haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
Ancient Greek
The origin of a name that reflects Europe’s cultural roots.
Scientific Latin
aÂma [haima] = blood a·matow [haimatos] = of blood lÒgow [logos]= reasoning
Scientific Latin
haematologicus (adjective) = related to blood
Modern English
haematologica (adjective, plural and neuter, used as a noun) = hematological subjects The oldest hematology journal, publishing the newest research results. 2016 JCR impact factor = 7.702
Haematologica, as the journal of the European Hematology Association (EHA), aims not only to serve the scientific community, but also to promote European cultural identify.
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation Editor-in-Chief Luca Malcovati (Pavia)
Managing Director Antonio Majocchi (Pavia)
Associate Editors Omar I. Abdel-Wahab (New York), Hélène Cavé (Paris), Simon Mendez-Ferrer (Cambridge), Pavan Reddy (Ann Arbor), Andreas Rosenwald (Wuerzburg), Monika Engelhardt (Freiburg), Davide Rossi (Bellinzona), Jacob Rowe (Haifa, Jerusalem), Wyndham Wilson (Bethesda), Paul Kyrle (Vienna), Swee Lay Thein (Bethesda), Pieter Sonneveld (Rotterdam)
Assistant Editors Anne Freckleton (English Editor), Cristiana Pascutto (Statistical Consultant), Rachel Stenner (English Editor), Kate O’Donohoe (English Editor), Ziggy Kennell (English Editor)
Editorial Board Jeremy Abramson (Boston); Paolo Arosio (Brescia); Raphael Bejar (San Diego); Erik Berntorp (Malmö); Dominique Bonnet (London); Jean-Pierre Bourquin (Zurich); Suzanne Cannegieter (Leiden); Francisco Cervantes (Barcelona); Nicholas Chiorazzi (Manhasset); Oliver Cornely (Köln); Michel Delforge (Leuven); Ruud Delwel (Rotterdam); Meletios A. Dimopoulos (Athens); Inderjeet Dokal (London); Hervé Dombret (Paris); Peter Dreger (Hamburg); Martin Dreyling (München); Kieron Dunleavy (Bethesda); Dimitar Efremov (Rome); Sabine Eichinger (Vienna); Jean Feuillard (Limoges); Carlo Gambacorti-Passerini (Monza); Guillermo Garcia Manero (Houston); Christian Geisler (Copenhagen); Piero Giordano (Leiden); Christian Gisselbrecht (Paris); Andreas Greinacher (Greifswals); Hildegard Greinix (Vienna); Paolo Gresele (Perugia); Thomas M. Habermann (Rochester); Claudia Haferlach (München); Oliver Hantschel (Lausanne); Christine Harrison (Southampton); Brian Huntly (Cambridge); Ulrich Jaeger (Vienna); Elaine Jaffe (Bethesda); Arnon Kater (Amsterdam); Gregory Kato (Pittsburg); Christoph Klein (Munich); Steven Knapper (Cardiff); Seiji Kojima (Nagoya); John Koreth (Boston); Robert Kralovics (Vienna); Ralf Küppers (Essen); Ola Landgren (New York); Peter Lenting (Le Kremlin-Bicetre); Per Ljungman (Stockholm); Francesco Lo Coco (Rome); Henk M. Lokhorst (Utrecht); John Mascarenhas (New York); Maria-Victoria Mateos (Salamanca); Giampaolo Merlini (Pavia); Anna Rita Migliaccio (New York); Mohamad Mohty (Nantes); Martina Muckenthaler (Heidelberg); Ann Mullally (Boston); Stephen Mulligan (Sydney); German Ott (Stuttgart); Jakob Passweg (Basel); Melanie Percy (Ireland); Rob Pieters (Utrecht); Stefano Pileri (Milan); Miguel Piris (Madrid); Andreas Reiter (Mannheim); Jose-Maria Ribera (Barcelona); Stefano Rivella (New York); Francesco Rodeghiero (Vicenza); Richard Rosenquist (Uppsala); Simon Rule (Plymouth); Claudia Scholl (Heidelberg); Martin Schrappe (Kiel); Radek C. Skoda (Basel); Gérard Socié (Paris); Kostas Stamatopoulos (Thessaloniki); David P. Steensma (Rochester); Martin H. Steinberg (Boston); Ali Taher (Beirut); Evangelos Terpos (Athens); Takanori Teshima (Sapporo); Pieter Van Vlierberghe (Gent); Alessandro M. Vannucchi (Firenze); George Vassiliou (Cambridge); Edo Vellenga (Groningen); Umberto Vitolo (Torino); Guenter Weiss (Innsbruck).
Editorial Office Simona Giri (Production & Marketing Manager), Lorella Ripari (Peer Review Manager), Paola Cariati (Senior Graphic Designer), Igor Ebuli Poletti (Senior Graphic Designer), Marta Fossati (Peer Review), Diana Serena Ravera (Peer Review)
Affiliated Scientific Societies SIE (Italian Society of Hematology, www.siematologia.it) SIES (Italian Society of Experimental Hematology, www.siesonline.it)
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
Information for readers, authors and subscribers Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org). Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org). Manuscripts should be submitted online at http://www.haematologica.org/. Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail online (www.haematologica.org/content/policies). Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties. Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google Play) is free. For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org). Rates of the International edition for the year 2018 are as following: Print edition
Institutional Euro 600
Personal Euro 150
Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org). Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature. Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955. Printing: Press Up, zona Via Cassia Km 36, 300 Zona Ind.le Settevene - 01036 Nepi (VT)
haematologica calendar of events
Journal of the European Hematology Association Published by the Ferrata Storti Foundation ESH Clinical Updates on CLL and Indolent Lymphoma European School of Haematology (ESH) Chairs: C Buske, C Wu, PL Zinzani March 2-4, 2018 Paris, France
1st European Myeloma Network Meeting Società Italiana di Ematologia (SIE) Chair: M Boccadoro April 19-21, 2018 Torino, Italy
ESH 4th International Conference on Hematologic Malignancies at Older Age: Biology and Therapy European School of Haematology (ESH) Chairs: B Löwenberg, B Ebert, P Fenaux, M Hallek March 9-11, 2018 Mandelieu-la Napoule, France
EHA-TSH Hematology Tutorial on Acute Leukemias April 28-29, 2018 Istanbul, Turkey
44th Annual Meeting of the European Society for Blood and Marrow Transplantation European Society for Blood and Marrow Transplantation (EBMT) Chairs: M Abecasis, E Carreras, J Guimarães, E Oliveira March 19-21, 2018 Lisbon, Portugal 5a REUNIÓN ANUAL DEL GRUPO PETHEMA LEUCEMIA MIELOBLÁSTICA AGUDA Pethema Foundation Chairs: MA Sanz, P Montesinos March 23, 2018 Madrid, Spain 8th EMIRATES HAEMATOLOGY CONFERENCE & 2nd PAN ARAB THALASSEMIA & HAEMOGLOBINOPATHIES CONFERENCE Emirates Society of Haematology (ESH) Chairs: AS Al Olama, KM Belhoul, RM Seliem, I Mirza April 12-14, 2018 Abu Dhabi, United Arab Emirates EHA-SWG Scientific Meeting on New Molecular Insights and Innovative Management Approaches for Acute Lymphoblastic Leukemia Chair: N Gökbuget April 12-14, 2018 Barcelona, Spain
The 4th World Congress on Controversies in Multiple Myeloma Chairs: M Mohty, A Nagler, T Facon May 3-5, 2018 Paris, France EHA Hematology Tutorial on Thalassemia May 10-11, 2018 Shiraz, Iran 23rd Congress of EHA June 14-17, 2018 Stockholm, Sweden EHA-SAH Hematology Tutorial on lymphoid Malignancies and Plasma Cell Dyscrasias September 14-15, 2018 Buenos Aires, Argentina EHA-SWG Scientific Meeting on Aging and Hematology Chair: D Bron October 12-14, 2018 Location TBC
Calendar of Events updated on February 19, 2018
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
Table of Contents Volume 103, Issue 3: March 2018 Cover Figure
Bone marrow smear showing increased erythropoietic activity, large cell size, nucleus-cytoplasmic asynchrony, and nuclear chromatin more open than expected in a patient with vitamin B12 deficiency. Courtesy of Prof. Rosangela Invernizzi.
Editorials 375
Can we forecast induction failure in acute myeloid leukemia? Felicitas Thol
377
Bortezomib for prevention of acute graft-versus-host disease: a conclusion reached Paul J. Martin
379
Use of desmopressin in the treatment of hemophilia A: towards a golden jubilee Pier Mannuccio Mannucci
Review Article 382
Extracellular vesicles in the hematopoietic microenvironment John T. Butler et al.
Articles Hematopoiesis
395
Glucocorticoids induce differentiation of monocytes towards macrophages that share functional and phenotypical aspects with erythroblastic island macrophages Esther Heideveld et al.
Red Cell Biology & Its Disorders
406
miR-144/451 represses the LKB1/AMPK/mTOR pathway to promote red cell precursor survival during recovery from acute anemia Xiao Fang et al.
Bone Marrow Failure
417
Hypomorphic FANCA mutations correlate with mild mitochondrial and clinical phenotype in Fanconi anemia Roberta Bottega et al.
Myelodysplastic Syndromes
427
Constitutional SAMD9L mutations cause familial myelodysplastic syndrome and transient monosomy 7 Victor B. Pastor et al.
Myeloproliferative Neoplasms
438
Benefits and pitfalls of pegylated interferon-Îą2a therapy in patients with myeloproliferative neoplasm-associated myelofibrosis: a French Intergroup of Myeloproliferative neoplasms (FIM) study Jean-Christophe Ianotto et al.
Chronic Myeloid Leukemia
447
CD36 defines primitive chronic myeloid leukemia cells less responsive to imatinib but vulnerable to antibody-based therapeutic targeting Niklas Landberg et al.
Acute Myeloid Leukemia
456
A 29-gene and cytogenetic score for the prediction of resistance to induction treatment in acute myeloid leukemia Tobias Herold et al.
Haematologica 2018; vol. 103 no. 3 - March 2018 http://www.haematologica.org/
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation Non-Hodgkin Lymphoma
466
Chemotherapy-induced differential cell cycle arrest in B-cell lymphomas affects their sensitivity to Wee1 inhibition Xiaoguang Wang et al.
477
Blood cytokine concentrations in pediatric patients with anaplastic lymphoma kinase-positive anaplastic large cell lymphoma Fabian Knรถrr et al.
486
Clinicopathological characteristics of T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies: a single-center case series of 25 patients and a review of the literature Marieke L. Nijland et al.
Chronic Lymphocytic Leukemia
497
In vitro and in vivo evidence for uncoupling of B-cell receptor internalization and signaling in chronic lymphocytic leukemia Eve M. Coulter et al.
Plasma Cell Disorders
506
Outcome and survival of myeloma patients diagnosed 2008-2015. Real-world data on 4904 patients from the Swedish Myeloma Registry Cecilie Hveding Blimark et al.
Plasma Cell Disorders
514
Melphalan 140 mg/m2 or 200 mg/m2 for autologous transplantation in myeloma: results from the Collaboration to Collect Autologous Transplant Outcomes in Lymphoma and Myeloma (CALM) study. A report by the EBMT Chronic Malignancies Working Party Holger W. Auner et al.
Stem Cell Transplantation
522
Bortezomib-based immunosuppression after reduced-intensity conditioning hematopoietic stem cell transplantation: randomized phase II results John Koreth et al.
Cell Therapy & Immunotherapy
531
In vivo IL-12/IL-23p40 neutralization blocks Th1/Th17 response after allogeneic hematopoietic cell transplantation Joseph Pidala et al.
Platelet Biology & Its Disorders
540
Variable impairment of platelet functions in patients with severe, genetically linked immune deficiencies Magdolna Nagy et al.
Coagulation & Its Disorders
550
Desmopressin in moderate hemophilia A patients: a treatment worth considering Janneke I. Loomans et al.
Letters to the Editor Letters are available online only at www.haematologica.org/content/103/3.toc
e94
PIEZO1-R1864H rare variant accounts for a genetic phenotype-modifier role in dehydrated hereditary stomatocytosis Immacolata Andolfo et al. http://www.haematologica.org/content/103/3/e94
e98
Macrocytosis and dysplastic anemia is associated with the cyclin-dependent kinase 4/6 inhibitor palbociclib in metastatic breast cancer Jesus Anampa et al. http://www.haematologica.org/content/103/3/e98
Haematologica 2018; vol. 103 no. 3 - March 2018 http://www.haematologica.org/
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation e103
Hematopoietic stem cell transplantation for patients with paroxysmal nocturnal hemoglobinuria previously treated with eculizumab: a retrospective study of 21 patients from SFGM-TC centers. Nicolas Vallet et al. http://www.haematologica.org/content/103/3/e103
e106
Unsatisfactory efficacy in randomized study of reduced-dose CPX-351 for medically less fit adults with newly diagnosed acute myeloid leukemia or other high-grade myeloid neoplasm Roland B. Walter et al. http://www.haematologica.org/content/103/3/e106
e110
High throughput sequencing in acute lymphoblastic leukemia reveals clonal architecture of central nervous system and bone marrow compartments Jack Bartram et al. http://www.haematologica.org/content/103/3/e110
e115
Loss of 5-hydroxymethylcytosine is a frequent event in peripheral T-cell lymphomas François Lemonnier et al. http://www.haematologica.org/content/103/3/e115
e119
Ibrutinib does not affect ristocetin-induced platelet aggregation evaluated by light transmission aggregometry in chronic lymphocytic leukemia patients Maria Adele Alberelli et al. http://www.haematologica.org/content/103/3/e119
e123
Risk of progression of monoclonal gammopathy of undetermined significance into lymphoplasmacytic malignancies: determining demographic differences in the USA Ronald S. Go et al. http://www.haematologica.org/content/103/3/e123
e126
Expression of programmed death-1 on lymphocytes in myeloma patients is lowered during lenalidomide maintenance Sophia Danhof et al. http://www.haematologica.org/content/103/3/e126
Comment Comment are available online only at www.haematologica.org/content/103/3.toc
e130
Comment on “MEK inhibition with trametinib and tyrosine kinase inhibition with imatinib in multifocal histiocytic sarcoma” Sophie Voruz et al. http://www.haematologica.org/content/103/3/e126
Case Reports Case Reports are available online only at www.haematologica.org/content/103/3.toc
e131
Partial tandem duplication of KMT2A (MLL) may predict a subset of myelodysplastic syndrome with unique characteristics and poor outcome Sarah M. Choi et al. http://www.haematologica.org/content/103/3/e131
e135
Venetoclax induced a complete response in a patient with immunoglobulin light chain amyloidosis plateaued on cyclophosphamide, bortezomib and dexamethasone Nelson Leung et al. http://www.haematologica.org/content/103/3/e135
e138
Acquired thrombotic thrombocytopenic purpura in a child: rituximab to prevent relapse. A pediatric report and literature review Sabrina Mariani et al. http://www.haematologica.org/content/103/3/e138
Haematologica 2018; vol. 103 no. 3 - March 2018 http://www.haematologica.org/
EDITORIALS Can we forecast induction failure in acute myeloid leukemia? Felicitas Thol Department of Hematology, Hemostasis, Oncology, and Stem Cell Transplantation, Hannover Medical School, Hannover, Germany E-mail: thol.felicitas@mh-hannover.de doi:10.3324/haematol.2018.187575
S
tandard induction therapy for fit patients with acute myeloid leukemia (AML) consists of a combination therapy with anthracycline and cytarabine. This classical regimen, typically called “7+3”, has not changed for several decades.1 While many patients achieve a complete remission (CR) with standard induction therapy, approximately 10-40% of patients fail to respond to induction treatment.2,3 These patients are classified as having primary refractory disease (RD) or treatment failure, defined as a failure to achieve CR or incomplete hematologic recovery (Cri) after two courses of induction treatment.4 Unfortunately, treatment of patients with RD is extremely challenging, as even with salvage therapy followed by allogeneic stem cell transplantation, patient outcomes remain poor.3 It is still difficult for hematologists to reliably predict RD in newly diagnosed AML patients prior to initiation of therapy. At time of diagnosis, we typically risk stratify our patients based on their cytogenetic and molecular profile. A very helpful classification was introduced by the European Leukemia Net (ELN) in 2010,5 (revised in 20174) and this currently includes three prognostic groups integrating cytogenetics as well as the mutational status of FLT3-ITD (including mutational load), NPM1, ASXL1, TP53, RUNX1, CEBPA (biallelic mutants). However, this risk stratification is geared towards the estimation of overall survival (OS) and eventfree survival (EFS), and not primarily towards forecasting RD.4 Although there is a strong correlation between treatment failure and OS, they still present different outcome measures.4,6 Several groups have attempted to develop specific scores to predict induction failure in AML. A reliable score primarily focusing on the likelihood of treatment failure rather than OS could improve patient care and treatment in many ways. If we could reliably predict that a patient would not respond to “7+3” treatment prior to induction therapy, we would be compelled to search for alternatives at the time of diagnosis, potentially sparing the patient from the toxicity of treatments that prove to be ineffective. As several new agents are being studied front line (e.g. FLT3 and IDH1/2 inhibitors with intensive chemotherapy, BCL2-inhibitors in combination with low-dose cytarabine or azacitidine, etc.) alternatives for “7+3” might soon become a reality. In addition, a reliable RD score could allow us to identify those patients who require an urgent donor search at the time of diagnosis.7-10 Furthermore, an RD score could become an important consideration when designing clinical trials that specifically target this high-risk patient group. In this issue of Haematologica, Herold et al. introduce a 29gene and cytogenetic score that can help to predict resistance to induction chemotherapy in adult AML patients.11 Importantly, this score was developed on the basis of various categories of prognostic markers, considering clinical characteristics, laboratory variables, cytogenetics, mutational status of 68 genes that are frequently mutated in AML, and the haematologica | 2018; 103(3)
expression profile of 29 genes known to be prognostic for AML. Their score estimates the likelihood of primary RD based on large independent clinical training sets. The first cohort (training set 1) included 407 patients of the AML Cooperative Group (AMLCG trials between 1999-2005), the second cohort (training set 2) consisted of 462 AML patients treated in the Haemato-Oncology Foundation for Adults in the Netherlands (HOVON) trials and the validation cohort was based on 210 AMLCG-2008 trial patients with the addition of 40 patients with RD from the AMLCG 1999 trial. The implementation of a large validation cohort is critical for assessing the reliability of any score, especially for clinical practice. The score was calculated as a weighted linear sum of the individual predictors. Interestingly, the final predictor by Herold et al. (predictive score 29 MRC or PS29MRC) included expression levels of 29 genes and the UK Medical Research Council (MRC) cytogenetic risk classification, while other parameters such as gene mutations were tested but were excluded from the final score.12 Importantly, this predictive classifier proved to be significant for RD, both as a continuous variable as well as a dichotomous variable that divides patients into high and low risk. In the multivariate analysis, only PS29MRC, age and TP53 mutations remained independently significant for RD prediction. While the predictor was primarily designed to be associated with RD on day 16 after induction chemotherapy, the score also proved to be strongly associated with survival. When examining different groups of the current ELN 2017 classification, the predictive power of the score was shown in the intermediate and the unfavorable ELN groups, while it could not be shown in the favorable genetic group (likely related to low RD rate in patients with favorable cytogenetics). The validation cohort nicely reproduced the data of the training cohort. All these aspects are suggestive of a very reliable predictive score. The area under receiver-operating characteristic curve (AUC) can be used as a measure for the predictive ability of a score, with an AUC of 0.7-0.8 classified as fair and less than we would desire for primary treatment decisions.13,14 The classifier by Herold et al. reached an AUC of 0.76 in the validation set. In contrast, Walter et al. developed a model for resistance prediction in AML based on the analysis of 4601 patients treated within European and US AML trials.13 They found that age, performance status, white blood cell count, secondary disease, cytogenetic risk and NPM1/FLT3-ITD mutational status were strongly associated independently with primary resistance. Unlike Herold et al., they did not include a complex mutational and gene expression profile in their analysis (Table 1). However, with their model, they achieved a similar AUC (0.78) to that of Herold et al. Krug et al. also developed a model based on a cohort of 1406 patients aged over 60 years diagnosed with AML but otherwise medically fit, and who underwent treatment with two intense induction chemotherapy cycles within the 375
Editorials
Table 1. Schematic overview of recent studies developing a model for response prediction to induction chemotherapy in intensively treated acute myeloid leukemia (AML) patients.
Publication (website for score)
Patient population
AUC
Variables considered
Variables in the final model
1079 adult patients (including 210 patients in VC)
0.76 (VC)
Cytogenetics risk according to MRC, expression data of 29 genes
Krug et al.15 CR+ ED (http://www.aml-score.org/)
1406 patients (TC) + 801 patients (VC) (only ≥ 60 years)
0.68 (VC)
Walter et al.14
4601 adult patients
0.78
Clinical characteristics, cytogenetics, laboratory variables, mutational status of 68 frequently mutated genes in AML, expression profile of 29 genes Body temperature, WBC, BM blasts, PB blasts, PB neutrophils, age, disease type, hemoglobin, platelet count, serum protein, ALT, bilirubin, BMI, extramedullary disease, fibrinogen, LDH, cytogenetics Age, PS,sex, WBC, platelet count, BM blast percentage, disease type, cytogenetic risk, FLT3-ITD and NPM1 mutation status Clinical data, cytogenetics, mutational data of 111 frequently mutated genes
Herold et al.11
Gerstung et al.16
Prediction for RD
RD
Not 1540 AML primarily RD (http://cancer.sanger.ac.uk/aml-multistage/)
N.A.
Body temperature, age, disease type, hemoglobin, platelet count, fibrinogen, LDH and cytogenetics
Age, PS, WBC, disease type, cytogenetic risk, FLT3-ITD/NPM1 mutation status
Age, sex, PS, WBC, platelet count, PB blasts, BM blasts, splenomegaly, disease type, hemoglobin, cytogenetics, mutational status of 58 genes
AUC: area under receiver-operating characteristic curve; CR: complete remission; RD: residual disease; ED: early death; VC: validation cohort; TC: training cohort; WBC: white blood cell count; BM: bone marrow; PB: peripheral blood; disease type: de novo leukemia versus leukemia secondary to cytotoxic treatment or an antecedent hematologic disease; ALT: alanine aminotransferase; BMI: Body Mass Index; LDH: serum concentration of lactate dehydrogenase; MRC: UK Medical Research Council; PS: Performance Status; N.A.: not applicable.
AML-CG.15 The validation cohort consisted of an independent cohort of 801 patients aged over 60 years. Their score was based on body temperature, age, secondary disease, hemoglobin, platelet count, fibrinogen, serum concentration of lactate dehydrogenase and cytogenetics. Instead of RD, the achievement of CR and early death were the primary outcome parameters of this score (Table 1). Using CR prediction, the model of Krug et al. had an AUC of 0.68 in the validation set.15 Gerstung et al. have also developed a prognostic algorithm based on a knowledge bank of 1540 AML patients whose cytogenetic, molecular profile, and clinical data were analyzed in detail.16,17 Here, a number of outcome parameters can be obtained (including death without remission, death without and after relapse, alive after relapse, alive in first CR and alive without CR), and RD can be indirectly calculated (Table 1). Thus, prediction of RD remains complex, and these scoring systems have yet to find their way into routine clinical practice. The questions of when and how we employ them for everyday clinical evaluation and treatment decisions remain. Here, feasibility and predictability must be considered. It will not be feasible to use a score requiring far more laboratory evaluation (e.g. microarray data, etc.) than is routinely performed. For example, gene expression analysis is not routinely performed in clinical practice and the time required might become relevant for patients with a high leukemic burden in need of urgent therapy. Furthermore, unlike sequencing, gene expression analysis is not covered by the healthcare systems of many countries. However, with the advances being made in technologies, such evaluation could quickly become more feasible. Just as important as feasibility is the level of predictability. We can only justify primarily basing our treat376
ment decisions on scoring systems with a sufficiently high predictability. That none of the proposed scoring systems reach an AUC close to 0.9, even when including all parameters currently known to be prognostic, underscores the challenges of reliably predicting patient outcome at the time of diagnosis. This is highlighted by Herold et al., who used all prognostic parameters currently considered relevant, studied these parameters extensively in the context of RD prediction, and thus, rightfully described an “obstacle” to achieving a higher AUC that is difficult to overcome. Herold et al. describe an innovative approach of how to tackle the pressing question of RD prediction. Independently of its clinical use, it can potentially help us to better understand the biology of primary refractory disease. It is still unknown why some patients with a molecularly more favorable risk profile still fail induction chemotherapy. The gene expression data that predict primary refractory disease might also lead the way to identifying novel targets for AML therapy. Even if the predictive classifier of Herold et al. may not find its way into clinical practice just yet, it carries the potential of becoming a tool for designing clinical trials and developing novel treatment strategies.
References 1. Dohner H, Weisdorf DJ, Bloomfield CD. Acute Myeloid Leukemia. N Engl J Med. 2015;373(12):1136-1152. 2. Burnett A. Treatment of acute myeloid leukemia: are we making progress? Hematology Am Soc Hematol Educ Program. 2012;2012:16. 3. Thol F, Schlenk RF, Heuser M, Ganser A. How I treat refractory and early relapsed acute myeloid leukemia. Blood. 2015;126(3):319-327. 4. Dohner H, Estey E, Grimwade D, et al. Diagnosis and management
haematologica | 2018; 103(3)
Editorials
5.
6.
7. 8.
9.
10.
of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-447. Dohner H, Estey EH, 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. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol. 2003;21(24):4642-4649. Stone RM, Mandrekar SJ, Sanford BL, et al. Midostaurin plus Chemotherapy for Acute Myeloid Leukemia with a FLT3 Mutation. N Engl J Med. 2017;377(5):454-464. Stein EM, DiNardo C, Altman JK, et al. Safety and Efficacy of AG221, a Potent Inhibitor of Mutant IDH2 That Promotes Differentiation of Myeloid Cells in Patients with Advanced Hematologic Malignancies: Results of a Phase 1/2 Trial. Blood. 2015;126:323. DiNardo CD, Pratz KW, Letai A, et al. Safety and preliminary efficacy of venetoclax with decitabine or azacitidine in elderly patients with previously untreated acute myeloid leukaemia: a non-randomised, open-label, phase 1b study. Lancet Oncol. 12 Jan 2018 [Epub ahead of print]. Wei AH, Tiong IS. Midostaurin, enasidenib, CPX-351, gemtuzumab ozogamicin, and venetoclax bring new hope to AML. Blood.
