1 (5%) 1 (5%) − −

− − − − 1 (5%)

1 (5%) 2 (9%) − 2 (9%) −

− 1 (5%) − − −

Data are n (%).

haematologica | 2017; 102(3)

607

J. Pantin et al.

ers would be predicted to fail to yield the minimally acceptable 2x106 CD34+ cells/kg after a 240 mg/kg dose of plerixafor while only 3/20 (15%) of those donors would fail to mobilize adequately after the 480 mg/kg of plerixafor. With a higher collection efficiency of 35-40%, 2/20 (10%) donors would still fail to yield a minimally acceptable CD34+ number following the 240 mg/kg dose, while all of the 20 donors would be expected to have an adequate CD34+ cell collection after the 480 mg/kg dose. Importantly, regardless of the CD34+ cell collection efficiency, the proportion of donors achieving a minimally acceptable apheresis collection would predictably be higher across the spectrum of apheresis volumes typically processed following stem cell mobilization (Online Supplementary Figure S3B). Furthermore, the total number of CD34+ cells collected following apheresis would be expected to be 20% greater after the 480 mg/kg dose than after the 240 mg/kg dose, with higher percentages of donors yielding grafts containing both >2x106 and >5x106 CD34+ cells/kg across typical collection efficiencies and volumes. These differences are clinically meaningful and could be the determining factor obviating the need to proceed with a second apheresis collection. For cost, donor safety and improved collection efficiency, a single collection is preferable to multiple collections, even if the duration of the procedure needs to be extended.26,27 Interestingly, in our study we observed that the time to achieve a higher peak peripheral CD34+ count was, on average, 2 h longer after high-dose plerixafor. Compared to the 240 mg/kg dose, the mean CD34+ counts in the circulation remained significantly higher between 8 and 24 h after the 480 mg/kg dose of plerixafor (Figure 3). These differences may be the consequence of altered pharmacokinetics that are observed with high-dose plerixafor, as we previously showed the plerixafor Cmax and AUC were higher after the 480 mg/kg dose than after the 240 mg/kg dose.16 Therefore, should a decision be made to utilize a higher dose of plerixafor to mobilize an allograft, this delay in peak circulating CD34+ cells would need to be factored into decisions related to the timing of apheresis procedures to optimize progenitor cell yield. Assuming cell harvesting is scheduled for 08:00, a 480 mg/kg dose of plerixafor would optimally be administered 8 h (24:00) the night before the apheresis procedure to allow for a collection that spans the time when CD34+ cells are peaking in the circulation. As shown in Figure 3B, the sustained increase in circulating CD34+ cells during the 8 to 24 h after high-dose plerixafor is administered would also provide advantages should circumstances lead to an unforeseen delay in initiating the apheresis collection. In this context, given these altered mobilization kinetics, an adequate CD34+ cell yield could still be collected even if the procedure was delayed 10 to 14 h after the 480 mg/kg dose of plerixafor had been administered.

Despite a significant increase in all measures of peripherally circulating CD34+ cell counts following the administration of the 480 mg/kg dose of plerixafor, we found no difference between conventional-dose or high-dose plerixafor with respect to mobilized CFU-E or CFU-GM measured from peripheral blood samples. This lack of observed difference was likely due to the fact that blood samples for this analysis were collected 6 h after plerixafor administration, which was found, retrospectively from this study, to be a time-point when circulating CD34+ cell numbers were similar between both dose 608

cohorts, as opposed to later time-points when the number of circulating CD34+ progenitor cells was significantly higher in the blood following high-dose plerixafor administration (Figure 3B). In our study, the enhanced efficacy of high-dose plerixafor on CD34+ cell mobilization was not affected by the donorâ&#x20AC;&#x2122;s age. The median age of the study subjects was 31 years (range, 19-48) and age, when analyzed in a multivariate analysis, did not correlate significantly with CD34+ cell mobilization. Besides CD34+ cells, we and others have reported that neutrophils, monocytes, B cells, and T cells are also mobilized with plerixafor.22 In this study, we also noted that total white blood cells, lymphocytes, monocytes, and granulocytes were significantly higher in the circulation over time following administration of the 480 mg/kg dose of plerixafor compared with the 240 mg/kg dose.