2017;130(23):2469-2474. 11. Herold T, Jurinovic V, Batcha AMN, et al. A 29-gene and cytogenetic score for the prediction of resistance to induction treatment in acute myeloid leukemia. Haematologica. 2017;103(3):000-000. 12. Grimwade D, Walker H, Oliver F, et al. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children's Leukaemia Working Parties. Blood. 1998;92(7):23222333. 13. Walter RB, Othus M, Burnett AK, et al. Resistance prediction in AML: analysis of 4601 patients from MRC/NCRI, HOVON/SAKK, SWOG and MD Anderson Cancer Center. Leukemia. 2015;29(2):312-320. 14. Walter RB, Othus M, Paietta EM, et al. Effect of genetic profiling on prediction of therapeutic resistance and survival in adult acute myeloid leukemia. Leukemia. 2015;29(10):2104-2107. 15. Krug U, Rollig C, Koschmieder A, et al. Complete remission and early death after intensive chemotherapy in patients aged 60 years or older with acute myeloid leukaemia: a web-based application for prediction of outcomes. Lancet. 2010;376(9757):2000-2008. 16. Gerstung M, Papaemmanuil E, Martincorena I, et al. Precision oncology for acute myeloid leukemia using a knowledge bank approach. Nat Genet. 2017;49(3):332-340. 17. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med. 2016;374(23):2209-2221.
Bortezomib for prevention of acute graft-versus-host disease: a conclusion reached Paul J. Martin Fred Hutchinson Cancer Research Center, Seattle and University of Washington, Seattle, WA, USA E-mail: pmartin@fredhutch.org doi:10.3324/haematol.2018.188052
P
revention of acute graft-versus-host disease (GvHD) after allogeneic hematopoietic cell transplantation (HCT) has long posed a major challenge for the field. In this issue, Koreth et al.1 report results of a randomized three-arm phase II trial testing two different immunosuppressive regimens using bortezomib to prevent acute GvHD after allogenic hematopoietic cell transplantation. Contrary to expectations, the results did not show major improvement in the experimental groups as compared to the control group. The impetus to explore the use of bortezomib to prevent GvHD came from its mechanism of action to prevent signaling through nuclear factor (NF)κB in activated T cells. In resting T cells, the inhibitor (I)-κB binds to NFκB as a complex that is sequestered in the cytoplasm.2 In activated T cells, ubiquitin moieties attach to I-κB, which is delivered to proteasomes. The NFκB molecules released from I-κB translocate to the nucleus where they activate the transcription of genes involved in immune responses. Among other possible mechanisms of action, bortezomib inhibits proteasome activity, allowing I-κB to prevent NFκB-mediated activation of T cells. Experimental results showed that administration of bortezomib early after allogeneic HCT could prevent acute GvHD in mice.1 These results led to a phase I/II study demonstrating that bortezomib can be combined with tacrolimus and methotrexate in a tolerable posttransplant immunosuppressive regimen after HCT with human leukocyte antigen (HLA)-mismatched donors. haematologica | 2018; 103(3)
Results of the completed study were published in 20123 and are summarized in Figure 1. The interpretation of the 2012 study was initially informed by historical experience showing a 46% incidence of grade II – IV GvHD in patients with HLA-mismatched unrelated donors.4 The 22% incidence of grade II – IV GvHD in the 2012 study was indeed encouraging when compared against this benchmark, although the comparability of demographic and treatment characteristics of patients in the phase I/II study and the historical group5 was not well documented. In the current study of HLA-matched unrelated HCT, the benchmark for grade II – IV GvHD was set at 40%.1 The observed 33% incidence of grade II – IV GvHD in patients treated with tacrolimus and methotrexate (Arm A) was somewhat lower than this benchmark, while the 29% incidence in patients treated with bortezomib added to tacrolimus and methotrexate (Arm B) was somewhat higher than the 22% incidence observed in HLA-mismatched recipients in the phase I/II study (Figure 1). As a result, the current study did not demonstrate a statistically significant improvement following the addition of bortezomib to tacrolimus and methotrexate in patients with HLA-matched unrelated donors. Results of the current study with HLA-matched unrelated recipients are similar to those observed in the BMT CTN 1203 PROGRESS study, which enrolled a mixed cohort of HLA-matched related and unrelated recipients and a small proportion of HLA-mismatched unrelated recipients. In this study, the day 180 cumulative inci377
Editorials
Figure 1. Summary of key results from Koreth et al.3 (JCO 2012) and from the current study.1 Results are shown for Arm A (tacrolimus and methotrexate), Arm B (bortezomib added to tacrolimus and methotrexate) and Arm C (bortezomib and tacrolimus with sirolimus substituted for methotrexate), as compared to tacrolimus and methotrexate used in the JCO 2012 study.3 Diamonds indicate the incidence frequencies of grades II – IV GvHD and III – IV GvHD, and bars indicate the 95% confidence intervals. All patients in the JCO 2012 study had HLA-mismatched unrelated donors. In Arms A, B and C of the current study, all patients had HLA-A, B, C, DRB1–matched unrelated donors. Data in the figure exclude GvHD that occurred after relapse of the pretransplant disease. Results for grade III – IV GvHD are shown on a log scale in order to visualize the lower limits of the 95% confidence intervals more clearly. GvHD: graft-versus-host disease; CI: confidence interval.
dence frequencies of grade II – IV GvHD were 31% in control patients treated with tacrolimus and methotrexate, and 26% in patients treated with bortezomib added to tacrolimus and methotrexate.6 Arm C of the current study1 tested a regimen of bortezomib and tacrolimus with the substitution of sirolimus for methotrexate when compared to Arm B. This substitution was motivated by a desire to enhance the survival and function of T-regulatory cells, thereby possibly facilitating the development of tolerance. The 15% incidence of grade II – IV GvHD in this group appears to be encouraging when compared against the 33% incidence in Arm A. More importantly, however, the results for grade III – IV GvHD were somewhat higher in Arms B and C than in Arm A. Here, the 9% incidence in Arm B and the 11% incidence in Arm C appear to be consistent with the ~10% incidence of grade III – IV GvHD observed with other current approaches, while the 2% incidence in Arm A appears to be much better than expected. In the BMT CTN 1203 PROGRESS study, for example, the cumulative incidence frequencies of grades III – IV GvHD were 13% in the controls treated with tacrolimus and methotrexate, and 8% in patients treated with bortezomib added to tacrolimus and methotrexate.6 The current results raise the question of whether data from the phase I/II trial truly justified the effort to investigate bortezomib-based immunosuppression in the subsequent three-arm trial and the BMT CTN 1203 PROGRESS trial. Enthusiasm for these studies was based primarily on a comparison of outcomes between patients enrolled in the phase I/II study and a group of 176 patients who received tacrolimus and sirolimus for immunosuppression after HLA-matched unrelated transplantation.3 In this comparison, the 22% incidence of grade II – IV GvHD in the phase I/II study did not differ statistically from the 11% incidence in the 176 patients with HLA-matched unrelated donors. The absence of statistical significance therein, however, cannot be interpreted as indicating that the incidence rates are similar. P-val378
ues >0.05 indicate only that the results were not demonstrably different between the two groups, given their respective sizes. In this case, the comparison actually showed a two-fold difference in the incidence of grade II – IV acute GvHD. A further question is whether grade II – IV GvHD actually represents the most appropriate primary endpoint in studies evaluating immunosuppressive regimens after allogeneic HCT. Results of a recent CIBMTR study showed that grade II acute GvHD has no statistically significant association with the risk of treatment failure defined as death or relapse, whereas grades III and IV acute GvHD were associated with increased risks of treatment failure.7 These observations suggest that clinical trials should focus on preventing grade III – IV acute GvHD, as opposed to grade II – IV acute GvHD. The prior report by Koreth et al.3 did not compare the incidence rates of grade III – IV acute GvHD between patients enrolled in the phase I/II study and the 176 patients who received tacrolimus and sirolimus for immunosuppression after HLA-matched unrelated transplantation. To some extent, we have become victims of our own success in our efforts to prevent grade III – IV acute GvHD. A 1:1 randomized trial would require approximately 200 patients per arm to test the difference between a 10% incidence and a 3% incidence at 80% power and a 0.05 two-side type-1 error. A larger effect size would require fewer patients. As an alternative, survival to 1 year without prior grade III – IV acute GvHD, chronic GvHD requiring systemic treatment, or relapse has become a very popular compound endpoint for acute GvHD prevention studies.7 Current typical estimates for this GvHD-free/relapse-free survival (GRFS) endpoint are in the 35% range, which leaves considerable room for improvement.7 However, grade III – IV acute GvHD makes the smallest contribution among the four components of this compound endpoint. Moreover, the risks of non-relapse mortality and relapse are heavily influenced by factors that are not associated with the risk of grade III haematologica | 2018; 103(3)
Editorials
– IV acute GvHD. For these reasons, the GRFS endpoint represents an unreliable surrogate for the ability of an intervention to prevent grade III – IV acute GvHD. The current report by Koreth et al.1 represents a wellcontrolled attempt to evaluate the merits of bortezomib for immunosuppression after HCT with reduced intensity conditioning regimens. The report is newsworthy, because the negative results do not support expectations that bortezomib-based regimens might have a major effect on the risk of grade III – IV acute GvHD, even with the substitution of sirolimus for methotrexate in Arm C. With the benefit of hindsight, one could question whether the expectations based on results of the phase I/II study were realistic. The experience from testing bortezomib yields important lessons for planning future trials. First, grade III – IV acute GvHD should be defined as the primary endpoint in trials designed to test an intervention intended to prevent acute GvHD. Second, the benchmark incidence of grade III – IV acute GvHD should be set at 10 – 15%, depending on the relationship and HLA-matching between the donor and recipient. Third, early phase trials should be designed to test whether an intervention can reduce the incidence of grade III – IV acute GvHD to 2% or less, as it will not be feasible to determine whether any smaller effect size holds true in a phase III trial. Early phase and later phase trials should have rules that discontinue enrollment when initial results indicate that the intervention is not likely to reach this benchmark of success. Finally, although the GRFS endpoint should not be used as the primary endpoint in trials of interventions
intended to prevent grade III – IV acute GvHD, it remains important to demonstrate that successful prevention of grade III – IV acute GvHD does not come at the expense of increasing the risk of non-relapse mortality or the risk of recurrent or progressive malignancy.
References 1. Koreth J, Kim HT, Lange PB, et al. Bortezomib-based immunosuppression after reduced-intensity conditioning hematopoietic stem cell transplantation: randomized phase II results. Haematologica. 2018;103(3):522-530. 2. Mohty M, Brissot E, Savani B and Gaugler B. Effects of bortezomib on the immune system: a focus on immune regulation. Biol Blood Marrow Transplant. 2014;19(10):1416-1420. 3. Koreth J, Stevenson KE, Kim HT, et al. Bortezomib-based graft-versus-host disease prophylaxis in HLA-mismatched unrelated donor transplantation. J Clin Oncol. 2012;30(26):3202-3208. 4. Koreth J, Stevenson KE, Kim HT, et al. Bortezomib, tacrolimus, and methotrexate for prophylaxis of graft-versus-host disease after reduced-intensity conditioning allogeneic stem cell transplantation from HLA-mismatched donors. Blood. 2009;114(18):3956-3959. 5. Ho V, Kim H, Windawai S, et al. HLA mismatch and clinical outcome after unrelated donor (URD) non-myeloablative hematopoietic stem cell transplantation (NST). Biol Blood Marrow Transplant. 2005;11(suppl 1):S14. 6. Bolaños-Meade J, Reshef R, Fraser R, et al. Novel approaches for graft-versus-host disease (GvHD) prophylaxis: Primary results of Progress I multicenter trial of matched allogeneic hematopoietic cell transplantation (alloHCT) using reduced intensity conditioning (RIC) BMT CTN 1203. Biol Blood Marrow Transplant. 2018 Epub ahead of print]. 7. Pasquini MC, Logan B, Jones RJ, et al. Blood and marrow transplant clinical trials network report on development of novel endpoints and selection of promising approaches for graft-versus-host disease prevention trials. Biol Blood Marrow Transplant. 2018 [Epub ahead of print]..
Use of desmopressin in the treatment of hemophilia A: towards a golden jubilee Pier Mannuccio Mannucci Scientific Direction, IRCCS Ca’ Granda Maggiore Policlinico Hospital Foundation and University of Milan, Italy E-mail: piermannuccio.mannucci@unimi.it doi:10.3324/haematol.2018.187567
T
hroughout the 1970s, the availability and safety of coagulation factors, employed for replacement therapy in patients with hemophilia (PWH), were very far from what they are nowadays. Plasma-derived concentrates of factor VIII (FVIII) and IX started to be industrially manufactured, but they were generally available in limited amounts in most countries, and thus used only for the acute treatment of bleeding episodes (so called ‘on-demand regimen’), but not for the treatment of choice, i.e., the prevention of bleeds by means of regularly spaced infusions (prophylaxis regimen). Most importantly, these products, produced from plasma pooled from thousand of donors, transmitted the hepatitis viruses with extremely high frequency; the agent causing hepatitis B and, more often, the so called non-A, non-B virus, which was only identified in 1989 as the hepatitis C virus. These bloodborne viruses heralded the bleak era of infection with the human immunodeficiency virus (HIV), haematologica | 2018; 103(3)
that started to contaminate plasma-derived coagulation factor concentrates at the end of the 1970s, eventually leading to the first appearance in PWH of the acquired immunodeficiency syndrome (AIDS) in 1982, which caused such a high toll of deaths during the 1980s and the 1990s. With this background, it is not surprising that in the 1970s and earlier a multitude of research efforts were directed towards the development of pharmacological alternatives to blood products. These agents were felt particularly necessary in patients with mild hemophilia A who, having measurable plasma levels of FVIII of 6% of normal or more, bleed much less frequently than those with severe disease and unmeasurable FVIII levels. In general they have little risk of mortality and morbidity, and their main clinical problem is excessive bleeding after trauma or surgery whereas, at variance with severe hemophilia, spontaneous bleeding episodes and joint 379
Editorials
bleeding are very rare. Early studies on pharmacological agents concentrated on epinephrine (adrenaline), which, administered to normal volunteers and patients with mild FVIII deficiency, was followed by a short-term plasma rise in the coagulant activity of FVIII.1,2 These seminal pioneer observations stimulated the search for pharmacological compounds devoid of the cardiovascular side effects, which obviously made the therapeutic use of adrenaline in patients impossible. An important step forward was made more than 40 years ago in the early 1970s, when Cash et al.3 and Mannucci et al.4 independently demonstrated that 1deamino-8-D-arginine vasopressin (DDAVP, desmopressin), a synthetic peptide analogue of the antidiuretic hormone vasopressin,5 raised plasma FVIII to 2-3 times above baseline levels following its infusion into normal volunteers and patients with mild hemophilia A, but not in those with severe disease and unmeasurable plasma FVIII. It remained to be demonstrated that the rise of the autologous FVIII released in plasma from storage sites upon the effect of desmopressin was as hemostatically effective as the allogeneic FVIII replaced by means of the infusion of plasmatic FVIII concentrates. To answer this clinically crucial question, desmopressin was used to prevent surgical bleeding, cautiously at first for minor procedures such as dental extractions, and subsequently for major procedures carried out in 23 patients with mild hemophilia A or von Willebrand disease.6 The results obtained in this first clinical study were excellent, because the procedures were done without undue bleeding, and without the need for resorting to allogeneic replacement therapy with blood products containing FVIII.6 These results were subsequently confirmed by several independent studies7-9 and desmopressin, designated as an essential drug by the World Health Organization, consequently acquired an established role in the management of patients with mild hemophilia A. Pertaining to the genuine role exerted by desmopressin in avoiding bloodborne infections associated with the use of plasma-derived FVIII products, the early adoption of this drug in Italy, where it was fist clinically employed in the late 1970s, led to a much lower prevalence of HIV infections in patients with mild hemophilia A,10 compared to those with mild hemophilia B who could only use plasma-derived products because they were unresponsive to desmopressin (factor IX is not increased) and in American patients with mild hemophilia A who started using desmopressin much later, at a time when the transmission of HIV was halted following the advent of heattreated plasma-derived coagulation factor concentrates and recombinant products.10 With this background, what is the significance of the enclosed report of Loomans et al.?11 In the frame of a study carried out internationally in a retrospective large cohort of 169 patients with moderate forms of hemophilia A, they investigated whether or not this group of patients with low but measurable FVIII in plasma were responsive to the administration of desmopressin. Patients with moderate hemophilia are defined as having plasma FVIII levels between 1% and 5%, i.e., much lower than in mild hemophilia (6% or more); thus, they bleed more frequently, but less than those with severe hemo380
philia A. In the original clinical study of desmopressin, no patient with moderate hemophilia was included, as it was felt that the expected 2/3-fold rise of FVIII induced cautiously by this compound would not allow the attainment of plasma levels high enough to secure surgical hemostasis, and therefore prevent bleeding.6 Subsequently, some patients with moderate hemophilia were indeed treated with desmopressin,8,9,12,13 but the cases and results were too few to dissipate the fear that the drug could not be efficaciously employed in this category of moderately severe patients. The value of the findings of Loomans et al. reported in this issue of Haematologica11 is that they help to fill this gap of knowledge. In as many as 40% of their moderately affected patients, low baseline FVIII levels reached after desmopressin plasma values of at least 30%, and 15% of them attained levels as high as 50% or more. Importantly, Loomans et al. also showed that it is possible to predict the good responders to desmopressin prior to infusion, because the degree of their FVIII increase was relatively proportional to the baseline pre-desmopressin plasma levels of this moiety. The main limitations of the findings reported in the study of Loomans et al., based upon patients recruited in 25 hemophilia treatment centers from three continents (Europe, North America and Australia), are that the authors only evaluated the post-desmopressin changes of a surrogate biomarker of hemostasis, such as the plasma levels of FVIII coagulant activity, but did not investigate whether or not these changes corresponded to a beneficial clinical effect on hemostasis. Thus, these novel findings should prompt the pursuance of a prospective clinical study designed to demonstrate whether the plasma increase of this surrogate laboratory marker is paralleled by the efficacy of the drug in preventing or treating bleeding episodes in patients with moderate hemophilia, who nowadays are usually treated with recombinant FVIII products. Regarding the latter there is no longer any concern about the onset of bloodborne infections, but another adverse effect of replacement therapy is still looming large: the development of alloantibodies which inactivate FVIII coagulant activity and thus render this therapy ineffective.14 As the autologous FVIII released by desmopressin is not seen as foreign by the recipient’s immune system, the increase of plasma FVIII levels effected by this drug does not elicit the onset of inhibitors that, albeit less frequently than in severe hemophilia, are a complication of allogeneic factor replacement, particularly when some at-risk FVIII gene mutations are present in patients with moderate hemophilia.15 All things considered, I agree with the declaration of Loomans et al., as stated in the title of the article herein. At variance with that which was hitherto believed, and after more than 40 years of clinical experience with desmopressin, this form of endogenous replacement therapy is also worth considering in patients with moderate hemophilia A.
References 1. Ingram GI. Increase in antihaemophilic globulin activity following infusion of adrenaline. J Physiol. 1961;156:217-224.
haematologica | 2018; 103(3)
Editorials
2. Egeberg O. Changes in the activity of antihemophilic a factor (f. Viii) and in the bleeding time associated with muscular exercise and adrenalin infusion. Scand J Clin Lab Invest. 1963;15:539-549. 3. Cash JD, Gader AM, da CJ. Proceedings: The release of plasminogen activator and factor VIII to lysine vasopressin, arginine vasopressin, I-desamino-8-d-arginine vasopressin, angiotensin and oxytocin in man. Br J Haematol. 1974;27(2):363-364. 4. Mannucci PM, Aberg M, Nilsson IM, Robertson B. Mechanism of plasminogen activator and factor VIII increase after vasoactive drugs. Br J Haematol. 1975;30(1):81-93. 5. Richardson DW, Robinson AG. Desmopressin. Ann Intern Med. 1985;103(2):228-239. 6. Mannucci PM, Ruggeri ZM, Pareti FI, Capitanio A. 1-Deamino-8-darginine vasopressin: a new pharmacological approach to the management of haemophilia and von Willebrands' diseases. Lancet. 1977;1(8017):869-872. 7. Warrier AI, Lusher JM. DDAVP: a useful alternative to blood components in moderate hemophilia A and von Willebrand disease. J Pediatr. 1983;102(2):228-233. 8. Mariani G, Ciavarella N, Mazzucconi MG, et al. Evaluation of the effectiveness of DDAVP in surgery and bleeding episodes in hemophilia and von Willebrand's disease. A study of 43 pateints. ClinLab Haematol 1984;6:229
haematologica | 2018; 103(3)
9. de la Fuente B, Kasper CK, Rickles FR, Hoyer LW. Response of patients with mild and moderate hemophilia A and von Willebrand's disease to treatment with desmopressin. Ann Intern Med. 1985;103(1):6-14. 10. Mannucci PM, Ghirardini A. Desmopressin: twenty years after. Thromb Haemost. 1997;78:958. 11. Loomans JI, Kruip MJHA, Carcao M, J et al, for the RISE consortium. Desmopressin in moderate hemophilia A patients: a treatment worth considering. Haematologica 2018;103(3):550-557. 12. Revel-Vilk S, Blanchette VS, Sparling C, Stain AM, Carcao MD. DDAVP challenge tests in boys with mild/moderate haemophilia A. Br J Haematol. 2002;117(4):947-951. 13. Stoof SCM, Sanders YV, Cnossen MH, de Maat MPM, Leebeek FWG, Kruip MJHA. Desmopressin response in hemophilia A patients with FVIII: C < 0.10 IU mL-1. J Thromb Haemost 2014;12(1):110-112. 14. Oldenburg J. Optimal treatment strategies for hemophilia: achievements and limitations of current prophylactic regimens. Blood 2015; 125(13):2038-44. 15. Eckhardt CL, Van Velzen AS, Peters M, et al. for the INSIGHT study group. Factor VIII gene (F8) mutation and risk of inhibitor development in nonsevere hemophilia A. Blood. 2013;122(11):1954â&#x20AC;&#x201C;1962.
381
REVIEW ARTICLE Ferrata Storti Foundation
Extracellular vesicles in the hematopoietic microenvironment John T. Butler,1,2 Sherif Abdelhamed1 and Peter Kurre1,3
Department of Pediatrics, PapĂŠ Family Pediatric Research Institute, Pediatric Blood & Cancer Biology Program, Oregon Health & Science University; 2Department of Biomedical Engineering, Oregon Health & Science University and 3OHSU Knight Cancer Institute, Oregon Health & Science University, Portland, OR, USA 1
Haematologica 2018 Volume 103(3):382-394
ABSTRACT
S
Correspondence: kurrepe@ohsu.edu
Received: November 13, 2017. Accepted: January 18, 2018. Pre-published: February 8, 2018. doi:10.3324/haematol.2017.183335 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/382 Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
382
elf-renewal and differentiation are defining characteristics of hematopoietic stem and progenitor cells, and their balanced regulation is central to lifelong function of both blood and immune systems. In addition to cell-intrinsic programs, hematopoietic stem and progenitor cell fate decisions are subject to extrinsic cues from within the bone marrow microenvironment and systemically. Yet, many of the paracrine and endocrine mediators that shape hematopoietic function remain to be discovered. Extracellular vesicles serve as evolutionarily conserved, constitutive regulators of cell and tissue homeostasis, with several recent reports supporting a role for extracellular vesicles in the regulation of hematopoiesis. We review the physiological and pathophysiological effects that extracellular vesicles have on bone marrow compartmental function while highlighting progress in understanding vesicle biogenesis, cargo incorporation, differential uptake, and downstream effects of vesicle internalization. This review also touches on the role of extracellular vesicles in hematopoietic stem and progenitor cell fate regulation and recent advances in therapeutic and diagnostic applications of extracellular vesicles in hematologic disorders.
Introduction To fulfill its critical systemic functions in oxygen delivery, coagulation and immune defense, hematopoiesis is regulated via integration of cell-intrinsic programs with extrinsic cues from the surrounding bone marrow (BM) microenvironment.1,2 Recent studies from infectious diseases, cardiovascular, and cancer fields demonstrate the existence of systemic crosstalk with BM cells which adds to the complexity of compartmental signaling, especially during injury responses.1,3 Cytokines, chemokines and other growth factors act as important mediators in a reasonably well-understood system by which the extrinsic ligands act on cells expressing the cognate receptor (Figure 1A). These in turn transmit signals to a network of cellular signaling pathways regulating hematopoiesis, including Wnt, Notch, transforming growth factor beta (TGF-β), phosphatidylinositol-3 kinase, and the mammalian target of rapamycin.4-7 Signaling by extrinsic mediators through any one of these pathways triggers activation of quiescent long-lived hematopoietic stem cells (HSCs). More recent studies of the leukemic microenvironment have revealed that tumor-derived paracrine factors also act on mesenchymal stromal cells, osteoprogenitors and endothelial cells within the BM, indirectly suppressing hematopoietic stem and progenitor cells (HSPCs).1,3,8 Thus, dynamic compartmental interactions shape physiological and pathophysiological regulation of BM function. Extracellular vesicle (EV) biogenesis is a constitutive cellular process, broadly conserved across evolution, with a role in development, homeostatic organismal function and tissue regeneration.9-11 EVs of various shapes and sizes have been demonstrated in every biofluid tested to date, with substantial variation in their structure, content and function.12 Protein, lipid and RNA components contribute to cell-cell crosstalk at a short distance, in a paracrine or endocrine manner via the bloodhaematologica | 2018; 103(3)
EVs in the hematopoietic microenvironment
stream (Figure 1B).10,12 However, given their complex cargo and poorly understood selectivity for cellular uptake, many phenotypic outcomes are not easily explained by conventional models of cell-cell crosstalk. The consequences of simultaneously transferring an unknown number of non-randomly assembled proteins and RNA to another cell defy the clear predictions that apply to more conventional receptor-ligand signaling. However, while an understanding of the molecular basis for EV crosstalk is in its infancy, the key principles of how EVs shape tissue function are beginning to emerge.12 Several groups have recently demonstrated that EVs contribute to the compartmental regulation of hematopoiesis in the BM.13,14 In this review, we present current evidence for the role of EVs in both homeostatic and pathogenic hematopoietic niches with emphasis on regulatory mechanisms, experimental outcomes and the critical open questions in the field.
Extracellular vesicles EVs are membrane-enclosed structures of varying size (30-10,000 nm) released from cells to mediate both local and distant intercellular communication. Platelet-derived vesicles were first identified by electron microscopy over 50 years ago,15 yet the full spectrum of subtypes and activities of EVs have only become a major focus of interest in recent years. In the early 1980s, it was reported that sheep reticulocytes selectively release transferrin receptor within EVs during programmed enucleation of the maturing red cell and were generally considered to simply reflect the export of cellular waste.16 Recent studies of EVs in the BM have shown that these vesicles serve to regulate hematopoiesis, participate in immune cell activation, and act as mediators of hemostatic functions.11,17,18 Hematologic malignancies such as leukemia, multiple myeloma or viral infections can coopt EV trafficking
Figure 1. Schematic representation of biogenesis of extracellular vesicles and unique aspects of their trafficking. (A) The conventional model of cellular crosstalk involves receptor-ligand interactions between secreted chemokines, cytokines and growth factors and cellular surface receptors. (B) EV-mediated crosstalk occurs through the trafficking of vesicle-associated protein, lipid and RNA components to proximal cells or to distal organs via the bloodstream in a “paracrine” or “endocrine” manner, respectively. (C) Exosomes are formed from the maturation of early endosomes into Rab7-containing late endosomes leading to the generation of intraluminal vesicles via the action of tetraspanin and ESCRT proteins which sort the endosomal constituents into distinct multivesicular bodies. Through the action of Rab27 and VPS33b, multivesicular bodies evade lysosome degradation and fuse with the plasma membrane to release 30-125 nm exosomes. Cells also release 50-1000 nm microvesicles that form through calcium-mediated budding of the plasma membrane, and during programed cell death, large (>1000 nm) apoptotic bodies. ApB: apoptotic bodies ESCRT: endosomal-sorting complex required for transport; GF: growth factors; ILV: intraluminal vesicle; MV: microvesicle; MVB: multivesiclular bodies; mTOR: mammalian target of rapamycin; PI3K; phosphatidylinositol-3 kinase; TGF-β: transforming growth factor beta; TGN: trans-Golgi network; TSPAN: tetraspanin; VPS33B: vacuolar protein sorting-associated protein 33B.
haematologica | 2018; 103(3)
383
J.T. Butler et al. Table 1. Types of extracellular vesicles.
Type of extracellular vesicles
Size (nm)
Biogenesis
Constituents
30-150
Early endosomes mature into late endosomes then, through the action of ESCRT, multivesicular bodies are formed containing intraluminal vesicles that fuse with the plasma membrane for release
Microvesicle
50-1000
Direct budding and cleavage of plasma membrane mediated by calcium influx, and remodeling of the cortical cytoskeleton
Microparticle ectosome Large vesicle
>1000
Cleavage of large cytoplasmic extensions from cell body
>1000
Cytoplasmic fragmentation during programmed cell death
Lipid membrane Nucleic acids Proteins Tetraspanins ALIX TSG101 Lipid membrane Nucleic acids Proteins Tetraspanins Lipid membrane Nucleic acids Proteins Organelles Organized cytoskeleton Lipid membrane Nucleic acids Proteins Organelles Nuclear fragments Apoptotic markers
Selection Detection Reference
Pseudonym Exosome Nanovesicle Nanoparticle
Large oncosome Apoptotic body
UC DG AC SEC
NTA Cryo-EM TEM SEM
[12,22,23, 24, 25, 26]
UC AC SEC
NTA Cryo-EM TEM SEM FM FC
[12, 27, 28]
FM
[31, 32]
CF FT FACS
CF FT FACS FC
[29, 30]
AC: affinity chromatography; CF: centrifugation; Cryo-EM: cryo-electron microscopy; DG: density gradient; ESCRT: endosomal-sorting complex required for transport; FACS: fluorescence activated cell sorting; FC: flow cytometry; FM: fluorescence microscopy; FT: filtration; NTA: nanoparticle tracking analysis; SEC: size-exclusion chromatography; SEM: scanning electron microscopy; TEM: transmission electron microscopy; UC: ultracentrifugation
mechanisms, upend these homeostatic processes and use EVs to reinforce tumor growth, chemotherapeutic resistance, invasion, metastasis and relapse.19-21 EVs can be broadly classified into four subtypes (Table 1) based upon vesicle size and method of cellular release: exosomes (30-150 nm), microvesicles (50-1000 nm), large vesicles (>1000 nm) and apoptotic bodies (>1000 nm).22 It is technically challenging to separate vesicle types, and no standardized method exists to date. Techniques utilized for EV purification often rely on size or density.12 However, there is overlap between exosomes and microvesicles in composition and function, and neither size-exclusion chromatography nor ultracentrifugation in density gradients for separation will yield pure populations.22 Moreover, due to overlap between these vesicles in size and miRNA carrier function with plasma abundant chylomicrons and very low density lipoproteins, EV dimension should be considered an arbitrary surrogate metric, and a more biologically informed classification would likely enhance reproducibility in the field, advance their detection and inform treatment strategies.
Exosomes The biogenesis of exosomes, the smallest type of EV, begins with the inward cleavage of the plasma membrane to form an endosome containing selectively enclosed cytoplasmic components within the lumen. As illustrated in Figure 1C, early endosomes, characterized by the presence of Rab5 protein, undergo maturation into Rab7 containing late endosomes which generate multiple intraluminal vesicles through the action of tetraspanins and endosomal sorting complex required for transport (ESCRT) 384
proteins.25 Together these proteins function to facilitate further inward cleavage and sorting of endosomal constituents into discrete intraluminal vesicles. These multivesicular bodies, through RAB27- and VPS33b-dependent mechanisms, evade lysosome degradation and fuse with the plasma membrane to release intraluminal vesicles as exosomes.23,24 Through this highly regulated endosomal process of formation, the size of exosomes is relatively constant as compared to the larger types of vesicle. In addition to tetraspanins, proteins ALG-2 interacting-protein X and tumor susceptibility gene 101 (ALIX and TSG101, respectively) are reported to be involved in the endosomal process, and are frequently used as markers for exosomes.12,22 Different cell types can release discrete subpopulations of exosomes, each with different proteomic properties and RNA cargo and divergent membrane protein composition.25,26
Microvesicles Intermediate-sized EVs are most frequently referred to as microvesicles, ectosomes, or if tumor-derived, oncosomes, which arise via direct budding and cleavage of the plasma. Microvesicles are spherical and span a broad range of sizes, being between 50 nm to 1000 nm in diameter. They are distinguished based on their formation and release, and do not utilize the endosomal/multivesicular body pathway.27 Instead, microvesicles are formed through a process that involves calcium influx and remodeling of the cortical cytoskeleton to release the membrane-enclosed cytosolic cargo.12 Viewed broadly, microvesicles do not appear to be formed in a consistent manner like exosomes. However, when restricted to a specific cell type, microvesicles may form in a uniform manner, as illustrated in one recent study haematologica | 2018; 103(3)
EVs in the hematopoietic microenvironment
of neutrophils that consistently shed two distinct narrowly defined vesicle populations of ~100 nm and ~500 nm, both budding at the limiting membrane.28
Large vesicles Large vesicles, also referred to as large oncosomes due to their tumor-derived origin, are a class of EVs that can reach up to 10 microns in size and contain intact organelles and an ordered cytoskeletal structure.29 Large vesicles are similar to apoptotic bodies in size and composition; however, unlike apoptotic bodies, large vesicles are formed from cleavage of cytoplasmic extensions from intact living cells. Large vesicles have been described in Bcell acute lymphoblastic leukemia and prostate cancer, and have been demonstrated within patientsâ&#x20AC;&#x2122; samples and in vitro cultures of cancer cell lines.29,30
Apoptotic bodies Apoptotic bodies emerge during the course of programmed cell-death, as nuclear karyorrhexis occurs with cytoplasm and surrounding plasma membrane beginning to bleb into fragments.31 Apoptotic bodies consist of an intact plasma membrane enclosing cytosolic components and can contain both organelles and nuclear fragments. These bodies are subsequently eliminated through phagocytosis by surrounding cells and degraded in phagolysosomes.31 It has been reported that apoptotic bodies can horizontally transfer DNA to phagocytic recipient cells. As an example of this, one study showed that Epstein-Barr virusinfected B-lymphocytes generate apoptotic bodies that carry viral DNA and aid in the transfer of the virus to uninfected cells.32
Vesicle fate Once released from the parent cell, EVs can follow multiple routes. Some cancer cells generate EVs that rupture soon after release from their parent cells, distributing enzymes such as vascular endothelial growth factor and matrix metalloproteases into the surrounding interstitial space in order to promote angiogenesis and support cancer invasion through metastatic dissemination.33,34 EVs released into the blood appear to have a short half-life in circulation. In one representative study of B16-BL6 melanoma-derived EVs packaged with luciferase and lactadherin, luciferase activity was lost within minutes of intravenous injection with an observed serum half-life of approximately 2 minutes followed by rapid redistribution into tissues.35 A broad range of mechanisms for cellular uptake have been identified for EVs, including membrane fusion, phagocytosis or receptor-mediated caveolin-, clathrin- or lipid raft-mediated endocytosis, all culminating with transport of the EV cargo directly into the intracellular compartment.36 The differences from study to study suggest that EV uptake is a variable process and likely dependent on the type of EV and the parent and recipient cells involved. Experiments have shown that uptake is prevented at lower temperatures, suggesting that internalization is energy dependent and does not occur as a passive process.37 The uptake of EVs can be partially blocked by treating vesicles with either heparan sulfate or proteinase K, indicating a role for proteoglycans and surface proteins, respectively, in gaining entry into the cell.37,38 Pre-treatment of cells with the actin-depolymerizing drug cytochalasin D prior to EV exposure prevents cytoskeletal remodeling and reduces EV internalization,39 The use of the dynamin 2 inhibitor dynahaematologica | 2018; 103(3)
sore, which abrogates caveolin/clathrin-mediated endocytosis, has also been shown to inhibit uptake of reticulocyte-derived exosomes by macrophages.40 These data taken together are suggestive of an endocytic process mediating vesicle internalization. Little is known about specific mechanisms of uptake within the hematopoietic niche, although one study reported that megakaryocytederived EVs gain entry into hematopoietic progenitors cells via lipid raft mediated endocytosis, macropinocytosis and membrane fusion.61 Further study is warranted in order to understand the cellular events by which HSPC and supportive cells of the bone marrow differentially regulate the process of EV entry. How EVs are specifically targeted to different cell types within the hematopoietic niche in order to regulate hematopoiesis remains largely unknown. Among the most abundant membrane-associated proteins found on EVs are tetraspanins, a large cell-surface protein superfamily that interacts with transmembrane proteins and cytosolic signaling molecules to facilitate the organization of these structures into microdomains.41 Tetraspanins have been linked to many functions: intracellular signaling through Gprotein coupled receptors and protein kinase C; migration and metastasis by interacting with integrins and vascular cell adhesion molecule; cell morphogenesis by direct binding of alpha-actinin and the induction of actin polymerization.42,43 EV-embedded tetraspanins are dependent on the type of their parent cell; however, CD9, CD63, CD81, CD82, and CD151 are enriched in EVs derived from a range of sources.22 CD9, a common tetraspanin used to identify EVs was previously described in association with c-kit/CD117, a tyrosine kinase receptor that is highly expressed on HSPCs.44 Tetraspanins such as CD37, CD53 and TSSC6 have been found exclusively on hematopoietic cells. It is known that these tetraspanins interact with hematopoietic-specific targets such as Src homology region 2 domain-containing phosphatase-1, the pattern recognition receptor dectin-1, MHC-I/II, integrin ι4β1, Tcell/NK-cell co-stimulatory CD2, as well as common signal transducers including phosphatidylinositol-3 kinase and protein kinase C.45 Hematopoietic-specific tetraspanins and integrins on the EV surface remain strong candidates in targeting vesicles to specific cell types within the hematopoietic niche. A recent study demonstrated that, once inside the target cell, EVs are sorted into the endosomal pathway, move quickly through the cytoplasm and then stall at the endoplasmic reticulum, before eventually fusing with lysosomes for degradation.46 The process of cargo release by internalized EVs remains to be clarified. As the principal compartment for translation within the cell, the endoplasmic reticulum is a likely site for the deposition of mRNA and miRNA cargo. This and the assembly of the RNA interference-silencing complex in the endoplasmic reticulum may potentially explain how EVs alter protein synthesis and change cellular behavior. The half-life of internalized EVs has not been well defined. In the same study, 293T-derived EVs remained intact for hours to days once inside primary fibroblasts, with 50-60% merging with lysosomes by 48 hours.46
Physiological regulation of hematopoiesis by extracellular vesicles The BM comprises hematopoietic and non-hematopoietic cells organized into specialized microenvironments 385
J.T. Butler et al.
that provide dynamic regulation of hematopoiesis to assure the adequate formation and function of mature blood cells from HSCs.1 Mesenchymal stem cells (MSCs), their osteoprogenitor cell progeny, as well as endothelial cells and adipocytes coordinately maintain hematopoiesis by regulating proliferation, quiescence, differentiation, and apoptosis of HSPCs through juxtacrine and paracrine activity.2 Changes in compartmental oxygen concentration, hemorrhage, chemotherapy and irradiation can all prompt the emergence of HSCs from quiescence,47,48 and
several lines of evidence suggest that EVs are involved in regulating BM function during homeostasis and in response to injury (Table 2A, Figure 2). Some of the earliest descriptions of EVs revealed their role as platelet-derived anti-hemophilic particles and in transferrin receptor release from sheep reticulocytes.9 Additionally, more recent evidence points to EVs as important physiological mediators of signaling across the immunological synapse.11,49 Yet, much less is known about how vesicles might contribute to steady-state
Figure 2. Current evidence for extracellular vesicle crosstalk in the homeostatic bone marrow microenvironment. (A) MSC-derived EVs signal to HSPCs through the TLR-4 pathway, resulting in myeloid biased expansion. (B) Megakaryocyte-derived MVs are internalized by HSPCs and increase differentiation of new megakaryocytes through RNA-mediated signaling. (C) Hypoxia induces erythroleukemia cells to release EVs containing miR-486 which increases erythroblastic differentiation by targeting Sirt1 in HSPCs. (D) G-CSF infusion stimulates the release of EVs containing miR-126 that act to down-regulate VCAM-1 in HSPCs, resulting in their mobilization out of the BM. (E) HSPCs autoregulate stem potential by packaging and releasing critical secretory proteins through the exosomal pathway via the action of VPS33B. ANGPTL-2/3; angiopoietin-like protein 2 and 3; BM: bone marrow; CMP: common myeloid progenitor; EB: erythroblast; EVs: extracellular vesicles; G-CSF: granulocyte colony-stimulating factor; GMP; granulocyte monocyte progenitor; HSPC: hematopoietic stem and progenitor cell; Mk: megakaryocytes; MkB: megakaryoblast; MSC: mesenchymal stem cell; miR: microRNA; MV: microvesicles; TLR-4: Toll-like receptor 4; TPO: thrombopoietin; VCAM-1: vascular cell adhesion molecule; VPS33B: vacuolar protein sorting-associated protein 33B.
386
haematologica | 2018; 103(3)
EVs in the hematopoietic microenvironment
hematopoietic function or during a regenerative BM response. EV release is very clearly subject to a range of cellular stimuli, including cytokine activation, ionizing radiation, and differences in tissue oxygen tension.50,51
Granulocyte colony-stimulating factor mobilization is one such stimulus that appears to increase vesicle release from hematopoietic progenitors.52 Following injury, EV release may promote the selective delivery of miRNAs
Table 2A. Physiological regulation of hematopoiesis by extracellular vesicles.
EV origin
Recipient cell
Cargo
Reticulocytes
Macrophage
Transferrin receptor
Molecular target and effect
MSCs
HSPCs
Megakaryocytes
HSPCs
Mk-RNA
Erythroleukemia cells G-CSF stimulated BM stroma
HSPCs
miR-486-5p
ICAM-1: binding/entry CD63: binding/entry CD18: binding/entry CD11b: binding/entry Sirt1: downregulation
Stroma ECs HSPCs Young mouse BM stromal cells
miR-126
VCAM-1: downregulation
miR-183-5p
HMOX1: downregulation
SCL: upregulation, HoxB4: upregulation GATA2: upregulation MAPK p24/44: phosphorylation Autocrine signaling loop: maintains stemness
Aged mouse BM cells
TLR4: binding/activation
Mouse embryonic stem cells
HSPCs
Wnt3, Oct4,
HSPCs
HSPCs
TPO, ANGPTL2, ANGPTL3
Functional event
Reference
Release and recycling of transferrin receptor during enucleation and maturation of erythrocytes TLR4 signaling results in myeloid biased expansion and skewed hematopoietic repopulation potential of HSPCs Selective differentiation of progenitors into functional megakaryocytes
[9]
[17]
[61]
Promotes erythroid differentiation in response to hypoxia Down regulation of VCAM1 leads to mobilization of HSPCs out of the niche and into peripheral blood Reduced proliferative ability of stromal cells and decreased osteogenic differentiation Expansion of HSPCs and expression of markers associated with early HSC states
[63, 64]
VPS33B mediated release of exosomes is required for maturation of secretory growth factors and maintaining cell stemness
[45]
Functional event
Reference
[52]
[65]
[66]
Table 2B. Pathophysiological regulation of hematopoiesis by extracellular vesicles.
EV origin
Recipient cell
Cargo
Molecular target and effect
AML blasts
HSCPs
miR-150/155
cMYB; downregulation
AML blasts
BM Stroma
AML and MDS cells
MDS patient MSCs CML cells
MSCs
CD34+ progenitor cells BM stroma
Melanoma cells BM progenitors
Suppression of cMYB in HSPC reduces clonogenicity and leads to down regulation of niche retention factor CXCL12 and mobilization of HSPCs to peripheral blood CXCL12: downregulation Down regulation of HSC-supportive SCF: downregulation factors and suppression of hematopoiesis IGF1: downregulation and osteolineage development by upregulating DKK1: upregulation Dkk1 expression in BM stroma miR-7977 PCBP1: downregulation Reduced HSC-supportive growth factors Jagged1: downregulation and hematopoiesis-supportive SCF: downregulation capacity of MSCs ANGPT1: downregulation miR-10a/15a P53: transcriptional dysregulation Alteration of HSCPs viability MDM2: transcriptional dysregulation and clonogenicity Amphiregulin EGFR: activation Alteration of BM microenvironment (EGFR-ligand) MMP9: upregulation leading to increase attachment IL8: upregulation and proliferative advantage of CML cells c-MET Mobilization of BM progenitors and upregulation of proinflammatory molecules at sites of macrophage trafficking leading to promotion of melanoma invasion and metastasis
[13, 51]
[14]
[70]
[71] [72]
[73]
AML: acute myeloid leukemia; ANGPT1: angiopoietin 1; ANGPTL2/3: angiopoietin-like protein 2/3; BM: bone marrow; CML: chronic myelogenous leukemia; EC: endothelial cell; EGFR: epithelial growth factor receptor; EV: extracellular vesicle; G-CSF: granulocyte colony-stimulating factor; HMOX: heme-oxygenase molecule 1; HSC: hematopoietic stem cell; HSPCs: hematopoietic stem and progenitor cells; ICAM1: intercellular adhesion molecule 1; IGF1: insulin-like growth factor 1; IL8: interleukin 8; Mk: megakaryocyte; MDS: myelodysplastic syndrome; miR: micro-ribonucleic acid; Mk: megakaryocyte; MSC: mesenchymal stem cell; MMP9: matrix metalloprotease 9; SCF: stem cell factor; TLR4: Toll-like receptor 4; TPO: thrombopoietin; VCAM1: vascular cell adhesion molecule 1; VPS33B: vacuolar protein sorting 33B.
haematologica | 2018; 103(3)
387
J.T. Butler et al.
and other cargo, and may explain the enhanced angiogenic and regenerative activity after hypoperfusion injury in distant tissues.53 Several groups have studied the release and function of EVs by BM stroma, such as endothelial cells and MSCs. Endothelial cells have been shown to generate EVs with pro-angiogenic effects through the actions of miR-126,54 and have been linked to age-related downregulation of osteogenic differentiation within the BM.55,56 More literature exists on the release and function of EVs from BMderived MSCs. Our group recently demonstrated the trafficking of EVs from BM-derived MSCs to hematopoietic cells influencing progenitor commitment.17 Other groups showed that MSC-derived EVs selectively promoted tumor growth in patients with multiple myeloma.21 Additionally MSC-derived EVs have been shown to regulate angiogenic activity in endothelial cells, supporting the notion that BM MSC-derived EVs can regulate specific cell populations both within and outside of the hematopoietic compartment.57 In a recent study we showed that murine HSPCs (KSL: c-kit+/sca-1+/lineage-depleted) exposed to BM MSCderived EVs in vitro prompted activation with myeloid progenitor biased expansion and a skewed hematopoietic repopulation potential.17 Remarkably, this process seemed to be dependent on Toll-like receptor signaling and could be specifically abrogated in HSPCs from TLR4 knockout
or MyD88 knockout animals (Figure 2A). EVs of all classes are also rich in lipid components, especially products of arachidonic acid metabolism, including prostaglandin E2.58 Considering the potent activity of prostaglandin E2 in regulating HSC expansion and engraftment,59 it is tempting to speculate that EV-bound prostaglandin E2 released by MSCs contributes to this activity.60 The EV-mediated influence on hematopoiesis is not limited to supportive stromal cells alone. Megakaryocytes have also been shown to impart regulatory control on HSPCs by releasing microvesicles to orchestrate specific cell-type commitment. Megakaryoctye-derived microvesicles are among the most abundant microvesicles in the circulation, and attach to HSPCs by interacting with ICAM-1, CD43, CD18 and CD11b epitopes. Upon cell surface contact, these microvesicles become internalized where megakaryocyte RNA appears to serve as the mediator of biological effects, as evidenced by a loss of function of megakaryocyte microvesicles following RNAase treatment. Functionally, the internalization of these megakaryocyte microvesicles was found to redirect the differentiation of HSPCs toward functional megakaryocytes with limited effects on the phenotype of endothelial or stromal cells (Figure 2B).61 Several studies have demonstrated the importance of EV miRNA in regulating erythropoietic differentiation of HSPCs in both murine and human models.62 One recent
Figure 3. Current evidence for extracellular vesicle crosstalk in the leukemic microenvironment (A) EVs from AML blasts traffic miR-155 to HSPCs and down-regulate critical transcription factor, c-MYB, resulting in reduced differentiation potential. (B) AML EVs reprogram MSCs and stromal cells, and downregulate niche retention factor CXCL12 resulting in mobilization of HSPCs from the BM. (C) AML and MDS EVs promote the loss of HSPC supportive factors, CXCL12, SCF, IGF-1 through the trafficking of miR-7797 to supportive stroma, leading to reduced HSPC viability and hematopoietic potential. AML: acute myelogenous leukemia; ANGPT-1: angiopoietin 1; BM: bone marrow; CXCL12: C-X-C motif chemokine 12; EVs: extracellular vesicles; HSPCs: hematopoietic stem and progenitor cells; IGF-1: insulin-like growth factor 1; MDS: myelodysplastic syndrome; miR: microRNA; MSC: mesenchymal stem cell; PCBP1: poly(rc) binding protein 1; SCF: stem cell factor.
388
haematologica | 2018; 103(3)
EVs in the hematopoietic microenvironment
report described that erythroleukemia cells respond to hypoxia by rapidly releasing exosomes containing miR486, a known regulator of erythroid differentiation, which targets Sirt1 in CD34+ HSPCs (Figure 2C).63 This confirmed and extended previous studies that had implicated the increased expression of miR-486-5p in supporting erythroid differentiation of CD34+ cells in vitro.64 Conversely, the inhibition of miR-486-5p has been found to suppress CD34+ cell growth in vitro and in vivo, and decrease erythroid differentiation and survival of erythroid cells. It is possible that a similar physiological mechanism might exist to regulate hypoxia-responsive erythropoiesis in order to increase the delivery of oxygen to starved tissues. EVs within the BM microenvironment have been shown to modulate the behavior of HSCs in other ways. For example, treatment with pharmacological concentrations of granulocyte colony-stimulating factor used to mobilize stem and progenitor cells for collection and subsequent transplantation causes an increase in EVs containing high levels of miR-126 within the BM. These EVs are internalized by stroma, HSPCs and endothelial cells, delivering miR-126 into the cell, where it acts to translationally suppress vascular cell adhesion molecule-1 (Figure 2D). This decrease in vascular cell adhesion molecule-1, along with other signaling events, results in reduced HSPC adhesion and a shift into the peripheral blood, for collection by leukapheresis.52 Experimentally, EV-contained miR-126 released from mobilized human CD34 cells conferred proangiogenic activity and promoted hindlimb ischemia repair.53 Another recent study found that aging and oxidative stress alter the miRNA content of EVs in the BM microenvironment leading to age-related stem cell dysfunction. The investigators showed that BM-derived EVs from aged mice contain abundant miR-183-5p which, when endocytosed by primary BM stromal cells from young mice, decreased proliferation and inhibited osteogenic differentiation by reducing heme oxygenase 1, an enzyme essential in heme catabolism.65 Microvesicles derived from mouse embryonic stem cells were found to contain high levels of transcripts associated with pluripotency (Wnt-3 and Oct-4), and when exposed to hematopoietic progenitors led to their expansion.66 Additionally, hematopoietic progenitors exposed to the microvesicles derived from mouse embryonic stem cells were found to upregulate the expression of early HSC markers (SCL, HoxB4 and GATA2) and showed phosphorylation of MAPK p24/44 and serinethreonine kinase AKT.66 Finally, HSCs may contribute to their own stemness in part through secretory signaling and autocrine loops, involving vacuolar protein sorting protein 33b (VPS33B)mediated release of exosomes as carriers of thrombopoietin and angiopoietin-like protein 2 and 3 (Figure 2E). Herein, the loss of VPS33B compromised HSC potential and reduced leukemogenicity in cancer models.23 This and other studies discussed in this section support the view that within the physiological BM microenvironment, HSPCs release and internalize EVs, and are broadly responsive to regulation by vesicle trafficking in order to maintain hematopoiesis.