Recent data from our group have shown that alterations in T-cell phenotype and cytokine gene expression profiles characteristic of G-CSF mobilization do not occur after mobilization with plerixafor.22 G-CSF mobilization has been shown to skew T cells toward a Th-2 type phenotype, which may increase the incidence of chronic graft-versus-host disease when G-CSF-mobilized allografts are used compared with bone marrow transplants.2830 In contrast, T cells mobilized with plerixafor do not display any changes in Th-1, Th-2, or Th-3 type cytokines compared to non-mobilized T cells. Recent studies of murine allogeneic transplantation models have also shown that the incidence of graft-versus-host disease may differ in recipients of allo-grafts mobilized with plerixafor alone or plerixafor plus G-CSF compared to G-CSF alone.22,31 These data suggest that T-cell-mediated events which occur after allogeneic transplantation may differ depending on whether donor allografts are mobilized with plerixafor or G-CSF. Given that high-dose plerixafor mobilized even greater numbers of lymphocytes than did conventional-dose plerixafor, it is possible that graft-versus-tumor effects, immune reconstitution, graft-versushost disease, and other variables affected by engrafting donor lymphoid cells could also differ in recipients of allogeneic transplants mobilized with high-dose versus conventional-dose plerixafor.

In conclusion, this study demonstrates that high-dose plerixafor can be administered safely and is superior to conventional-dose plerixafor in mobilizing CD34+ cells in healthy donors. The enhanced mobilizing effect of highdose plerixafor was most evident in subjects who had the greatest need for this effect, namely those who mobilized poorly with conventional-dose plerixafor. Our data suggest that mobilization of allogeneic stem cell donors with high-dose plerixafor would improve the chances of using a single apheresis procedure to collect a sufficient number of CD34+ cells for allo-grafting and would likely result in graft collections containing higher CD34+ cell numbers compared to those of donors mobilized with conventional-dose plerixafor. Our findings warrant further studies to explore the clinical impact of high-dose plerixafor use for allogeneic stem cell transplantation.

Acknowledgements This research was supported by the Intramural Research Program of the National Heart, Lung and Blood Institute and the Clinical Center, National Institutes of Health. Sanofi US supported this trial by supplying the study drug. haematologica | 2017; 102(3)

Stem cell mobilization using high-dose plerixafor

References 1. Gyger M, Stuart RK, Perreault C. Immunobiology of allogeneic peripheral blood mononuclear cells mobilized with granulocyte-colony stimulating factor. Bone Marrow Transplant. 2000;26(1):1-16.

2. 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.

3. Brissot E, Chevallier P, Guillaume T, et al.

Factors predicting allogeneic PBSCs yield after G-CSF treatment in healthy donors.

Bone Marrow Transplant. 2009;44(9):613615.

4. Visani G, Lemoli RM, Tosi P, et al.

Fludarabine-containing regimens severely impair peripheral blood stem cells mobilization and collection in acute myeloid leukaemia patients. Br J Haematol. 1999;105 (3):775-779.

5. Pastore D, Specchia G, Mestice A, et al.

Good and poor CD34+ cells mobilization in acute leukemia: analysis of factors affecting the yield of progenitor cells. Bone Marrow Transplant. 2004;33(11):1083-1087.

6. Goldschmidt H, Hegenbart U, Wallmeier M, Hohaus S, Haas R. Factors influencing collection of peripheral blood progenitor cells following high-dose cyclophosphamide and granulocyte colony-stimulating factor in patients with multiple myeloma. Br J Haematol. 1997;98(3):736-744.

7. Kumar S, Dispenzieri A, Lacy MQ, et al.

Impact of lenalidomide therapy on stem cell mobilization and engraftment post-peripheral blood stem cell transplantation in patients with newly diagnosed myeloma.

Leukemia. 2007;21(9):2035-2042.

8. Aiuti A, Webb IJ, Bleul C, Springer T, Gutierrez-Ramos JC. The chemokine SDF1 is a chemoattractant for human CD34+ hematopoietic progenitor cells and provides a new mechanism to explain the mobilization of CD34+ progenitors to peripheral blood. J Exp Med. 1997;185(1): 111-120.