Pathophysiological regulation of hematopoiesis by extracellular vesicles Aside from the role of EVs in the cellular crosstalk in the BM under physiological conditions, EV trafficking also haematologica | 2018; 103(3)
plays a distinct role in deregulating hematopoiesis in injury and disease states, such as hematologic malignancies and extramedullary cancers (Table 2B).13,14,48 For example, MSC-derived EVs appear to contribute to marrow repair after radiation damage, restoring HSPC proliferation and engraftment with partial restoration of peripheral blood counts after intravenous injection of MSC-derived EVs.48 We reported that acute myeloid leukemia (AML) blasts rely on EVs for the transfer of miR-150 and miR-155, which target cMyb, a highly expressed transcription factor in progenitor cells to suppress HSPC clonogenicity. The coincident downregulation of the niche retention factor CXCL12 in those studies led to HSPC mobilization into the peripheral blood (Figure 3A).13,51 These observations were extended more recently by others showing that AML EVs not only downregulate HSC-supporting factors (CXCL12, stem cell factor, and insulin-like growth factor 1) (Figure 3B), but simultaneously suppress hematopoiesis and osteolineage development by upregulating Dkk1 expression in BM stromal cells.14 On the other hand, one study showed that AML EVs increased the number of HSCs by enhancing their survival while retaining their clonogenicity and stemness with no change in the hematopoietic CD34+, CD34+CD38â&#x2C6;&#x2019;, CD90+, and CD117+ phenotypes.67 Illustrating one of the key challenges in understanding HSPC regulation by EVs, neither of the two latter studies identified the specific EV component responsible. We and others previously showed that EVs released by steady-state or reprogrammed malignant stroma carry cytokines.17,68,69 Because most analyses of secreted cytokines do not separate vesicle-bound and vesicle-free cytokine activity it is entirely possible that some of the known cytokine activities that regulate HSPC in the leukemic niche reflect EV-mediated trafficking. Other hematologic disorders affect hematopoiesis indirectly by altering the function of the supportive nonhematopoietic stroma. Both AML and myelodysplastic syndrome cells were shown to reduce the hematopoiesissupportive capacity of MSCs by delivering miR-7977 via EVs. After uptake by MSCs, the EV-trafficked miR-7797 suppresses hematopoietic growth factors (jagged-1, stem cell factor and angiopoietin-1) by targeting the poly (rC) binding protein 1 post-transcriptional regulator (Figure 3C).70 MSCs from patients with myelodysplastic syndrome were also shown to release EVs that traffic miR10a and miR-15a to CD34+ progenitor cells, causing the transcriptional regulation of MDM2 and P53 genes, altering HSPC viability and clonogenicity.71 EVs released from chronic myelogenous leukemia cells have also been implicated in altering the BM microenvironment by activating epithelial growth factor receptor signaling in stromal cells. Chronic myelogenous leukemia exosomes were shown to contain amphiregulin, an epithelial growth factor receptor-activating ligand that leads to the downstream expression of matrix metalloproteinase-9 and interleukin-8, giving leukemic cells an adhesive and proliferative advantage within the hematopoietic niche.72 Extramedullary cancers, such as melanoma, also use EVs for the endocrine regulation of BM progenitors. For example, one study showed that melanoma EVs mobilize BM progenitors by targeting the receptor tyrosine kinase, cMET, in turn upregulating pro-inflammatory molecules at sites of macrophage trafficking to promote their invasion and metastasis in distant organs.73 389
J.T. Butler et al.
Diagnostic and therapeutic application of extracellular vesicles in hematologic disorders Vesicles are continuously shed by a wide range of blood cells.74 In addition to their short-range biological effects, the small size and rapid equilibration of EVs between tissues and the bloodstream have fueled interest in developing minimally invasive biomarkers based on EVs and their cargo. Identification of disease-specific markers based on circulating EVs may aid in early detection and post-remission monitoring for several types of hematologic disorders (Figure 4C). A number of studies indicate the feasibility of developing circulating EVs as a minimally invasive platform for the analysis of miRNA and protein content profiles to detect and classify hematologic malignancies and non-malignant hematopoietic disorders.75,76 For example, antibody microarray profiling of the membrane protein content of plasma EVs from patients with chronic lymphoid leukemia showed elevated levels of CD5, CD19, CD31, CD44, CD55, CD62L, CD82, HLA-A, HLA-B, HLA-C, and HLA-DR and low levels of CD21, CD49c, and CD63.77 The utility of EVs as biomarkers in chronic lymphoid leukemia is also supported by the observation of high levels of CD19 and CD37 relative to the levels in healthy controls.78 In patients with newly diagnosed AML, plasma EVs were rich in myeloblastic markers (CD34, CD33 and CD117), but also TGF-β1 protein, and MHC class I chain-related genes (MICA and MICB).69 The dynamic range of plasma EV-TGF-β1 readily separated AML patients by diagnosis, early and late remission status. In multiple myeloma, EVs bearing CD38, CD138, CD44 and CD147 allowed stratification of patients by disease phase and therapy response.79 There is a range of platforms for RNA amplification and the high incidence of relapse in AML patients has driven efforts to mine EV RNA content for highly sensitive detection of minimal residual disease and emergent drug resistance. EV miRNA appears to be a particularly promising minimally invasive biomarker platform. When EV cargo loading leads to the selective enrichment of some and exclusion of other cellular miRNA, the resulting highly selective vesicle miRNA profiles offer a potentially significant advantage over analysis of more diverse and abundant “free” circulating miRNA, which is often complexed with small lipoprotein particles. This improved signal to noise ratio led a number of groups to survey EV miRNA as a highly dynamic biomarker tool in hematologic malignancies (Table 3), and several studies have shown that AML EVs contain characteristic miRNA profiles.80,81 A particularly intriguing aspect of circulating EV miRNA is that it represents contributions of (occasionally identical) miRNA contained in EVs from multiple cellular sources. Such a compartmental biomarker reflects EV miRNA contributions from leukemic clones and the surrounding BM stromal cells.68,81 The use of EVs as biomarkers is not limited to hematologic malignancies; it may also be prognostically useful for benign hematologic disorders such as sickle-cell anemia. Plasma EVs from patients with sickle-cell anemia showed a distinct signature of miRNA that not only distinguished patients from healthy donors, but also coincided with the stage of the disease in the patients.82 In a study of patients with BM failure, investigators showed distinct profiles in patients with aplastic anemia and myelodysplastic syndrome and a biomarker response to successful immunosuppressive therapy.83 While many believe that EV miRNA under consideration as hematologic biomarkers may pri390
marily serve a biological role in endocrine cell-cell communication between distant tissues and cells, an alternative, and no less intriguing, possibility is the deliberate secretion of protein and RNA into EVs as a way of preventing regulation of the originating cell.84 Finally, autologous EVs from any number of accessible and in vitro expandable cells, including MSCs, T cells, and NK cells, can be a potentially exciting prospect for EV therapeutics, especially if approaches for selective loading and organ targeting can be developed. Human MSC EVs, for example, were shown in vitro to shuttle miR-155 and miR146 that exerted immunomodulatory effects through suppression of NK-, B-, and T-cell activity.85 In a murine model, CD73-bearing MSC EVs effectively reversed graft-versushost disease through the promotion of adenosine metabolism that in turn suppressed Th1-mediated inflammation immune suppression.86 As noted earlier, CD34-derived EVs appear to reverse hindlimb ischemia in animal models.53 A recent human study relied on MSC-derived EVs to treat graft-versus-host disease, based on prior successful work using pooled MSCs as a promising therapy for refractory graft-versus-host disease, a complication of allogeneic HSC transplantation.87 Presumably based on the high concentration of interleukin-10, TGF-β, and HLA-G, EV injection resulted in a significant reduction of the patient’s inflammatory response and improved the symptoms of graft-versushost disease in multiple organ systems. One recent study showed that Rab27 alpha/beta double knockout (RAB27DKO) mice, with impaired exosome release, have increased levels of cytokines and myeloproliferation, consistent with chronic inflammation. Grafting these mice with wild-type HSCs, or injecting EVs produced by granulocyte-macrophage colony-stimulating fac-
Table 3. Extracellular vesicle miRNA, selective roles in homeostasis and hematologic malignancies.
Function
Homeostasis/ disease
miRNA
Reference
Support erythroid differentiation Regulate osteogenic differentiation Regulate HSPC clonogenicity and mobilization Regulate hematopoietic function Control HSPC viability and clonogenicity Biomarker Biomarker
Homeostasis
miR-486
63
Homeostasis
miR-183-5p
65
AML
miR-150 and miR-155
24,99,100
AML and MDS
miR-7977
70
MDS
miR-10a, and miR-15a
71
Biomarker
Biomarker
AML CLL
miR-155 and miR-1246 miR-20a, miR-29, miR-150, miR-155, and miR-202-3p, miR-223 Hodgkin miR-21-5p, miR-24-3p, lymphoma miR-127-3p, miR-155-5p, and let-7a-5p MM miR-15, and miR-18a, miR-21, miR-135b and let-7b
81 101
102
103
AML: acute myeloid leukemia; CLL: chronic lymphocytic leukemia; miR: micro-ribonucleic acid; HSPC: hematopoietic stem and progenitor cell; MDS: myelodysplastic syndrome; MM: multiple myeloma.
haematologica | 2018; 103(3)
EVs in the hematopoietic microenvironment
tor-expanded wild-type HSCs ameliorated the inflammatory state. While Rab27DKO mice showed no response to lipopolysaccharide, response could be restored following exposure to wild-type hematopoietic EVs, but not EVs harvested from miR-155-/- cells, indicating that vesicle trafficking of miR-155 regulates the innate immune response.88
Perspective and open questions HSCs are competitively displaced from the hematopoietic niche in several types of cancer. Yet, the successive loss of HSCs from the BM is not explained by mere physical displacement, but rather occurs even at a disproportionately low tumor burden or with extramedullary tumor location. The underlying mechanisms of how cancer cells are able to disrupt the hematopoietic niche still remain unclear.8,89,90 Cellular competition was historically characterized in Drosophila as a non-cell autonomous mechanism involving p53 as a rheostat, whereby healthy (low p53) cells can competitively eliminate damaged and func-
tionally compromised neighboring cells (Figure 4B).91,92 It is thus tempting to speculate that a process such as cell competition between healthy and leukemic cells, the senescence-associated secretory phenotype, or even submicroscopic pre-cancerous changes in apparently healthy cells adjacent to tumor tissue (the so called “field effects”), result from EV trafficking.92,93 The described interaction between EVs and other cells through integrin receptor on the cell surface constitutes a potential candidate mechanism for the disruption of HSC retention in the BM niche.3,13,94 More broadly, the cellular context of vesicle transfer between cells in the hematopoietic niche, and how different routes of delivery affect target cell response are areas requiring urgent clarification. For example, it is evident that EV crosstalk occurs through the release of free vesicles into the interstitial space to interact with target cells in a paracrine or endocrine manner. However, vesicle transfer also appears to utilize cytoplasmic extensions (variously referred to as cytonemes, nanotubes, or invadopodia) which deliver contents directly into adjacent cells. These alternative modes of delivery make it difficult to cleanly segregate contact-dependent effects from those
Figure 4. Unresolved aspects of extracellular vesicle biology in the regulation of hematopoiesis. (A) EVs have been proposed to enter recipient cells through lipid raft-mediated internalization, endocytosis, phagocytosis, membrane fusion, caveolin-mediated endocytosis and macropinocytosis. (B) Exosome-mediated crosstalk may explain the intercellular competition of neighboring cells where the “winner” HSPC outcompetes the less fit HSPC through a P53-dependent mechanism. (C) Vesicles contain cargo comprised of uniquely packaged proteins, miRNAs and RNAs which serve as promising biomarkers for disease detection. (D) Vesicles from HSPCs and other cells of the bone marrow niche have been shown to exhibit preferential targeting to specific recipient cells for entry. (E) Cytonemes (filopodia, invadopodia, tunneling nanotubes) are cytoplasmic extensions that serve as modes of exosomal transfer to adjacent bystander cells. EVs: extracellular vesicles; HSPC: hematopoietic stem and progenitor cell.
haematologica | 2018; 103(3)
391
J.T. Butler et al.
relying on “free” vesicle exchange (Figure 4E).46,95 Clearly, the development of experimental approaches that more closely resemble the physiological context should be a priority for future studies. In addition to hematopoietic regulation, the vesicles released by BM cells may play a central role in the establishment and propagation of pathophysiological events outside the medullary space, ranging from promoting metastatic dissemination of melanoma cells, to priming inflammatory responses after cardiac injury.73,96 A mechanistic understanding of the EV-mediated crosstalk between tissues is currently lacking, and EV as carriers of cytokines, bioactive lipids and several classes of RNA may deserve greater consideration when systemic conditions affect BM function. For example, given the EV-mediated crosstalk between lymphocytes and antigen-presenting cells,11 or priming of an inflammatory phenotype in the BM by EV miRNA,97 it is not inconceivable that systemic inflammatory effects after cardiovascular injury or chronic stress conditions are similarly induced by EVs and their cargo.96 Such a systemic communication model finds further support in reports of BM-derived EV trafficking to the brain during experimentally induced systemic inflammation.98 Finally, it is now widely accepted that EVs contribute to pathophysiological regulation, and the suppression of EV release in disease states may offer therapeutic benefit. However, while several of the molecular mechanisms involved in EV release have been described,24,99 broad suppression of EV release is an unlikely therapeutic goal given the role EVs play in maintaining homeostasis. Rather, a nuanced understanding of cell-specific biogenesis, cargo incorporation, and EV-recipient cell affinity may offer the insight necessary for more targeted and disease-specific
References 1. Morrison SJ, Scadden DT. The bone marrow niche for haematopoietic stem cells. Nature. 2014;505(7483):327–334. 2. Pietras EM, Warr MR, Passegué E. Cell cycle regulation in hematopoietic stem cells. J Cell Biol. 2011;195(5):709–720. 3. Schepers K, Campbell TB, Passegué E. Normal and leukemic stem cell niches: insights and therapeutic opportunities. Cell Stem Cell. 2015;16(3):254–267. 4. Blank U, Karlsson S. TGF-β signaling in the control of hematopoietic stem cells. Blood. 2015;125(23):3542–3550. 5. Yamazaki S, Iwama A, Takayanagi S-I, et al. Cytokine signals modulated via lipid rafts mimic niche signals and induce hibernation in hematopoietic stem cells. EMBO. 2006;25(15):3515–3523. 6. Fleming HE, Janzen V, Celso Lo C, et al. Wnt signaling in the niche enforces hematopoietic stem cell quiescence and is necessary to preserve self-renewal in vivo. Cell Stem Cell. 2008;2(3):274–283. 7. Yilmaz OH, Morrison SJ. The PI-3kinase pathway in hematopoietic stem cells and leukemia-initiating cells: a mechanistic difference between normal and cancer stem cells. Blood Cells Mol Dis. 2008;41(1):73–76. 8. Colmone A, Amorim M, Pontier AL, Wang
392
approaches. Further research into the identity of vesicle surface molecules that govern target cell specificity and route of cellular uptake will be critical in mapping the role that EVs play in regulating hematopoiesis (Figure 4A,D). Classifying these surface molecules may also prove useful for harnessing the potential of vesicles to deliver targeted therapeutics within the BM, or blocking the action of cancer-derived EVs, in order to advance the treatment of hematologic disease. Additionally, the intracellular events that follow uptake are particularly poorly understood. Answering questions about miRNA copy number per EV, or intracellular EV processing, cargo unloading and vesicle degradation will be crucial to realize the therapeutic potential for modified EVs. Finally, understanding how different components of a given vesicle cooperatively alter the behavior of a cell, and whether vesicles with an identical cell origin can differentially regulate multiple cell types in the niche, will be key to utilizing this complex biological process in order to create realistic therapeutics. In sum, EVs offer fundamental new insight into the biology of HSC regulation, as well as translational opportunities for mitigating injury, and opposing malignancy. Due to their constitutive role in regulating specific cell populations within the marrow niche, and unique signatures, EVs could prove to be a powerful tool for advancing hematology, and be exploited to improve diagnosis, disease monitoring and therapy. Acknowledgments We apologize to those investigators whose work we were unable to cite for space limitations. We recognize funding through the Hyundai Hope on Wheels Program (PK), Max Blue Butterfly Campaign (PK), the Medical Research Foundation of Oregon (SA) and we gratefully acknowledge contributions from Ben Doron.
S, Jablonski E, Sipkins DA. Leukemic cells create bone marrow niches that disrupt the behavior of normal hematopoietic progenitor cells. Science. 2008;322(5909):1861– 1865. 9. Johnstone RM. Exosomes biological significance: a concise review. Blood Cells Mol Dis. 2006;36(2):315–321. 10. Thomou T, Mori MA, Dreyfuss JM, et al. Adipose-derived circulating miRNAs regulate gene expression in other tissues. Nature. 2017;542(7642):450–455. 11. Mittelbrunn M, Gutiérrez-Vázquez C, Villarroya-Beltri C, et al. Unidirectional transfer of microRNA-loaded exosomes from T cells to antigen-presenting cells. Nat Commun. 2017;2:282. 12. Yáñez-Mó M, Siljander PR-M, Andreu Z, et al. Biological properties of extracellular vesicles and their physiological functions. J Extracell Vesicles. 2015;4(0):27066. 13. Hornick NI, Doron B, Abdelhamed S, et al. AML suppresses hematopoiesis by releasing exosomes that contain microRNAs targeting c-MYB. Sci Signal. 2016;9(444):ra88. 14. Kumar B, Garcia M, Weng L, et al. Acute myeloid leukemia transforms the bone marrow niche into a leukemia-permissive microenvironment through exosome secretion. Leukemia. 2017 Aug 17. [Epub ahead of print] 15. Wolf P. The nature and significance of
platelet products in human plasma. Br J Haematol. 1967;13(3):269–288. 16. Pan BT, Johnstone RM. Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: selective externalization of the receptor. Cell. 1983;33(3):967– 978. 17. Goloviznina NA, Verghese SC, Yoon YM, Taratula O, Marks DL, Kurre P. Mesenchymal stromal cell-derived extracellular vesicles promote myeloid-biased multipotent hematopoietic progenitor expansion via Toll-like receptor engagement. J Biol Chem. 2016;291(47): 24607–24617. 18. Geddings JE, Hisada Y, Boulaftali Y, et al. Tissue factor-positive tumor microvesicles activate platelets and enhance thrombosis in mice. J Thromb Haemost. 2016;14(1):153– 166. 19. Al-Nedawi K, Meehan B, Kerbel RS, Allison AC, Rak J. Endothelial expression of autocrine VEGF upon the uptake of tumorderived microvesicles containing oncogenic EGFR. Proc Natl Acad Sci USA. 2009;106 (10):3794–3799. 20. Hoshino A, Costa-Silva B, Shen T-L, et al. Tumour exosome integrins determine organotropic metastasis. Nature. 2015;527 (7578):329–335. 21. Roccaro AM, Sacco A, Maiso P, et al. BM mesenchymal stromal cell-derived exosomes facilitate multiple myeloma progres-
haematologica | 2018; 103(3)
EVs in the hematopoietic microenvironment sion. J Clin Invest. 2013;123(4):1542–1555. 22. Kowal J, Arras G, Colombo M, et al. Proteomic comparison defines novel markers to characterize heterogeneous populations of extracellular vesicle subtypes. Proc Natl Acad Sci USA. 2016;113(8):E968–977. 23. Gu H, Chen C, Hao X, et al. Sorting protein VPS33B regulates exosomal autocrine signaling to mediate hematopoiesis and leukemogenesis. J Clin Invest. 2016;126(12):45374553. 24. Ostrowski M, Carmo NB, Krumeich S, et al. Rab27a and Rab27b control different steps of the exosome secretion pathway. Nat Cell Biol. 2010;12(1):19–30 25. Willms E, Johansson HJ, Mäger I, et al. Cells release subpopulations of exosomes with distinct molecular and biological properties. Sci Rep. 2016;6(1):22519. 26. Smith ZJ, Lee C, Rojalin T, et al. Single exosome study reveals subpopulations distributed among cell lines with variability related to membrane content. J Extracell Vesicles. 2015;4:28533. 27. Raposo G, Stoorvogel W. Extracellular vesicles: exosomes, microvesicles, and friends. J Cell Biol. 2013;200(4):373–383. 28. Timár CI, Lőrincz ÁM, Csépányi-Kömi R, et al. Antibacterial effect of microvesicles released from human neutrophilic granulocytes. Blood. 2013;121(3):510–518. 29. Johnson SM, Dempsey C, Parker C, Mironov A, Bradley H, Saha V. Acute lymphoblastic leukaemia cells produce large extracellular vesicles containing organelles and an active cytoskeleton. J Extracell Vesicles. 2017;6(1):e1294339. 30. Di Vizio D, Morello M, Dudley AC, et al. Large oncosomes in human prostate cancer tissues and in the circulation of mice with metastatic disease. Am J Pathol. 2012;181(5):1573–1584. 31. Taylor RC, Cullen SP, Martin SJ. Apoptosis: controlled demolition at the cellular level. Nat Rev Mol Cell Biol. 2008;9(3):231–241. 32. Holmgren L, Szeles A, Rajnavölgyi E, et al. Horizontal transfer of DNA by the uptake of apoptotic bodies. Blood. 1999;93(11):3956– 3963. 33. Hakulinen J, Sankkila L, Sugiyama N, Lehti K, Keski-Oja J. Secretion of active membrane type 1 matrix metalloproteinase (MMP-14) into extracellular space in microvesicular exosomes. J Cell Biochem. 2008;105(5):1211–1218. 34. Becker A, Thakur BK, Weiss JM, Kim HS, Peinado H, Lyden D. Extracellular vesicles in ccncer: cell-to-cell mediators of metastasis. Cancer Cell. 2016;30(6):836–848. 35. Takahashi Y, Nishikawa M, Shinotsuka H, et al. Visualization and in vivo tracking of the exosomes of murine melanoma B16-BL6 cells in mice after intravenous injection. J Biotechnol. 2013;165(2):77–84. 36. Mulcahy LA, Pink RC, Carter DR. Routes and mechanisms of extracellular vesicle uptake. J Extracell Vesicles. 2014;3(1):24641. 37. Christianson HC, Svensson KJ, van Kuppevelt TH, Li J-P, Belting M. Cancer cell exosomes depend on cell-surface heparan sulfate proteoglycans for their internalization and functional activity. Proc Natl Acad Sci USA. 2013;110(43):17380–17385. 38. Escrevente C, Keller S, Altevogt P, Costa J. Interaction and uptake of exosomes by ovarian cancer cells. BMC Cancer. 2011;11 (1):108. 39. Svensson KJ, Christianson HC, Wittrup A, et al. Exosome uptake depends on ERK1/2heat shock protein 27 signaling and lipid Raft-mediated endocytosis negatively regu-
haematologica | 2018; 103(3)
lated by caveolin-1. J Biol Chem. 2013;288(24):17713–17724. 40. Barrès C, Blanc L, Bette-Bobillo P, et al. Galectin-5 is bound onto the surface of rat reticulocyte exosomes and modulates vesicle uptake by macrophages. Blood. 2010;115(3):696–705. 41. Zöller M. Tetraspanins: push and pull in suppressing and promoting metastasis. Nature. 2009;9(1):40–55. 42. Yáñez-Mó M, Barreiro O, Gordon-Alonso M, Sala-Valdés M, Sánchez-Madrid F. Tetraspanin-enriched microdomains: a functional unit in cell plasma membranes. Trends Cell Biol. 2009;19(9):434–446. 43. Hemler ME. Tetraspanin functions and associated microdomains. Nat Rev Mol Cell Biol. 2005;6(10):801–811. 44. Anzai N, Lee Y, Youn B-S, et al. C-kit associated with the transmembrane 4 superfamily proteins constitutes a functionally distinct subunit in human hematopoietic progenitors. Blood. 2002;99(12):4413–4421. 45. Beckwith KA, Byrd JC, Muthusamy N. Tetraspanins as therapeutic targets in hematological malignancy: a concise review. Front Physiol. 2015;6:91. 46. Heusermann W, Hean J, Trojer D, et al. Exosomes surf on filopodia to enter cells at endocytic hot spots, traffic within endosomes, and are targeted to the ER. J Cell Biol. 2016;213(2):173–184. 47. Kucharzewska P, Christianson HC, Welch JE, et al. Exosomes reflect the hypoxic status of glioma cells and mediate hypoxia-dependent activation of vascular cells during tumor development. Proc Natl Acad Sci USA. 2013;110(18):7312–7317. 48. Wen S, Dooner M, Cheng Y, et al. Mesenchymal stromal cell-derived extracellular vesicles rescue radiation damage to murine marrow hematopoietic cells. Leukemia. 2016;30(11):2221–2231. 49. Gutiérrez-Vázquez C, Villarroya-Beltri C, Mittelbrunn M, Sánchez-Madrid F. Transfer of extracellular vesicles during immune cellcell interactions. Immunol Rev. 2013;251(1):125–142. 50. Tadokoro H, Umezu T, Ohyashiki K, Hirano T, Ohyashiki JH. Exosomes derived from hypoxic leukemia cells enhance tube formation in endothelial cells. J Biol Chem. 2013;288(48):34343–34351. 51. Huan J, Hornick NI, Goloviznina NA, et al. Coordinate regulation of residual bone marrow function by paracrine trafficking of AML exosomes. Leukemia. 2015;29(12): 2285–2295. 52. Salvucci O, Jiang K, Gasperini P, et al. MicroRNA126 contributes to granulocyte colony-stimulating factor-induced hematopoietic progenitor cell mobilization by reducing the expression of vascular cell adhesion molecule 1. Haematologica. 2012;97(6):818–826. 53. Mathiyalagan P, Liang Y, Kim D, et al. Angiogenic mechanisms of human CD34(+) stem cell exosomes in the repair of ischemic hindlimb. Circ Res. 2017;120(9):1466–1476. 54. Lombardo G, Dentelli P, Togliatto G, et al. Activated Stat5 trafficking via endothelial cell-derived extracellular vesicles controls IL3 pro-angiogenic paracrine action. Sci Rep. 2016;6(1):25689. 55. Weilner S, Schraml E, Wieser M, et al. Secreted microvesicular miR-31 inhibits osteogenic differentiation of mesenchymal stem cells. Aging Cell. 2016;15(4):744–754. 56. Hromada C, Mühleder S, Grillari J, Redl H, Holnthoner W. Endothelial extracellular vesicles-promises and challenges. Front
Physiol. 2017;8:275. 57. Gong M, Yu B, Wang J, et al. Mesenchymal stem cells release exosomes that transfer miRNAs to endothelial cells and promote angiogenesis. Oncotarget. 2017;8(28):4520045212. 58. Record M, Carayon K, Poirot M, SilventePoirot S. Exosomes as new vesicular lipid transporters involved in cell-cell communication and various pathophysiologies. Biochim Biophys Acta. 2014;1841(1):108– 120. 59. Hoggatt J, Singh P, Sampath J, Pelus LM. Prostaglandin E2 enhances hematopoietic stem cell homing, survival, and proliferation. Blood. 2009;113(22):5444–5455. 60. Madrigal M, Rao KS, Riordan NH. A review of therapeutic effects of mesenchymal stem cell secretions and induction of secretory modification by different culture methods. J Transl Med. 2014;12(1):260. 61. Jiang J, Kao C-Y, Papoutsakis ET. How do megakaryocytic microparticles target and deliver cargo to alter the fate of hematopoietic stem cells? J Control Release. 2017;247:1–18. 62. O'Connell RM, Chaudhuri AA, Rao DS, Gibson WSJ, Balazs AB, Baltimore D. MicroRNAs enriched in hematopoietic stem cells differentially regulate long-term hematopoietic output. Proc Natl Acad Sci USA. 2010;107(32):14235–14240. 63. Shi X-F, Wang H, Kong F-X, et al. Exosomal miR-486 regulates hypoxia-induced erythroid differentiation of erythroleukemia cells through targeting Sirt1. Exp Cell Res. 2017;351(1):74–81. 64. Wang L-S, Li L, Li L, et al. MicroRNA-486 regulates normal erythropoiesis and enhances growth and modulates drug response in CML progenitors. Blood. 2015;125(8):1302–1313. 65. Davis C, Dukes A, Drewry M, et al. MicroRNA-183-5p increases with age in bone-derived extracellular vesicles, suppresses bone marrow stromal (stem) cell proliferation, and induces stem cell senescence. Tissue Eng Part A. 2016.23(2122);1231-1240. 66. Ratajczak J, Miekus K, Kucia M, et al. Embryonic stem cell-derived microvesicles reprogram hematopoietic progenitors: evidence for horizontal transfer of mRNA and protein delivery. Leukemia. 2006;20(5):847– 856. 67. Razmkhah F, Soleimani M, Mehrabani D, et al. Leukemia microvesicles affect healthy hematopoietic stem cells. Tumor Biol. 2017;39(2):101042831769223. 68. Viola S, Traer E, Huan J, et al. Alterations in acute myeloid leukaemia bone marrow stromal cell exosome content coincide with gains in tyrosine kinase inhibitor resistance. Br J Haematol. 2015;172(6):938-986 69. Szczepanski MJ, Szajnik M, Welsh A, Whiteside TL, Boyiadzis M. Blast-derived microvesicles in sera from patients with acute myeloid leukemia suppress natural killer cell function via membrane-associated transforming growth factor-beta1. Haematologica. 2011;96(9):1302–1309. 70. Horiguchi H, Kobune M, Kikuchi S, et al. Extracellular vesicle miR-7977 is involved in hematopoietic dysfunction of mesenchymal stromal cells via poly(rC) binding protein 1 reduction in myeloid neoplasms. Haematologica. 2016;101(4):437–447. 71. Muntión S, Ramos TL, Diez-Campelo M, et al. Microvesicles from mesenchymal stromal cells are involved in HPC-microenvironment crosstalk in myelodysplastic patients.
393
J.T. Butler et al. Wagner W, editor. PLoS One. 2016;11(2): e0146722. 72. Corrado C, Saieva L, Raimondo S, Santoro A, De Leo G, Alessandro R. Chronic myelogenous leukaemia exosomes modulate bone marrow microenvironment through activation of epidermal growth factor receptor. J Cell Mol Med. 2016;20(10):1829–1839. 73. Peinado H, Alečković M, Lavotshkin S, et al. Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nat Med. 2012;18(6):883–891. 74. Burnouf T, Chou M-L, Goubran H, Cognasse F, Garraud O, Seghatchian J. An overview of the role of microparticles/ microvesicles in blood components: are they clinically beneficial or harmful? Transfus Apher Sci. 2015;53(2): 137–145. 75. Caivano A, Laurenzana I, De Luca L, et al. High serum levels of extracellular vesicles expressing malignancy-related markers are released in patients with various types of hematological neoplastic disorders. Tumour Biol. 2015;36(12):9739–9752. 76. Caivano A, La Rocca F, Laurenzana I, et al. Extracellular vesicles in hematological malignancies: from biology to therapy. Int J Mol Sci. 2017;18(6):1183. 77. Belov L, Matic KJ, Hallal S, Best OG, Mulligan SP, Christopherson RI. Extensive surface protein profiles of extracellular vesicles from cancer cells may provide diagnostic signatures from blood samples. J Extracell Vesicles. 2016;5(1):e25355. 78. De Luca L, D'Arena G, Simeon V, et al. Characterization and prognostic relevance of circulating microvesicles in chronic lymphocytic leukemia. Leuk Lymphoma. 2017;58(6):1424–1432. 79. Harshman SW, Canella A, Ciarlariello PD, et al. Proteomic characterization of circulating extracellular vesicles identifies novel serum myeloma associated markers. J Proteomics. 2016;136:89–98. 80. Valadi H, Ekström K, Bossios A, Sjöstrand M, Lee JJ, Lötvall JO. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol. 2007;9(6):654– 659.
394
81. Hornick NI, Huan J, Doron B, et al. Serum exosome microRNA as a minimally-invasive early biomarker of AML. Sci Rep. 2015;5:11295. 82. Khalyfa A, Khalyfa AA, Akbarpour M, et al. Extracellular microvesicle microRNAs in children with sickle cell anaemia with divergent clinical phenotypes. Br J Haematol. 2016;174(5):786–798. 83. Hosokawa K, Kajigaya S, Feng X, et al. A plasma microRNA signature as a biomarker for acquired aplastic anemia. Haematologica. 2017;102(1):69–78. 84. Lin M-C, Chen S-Y, Tsai H-M, et al. PGE2 /EP4 signaling controls the transfer of the mammary stem cell state by lipid rafts in extracellular vesicles. Stem Cells. 2017;35 (2):425–444. 85. Di Trapani M, Bassi G, Midolo M, et al. Differential and transferable modulatory effects of mesenchymal stromal cell-derived extracellular vesicles on T, B and NK cell functions. Sci Rep. 2016;6(1):24120. 86. Amarnath S, Foley JE, Farthing DE, et al. Bone marrow-derived mesenchymal stromal cells harness purinergenic signaling to tolerize human Th1 cells in vivo. Stem Cells. 2015;33(4):1200–1212. 87. Kordelas L, Rebmann V, Ludwig A-K, et al. MSC-derived exosomes: a novel tool to treat therapy-refractory graft-versus-host disease. Leukemia. 2014;28(4):970–973. 88. Alexander M, Ramstead AG, Bauer KM, et al. Rab27-dependent exosome production inhibits chronic inflammation and enables acute responses to inflammatory stimuli. J Immunol. 2017;199(10):3559–3570. 89. Boyd AL, Campbell CJ, Hopkins CI, et al. Niche displacement of human leukemic stem cells uniquely allows their competitive replacement with healthy HSPCs. J Exp Med. 2014;211(10):1925–1935. 90. Shiozawa Y, Pedersen EA, Havens AM, et al. Human prostate cancer metastases target the hematopoietic stem cell niche to establish footholds in mouse bone marrow. J Clin Invest. 2011;121(4):1298–1312. 91. Yu X, Harris SL, Levine AJ. The regulation of exosome secretion: a novel function of the p53 protein. Cancer Res. 2006;66(9):4795– 4801. 92. Bondar T, Medzhitov R. p53-mediated
hematopoietic stem and progenitor cell competition. Stem Cell. 2010;6(4):309–322. 93. Slaughter DP, Southwick HW, Smekjal W. Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin. Cancer. 1953;6(5):963– 968. 94. Dhawan A, Bonin von M, Bray LJ, et al. Functional interference in the bone marrow microenvironment by disseminated breast cancer cells. Stem Cells. 2016;34(8):2224– 2235. 95. Jung Y, Wang J, Havens A, et al. Cell-to-cell contact is critical for the survival of hematopoietic progenitor cells on osteoblasts. Cytokine. 2005;32(3-4):155– 162. 96. Heidt T, Sager HB, Courties G, et al. Chronic variable stress activates hematopoietic stem cells. Nat Med. 2014;20(7):754–758. 97. Alexander M, Hu R, Runtsch MC, et al. Exosome-delivered microRNAs modulate the inflammatory response to endotoxin. Nat Commun. 2015;6:7321. 98. Ridder K, Keller S, Dams M, et al. Extracellular vesicle-mediated transfer of genetic information between the hematopoietic system and the brain in response to inflammation. PLoS Biol. 2014; 12(6):1001874. 99. Dai J, Lu Y, Wang C, et al. Vps33b regulates Vwf-positive vesicular trafficking in megakaryocytes. J Pathol. 2016;240(1):108– 119. 100. Zhi F, Cao X, Xie X, et al. Identification of circulating microRNAs as potential biomarkers for detecting acute myeloid leukemia. PLoS One. 2013;8(2):56718. 101. Yeh Y-Y, Ozer HG, Lehman AM, et al. Characterization of CLL exosomes reveals a distinct microRNA signature and enhanced secretion by activation of BCR signaling. Blood. 2015;125(21):3297–3305. 102. van Eijndhoven MA, Zijlstra JM, Groenewegen NJ, et al. Plasma vesicle miRNAs for therapy response monitoring in Hodgkin lymphoma patients. JCI Insight. 2016;1(19):89631. 103. Manier S, Liu C-J, Avet-Loiseau H, et al. Prognostic role of circulating exosomal miRNAs in multiple myeloma. Blood. 2017;129(17):2429–2436.
haematologica | 2018; 103(3)
ARTICLE
Hematopoiesis
Glucocorticoids induce differentiation of monocytes towards macrophages that share functional and phenotypical aspects with erythroblastic island macrophages
Ferrata Storti Foundation
Esther Heideveld,1 Lea A. Hampton-Oâ&#x20AC;&#x2122;Neil,2 Stephen J. Cross,3 Floris P.J. van Alphen,4 Maartje van den Biggelaar,4,5 Ashley M. Toye2,6,7 and Emile van den Akker1
Sanquin Research, Department of Hematopoiesis, Amsterdam and Landsteiner Laboratory, Academic Medical Center, University of Amsterdam, the Netherlands; 2 Department of Biochemistry, School of Medical Sciences, Bristol, UK; 3Wolfson Bioimaging Facility, School of Medical Sciences, Bristol, UK; 4Sanquin Research, Department of Research Facilities, Amsterdam, the Netherlands; 5Sanquin Research, Department of Plasma Proteins, Amsterdam, the Netherlands 6Bristol Institute for Transfusion Sciences, NHS Blood and Transplant, Filton, Bristol, UK and 7National Institute for Health Research (NIHR) Blood and Transplant Research Unit in Red Blood Cell Products, University of Bristol, UK 1
Haematologica 2018 Volume 103(3):395-405
ABSTRACT
T
he classical central macrophage found in erythroblastic islands plays an important role in erythroblast differentiation, proliferation and enucleation in the bone marrow. Convenient human in vitro models to facilitate the study of erythroid-macrophage interactions are desired. Recently, we demonstrated that cultured monocytes/macrophages enhance in vitro erythropoiesis by supporting hematopoietic stem and progenitor cell survival. Herein, we describe that these specific macrophages also support erythropoiesis. Human monocytes cultured in serum-free media supplemented with stem cell factor, erythropoietin, lipids and dexamethasone differentiate towards macrophages expressing CD16, CD163, CD169, CD206, CXCR4 and the phagocytic TAM-receptor family. Phenotypically, they resemble both human bone marrow and fetal liver resident macrophages. This differentiation is dependent on glucocorticoid receptor activation. Proteomic studies confirm that glucocorticoid receptor activation differentiates monocytes to anti-inflammatory tissue macrophages with a M2 phenotype, termed GC-macrophages. Proteins involved in migration, tissue residence and signal transduction/receptor activity are upregulated whilst lysosome and hydrolase activity GO-categories are downregulated. Functionally, we demonstrate that GC-macrophages are highly mobile and can interact to form clusters with erythroid cells of all differentiation stages and phagocytose the expelled nuclei, recapitulating aspects of erythroblastic islands. In conclusion, glucocorticoid-directed monocyte differentiation to macrophages represents a convenient model system to study erythroid-macrophage interactions. Introduction In human bone marrow (BM) and fetal liver (FL), the production of erythrocytes through erythropoiesis occurs on erythroblastic islands.1,2 These erythroblastic islands consist of a central macrophage surrounded by erythroid cells at different stages of terminal differentiation and support proliferation, differentiation and phagocytose the extruded nuclei (or pyrenocytes) of erythroid cells.2-6 Chow et al. described that mouse CD169+ (SIGLEC1) BM resident macrophages display a dual role promoting erythropoiesis and retention of hematopoietic stem and progenitor cells (HSPC).7,8 Their absence leads to the mobilization of HSPC, reduced BM erythropoiesis and the inability to properly respond to anemia.7-10 It is, however, unclear whether CD169 identifies different macrophage populations or indicates an intrinsic dual role for the same tissue macrophage. FL macrophages that are unable to interact with erythroblasts due to disruption of the retinoblastoma tumor suphaematologica | 2018; 103(3)
Correspondence: e.vandenakker@sanquin.nl
Received: August 22, 2017. Accepted: December 27, 2017. Pre-published: December 28, 2017. doi:10.3324/haematol.2017.179341 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/395 Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
395
E. Heideveld et al.
pressor gene in mice lead to embryonic death as erythroblasts fail to enucleate.11 These data show that in vivo, macrophages are important in regulating erythropoiesis in adults and during development. Previously, we found that blood-derived monocytes induced to differentiate using stem cell factor (SCF), erythropoietin (EPO) and glucocorticoids enhance in vitro erythropoiesis by supporting HSPC survival.12 These macrophages display a tissue-resident profile expressing CD14 (lipopolysaccharide [LPS]-receptor), CD16 (FcγRIII), scavenger receptor CD163, CD169, CD206 (mannose receptor), CXCR4 and minimal expression of dendritic cell-specific intercellular adhesion molecule 3-grabbing non-integrin (DC-SIGN).12 We hypothesized that these cultured monocyte-derived macrophages may have a similar role as mouse CD169+ macrophages in both hematopoiesis and erythropoiesis. This would provide an easy-to-use in vitro human model system to mimic erythroblastic islands allowing for the study of functional interactions between macrophages and erythroid cells, which is currently limited to harvesting BM or involves genetic modification.13 A better understanding of the mechanism(s) through which human macrophages interact and regulate erythroblast maturation and enucleation is important in order to understand the pathology of erythropoietic disorders, such as erythrocytosis in polycythemia vera or erythrophagocytosis in several types of hemolytic anemia, as well as to improve in vitro erythroid differentiation protocols for erythrocyte production.14,15 In mice BM, erythroblasts are bound to macrophages via interactions between integrin-α4β1 on erythroblasts and VCAM1 on macrophages, and blocking these molecules disrupts erythroblastic islands.16 Chow et al. described human BM macrophages as also expressing VCAM1. However, Ulyanova et al. have shown that Vcam-/- mice do not display an erythroid phenotype during homeostasis or phenylhydrazide-induced stress.17 During terminal differentiation erythroblasts enucleate, resulting in reticulocytes and pyrenocytes. The latter are also still encapsulated by plasma membrane. In mice, clearance of pyrenocytes occurs via TAM-receptors on the central macrophages that recognize and bind phosphatidylserine (PS) exposed on pyrenocytes resulting in phagocytosis in a protein Sdependent manner.18,19 The TAM-receptor family of tyrosine kinases (TYRO3, AXL, and MERTK) play an important role in the phagocytic ability of macrophages as triple knock-out mice fail to clear apoptotic cells in multiple tissues. These mice develop normally, but eventually develop autoimmunity, such as systemic lupus erythematosus (SLE).20 This is in line with studies showing that SLE has been associated with failure of macrophages to phagocytose apoptotic cells and pyrenocytes in both humans and mice.21-24 In addition, anemia is found in about 50% of SLE patients; Toda et al. showed that embryos suffer from severe anemia caused by failure of macrophages to phagocytose pyrenocytes.25 These data indicate that macrophages are essential during all stages of erythropoiesis, including enucleation, and display inherent features that are indispensable to the functionality of these macrophages. Herein, we show that peripheral blood monocytes can be differentiated to erythropoiesis-supporting macrophages that interact with erythroid cells, phagocytose pyrenocytes and phenotypically resemble human CD169+ BM and FL macrophages. 396
Methods Human materials Human blood, BM and FL mononuclear cells were purified by density separation, following manufacturer’s protocol. Regarding blood, informed consent was given in accordance with the Declaration of Helsinki, the Dutch National and Sanquin Internal Ethic Boards, and by the Bristol Research Ethics Committee (REC; 12/SW/0199). Following informed consent, adult BM aspirates were obtained from the sternum of patients undergoing cardiac surgery, and approved by the Medical Ethical Review Board of the AMC (MEC:04/042#04.17.370). Fetal tissues (week 15-22) were obtained from elective abortions contingent on informed consent and approval by the Medical Ethical Commission of the Erasmus University Medical Center Rotterdam (MEC-2006-202).
Cell culture CD14 and CD34 MicroBeads (Miltenyi Biotec, Gladbach, Germany) were used for cell isolation from peripheral blood. CD14+ monocytes were cultured at 1.5-3x106 cells/well (CASY® Model TTC, Schärfe System GmbH, Reutlingen, Germany) in a 12-well plate as described.12 Cells were treated with 1-20μM mifepristone (Sigma-Aldrich, Munich, Germany) directly after isolation or 4-24 hours after three days of culture. CD34+ cells were differentiated towards erythroblasts,12 with the addition of 1ng/ml IL-3 (R&D systems, Abingdon, UK) at the start of culture. Media was replenished every two days. After 8-10 days, cells were differentiated towards reticulocytes by removing dexamethasone, increasing EPO (10U/ml, ProSpec; East Brunswick, NJ, USA) and adding heparin (5U/ml, LEO Pharma B.V., Breda, The Netherlands), 5% pooled AB+ plasma and holotransferrin (700µg/ml, Sanquin, Amsterdam, The Netherlands). Every other day, half the media was replenished. For co-culture experiments, CD14+ cells were differentiated with (GC-macrophages) or without dexamethasone for three days and co-cultured with erythroblasts (day 8-10 of culture; ratio 1:1.5) or more differentiated erythroid cells (day 6 of differentiation; ratio 1:4) for 24 hours.
Flow cytometry Cells were washed in phosphate-buffered saline (PBS) and resuspended in 1% bovine serum albumin (BSA)/PBS. Cells were incubated with primary antibodies for 30min at 4C, measured on LSRII or LSRFortessa (both BD Biosciences, Oxford, UK) and analyzed using FlowJo software (FlowJo v10; Tree Star, Inc., Ashland, OR, USA) (antibodies listed in Online Supplementary Methods).
Mass spectrometry See Online Supplementary Methods.
ImageStreamX and IncuCyte GC-macrophages or unstimulated cells were incubated with 100μg/ml fluorescein isothiocyanate (FITC)-labeled zymosan (S. cerevisiae; MP Biomedicals, Solon, OH, USA) for 40min at 37C. Zymosan was removed and cells were fixed in 4% paraformaldehyde (PFA) for 20min at 4C. Cells were transferred to 1% BSA/PBS and stained with human leukocyte antigen-antigen Drelated R-phycoerythrin (HLA-DR PE; BD Biosciences). Furthermore, erythroid cells at day seven of differentiation were stained with Deep Red Anthraquinone 5 (DRAQ5; Abcam, Cambridge, UK). Imaging was performed on the ImageStreamX (Amnis Corporation, Seattle, WA, USA) and images were analyzed using IDEAS Application v6.1 software (Amnis Corporation). For IncuCyte experiments see Online Supplementary Methods. haematologica | 2018; 103(3)
Modeling human erythroid-macrophage interactions
Cytospins Cells were cytospun using Shandon Cytospin II (Thermo Scientific), dried and fixed in methanol. Cells were stained with benzidine and Differential Quik Stain Kit (PolySciences, Warrington, PA, USA) following manufacturer’s instructions. Slides were dried, embedded in Entellan (Merck, Darmstadt, Germany) and images were taken (Leica DM-2500, Germany).
Reverse transcription polymerase chain reaction analysis Reverse transcription polymerase chain reaction (RT-PCR) was performed as previously described.12 Values were normalized using S18 and HPRT as a reference gene and calibrated relative to expression of CD14+ monocytes at day 0 (primers listed in Online Supplementary Methods).
A
B
C
D
E
Figure 1. Glucocorticoid receptor activation directs CD14+ monocytes towards a tissue resident macrophage phenotype. (A-D) Distribution graphs displaying the relative geometric mean fluorescence intensity (MFI) of CD16, CD163, CD169 and CXCR4 on human monocytes (n=3-6) cultured for three days under various conditions (EPO, SCF, lipids or dexamethasone). MFI was normalized to change to isotype control and presented as fold change (fc). Mean ± SEM (two-way ANOVA, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001). (E) Relative expression of CD16, CD163, CD169, CXCR4, CD206 and DC-SIGN on CD14+ monocytes (n=3) directly after isolation from mononuclear cells (D0) and after culture in the presence or absence of dexamethasone (Dex) and/or mifepristone (Mif). MFI was normalized to isotype control and displayed as a fold change to day 0. Mean ± SEM (ratio paired t-test, *P<0.05, **P<0.01). EPO: erythropoietin; SCF: stem cell factor; ns: not significant; ND: not detected.
haematologica | 2018; 103(3)
397
E. Heideveld et al.
Results Glucocorticoid stimulation directs monocyte differentiation to CD16+CD163+CD169+CXCR4+CD206+ macrophages We previously found that purified peripheral blood CD14+ monocytes cultured in EPO, SCF, lipids and dexamethasone differentiate within three days into CD163, CD169, CXCR4 and CD16-positive macrophages that, upon co-culture with CD34+ cells, significantly increase the erythroid yield.12 However, it remained unclear as to which growth factors were crucial to differentiate monocytes to
A
B
C
D
macrophages supporting erythropoiesis. Therefore, we examined which growth factors or supplements determined this differentiation cue. Flow cytometry analysis showed that dexamethasone, exclusively, induces high expression of CD16 and CD163 in macrophages. The addition of EPO, SCF or lipids does not contribute to this high expression (Figure 1A,B). CXCR4 expression was already upregulated in the absence of dexamethasone but was further increased upon stimulation with dexamethasone and lipids, whilst the expression of tissue residency marker CD169 was also upregulated but occurred in a dexamethasone-independent manner (Figure 1C,D). Online
E
Figure 2. Proteome analysis of CD14+ monocytes cultured in the presence or absence of dexamethasone revealed two distinct macrophage populations. (A) Principal component analysis of GC-macrophages (red) versus non-glucocorticoid stimulated cells (blue) of four donors (indicated A-D). (B) Volcano plot (false discovery rate 0.05 S0 0.4) showing P-values (-log) versus difference of cells cultured for three days in the presence or absence of dexamethasone. (C) Heatmap of differentially expressed proteins based on Z-scored label-free quantification values. (D) Interaction analysis based on STRING (all interactions) of upregulated (red) and downregulated (blue) proteins. (E) Enrichment analysis using BiNGO and enrichment mapper in GC-macrophages with upregulated (red) and downregulated (blue) processes.
398
haematologica | 2018; 103(3)
Modeling human erythroid-macrophage interactions
Supplementary Figure S1A depicts distinct morphological changes upon dexamethasone-induced differentiation between freshly isolated CD14+ monocytes and cultured CD14+ cells. Monocytes were incubated with mifepristone, which blocks glucocorticoid receptor activation. Membrane and messenger ribonucleic acid (mRNA) expression of CD16, CD163, and CD206 was significantly reduced by mifepristone treatment, and thus dependent on glucocorticoid receptor transcriptional control (Figure 1E and Online Supplementary Figure S1B,C). Although neither Figure 1C nor Figure 1E show an effect of dexamethasone on the fluorescence intensity of CD169, mRNA levels of CD169 were clearly increased upon stimulation of the glucocorticoid receptor and reduced when cells were treated with mifepristone. In contrast, CXCR4 mRNA levels did not change upon mifepristone treatment, but membrane expression was increased (Online Supplementary Figure S1B). Monocyte differentiation increases expression of DC-SIGN independently of dexamethasone, albeit to
expression levels that are significantly lower compared to dendritic cells (Figure 1E and Online Supplementary Figure S1C).26 Note that cultured monocytes in all conditions are a homogeneous population, as single peaks observed in histograms and multi-color flow cytometry data revealed that monocytes stimulated with glucocorticoids are CD16+CD163+CD169+CXCR4+CD206+ cells (Online Supplementary Figure S1C,D). Interestingly, flow cytometry data revealed that monocytes that have been differentiated for three days in the presence of dexamethasone were unable to change their phenotype after 4 or 24 hours of mifepristone treatment. Only CD163 expression was slightly reduced after 24 hours mifepristone treatment (Online Supplementary Figure S1E). The data indicates that glucocorticoid stimulation initiates an irreversible differentiation program of monocytes towards macrophages CD16+CD163+CD169+CXCR4+CD206+ which is maintained for at least 17 days of culture (Online Supplementary Figure S2A,B).