9. 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):1095-1102.

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

haematologica | 2017; 102(3)

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

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

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.

Brave M, Farrell A, Ching Lin S, et al. FDA review summary: Mozobil in combination with granulocyte colony-stimulating factor to mobilize hematopoietic stem cells to the peripheral blood for collection and subsequent autologous transplantation.

Oncology. 2010;78(3-4):282-288.

Donahue RE, Jin P, Bonifacino AC, et al.

Plerixafor (AMD3100) and granulocyte colony-stimulating factor (G-CSF) mobilize different CD34+ cell populations based on global gene and microRNA expression signatures. Blood. 2009;114(12):2530-2541.

Devine SM, Vij R, Rettig M, et al. Rapid mobilization of functional donor hematopoietic cells without G-CSF using AMD3100, an antagonist of the CXCR4/SDF-1 interaction. Blood. 2008;11 2(4):990-998.

Larochelle A, Krouse A, Metzger M, et al.

AMD3100 mobilizes hematopoietic stem cells with long-term repopulating capacity in nonhuman primates. Blood. 2006;107(9): 3772-3778.

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.

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.

Stewart DA, Smith C, MacFarland R, Calandra G. Pharmacokinetics and pharmacodynamics of plerixafor in patients with non-Hodgkin lymphoma and multiple myeloma. Biol Blood Marrow Transplant.

2009;15(1):39-46.

Hubel K, Liles WC, Broxmeyer HE, et al.

Leukocytosis and mobilization of CD34+ hematopoietic progenitor cells by AMD3100, a CXCR4 antagonist. Support Cancer Ther. 2004;1(3):165-172.

Sutherland DR, Anderson L, Keeney M, Nayar R, Chin-Yee I. The ISHAGE guidelines for CD34+ cell determination by flow

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

cytometry. International Society of Hematotherapy and Graft Engineering. J Hematother. 1996;5(3):213-226.

Armitage S, Hargreaves R, Samson D, Brennan M, Kanfer E, Navarrete C. CD34 counts to predict the adequate collection of peripheral blood progenitor cells. Bone Marrow Transplant. 1997;20(7):587-591.

Lundqvist A, Smith AL, Takahashi Y, et al.

Differences in the phenotype, cytokine gene expression profiles, and in vivo alloreactivity of T cells mobilized with plerixafor compared with G-CSF. J Immunol. 2013;191(12): 6241-6249.

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

Rosenbaum ER, O'Connell B, Cottler-Fox M. Validation of a formula for predicting daily CD34(+) cell collection by leukapheresis. Cytotherapy. 2012;14(4):461-466.

Singhal S, Powles R, Treleaven J, et al. A low CD34+ cell dose results in higher mortality and poorer survival after blood or marrow stem cell transplantation from HLA-identical siblings: should 2 x 10(6) CD34+ cells/kg be considered the minimum threshold? Bone Marrow Transplant. 2000;26(5):489-496.

Winters JL. Complications of donor apheresis. J Clin Apher. 2006;21(2):132-141.

Bolan CD, Carter CS, Wesley RA, et al.

Prospective evaluation of cell kinetics, yields and donor experiences during a single largevolume apheresis versus two smaller volume consecutive day collections of allogeneic peripheral blood stem cells. Br J Haematol.

2003;120(5):801-807.

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

Pharmacologic doses of granulocyte colonystimulating factor affect cytokine production by lymphocytes in vitro and in vivo.

Blood. 2000;95(7):2269-2274.

Franzke A, Piao W, Lauber J, et al. G-CSF as immune regulator in T cells expressing the G-CSF receptor: implications for transplantation and autoimmune diseases. Blood.

2003;102(2):734-739.

Anasetti C, Logan BR, Lee SJ, et al.

Peripheral-blood stem cells versus bone marrow from unrelated donors. N Engl J Med. 2012;367(16):1487-1496.

Arbez J, Saas P, Lamarthee B, et al. Impact of donor hematopoietic cells mobilized with G-CSF and plerixafor on murine acute graftversus-host-disease. Cytotherapy. 2015;17 (7):948-955.

609