A
B
C
E
haematologica | 2018; 103(3)
D
Figure 3. GC-macrophages form erythroblast clusters with increased frequency and erythroblast composition. (A) Expression of integrins ITGA4 and ITGB1 and adhesion molecules ICAM1, PECAM, and VCAM1 on GC-macrophages (Mφ) (n=6) and erythroblasts (EBL) at day 1 and 7 of differentiation (n=3-4). Mean fluorescence intensity (MFI) has been normalized to the isotype control. Mean ± SEM (unpaired t-test, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001). (B) Scaled cell-displacement vector diagram (left; 20 representative macrophages in both conditions) and box-and-whisker plot (right; 68 representative macrophages in -Dex and 21 in +Dex) after three days of culture in the absence or presence of dexamethasone. (Welch's unpaired t-test, **P<0.01, n=5). (C-D) Co-culture of GCmacrophages or unstimulated cells with erythroblasts (unpaired t-test of 1153 (-Dex) and 749 (+Dex) macrophages, **P<0.01, ****P<0.0001, n=5). Images were taken every hour during 64 hours of analysis. (C) Plot showing the average erythroblast-macrophage links for each macrophage. Mean ± SD. (D) 5-95% box plot showing the maximum number of links per macrophage. Mean is indicated by crosses. (E) Representative images of cytospins of GC-macrophages (+Dex) or unstimulated cells (-Dex) co-cultured with erythroblasts for 24 hours (in 50x magnification, panels i-ii or 100x magnification, panels iii-v; n=4). Dex: dexamethasone; ND: not detected.
399
E. Heideveld et al.
Proteomics data revealed GC-macrophages display a distinct anti-inflammatory profile To gain further insights into the dexamethasoneinduced monocyte differentiation process, we performed mass spectrometry-based quantitative proteomics on these cells after three days of differentiation and compared this to non-glucocorticoid stimulated monocytes. A total of 3,210 proteins were quantified, and principal com-
ponent analysis clearly separated glucocorticoid-stimulated from non-stimulated cells (Figure 2A and Online Supplementary Table S1). Glucocorticoid stimulation induced a distinct expression pattern compared to nonglucocorticoid stimulated monocytes, as visualized in the volcano plot and corresponding heatmap of the 169 differentially expressed proteins for individual donors (Figure 2B,C). Note that the expression of CD163 and CD206
A
B
C
D
G
400
E
F
Figure 4. GC-macrophages can bind erythroid cells and phagocytose pyrenocytes. (A) Relative mRNA expression of TAM-receptor family members MERTK, AXL and TYRO3 on CD14+ cells (D0) cultured for three days (D3) in the presence or absence of dexamethasone (Dex). 20μM mifepristone (Mif) was added for three days or after three days for 4 hours (n=4). Mean ± SEM (ratio paired t-test, *P<0.05, ****P<0.0001). (B) Representative ImageStreamX images of zymosan (green) phagocytosed by HLA-DR (red) positive unstimulated macrophages (-Dex) and GCmacrophages (+Dex) (left), and corresponding 10-90% box plot showing the number of zymosan particles phagocytosed (right) (unpaired t-test of 1285 -Dex and 530 +Dex macrophages, ****P<0.0001, n=3). (C-F) GC-macrophages and unstimulated cells were co-cultured for 24 hours with day 6 differentiated erythroid cells (unpaired t-test of 370 –Dex and 313 +Dex macrophages, ****P<0.0001, n=3). (C) Representative images of cytospins (in 50x magnification, panels i-ii or 100x magnification, panels iii-v). Macrophages bind nucleated erythroid cells (large arrow), reticulocytes (arrowhead) and phagocytose pyrenocytes (small arrow) and some erythroid cells during differentiation (asterisk). 1090% box plots showing the number of nucleated cells (D) or reticulocytes (E) bound to macrophages (Mφ). (F) Scatter plot showing the number of pyrenocytes bound to or phagocytosed by macrophages. Mean ± SD. (G) Graph showing the binding of CD235a+ differentiated erythroid cells to GC-macrophages versus unstimulated cells. Corresponding histogram showing geometric mean of CD235a in FITC (n=4). Mean ± SEM (paired t-test, **P<0.01). HLA-DR: human leukocyte antigen – antigen D-related; MFI: mean fluorescence intensity; ns: not significant; BF: bright-field.
haematologica | 2018; 103(3)
Modeling human erythroid-macrophage interactions
(MRC1) was highly induced after glucocorticoid receptor activation, corroborating the flow cytometry experiments. The most differentially expressed proteins (n=169) were mapped to evaluate specific upregulation or downregulation of functionality-linked protein networks, based on the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) analysis (Figure 2D). CD163 and CD206 are part of an interactome protein node that is specifically upregulated in dexamethasone-induced macrophages, and includes M2 macrophage markers CSF1R, stabilin-1 (STAB1) and complement proteins C3AR1, C1QC, and FcγRIIa (CD32) which has been associated with high phagocytic capacity of the cells. Moreover, VSIG4 was upregulated in dexamethasoneinduced macrophages, which is restricted to resting tissue macrophages,27 while ABCA1 was also upregulated, which has been highly associated with hemoglobin-asso-
A
B
ciated macrophages.28 In addition, proteins with a positive regulation of cell migration and motility, including DAB2, ADAM9, Serpine1 (PAI1) and CD81, are upregulated in dexamethasone-induced macrophages. Furthermore, a whole range of signaling receptors were upregulated, amongst which are TGF and IFNγ receptors (TGFBR1 and IFNGR1) and IL13RA1. These proteins belong to processes that are enriched as GO-term, e.g., membrane part, signal transducer activity, transmembrane receptor activity and molecular transduced activity. In addition, many immune regulatory processes are also enriched (Figure 2E and Online Supplementary Table S2). Interestingly, members of the cathepsin family involved in antigen presentation (e.g., CTSC, CTSL1, CTSD and CTSS) were downregulated. A range of pro-inflammatory proteins, clustered within an interactome node, were downregulated; these include lysosomal enzymes HEXA and HEXB, MANBA,
C
D
F E
Figure 5. CD163+ macrophage populations in human BM and FL. (A) Representative dot plots of erythroid cells characterized by the expression of CD71high and CD235a in total human BM (n=4) and FL mononuclear cells (n=4). (B) Graph belonging to panel A. Mean ± SD (unpaired t-test, ***P<0.001). (C) Graph showing the percentage of CD163+ cells present in the total cell population in BM (n=7) and FL (n=5). Mean ± SEM (unpaired t-test, **P<0.01). (D) Characterization of CD163+ macrophages in human BM (n=3-7) and FL (n=3-5) based on the expression of CD14, CD16, CD163, CD169, CXCR4, CD206 and VCAM1. Mean fluorescence intensity (MFI) has been normalized to the isotype control. Mean ± SEM (unpaired t-test, **P<0.01). (E) Representative dot plots of erythroid-macrophage clusters formation of erythroid cells (CD71highCD235a+) with CD163+ BM (n=4) and FL (n=4) macrophages. (F) Graph belonging to panel E. Mean ± SD (unpaired t-test, *P<0.05). BM: bone marrow; FL: fetal liver.
haematologica | 2018; 103(3)
401
E. Heideveld et al.
saponin PSAP and GLB1, in addition to other lysosome/hydrolase activity-related GO-categories (Figure 2D,E). In addition, GO-categories associated with lipid metabolic processes were also downregulated in GCmacrophages. Furthermore, CHI3L1 and CD44 are highly upregulated in non-glucocorticoid stimulated cells (Figure 2B). CHI3L1 is described as a pro-inflammatory factor,29,30 while CD44 has been expressed on pro-inflammatory tissue macrophages.31 In conclusion, CD14+ monocytes that have been differentiated in the presence of dexamethasone display a distinct anti-inflammatory proteomic profile and are further denoted as GC-macrophages, while unstimulated cells have a more inflammatory profile.
GC-macrophages are motile and bind erythroblasts GC-macrophages may, besides supporting the erythroid yield, also regulate terminal differentiation of erythroblasts, recapitulating aspects of erythroblastic islands. In mice, it has been shown that BM central macrophages can bind erythroblasts through various interactions: VCAM1integrin-α4β1,16,32 integrin-α5β1-ICAM4,33,34 erythroblast macrophage protein (EMP)-EMP,4,35 or EphrinB2EphrinB4.36 Flow cytometry data revealed that GCmacrophages express common cell adhesion molecules (CAM), such as integrins (α4 [ITGA4], β1,2 [ITGB1, ITGB2/CD18] and αL,M,X [ITGAL/CD11a, ITGAM/CD11b, ITGAX/CD11c]), the immunoglobulin (Ig) superfamily (ICAM1, PECAM, VCAM1) and E- and Lselectin (Figure 3A and Online Supplementary Figure S3A). Most of these CAM could be identified in the proteomics data, including ICAM3, integrin-β5 and α5, however, VCAM1, selectins and EMP were not detected (Online Supplementary Table S1). With the exception of integrin-β5, these CAM were not differentially expressed between GC-macrophages and non-glucocorticoid stimulated cells. Erythroblasts expressed similar ITGA4 levels compared to GC-macrophages, but exhibited a 10-fold reduction in ITGB1 expression and low expression of ICAM1 and PECAM, whereas VCAM1 was not detected (Figure 3A). When differentiating erythroblasts towards reticulocytes (Online Supplementary Figure S3B,C), the expression of CAM was reduced, as expected, which potentially indicates a lower binding affinity of erythroid cells to macrophages during erythroid differentiation. Next, we investigated whether GC-macrophages interact in vitro with erythroid cells compared to non-glucocorticoid stimulated monocytes. Indeed, live imaging cells for 2.5 days showed that GC-macrophages are highly motile and nonstimulated macrophages are non-motile (Figure 3B), a finding which corroborates the increased expression of cell migration and motility proteins (Figure 2D) whilst engaging twice as many erythroblasts (0.5 vs. 0.3, P<0.0001) at every time point measured (Figure 3C,D). In addition, cytospins of macrophages co-cultured for 24 hours with erythroblasts showed that the number of macrophages binding erythroblasts as well as the number of erythroblasts bound was increased on GCmacrophages compared to non-GC macrophages (Figure 3E and Online Supplementary Figure S3D). Nonetheless, no difference in interaction duration between erythroblasts and macrophages from both conditions was observed (Online Supplementary Figure S3E), suggesting that the unstimulated cells possess some machinery to interact with erythroblasts. In conclusion, GC-macrophages are motile, express a variety of CAM and form erythroblast 402
interactions with increased frequency and numbers per macrophage compared to cells cultured in the absence of dexamethasone.
GC-macrophages express TAM-receptor family members and phagocytose pyrenocytes As CD169+CD163+ macrophages promote erythrowe decided to examine whether poiesis,8 GC-macrophages can provide a similar functional role in vitro. In mice, pyrenocytes are phagocytosed by central macrophages in a Mer tyrosine kinase (MERTK)-dependent manner.18 RT-PCR showed that GC-macrophages upregulate both MERTK and AXL mRNA compared to freshly isolated and non-glucocorticoid stimulated monocytes (Figure 4A). MERTK expression was inhibited by mifepristone treatment during the first three days of culture, whereas AXL was not, suggesting that AXL expression is induced via a trans-regulated process while MERTK needs the transcriptional activity of the glucocorticoid receptor. Note that TYRO3 levels are dexamethasoneindependently increased. Besides TAM-receptors, other PS-receptors on macrophages have been reported to be involved in clearing apoptotic bodies, such as TIM337 (T-cell Ig and mucin-domain containing-3), STAB38 and CD300A39 (CMRF35-like molecule 8). TIM3 mRNA levels were increased, albeit independently of dexamethasone (Online Supplementary Figure S4A). This was confirmed by mass spectrometry, as peptides corresponding to TIM3 were identified in GC-macrophages (HAVCR2 in Online Supplementary Table S1). CD300A and STAB1 were also identified, of which STAB1 was significantly increased in GC-macrophages compared to unstimulated cells. Interestingly, proteomics data showed that lactadherin, a PS-binding glycoprotein which stimulates phagocytosis of red blood cells by macrophages,40 was significantly induced in GC-macrophages compared to unstimulated cells. RT-PCR confirmed increased lactadherin mRNA levels, but this was dexamethasone-independent (Online Supplementary Figure S4B). Moreover, both GCmacrophages and unstimulated cells express DNASE2, a crucial protein required to degrade DNA within phagocytosed apoptotic bodies or pyrenocytes in macrophages.41 Expression of TAM-receptors and other PS-receptors on GC-macrophages may be a prerequisite to phagocytose particles, cells or pyrenocytes in case of erythropoiesis. Figure 4B shows that the number of GC-macrophages that phagocytose particles, in addition to the amount of zymosan particles per macrophage, is higher (73% vs. 45%, 2.3 vs. 1.7, respectively) compared to unstimulated cells. Subsequently, both unstimulated cells and GCmacrophages were co-cultured with a mixture of differentiating erythroblasts, reticulocytes and pyrenocytes (Online Supplementary Figure S3B,C) for 24 hours. Cytospin analysis showed that both GC-macrophages and unstimulated cells bind erythroid cells (Figure 4C), however, increased numbers of nucleated cells, reticulocytes and pyrenocytes bind to GC-macrophages compared to unstimulated cells (Figure 4D-F and Online Supplementary Figure S4C). Note that all nucleated erythroid cells are specifically aligned with their nucleus towards the macrophage as observed in vivo (Figure 4C). Pyrenocytes, however, were almost solely phagocytosed by GCmacrophages (Figure 4F and Online Supplementary Figure S4D). Importantly, GC-macrophages and unstimulated cells did not overtly phagocytose nucleated cells or reticuhaematologica | 2018; 103(3)
Modeling human erythroid-macrophage interactions
locytes (Online Supplementary Figure S4E,F). Flow cytometry data showed that indeed both GC-macrophages and unstimulated cells can bind erythroid cells, however, increased cluster formation was found for GCmacrophages compared to unstimulated cells (Figure 4G). These results demonstrate that GC-macrophages functionally resemble specific aspects of macrophages within the erythroblastic island by binding erythroblasts and reticulocytes and phagocytosing pyrenocytes.
GC-macrophages share characteristics with CD163+ macrophages found in human BM and FL To investigate whether GC-macrophages share phenotypical characteristics with macrophages found in the two major erythropoietic organs during human development and adulthood (FL and BM, respectively), mononuclear cells of both organs were analyzed. Between week 15 and 22 of human development, the FL is primarily undertaking erythropoiesis, representing a median of 85% of the total number of mononuclear cells compared to 29% in BM, with increased frequencies of CD71+CD235– pro-erythroblasts in FL (Figure 5A,B). To prevent the presence of free immunogenic pyrenocytes and to support erythroid cell requirements in the developing embryo, it is anticipated that the FL contains significant amounts of erythroblastic islands and, thus, supporting macrophages. Indeed, Figure 5C shows a 6.5-fold increase in CD163+ FL macrophages compared to BM (3.3% vs. 0.5%). Further characterization shows only subtle differences in expression of macrophage markers (Figure 5D and Online Supplementary Figure S5A,B), as both macrophage populations express high levels of CD163 and CD14 and have intermediate levels of CD169, CD206 and VCAM1. CD163+ BM macrophages tend to express more CXCR4, whereas CD163+ FL macrophages have higher expression of CD16. Online Supplementary Table S3 displays the comparison between the mean fluorescence intensity (MFI) of BM, FL, non-stimulated and GC-macrophages and reveals that GC-macrophages phenotypically recapitulate macrophages found in the FL and BM. GC-macrophages are more similar to BM macrophages (CD16 and CXCR4 expression), however, they also share features of FL macrophages (CD206 expression). Unstimulated cells do not express VCAM1, and have low expression of CD206, CD163, CD14 and CD16. Figure 5E,F shows that both BM and FL CD163+ macrophages bind erythroid cells (46% in BM vs. 83% in FL), indicating that CD163 purifies erythroid-supporting macrophages. Interestingly, FL macrophages have increased interactions with CD71+CD235a+ cells compared to BM. The similarity of marker expression levels of BM, FL and GC-macrophages and the fact that all three populations form erythroid clusters suggest that GC-macrophages share phenotypic and functional characteristics with in vivo erythroid-supporting macrophages. GC-macrophages could thus be used as a substitute in vitro model to study the supportive effects of macrophages on erythropoiesis.
Discussion We have previously shown that monocyte-derived macrophages can support erythropoiesis by increased survival of HSPC.12 Herein, we show that these macrophages, derived from CD14+ monocytes, are differentiated in a gluhaematologica | 2018; 103(3)
cocorticoid-dependent manner (termed GCmacrophages), interact with erythroid cells of all stages and phagocytose the extruded pyrenocytes. Besides these functional aspects, GC-macrophages also share phenotypic characteristics with resident macrophages from both human BM and FL, among which there is high expression of CD163 and CD206. Interestingly, CD163+ BM cells appear to be more heterogeneous compared to FL cells. GC-macrophages also phenotypically resemble macrophages described recently by Belay et al., who employed a lentivirally introduced small molecule responsive Mpl-based cell growth switch that enabled cord blood or BM CD34+ cells to be differentiated to erythroidsupporting macrophages.13 Similar to GC-macrophages, these cells express CD14, CD163, CD169, CD206, VCAM1, ITGAM and ITGAX. Herein, we show that these macrophages can also be differentiated from peripheral blood monocytes using dexamethasone, without the need for genetic manipulation. Falchi et al. showed that in erythroid culture conditions, CD34+ cells can also differentiate to macrophages that interact with erythroid cells, however, we can exclude this differentiation pathway as the purified CD14+ monocytes we used to differentiate macrophages from peripheral blood did not show hematopoietic colony potential or CD34+ contamination.12 The erythroid system is renowned for its rapid response to systemic decreases in oxygen pressure. Together with elevated EPO levels, glucocorticoid levels also increase upon exposure to high altitude.42 EPO, SCF and glucocorticoids induce erythroblasts to proliferate whilst inhibiting differentiation.43-46 Elevated systemic EPO and glucocorticoids as a response to low-oxygen stress leads to increased erythroid output due to augmented survival and proliferation of BM erythroblasts. To accommodate this increased erythropoiesis, we hypothesize that the number of central macrophages must also be increased or alternatively these cells would have to engage with more erythroblasts. Our flow cytometry and cytospin data confirmed that GC-macrophages interact with erythroid cells of all stages, be that as it may, this does not provide information on the longevity of the interactions, as these could be transient, as previously implied.47 Via live cell imaging we analyzed the interaction between GC-macrophages and erythroblasts, which revealed that GC-macrophages are more mobile compared to cells that were cultured in the absence of dexamethasone, and that this mobility, or “macrophage ranging”, results in more interactions with erythroblasts. Higher mobility was accompanied by an increased expression of proteins involved in migration and motility. High motility has previously been observed in CD34+ differentiated macrophages stimulated with dexamethasone.47 Motility is an important functional aspect, as erythroblastic islands in vivo form away from sinusoids and migrate to the sinusoidal endothelium to release reticulocytes into the circulation.48,49 Interestingly, this work also demonstrates that non-glucocorticoid-stimulated monocytes can interact with erythroblasts, as they form interactions for the same length of time (1.8 hours on average) when they encounter erythroblasts. This suggests that both populations express receptors that allow engagement and interaction with erythroblasts, however, GCmacrophages have significantly more interactions with erythroblasts per macrophage and bind a higher number of erythroblasts. Surprisingly, GC-macrophages display low expression of VCAM1, suggesting that erythroblast 403
E. Heideveld et al.
interactions may also occur in a VCAM1-independent manner. Indeed, Ulyanova et al. reported that Vcam1-/mice do not display a compromised erythroid stress response in spleen and BM.17 Whether another interaction substitutes for VCAM1 would need to be determined. The presented monocyte differentiation methodology has potential to be exploited as an imaging platform to delineate the hierarchy of contributions of various receptors within the macrophage-erythroblasts in BM and GCmacrophages in future studies. We have also demonstrated, using proteomics and imaging, that GC-macrophages actively phagocytose pyrenocytes and express the correct putative machinery to recognize pyrenocytes. The mechanism(s) through which macrophages recognize reticulocytes but phagocytose pyrenocytes are ill-defined in human erythropoiesis. Our proteomic study and RT-PCR data demonstrate that GCmacrophages express all TAM-receptors, including MERTK and other PS-receptors, which may be used by GC-macrophages to take up pyrenocytes. This work, alongside our ability to manipulate erythroblast protein expression, now provides an excellent accessible model system to mechanistically understand how macrophages promote erythropoiesis and eventually target pyrenocytes for phagocytosis and destruction. Furthermore, it is interesting to note that GC-macrophages interact preferably to the polarized nuclear side of erythroid cells as observed in BM erythroblastic islands. In general, proteomic analysis revealed an array of processes and proteins that are differentially regulated between GC-macrophages and unstimulated cells. The data will allow further studies to delineate essential pathways that are key to glucocorticoidstimulated differentiation of monocytes towards erythroid-supporting GC-macrophages. This is probably the concerted action of multiple pathways. Finally, our observations have important implications for our understanding of the dynamics of the macrophage populations in human BM. We characterized both human BM and FL macrophages and found that CD163+ FL macrophages define a homogeneous population. In contrast, CD163+ BM macrophages show a more heterogeneous population, reflecting that CD163+ cells represent a mixed population of myeloid cells. Both human BM and FL CD163+ macrophages are capable of binding erythroid cells, however, this percentage is lower in BM (46%) compared to FL (83%). The FL is primarily performing erythropoiesis at week 15-22 of embryonic development, which suggests that CD163 purifies mainly central macrophages. The reduced erythroid-macrophage clusters in BM may reflect a more heterogeneous CD163+ population with possibly different functions. Changes observed in marker expression of macrophages in both organs could thus be due to this heterogeneity in the BM population. CD163
References 1. Lee SH, Crocker PR, Westaby S, et al. Isolation and immunocytochemical characterization of human bone marrow stromal macrophages in hemopoietic clusters. J Exp Med. 1988;168(3):1193-1198. 2. Mohandas N, Prenant M. Three-dimensional model of bone marrow. Blood. 1978;51(4):633-643.
404
isolation in combination with single cell RNA-sequencing may discriminate these different populations and identify specific discriminatory cell surface markers to allow for functional experiments. Albeit for decades it was believed that all macrophages originate from monocytes,50 recent parabiosis and fatemapping studies showed that most resident macrophages are maintained independently of monocytes.51 However, Theurl et al. showed that resident Kupffer cells in the liver contain a mixture of de novo hematopoiesis-derived and embryonic-derived macrophages. They identified an ondemand mechanism to facilitate quick and transient increases in cells that can function as Kupffer cells but originate from classical monocytes.52 Taken together with our work, we hypothesize a new scenario in which specific macrophages originate from different sources depending on the need of a specific tissue or process. These processes may also occur in other tissues in response to stress, like the BM. The origin and homeostasis of human BM resident macrophages is presently ill-defined, if described at all. Elevated glucocorticoid levels may lead to direct differentiation of monocytes and elevated numbers of nursing central macrophages to facilitate the increased erythroid output in analogy to Kupffer cells. Active research is aimed at unraveling the origin of tissue resident macrophages, which is important in order to understand not only homeostatic but also pathogenic erythropoiesis in which a driving role of macrophages has been implicated, such as polycythemia vera and β-thalassemia. Herein, we provide evidence that monocytes can indeed differentiate in vitro to macrophages that support erythropoiesis, providing a model to study such erythroid-macrophage interactions. Funding This work was supported by grants from the Landsteiner Foundation (LSBR1141; EA and EH and LSBR1517; MB), a NHS Blood and Transplant (NHSBT) R&D grant (WP15-05; AMT), National Institute for Health Research (NIHR) for a Blood and Transplant Research Unit in Red Blood Cell Products at the University of Bristol in partnership with NHSBT (AMT and LAH-O) and the Wellcome Trust (105385/Z/14/Z; LAHO and ISSF; SJC). This article presents independent research partly funded by the NIHR. The views expressed are those of the authors and not necessarily the NHS, the NIHR or the Department of Health. Acknowledgments The authors would like to thank the staff of the CASA clinic in Leiden for collecting human fetal tissues and Dr. Tom Cupedo, Natalie Papazian and Martijn Bogaerts from the Erasmus Medical Center, Rotterdam, for providing human fetal liver material. We also thank the Central Facility of Sanquin for technical assistance regarding ImageStreamX.
3. Hanspal M, Smockova Y, Uong Q. Molecular identification and functional characterization of a novel protein that mediates the attachment of erythroblasts to macrophages. Blood. 1998;92(8):29402950. 4. Soni S, Bala S, Gwynn B, Sahr KE, Peters LL, Hanspal M. Absence of erythroblast macrophage protein (Emp) leads to failure of erythroblast nuclear extrusion. J Biol
Chem. 2006;281(29):20181-20189. 5. Hanspal M, Golan DE, Smockova Y, et al. Temporal synthesis of band 3 oligomers during terminal maturation of mouse erythroblasts. Dimers and tetramers exist in the membrane as preformed stable species. Blood. 1998;92(1):329-338. 6. Sawada K, Krantz SB, Dessypris EN, Koury ST, Sawyer ST. Human colony-forming units-erythroid do not require accessory
haematologica | 2018; 103(3)
Modeling human erythroid-macrophage interactions
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
cells, but do require direct interaction with insulin-like growth factor I and/or insulin for erythroid development. J Clin Invest. 1989;83(5):1701-1709. Chow A, Lucas D, Hidalgo A, et al. Bone marrow CD169+ macrophages promote the retention of hematopoietic stem and progenitor cells in the mesenchymal stem cell niche. J Exp Med. 2011;208(2):261-271. Chow A, Huggins M, Ahmed J, et al. CD169(+) macrophages provide a niche promoting erythropoiesis under homeostasis and stress. Nat Med. 2013;19(4):429436. Winkler IG, Sims NA, Pettit AR, et al. Bone marrow macrophages maintain hematopoietic stem cell (HSC) niches and their depletion mobilizes HSCs. Blood. 2010;116(23):4815-4828. Heideveld E, van den Akker E. Digesting the role of bone marrow macrophages on hematopoiesis. Immunobiology. 2017; 222(6):814-822. Iavarone A, King ER, Dai XM, Leone G, Stanley ER, Lasorella A. Retinoblastoma promotes definitive erythropoiesis by repressing Id2 in fetal liver macrophages. Nature. 2004;432(7020):1040-1045. Heideveld E, Masiello F, Marra M, et al. CD14+ cells from peripheral blood positively regulate hematopoietic stem and progenitor cell survival resulting in increased erythroid yield. Haematologica. 2015;100(11):1396-1406. Belay E, Hayes BJ, Blau CA, Torok-Storb B. Human cord blood and bone marrow CD34+ cells generate macrophages that support erythroid islands. PLoS One. 2017; 12(1):e0171096. Ramos P, Casu C, Gardenghi S, et al. Macrophages support pathological erythropoiesis in polycythemia vera and beta-thalassemia. Nat Med. 2013;19(4):437-445. Jacobsen RN, Perkins AC, Levesque JP. Macrophages and regulation of erythropoiesis. Curr Opin Hematol. 2015;22(3):212-219. Sadahira Y, Yoshino T, Monobe Y. Very late activation antigen 4-vascular cell adhesion molecule 1 interaction is involved in the formation of erythroblastic islands. J Exp Med. 1995;181(1):411-415. Ulyanova T, Phelps SR, Papayannopoulou T. The macrophage contribution to stress erythropoiesis: when less is enough. Blood. 2016;128(13):1756-1765. Toda S, Segawa K, Nagata S. MerTK-mediated engulfment of pyrenocytes by central macrophages in erythroblastic islands. Blood. 2014;123(25):3963-3971. Yoshida H, Kawane K, Koike M, Mori Y, Uchiyama Y, Nagata S. Phosphatidylserinedependent engulfment by macrophages of nuclei from erythroid precursor cells. Nature. 2005;437(7059):754-758. Lu Q, Lemke G. Homeostatic regulation of the immune system by receptor tyrosine kinases of the Tyro 3 family. Science. 2001;293(5528):306-311. Gaipl US, Voll RE, Sheriff A, Franz S, Kalden JR, Herrmann M. Impaired clearance of dying cells in systemic lupus erythematosus. Autoimmun Rev. 2005;4(4):189-194. Munoz LE, Janko C, Schulze C, et al.
haematologica | 2018; 103(3)
23. 24. 25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Autoimmunity and chronic inflammation two clearance-related steps in the etiopathogenesis of SLE. Autoimmun Rev. 2010;10(1):38-42. Nagata S. Apoptosis and autoimmune diseases. Ann N Y Acad Sci. 2010;1209:10-16. Rothlin CV, Lemke G. TAM receptor signaling and autoimmune disease. Curr Opin Immunol. 2010;22(6):740-746. Toda S, Nishi C, Yanagihashi Y, Segawa K, Nagata S. Clearance of apoptotic cells and pyrenocytes. Curr Top Dev Biol. 2015; 114:267-295. Ciudad MT, Sorvillo N, van Alphen FP, et al. Analysis of the HLA-DR peptidome from human dendritic cells reveals high affinity repertoires and nonconventional pathways of peptide generation. J Leukoc Biol. 2017;101(1):15-27. Vogt L, Schmitz N, Kurrer MO, et al. VSIG4, a B7 family-related protein, is a negative regulator of T cell activation. J Clin Invest. 2006;116(10):2817-2826. Finn AV, Nakano M, Polavarapu R, et al. Hemoglobin directs macrophage differentiation and prevents foam cell formation in human atherosclerotic plaques. J Am Coll Cardiol. 2012;59(2):166-177. Krause SW, Rehli M, Kreutz M, Schwarzfischer L, Paulauskis JD, Andreesen R. Differential screening identifies genetic markers of monocyte to macrophage maturation. J Leukoc Biol. 1996;60(4):540-545. Libreros S, Garcia-Areas R, Keating P, Carrio R, Iragavarapu-Charyulu VL. Exploring the role of CHI3L1 in "premetastatic" lungs of mammary tumor-bearing mice. Front Physiol. 2013;4:392. Hoban MD, Cost GJ, Mendel MC, et al. Correction of the sickle cell disease mutation in human hematopoietic stem/progenitor cells. Blood. 2015;125(17):2597-2604. Ulyanova T, Scott LM, Priestley GV, et al. VCAM-1 expression in adult hematopoietic and nonhematopoietic cells is controlled by tissue-inductive signals and reflects their developmental origin. Blood. 2005; 106(1):86-94. Lee G, Lo A, Short SA, et al. Targeted gene deletion demonstrates that the cell adhesion molecule ICAM-4 is critical for erythroblastic island formation. Blood. 2006; 108(6):2064-2071. Ulyanova T, Jiang Y, Padilla S, Nakamoto B, Papayannopoulou T. Combinatorial and distinct roles of alpha(5) and alpha(4) integrins in stress erythropoiesis in mice. Blood. 2011;117(3):975-985. Soni S, Bala S, Hanspal M. Requirement for erythroblast-macrophage protein (Emp) in definitive erythropoiesis. Blood Cells Mol Dis. 2008;41(2):141-147. Suenobu S, Takakura N, Inada T, et al. A role of EphB4 receptor and its ligand, ephrin-B2, in erythropoiesis. Biochem Biophys Res Commun. 2002;293(3):11241131. Ocana-Guzman R, Torre-Bouscoulet L, Sada-Ovalle I. TIM-3 regulates distinct functions in macrophages. Front Immunol. 2016;7:229. D'Souza S, Park SY, Kim IS. Stabilin-2 acts as an engulfment receptor for the phosphatidylserine-dependent clearance of pri-
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50. 51.
52.
mary necrotic cells. Biochem Biophys Res Commun. 2013;432(3):412-417. Simhadri VR, Andersen JF, Calvo E, Choi SC, Coligan JE, Borrego F. Human CD300a binds to phosphatidylethanolamine and phosphatidylserine, and modulates the phagocytosis of dead cells. Blood. 2012; 119(12):2799-2809. Dasgupta SK, Abdel-Monem H, Guchhait P, Nagata S, Thiagarajan P. Role of lactadherin in the clearance of phosphatidylserine-expressing red blood cells. Transfusion. 2008;48(11):2370-2376. Kawane K, Fukuyama H, Kondoh G, et al. Requirement of DNase II for definitive erythropoiesis in the mouse fetal liver. Science. 2001;292(5521):1546-1549. Humpeler E, Skrabal F, Bartsch G. Influence of exposure to moderate altitude on the plasma concentraton of cortisol, aldosterone, renin, testosterone, and gonadotropins. Eur J Appl Physiol Occup Physiol. 1980;45(2-3):167-176. van den Akker E, Satchwell TJ, Pellegrin S, Daniels G, Toye AM. The majority of the in vitro erythroid expansion potential resides in CD34- cells, outweighing the contribution of CD34+ cells and significantly increasing the erythroblast yield from peripheral blood samples. Haematologica. 2010;95(9):1594-1598. Leberbauer C, Boulme F, Unfried G, Huber J, Beug H, Mullner EW. Different steroids co-regulate long-term expansion versus terminal differentiation in primary human erythroid progenitors. Blood. 2005; 105(1):85-94. von Lindern M, Deiner EM, Dolznig H, et al. Leukemic transformation of normal murine erythroid progenitors: v- and cErbB act through signaling pathways activated by the EpoR and c-Kit in stress erythropoiesis. Oncogene. 2001;20(28):36513664. Bauer A, Tronche F, Wessely O, et al. The glucocorticoid receptor is required for stress erythropoiesis. Genes Dev. 1999; 13(22):2996-3002. Falchi M, Varricchio L, Martelli F, et al. Dexamethasone targeted directly to macrophages induces macrophage niches that promote erythroid expansion. Haematologica. 2015;100(2):178-187. Yokoyama T, Kitagawa H, Takeuchi T, Tsukahara S, Kannan Y. No apoptotic cell death of erythroid cells of erythroblastic islands in bone marrow of healthy rats. J Vet Med Sci. 2002;64(10):913-919. Yokoyama T, Etoh T, Kitagawa H, Tsukahara S, Kannan Y. Migration of erythroblastic islands toward the sinusoid as erythroid maturation proceeds in rat bone marrow. J Vet Med Sci. 2003;65(4):449-452. van Furth R, Cohn ZA. The origin and kinetics of mononuclear phagocytes. J Exp Med. 1968;128(3):415-435. Hashimoto D, Chow A, Noizat C, et al. Tissue-resident macrophages self-maintain locally throughout adult life with minimal contribution from circulating monocytes. Immunity. 2013;38(4):792-804. Theurl I, Hilgendorf I, Nairz M, et al. Ondemand erythrocyte disposal and iron recycling requires transient macrophages in the liver. Nat Med. 2016;22(8):945-951.
405
ARTICLE
Red Cell Biology & Its Disorders
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):406-416
miR-144/451 represses the LKB1/AMPK/mTOR pathway to promote red cell precursor survival during recovery from acute anemia
Xiao Fang,1,2,* Feiyang Shen,2,* Christophe Lechauve,3 Peng Xu,3 Guowei Zhao,3 Jacobi Itkow,2 Fan Wu,2 Yaying Hou,2 Xiaohui Wu,2,4 Lingling Yu,2,4 Huiqing Xiu,2 Mengli Wang,2 Ruiling Zhang,2 Fangfang Wang,2 Yanqing Zhang,2 Daxin Wang,1 Mitchell J. Weiss3 and Duonan Yu2,5,6,7
1 Clinical Medical College of Yangzhou University, Yangzhou, China; 2Jiangsu Key Laboratory of Experimental & Translational Non-Coding RNA Research, University School of Medicine, China; 3Department of Hematology, St. Jude Children's Research Hospital, Memphis, TN, USA; 4Department of Pediatrics, Jingjiang Peopleâ&#x20AC;&#x2122;s Hospital, Yangzhou University, Jingjiang, China; 5Institute of Comparative Medicine, Yangzhou University, China; 6Institute of Translational Medicine, Yangzhou University School of Medicine, Yangzhou, China and 7Jiangsu Co-Innovation Center for Prevention and Control of Important Animal Infectious Disease and Zoonosis, Yangzhou, China *
XF and FS contributed equally to this work.
ABSTRACT
T
Correspondence: dnyu@yzu.edu.cn
Received: July 27, 2017. Accepted: December 20, 2017. Pre-published: December 21, 2017.
he microRNAs miR-144 and -451 are encoded by a bicistronic gene that is strongly induced during red blood cell formation (erythropoiesis). Ablation of the miR-144/451 gene in mice causes mild anemia under baseline conditions. Here we show that miR-144/451-/- erythroblasts exhibit increased apoptosis during recovery from acute anemia. Mechanistically, miR-144/451 depletion increases the expression of the miR-451 target mRNA Cab39, which encodes a co-factor for the serine-threonine kinase LKB1. During erythropoietic stress, miR-144/451-/erythroblasts exhibit abnormally increased Cab39 protein, which activates LKB1 and its downstream AMPK/mTOR effector pathway. Suppression of this pathway via drugs or shRNAs enhances survival of the mutant erythroblasts. Thus, miR-144/451 facilitates recovery from acute anemia by repressing Cab39/AMPK/mTOR. Our findings suggest that miR-144/451 is a key protector of erythroblasts during pathological states associated with dramatically increased erythropoietic demand, including acute blood loss and hemolytic anemia.
doi:10.3324/haematol.2017.177394
Introduction
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/406
Dysregulation of microRNAs (miRNAs) is implicated in the pathophysiology of many human diseases including cancer, cardiovascular disease, and anemia.1,2 There is increasing evidence that miRNAs regulate red blood cell (RBC) formation (erythropoiesis) by controlling the proliferation and differentiation of RBC precursors, termed erythroblasts.3 For example, miR-126 negatively regulates erythropoiesis by repressing mRNA encoding the tyrosine phosphatase PTPN9, which is required for erythroblast proliferation,4 whereas ectopic expression of miR-27a, miR-24, and miR-146b in CD34+ hematopoietic progenitor cells promotes erythroid maturation by repressing GATA-2 or increasing GATA-1, which activates the GATA switch, a key step in erythropoiesis.5,6 However, the role of miRNAs and their targets in regulating erythropoiesis is not fully understood. The bicistronic miRNA locus encoding miRs-144 and -451 is strongly induced during erythropoiesis in zebrafish, mice, and humans.7-9 Chromatin immunoprecipitation (ChIP) and gene complementation studies show that miR-144/451 transcription is activated by GATA-1,10 a transcription factor that regulates many aspects of erythropoiesis, including precursor proliferation, maturation, and survival. Remarkably, miR-451 accounts for approximately 50% of the total miRNA pool in mouse fetal liver (FL) erythroblasts.11 Unlike most miRNAs, miR-451 biogenesis occurs independently of the RNA III enzyme Dicer. Rather, it is Argonaut 2 (Ago2)
Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
406
haematologica | 2018; 103(3)
miR-451 inhibits Cab39 for stress erythropoiesis
that catalyzes the cleavage of pre-miR-451 hairpins.12 Inhibition of miR-144/451 blocks erythropoiesis in tissue culture models.8,10,13-15 Fewer studies have been performed with in vivo models.13-15 Moreover, the phenotypes observed after manipulating miR-144/451 expression vary according to the model used and the mode of gene manipulation. For example, miR-144/451 inhibition appears to exert a greater effect on erythroblasts in culture than on those in vivo, suggesting that the phenotype depends on the cell environment (X Fang et al., unpublished data, 2017). Along with others, we have also demonstrated that miR144/451 gene knockout (KO) mice exhibit mild baseline anemia that worsens upon oxidative stress.13,15 Similarly, loss of miR-451 in zebrafish renders erythroid precursors sensitive to oxidant stress.15 The anti-oxidant role of miR144/451 during erythropoiesis is at least partially dependent on suppression of the mRNA target Ywhaz, which encodes the cytoplasmic adaptor protein 14-3-3ζ.13,15 miR144/451 depletion increases 14-3-3ζ protein, which sequesters the transcription factor FoxO3 in the cytoplasm, thereby reducing expression of several target genes that encode anti-oxidant proteins. This mechanism explains the hypersensitivity of miR-144/451−/− RBCs to oxidant stress but is unlikely to account for all the activities of these miRs. For example, miR-144/451−/− mice exhibit ineffective erythropoiesis at baseline and delayed recovery after anemia caused by oxidant stress13-15 via unknown mechanisms. We discovered that erythroblasts isolated from miR-144/451−/− FL during embryonic gestation or bone marrow and spleen during acute anemia exhibit increased apoptosis compared to wild-type (WT) counterparts. This effect is mediated by derepression of the direct miR-451 target mRNA Cab39 followed by activation of the downstream LKB1/AMPK/mTOR pathway. Thus, miR-144/451 enhances physiological responses to acute anemia by promoting the survival of RBC precursors.
Methods Animals miR-144/451 KO mice were described previously.15 p53ER knock-in (KI) mice were kindly provided by Gerard Evan (University of Cambridge, UK).16
Cell culture and treatment G1E and G1E-ER4 erythroid cells were grown in culture as previously described.17 The isolation of erythroid progenitors from embryonic day 14.5 (E14.5) FLs, the growth of erythroid progenitors in maturation medium, and the retroviral infection of erythroid cells in expansion medium have all been described previously.18 Details of drug treatments of cells are described in the Online Supplementary Appendix.
Protein and miRNA expression Western blot and real-time PCR analyses for gene expression were described in the Online Supplementary Appendix.
Fluorescence-activated cell sorting (FACS) The expression of RBC surface markers and cell death were analyzed with an LSRII or LSRFortessa Cell Analyzer System (BD Biosciences). The Annexin V Early Apoptosis Detection Kit (cat. n. 553786) was obtained from BD Biosciences. The nucleation of erythroid cells was quantitated by staining with Hoechst 33342 haematologica | 2018; 103(3)
(Sigma); cell viability was quantitated by staining cells with death markers 7AAD, propidium iodide (PI) or Live/Dead® Near-IR Fixable Dead Cell Stain (Invitrogen). Erythroid subpopulations were sorted on the basis of CD71/Ter119 expression (BD Biosciences).
Dual-luciferase reporter assay Construction of the plasmids for luciferase assays is described in the Online Supplementary Appendix. The dual-luciferase reporter assay was performed as previously described.15
Retroviral shRNA delivery The construction of retroviral plasmids and transduction of cells are described in the Online Supplementary Appendix.
Hematologic analysis For hematocrit (HCT) and reticulocyte counts, blood from adult mice was sampled retro-orbitally, anticoagulated with EDTA, and analyzed on a Hemavet HV950FS analyzer (Drew Scientific, Dallas, TX, USA). Heparinized glass microhematocrit tubes (Globe Scientific, Paramus, NJ, USA) were used for manual spun hematocrits. Reticulocytes were counted using Retic-COUNT reagent (BD Biosciences) or Ter119/CD71 staining and analyzed on a FACSCalibur Flow Cytometer (BD Biosciences). For phenylhydrazine (PHZ) treatment, mice were injected intraperitoneally (63 mg/kg). HCT and reticulocyte counts were analyzed for ten consecutive days thereafter. To eradicate erythroid progenitors in adult mice, 5-FU (SigmaAldrich) at a single dose of 150 mg/kg was injected intraperitoneally, and the HCT and reticulocyte counts were analyzed for 25 consecutive days thereafter. Animal survival was monitored every day. To generate acute anemia by bleeding, 400 μL of blood was drawn retro-orbitally every day. The HCT and reticulocyte counts were analyzed three days later.
Microarray analysis CD71+/Ter119+/FSChigh nucleated bone marrow cells from miR-144/451−/− mice and WT controls were purified by flow cytometry, and samples were processed for microarray analysis using the GeneChip Mouse Genome 430 2.0 Array (Affymetrix), as described previously.15 The database for G1E-ER4 erythroid maturation has been described in an earlier paper.19
Statistical analysis Statistical analyses were performed using Microsoft Office Excel 2011 (Microsoft Corporation, Redmond, WA, USA). Graphs were created using Adobe Photoshop CS6 (Adobe Systems Inc., San Jose, CA, USA). Data from triplicate experiments or 3 different samples are presented as the mean ± standard deviation. The differences were assessed by two-tailed Student t-tests. P<0.05 was considered statistically significant. All experiments were repeated at least 3 times.
Results Increased apoptosis of miR-144/451−/− erythroblasts during erythropoietic stress To study the effects of miR-144/451 on erythropoiesis, we cultured equal numbers of FL erythroid precursors from embryonic day (E) 14.5 KO or WT embryos in media that facilitated their expansion or terminal maturation18 (Figure 1A). After 48 hours (h) in expansion medium, miR-144/451−/− erythroblasts exhibited reduced cell num407
X. Fang et al.
bers (Figure 1B) and increased apoptosis, which was quantified by Annexin V staining (Figure 1C and D). As expected, KO erythroblasts lacked miR-451 expression (Figure 1E). Both WT and KO cells maintained their erythroblast identity after 48 h in culture (Figure 1F). Similarly to the
A
effects observed in expansion medium, miR-144/451−/− erythroid progenitors exhibited reduced cell numbers and increased apoptosis after 48 h culture in maturation medium (Figure 1G and H). Importantly, erythroid precursors isolated directly from miR-144/451−/− E14.5 FLs were
B
C D
E F
H
G
J I
Figure 1. Increased apoptosis of miR-144/451−/− erythroblasts during erythropoietic stress. (A) miR-144/451−/− progenitor cells were isolated from E14.5 fetal liver (FL) erythroblasts and grown in culture in expansion medium or maturation medium. (B) Cell proliferation rates of FL erythroid blasts in expansion medium for 24 and 48 hours (h). n=3 FLs. **P<0.01 (t-test). Experiments were repeated 3 times. (C) Flow cytometry-based analysis of apoptosis using Annexin V. (D) Quantitative analysis of the flow cytometry data from (C). Data from 3 independent experiments. **P<0.01 (t-test). (E) Quantitative PCR analysis of miR-451 expression in FL progenitors grown in expansion medium for different lengths of time. Progenitor cells from miR-144/451−/− FLs, mouse B-lymphoma cell line 38B9, and normal B lymphocytes sorted from mouse bone marrow were used as negative controls. Note: the high miR-451 expression in the 0-h culture suggests the erythroid identity of the progenitors, whereas the lack of a significant increase in miR-451 after 48 h in culture suggests that there is no further differentiation of erythroid progenitors grown in culture. (F) FL progenitors grown in expansion medium for 48 h. Cells were cytospun onto slides and stained with May-Grunwald-Giemsa. (G) Cell proliferation rates of FL erythroid cells in maturation medium for 24 and 48 h. N=3. **P<0.01 (t-test). (H) Percentage of apoptotic cells in maturation medium based on flow cytometric analysis. Data represent 3 independent experiments. **P<0.01 (t-test). (I) Cell numbers in whole E14.5 FLs without in vitro expansion. WT n=10, KO n=11. *P<0.05 (t-test). (J) Percentage of apoptotic cells in different regions gated by CD71/Ter119 staining and FSC intensity. WT n=7, KO n=6. *P<0.05 (t-test).
408
haematologica | 2018; 103(3)
miR-451 inhibits Cab39 for stress erythropoiesis
reduced in number and exhibited increased apoptosis compared to controls (Figure 1I and J). However, we detected no increase in apoptosis of erythroblasts isolated directly (not cultured) from spleen or bone marrow from adult miR-144/451−/− mice (Online Supplementary Figure S1A and B). Fetal liver erythropoiesis is considered a “stress
A
state” because production demands are extremely high compared to steady state bone marrow erythropoiesis in adults.20 Thus, miR-144/451 may protect erythroblasts from apoptosis during erythropoietic stress associated with increased demands for RBC production. Consistent with this, we had previously noted that recovery from
B
C
D
E F
G H
I
Figure 2. Apoptotic erythroblasts in miR-144/451−/− bone marrow increase after administration of 5-fluorouracil (5-FU) in adult mice. (A) Survival rates after 5-FU treatment for n=38 (wild-type, WT) and n=36 (miR-144/451−/− , KO) mice. (B) Increased hemolysis of miR-144/451−/− erythrocytes after exposure to 5-FU, as determined by serial hematocrit measurement. N=12 miR-144/451−/− and n=12 WT mice were used. *P<0.05; **P<0.01 (t-test). (C) Flow cytometric analysis of Ter119+CD71+ reticulocytes in circulating blood after 5-FU treatment. N=5 mice of each genotype were used. **P<0.01 (t-test). Note: for the first eight days there were relatively more reticulocytes in miR-144/451−/− blood as compared to WT blood, but significantly fewer during days 8 to 12. The number of reticulocytes in miR144/451−/− blood dramatically increased around day 14, and much higher levels were sustained than in WT animals. (D) Flow cytometric analysis of Ter119+CD71+ erythroid cells in bone marrow after 5-FU administration for 7-11 days. Note: the appearance of Ter119+CD71+ erythroid cells in bone marrow was delayed relative to that in WT mice, indicating a maturation arrest and/or sustained apoptosis of erythroid cells in miR-144/451−/− mice. (E) Quantitated analysis of Ter119+CD71+ erythroid cells in bone marrow after 5-FU administration for 7-11 days. There were n=5 mice of each genotype at each time point. **P<0.01 (t-test). (F) Flow cytometric analysis of nucleated cell numbers in bone marrow after 5-FU administration. All cells shown in the windows were Ter119+. Gated regions represent nucleated erythroblasts (Hoechst+FSChigh). (G) Quantitative analysis of flow cytometry data from (F). There were n=5 mice of each genotype at each time point. **P<0.01 (ttest). Note: there were far fewer nucleated erythroid cells in miR-144/451−/− bone marrow during days 9 to 11, indicating a maturation arrest and/or sustained apoptosis. (H) Flow cytometric analysis of early apoptosis using Annexin V. Ter119+/Hoechst+ cells are nucleated erythroid cells (left). Near-IR cell death marker−/Annexin V+ cells are early apoptotic cells (right). (I) Quantitative analysis of flow cytometry data from (H). We used n=5 mice of each genotype at day 11 after 5-FU treatment. **P<0.01 (t-test).
haematologica | 2018; 103(3)
409
X. Fang et al.
(PHZ)-induced hemolytic anemia was delayed in miR144/451−/− mice.15 To examine further the effects of erythropoietic stress on adult miR-144/451−/− mice, we treated them with a single dose of 5-fluorouracil (5-FU), which destroys committed hematopoietic progenitors, including erythroblasts. Compared to WT mice, miR-144/451−/− mice exhibited increased mortality after 5-FU treatment (55.6% vs. 86.8% survival) (Figure 2A), which was associated with a greater decline in their hematocrit (Figure 2B). During the recovery phase after 5-FU treatment, the mutant mice exhibited higher levels of late stage circulating erythroid precursors (reticulocytes) (Figure 2C), probably in response to the more severe anemia (Figure 2B). However, the time to maximal reticulocyte response was delayed by several days in KO mice compared to WT controls (Figure 2C). Similarly, the emergence of bone marrow erythroid precursors was delayed in KO mice (Figure 2D-G). Moreover, bone marrow erythroblasts of miR-144/451−/− mice with 5FU treatment exhibited increased apoptosis compared to WT controls (Figure 2H and I). We also observed increased apoptosis of miR-144/451−/− bone marrow and splenic erythroblasts during recovery from PHZ-induced anemia (Figure 3A and B) or phlebotomy (Figure 3C and D). Thus, miR-144/451−/− erythroblasts exhibit increased apoptosis
under conditions of increased physiological demand for RBC production, i.e. under erythropoietic stress.
miR-451 targets Cab39 mRNA in erythroblasts Several mRNAs previously identified as miR-451 target mRNAs, including Myc, Ywhaz/14-3-3ζ, and Cab39 (Figure 4A), encode general regulators of cell survival, proliferation, and maturation.13,15,21 Cab39 is an obligatory cofactor for the serine/threonine kinase LKB1, a tumor suppressor that regulates responses to metabolic stress, in part by activating AMP-activated protein kinase (AMPK).22 miR-451 drives human glioma cell expansion by inhibiting this pathway via direct repression of Cab39.21 Therefore, we investigated whether miR-451 repression of Cab39 regulates erythroblast survival during erythropoietic stress. Compared to controls, Cab39 mRNA and protein were up-regulated in miR-144/451−/− erythroblasts in spleen, bone marrow (Figure 4B and C) and FL (Figure 4D) compared to WT erythroblasts in the same tissues. Retroviral vector-mediated expression of miR-451 in the erythroid cell line G1E17 reduced Cab39 protein by approximately 50% (Figure 4E). The seed sequence of miR-451 is complementary to a conserved sequence within the 3′ untranslated region (UTR) of human and mouse Cab39 mRNA (Figure 4A). To verify whether miR-451 inhibited Cab39
A
B
C
D
Figure 3. Apoptosis of miR-144/451−/− bone marrow and spleen erythroblasts is increased under other stress conditions. (A and B) Results of flow cytometric analysis of nucleated erythroid cell apoptosis in spleen (A) and bone marrow (B) after phenylhydrazine (PHZ) administration. miR-144/451−/− mice (n=5) and wild-type (WT) littermates (n=5) were treated with PHZ (120 mg/kg) on days 0 and 1. Splenocytes and bone marrow cells were harvested four days after PHZ injection. Ter119+Hoechst+ cells are nucleated erythroid cells. Annexin V+PI–near IR− cells are early apoptotic cells. Ter119– cells are non-erythroid cells. **P<0.01 (t-test). (C and D) Flow cytometric analysis of early apoptosis rates of spleen (C) and bone marrow (D) nucleated erythroblasts after bleeding. miR-144/451−/− mice (n=5) and WT littermates (n=5) had 0.4 mL of blood drawn on two consecutive days. Splenocytes and bone marrow cells were harvested on day 3. **P<0.01 (t-test). Note: more early apoptotic erythroblasts were seen in miR-144/451−/− mice.
410
haematologica | 2018; 103(3)
miR-451 inhibits Cab39 for stress erythropoiesis
mRNA expression via direct interaction with this region, we fused the 3′-UTR of Cab39 mRNA to the coding sequence of luciferase cDNA (Figure 4F). In 293T cells, luciferase reporter activity was inhibited approximately 200-fold after co-expression of miR-451. Mutations in the Cab39 mRNA 3’ UTR that disrupt complementarity to the miR-451 seed sequence abrogated repression of reporter activity. Together, these findings verify that miR-451 inhibits Cab39 mRNA expression directly and that this interaction occurs during erythropoiesis.
Activation of the Cab39/AMPK/mTOR pathway in miR-144/451−/− erythroblasts Cab39 binds LKB1 and STRAD to activate AMPK by phosphorylating the protein at Thr172 (Figure 5A).22,23 Of note, LKB1 and AMPK mRNAs are up-regulated during normal mouse and human erythroid maturation (Online Supplementary Figure S2A-E).19,24 Cab39 was up-regulated in cultured miR-144/451−/− FL erythroblasts compared to controls (Figures 4D and 5B and C). These cells also exhibited strongly increased phosphorylation of AMPK at Thr172, and to a lesser extent, upregulation of total AMPK protein (Figure 5B and D). AMPK can inhibit the mTOR pathway to inhibit cell growth and either induce or suppress apoptosis, depending on cellular context.25 To investigate this
A
B
C D
E
F
haematologica | 2018; 103(3)
mechanism in miR-144/451−/− erythroblasts, we performed Western blot studies to interrogate upstream and downstream effectors.26,27 miR-144/451−/− FL erythroblasts grown in culture exhibited elevated phospho-Raptor and phospho-TSC2, along with reduced phosphorylation of p70S6K, S6, and eIF4B (Figure 5B and E-I), consistent with suppression of the mTOR by activated AMPK (Figure 5A). We observed similar patterns in primary erythroblasts from E14.5 FLs (Online Supplementary Figure S3). Together, these findings indicate that the loss of miR-144/451 during FL erythropoiesis derepresses the miR-451 target Cab39, resulting in mTOR repression.
Attenuation of Cab39/AMPK/mTOR signaling rescues erythroid apoptosis after miR-144/451 depletion To determine whether miR-144/451 regulates survival of FL erythroblasts via Cab39/AMPK/mTOR, we designed shRNA-expressing retroviruses to knock down components of this pathway. We grew FL erythroblasts in expansion medium, infected them with individual shRNA retroviruses, induced erythroid maturation for 24 h, and then examined apoptosis in the cells. Inhibiting Cab39 expression in miR-144/451−/− erythroblasts by approximately 5570% with 2 different shRNAs reduced apoptosis by approximately 45% (Figure 6A-C). In contrast, no signifi-
Figure 4. miR-451 targets Cab39 mRNA in erythroblasts. (A) Nucleotide sequence alignments showing partial complementarity between the mouse and human 3′-UTRs of Cab39 mRNAs and miR-451. The miR-451 seed-sequence recognition sites in the mRNAs are boxed in white. (B) Fold change in the Cab39 mRNA level from quantitative RT-PCR analysis. Ter119+ nucleated cells were sorted from bone marrow, and a microarray was performed using an Affymetrix GeneChip.15 N=3 mice were used. **P<0.01 (t-test). (C) Cab39 protein levels in both spleen and bone marrow miR-144/451−/− erythroblasts were increased relative to those in wild-type (+/+, WT) controls (top). Ter119+ cells were purified from bone marrow, and whole-cell lysates were analyzed via Western blots with the Cab39 antibody. Representative results for 2 mice of each genotype are shown. (Bottom) Quantitative image analysis from multiple experiments. (D) Cab39 protein levels in miR-144/451−/− fetal liver (FL) erythroblasts were increased relative to those in WT controls (top). (Bottom) Quantitative image analysis for 3 WT and 6 miR-144/451−/− (KO) FLs. (E) G1E proerythroblast cells were transduced with retrovirus encoding miR451 or empty vector as control. Transduced cells were selected with puromycin and analyzed for Cab39 via Western blots. (Top) Results of a representative experiment and (Bottom) of a quantitative analysis of the Western blot signal intensity from 3 independent experiments. **P<0.01 (t-test). (F) Interaction between miR-451 and the Cab39 3′-UTR inhibits the expression of a linked reporter gene. Firefly luciferase cDNA was fused to the normal 3′UTR of Cab39 cDNA or a mutant version (mt) containing a 3-bp mutation within the region complementary to the miR-451 seed sequence. The reporter constructs were cloned into an expression vector and transfected into 293T cells, along with an miR-451 expression construct and a constitutively active Renilla luciferase control plasmid. Luciferase activities were determined 24 hours post transfection. (Bottom) Bars represent the Firefly/Renilla luciferase activity; levels from the reporter vector lacking the Cab39 3′UTR were assigned an arbitrary value of 1. Results are given as the average of 3 separate experiments. **P<0.01 (t-test). Hsa: human; Mmu: mouse.
411
X. Fang et al.
cant change in apoptosis of WT erythroblasts occurred after shRNA suppression of Cab39 (Figure 6B and C). Similarly, shRNA inhibition of AMPKα and TSC2 inhibited apoptosis of miR-144/451−/− erythroblasts but not WT controls (Figure 6D-G). To further examine the effects of AMPK/mTOR signaling during WT and miR-144/451−/− FL erythropoiesis, we used drugs to manipulate the pathway. Consistent with the results of shRNA studies, the AMPK inhibitor Compound C (CC) reduced apoptosis significantly in miR144/451−/−, but not WT erythroblasts (Online Supplementary Figure S4A and B). Conversely, inhibiting mTOR activity with rapamycin or activating AMPK with AICAR induced apoptosis in WT erythroblasts (Online Supplementary Figure S4C-F). Overall, our results with shRNAs and pharmacological inhibitors indicate that miR-451 facilitates fetal erythroblast survival by inhibiting expression of Cab39, resulting in suppression of LKB1 and AMPK and activation of the downstream mTOR pathway.
A
C
F
H
412
The increased apoptosis of miR-144/451−/− erythroblasts is p53-dependent Depending on cell context, mTOR can regulate p53 positively or negatively to alter rates of apoptosis.28-30 Thus, we investigated p53 levels and effector functions in miR144/451−/− erythroblasts that exhibit reduced mTOR activity. p53 level was increased in miR-144/451−/− E14.5 FL erythroblasts (Figure 7A). To investigate the functional implications of this finding, we crossed miR-144/451−/− mice with “p53 knock-in (KI)” mice, in which both alleles of the normal p53 gene are replaced by a cDNA encoding a 4hydroxytamoxifen (4-OHT)-dependent form of the fusion protein, p53-estrogen receptor.16,31 There is a lack of endogenous p53 activity in the p53 KI mice unless 4-OHT is applied. In the absence of 4-OHT, loss of p53 function rescued the deficient erythroblast numbers in E14.5 FL from miR-144/451−/− mice (Figure 7B and C). Moreover, apoptosis of E14.5 miR-144/451−/− FL erythroblasts was significantly reduced in the absence of p53 activity (Figure
B
D
E
G
I
Figure 5. Activation of the Cab39/AMPK/mTOR pathway in miR-144/451−/− erythroblasts. (A) Schematic illustration of the major signaling components in the Cab39/AMPK/mTOR molecular pathway. Cab39 expression leads to increased LKB1 activity in the Cab39/LKB1/STRAD protein complex, which activates its substrate AMPK. AMPK activation is vital for modulation of the mTOR signaling cascade and for potential stabilization of p53 activity, which subsequently governs the fate of the target cells, including their apoptosis, survival, and/or proliferation. (B) Western blot analysis of the expression of Cab39, p-AMPKα, p-Raptor, pTSC2, p-p70S6K, p-S6, and p-eIF4B, along with their non-phosphorylated counterparts, in fetal liver (FL) erythroblasts grown in maturation medium for 24 hours, when all the cells were nucleated erythroblasts. 14-33ζ was used as a positive control, and actin was used as a sample loading control. Quantitative analyses of the protein intensity are shown in (CI) as follows: (C) Cab39, (D) pAMPKα/AMPKα, (E) pRaptor/Raptor, (F) p-TSC2/TSC2, (G) p-p70S6K/p70S6K, (H) p-S6/S6, and (I) p-eIF4B/eIF4B. Signals were normalized to actin. Data are the mean values from 3 separate experiments. *P<0.05; **P<0.01 (t-test).
haematologica | 2018; 103(3)
miR-451 inhibits Cab39 for stress erythropoiesis
7D and E). Together, these results demonstrate that increased apoptosis of miR-144/451−/− erythroblasts is p53dependent.
Discussion Although the biological functions of miR-144/451 have been studied extensively, few studies have been performed in animal models. Moreover, less is known about the miR-144/451-regulated molecular pathways underlying the phenotypes observed after miR-144/451 depletion. Our results provide further explanation, in addition to that of elevated oxidative stress caused by aberrant 14-3-3ζ accumulation and consequent FoxO3 sequestration, for the defective erythropoiesis and hemolytic anemia seen in miR-144/451−/− mice.13,15 Upon various erythropoietic stresses, miR-144/451 depletion up-regulated expression of the miR-451 target Cab39 with consequent activation of
A
the AMPK/mTOR pathway, leading to increased erythroid apoptosis. Manipulating both AMPK and mTOR activities altered the apoptotic rate in miR-144/451−/− erythroblasts. In contrast, reduced miR-451 levels in glioma cells as an adaptation to metabolic stress derepress Cab39 to activate the LKB1/AMPK/mTOR pathway and enhance survival.21 Thus, miR-451 may employ a common pathway to regulate stress responses in different cell types, but the net effects are context-dependent. During erythropoiesis, mTOR activity is relatively high, and the pathway appears to exert a positive effect on precursor expansion and protection against erythropoietic stress,32 consistent with the current study. Contrasting effects of AMPK activity on apoptosis have been observed during various cellular stresses.25 In some cases, AMPK functions to balance cellular redox state and promote survival during metabolic or genotoxic stress.33,34 In other cases, AMPK activation during stress causes increased apoptosis.35,36 In this study, we showed that in
B
Cab39 shRNA
C
D
E F
G
haematologica | 2018; 103(3)
Figure 6. Attenuation of Cab39/AMPK/mTOR signaling rescues erythroid apoptosis after miR144/451 depletion. (A) shRNA suppression of Cab39 in miR-144/451−/− erythroblasts as shown by Western blot. (B) Apoptosis was measured by flow cytometry, using Annexin V as an indicator. Cab39 shRNA-expressing nucleated erythroblasts were indicated as Ter119+Hoechst+ cells. (C) The apoptotic index in nucleated erythroid cells after knockdown of Cab39. Luciferase (luc) shRNA was used as a negative knockdown control. **P<0.01 (t-test). N=5 fetal liver (FL) erythroblasts of each genotype were used. (D) shRNA suppression of AMPKα in miR-144/451−/− erythroblasts, as shown by Western blot. (E) The percentage of apoptotic nucleated erythroblasts after knockdown of AMPKα. **P<0.01 (t-test). N=4 FLs of each genotype were used. (F) shRNA suppression of TSC2 in miR-144/451−/− erythroblasts, as shown by Western blot. (G) The apoptotic rates of nucleated erythroblasts after knockdown of TSC2. **P<0.01 (t-test). N=4 FLs of each genotype were used.
413
X. Fang et al.
miR-144/451−/− erythroblasts, AMPK is activated by overexpressed Cab39 and promotes apoptosis by inhibiting mTOR activity. It is unclear how levels of stress affect the regulatory role of AMPK in apoptosis, i.e. does AMPK favor cell survival under mild stress but enhance apoptosis under severe stress? In this study, we found that loss of miR-144/451 induces apoptosis in a p53-dependent fashion. Some evidence from a study in Diamond-Blackfan anemia (DBA) has shown that haploinsufficiency of ribosomal protein subunits can induce erythroblast apoptosis via p53-dependent mechanisms.37 Stabilization of p53 by the MDM2 antagonist SAR405838 induces major hematopoietic defects including erythroid precursor apoptosis in vitro.38 How p53 is activated after miR-144/451 depletion remains to be elucidated. In some cell types, AMPK induces apoptosis by activating p53 through phosphorylation,34,35 whereas others have reported that p53 activates AMPK/mTOR signaling to suppress cell growth by targeting sestrin 1 and sestrin 2 upon stress.39 In addition, several groups report an inhibitory role of mTOR activity on p53 function.28, 29 It is possible that a positive feedback loop involving AMPK, p53, and mTOR signaling regulates apoptosis of miR144/451−/− erythroid cells under stress conditions, although further investigations are required to fully define the process. Anemia is a common complication after orthotopic kidney transplantation due to multifactorial effects including iron deficiency, reduced erythropoietin production, and chronic or acute inflammation. Additionally, mTOR inhibi-
A
tion by sirolimus is a possible risk factor for the development of anemia in kidney transplant recipients.40,41 There is also a substantial risk of anemia from the mTOR inhibitor everolimus in cancer therapy.42 Interestingly, studies in animal models and cell cultures demonstrate that treatment with mTOR inhibitors reduces RBC size, independent of alterations in kidney function.32,43 Consistent with this finding, miR-144/451−/− mice exhibit microcytic anemia,15 perhaps due to mTOR inhibition. Moreover, the current study shows that these mice exhibit enhanced erythroblast apoptosis with various erythropoietic stresses including developmental expansion of FL, hemolysis, acute blood loss, and precursor depletion by 5-FU, a chemotherapeutic drug. Accelerated apoptosis in the absence of miR-144/451 is caused by overexpression of the miR-451 target Cab39, which activates AMPK, thus inhibiting mTOR signaling. Therefore, our data explain why erythroid apoptosis is an underlying mechanism of profound anemia when the enzymatic activity of Cab39/LKB1/AMPK is significantly increased and mTOR signaling is perturbed. A mechanistic understanding of the differences between steady state and stress erythropoiesis could be of benefit in multiple clinical settings.44 Moreover, these processes are likely to be impacted differently by various disease states. Common causes of acquired anemia include dietary iron deficiency, malaria, chronic infectious diseases, autoimmune or rheumatological disorders, chemotherapy, and chronic kidney disease. Genetic causes of anemia include DBA and hemoglobinopathies such as sickle cell disease (SCD) and thalassemia.45 Of note,
B
C D
E
414
Figure 7. The increased apoptosis of miR-144/451−/− erythroblasts is p53-dependent. (A) Western blot showing the protein levels of the tumor suppressor p53 in wild-type (WT) and miR-144/451−/− (KO) E14.5 fetal liver (FL) cells. Lineage negative-selected WT and KO E14.5 FL cells were grown in culture in erythroid maturation medium, and cells were harvested for Western blot analysis for p53 after 24 hours in culture. Actin was used as the loading control. (B) Gross view of the cell pellets from miR-144/451−/− and miR-144/451−/−/p53 knock-in (KI) double-mutant E14.5 FLs. (C) Total erythroblast number for whole E14.5 FLs from miR-144/451−/− and miR-144/451−/−/p53 KI double-mutant mice. miR-144/451−/− mice were crossed with p53-deficient KI mice.16,31 N=3 WT mice, n=6 KO mice, and n=6 double-mutant mice were used. **P<0.01 (t-test). (D) Flow cytometric analysis of apoptosis for miR-144/451−/− and miR-144/451−/−/p53 KI double-mutant E14.5 FL cells. WT FLs were used as controls. PI−Annexin V＋ labeling was taken to indicate early apoptotic cells. (E) Quantitative analysis of flow cytometric data from (D).
haematologica | 2018; 103(3)
miR-451 inhibits Cab39 for stress erythropoiesis
while numerous studies, including the current one, reveal a positive effect of mTOR on erythropoiesis, other studies show that mTOR inhibition may have beneficial effects in some forms of anemia including SCD and thalassemia,46,47 again emphasizing context- or disease-dependent functions for this pathway. Our mouse model makes it possible to investigate further the roles of miR-144/451, including its effects on mTOR, in physiological adaptations to various red cell disorders. miR-144/451 is a bicistronic gene whose expression is directly controlled by GATA1 in erythroid cells.10,15 Interestingly, our research and that of others had previously observed that: 1) miR-144 level is always lower than miR-451 level in both fetal and adult erythroid cells;10,14 and 2) the expression of miR-144 is ubiquitous whereas the expression of miR-451 is much more constrained in hematopoietic compartments during embryonic development.48,49 These data suggest that overlapping and independent mechanisms regulate the differential expression of miR-144 and miR-451. The current study focuses on an miR-451-dependent mechanism for regulating stress erythropoiesis. However, it is not clear whether miR-144 impacts this process. Of note, suppression of miR-144
References 1. Hosokawa K, Muranski P, Feng X, et al. Identification of novel microRNA signatures linked to acquired aplastic anemia. Haematologica. 2015;100(12):1534-1545. 2. Mendell JT, Olson EN. MicroRNAs in stress signaling and human disease. Cell. 2012;148(6):1172-1187. 3. Hattangadi SM, Wong P, Zhang L, Flygare J, Lodish HF. From stem cell to red cell: regulation of erythropoiesis at multiple levels by multiple proteins, RNAs, and chromatin modifications. Blood. 2011;118(24):62586268. 4. Huang X, Gschweng E, Van Handel B, Cheng D, Mikkola HK, Witte ON. Regulated expression of microRNAs126/126* inhibits erythropoiesis from human embryonic stem cells. Blood. 2011; 117(7):2157-2165. 5. Wang F, Zhu Y, Guo L, et al. A regulatory circuit comprising GATA1/2 switch and microRNA-27a/24 promotes erythropoiesis. Nucleic Acids Res. 2014;42(1):442-457. 6. Zhai PF, Wang F, Su R, et al. The regulatory roles of microRNA-146b-5p and its target platelet-derived growth factor receptor α (PDGFRA) in erythropoiesis and megakaryocytopoiesis. J Biol Chem. 2014; 289(33):22600-22613. 7. Fu YF, Du TT, Dong M, et al. Mir-144 selectively regulates embryonic α-hemoglobin synthesis during primitive erythropoiesis. Blood. 2008;113(6):1340-1349. 8. Pase L, Layton JE, Kloosterman WP, Carradice D, Waterhouse PM, Lieschke GJ. miR-451 regulates zebrafish erythroid maturation in vivo via its target gata2. Blood. 2009;113(8):1794-1804. 9. Zhan M, Miller CP, Papayannopoulou T, Stamatoyannopoulos G, Song CZ. MicroRNA expression dynamics during murine and human erythroid differentiation. Exp Hematol. 2007;35(7):1015-1025. 10. Dore LC, Amigo JD, Dos Santos CO, et al.
haematologica | 2018; 103(3)
11.
12.
13.
14.
15.
16.
17.
18.
19.
20. 21.
inhibits erythropoiesis in cultured human CD34+ cells.50 In future studies, it will be interesting to investigate the effects of miR-144 and miR-451 on erythropoiesis separately by examining mice that harbor single miR-specific mutations. Acknowledgments The authors would like to thank Keith Laycock (St. Jude Children's Research Hospital) and Yuanjun Yu for language editing. Funding This work is supported by the National Natural Science Foundation of China (grant n. 81470277 to DY), a grant from the Ministry of Finance of China for the “Biology & Medical Science Innovation Team” program (to DY), and the Priority Academic Program Development of Jiangsu Higher Education Institution (Veterinary Medicine) (to DY) and R01 DK092318 (to MW). XF is also supported by the National Natural Science Foundation of China (grant n. 81402484), the Jiangsu Provincial Natural Science Foundation (grant n. BK20140497) and Natural Science Fund for Colleges and Universities in Jiangsu Province (grant n. 14KJB310024).
A GATA-1-regulated microRNA locus essential for erythropoiesis. Proc Natl Acad Sci USA. 2008;105(9):3333-3338. Zhang L, Flygare J, Wong P, Lim B, Lodish HF. miR-191 regulates mouse erythroblast enucleation by down-regulating Riok3 and Mxi1. Genes Dev. 2011;25(2):119-124. Yang JS, Maurin T, Robine N, et al. Conserved vertebrate mir-451 provides a platform for Dicer-independent, Ago2mediated microRNA biogenesis. Proc Natl Acad Sci USA. 2010;107(34):15163-15168. Patrick DM, Zhang CC, Tao Y, et al. Defective erythroid differentiation in miR451 mutant mice mediated by 14-3-3zeta. Genes Dev. 2010;24(15):1614-1619. Rasmussen KD, Simmini S, AbreuGoodger C, et al. The miR-144/451 locus is required for erythroid homeostasis. J Exp Med. 2010;207(7):1351-1358. Yu D, dos Santos CO, Zhao G, et al. miR451 protects against erythroid oxidant stress by repressing 14-3-3ζ. Genes Dev. 2010;24(15):1620-1633. Yu D, Carroll M, Thomas-Tikhonenko A. p53 status dictates responses of B lymphomas to monotherapy with proteasome inhibitors. Blood. 2007;109(11):4936-4943. Gregory T, Yu C, Ma A, Orkin SH, Blobel GA, Weiss MJ. GATA-1 and erythropoietin cooperate to promote erythroid cell survival by regulating bcl-xL expression. Blood. 1999;94(1):87-96. Khandros E, Thom CS, D'Souza J, Weiss MJ. Integrated protein quality-control pathways regulate free α-globin in murine β-thalassemia. Blood. 2012;119(22):52655275. Cheng Y, Wu W, Kumar SA, et al. Erythroid GATA1 function revealed by genome-wide analysis of transcription factor occupancy, histone modifications, and mRNA expression. Genome Res. 2009;19(12):2172-2184. Palis J, Segel GB. Developmental biology of erythropoiesis. Blood Rev. 1998;12(2): 106-114. Godlewski J, Nowicki MO, Bronisz A, et
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
al. MicroRNA-451 regulates LKB1/AMPK signaling and allows adaptation to metabolic stress in glioma cells. Mol Cell. 2010; 37(5):620-632. Shackelford DB, Shaw RJ. The LKB1AMPK pathway: metabolism and growth control in tumour suppression. Nat Rev Cancer. 2009;9(8):563-575. Hawley SA, Boudeau J, Reid JL, et al. Complexes between the LKB1 tumor suppressor, STRAD α/β and MO25 α/β are upstream kinases in the AMP-activated protein kinase cascade. J Biol. 2003;2(4):28. An X, Schulz VP, Li J, et al. Global transcriptome analyses of human and murine terminal erythroid differentiation. Blood. 2014; 123(22):3466-3477. Dasgupta B, Chhipa RR. Evolving lessons on the complex role of AMPK in normal physiology and cancer. Trends Pharmacol Sci. 2016;37(3):192-206. Ginion A, Auquier J, Benton CR, et al. Inhibition of the mTOR/p70S6K pathway is not involved in the insulin-sensitizing effect of AMPK on cardiac glucose uptake. Am J Physiol Heart Circ Physiol. 2011; 301(2):H469-H477. Laplante M, Sabatini DM. mTOR signaling in growth control and disease. Cell. 2012; 149(2):274-293. Kwon S, Jeon JS, Ahn C, Sung JS, Choi I. Rapamycin regulates the proliferation of Huh7, a hepatocellular carcinoma cell line, by up-regulating p53 expression. Biochem Biophys Res Commun. 2016;479(1):74-79. Lee JY, Nakada D, Yilmaz OH, et al. mTOR activation induces tumor suppressors that inhibit leukemogenesis and deplete hematopoietic stem cells after Pten deletion. Cell Stem Cell. 2010;7(5):593-605. Wang C, Gao D, Guo K, et al. Novel synergistic antitumor effects of rapamycin with bortezomib on hepatocellular carcinoma cells and orthotopic tumor model. BMC Cancer. 2012;12:166. Christophorou MA, Martin-Zanca D, Soucek L, et al. Temporal dissection of p53
415
X. Fang et al.
32.
33.
34.
35.
36.
37.
416
function in vitro and in vivo. Nat Genet. 2005;37(7):718-726. Knight ZA, Schmidt SF, Birsoy K, Tan K, Friedman JM. A critical role for mTORC1 in erythropoiesis and anemia. Elife. 2014; 3:e01913. Jeon SM, Chandel NS, Hay N. AMPK regulates NADPH homeostasis to promote tumour cell survival during energy stress. Nature. 2012;485(7400):661-665. Jones RG, Plas DR, Kubek S, et al. AMPactivated protein kinase induces a p53dependent metabolic checkpoint. Mol Cell. 2005;18(3):283-293. Okoshi R, Ozaki T, Yamamoto H, et al. Activation of AMP-activated protein kinase induces p53-dependent apoptotic cell death in response to energetic stress. J Biol Chem. 2008;283(7):3979-3987. Tzatsos A, Tsichlis PN. Energy depletion inhibits phosphatidylinositol 3-kinase/Akt signaling and induces apoptosis via AMPactivated protein kinase-dependent phosphorylation of IRS-1 at Ser-794. J Biol Chem. 2007;282(25):18069-18082. Chakraborty A, Uechi T, Kenmochi N. Guarding the 'translation apparatus': defective ribosome biogenesis and the p53 signaling pathway. Wiley Interdiscip Rev
RNA. 2011;2(4):507-522. 38. Mahfoudhi E, Lordier L, Marty C, et al. P53 activation inhibits all types of hematopoietic progenitors and all stages of megakaryopoiesis. Oncotarget. 2016;7(22):3198031992. 39. Budanov AV, Karin M. p53 target genes sestrin1 and sestrin2 connect genotoxic stress and mTOR signaling. Cell. 2008; 134(3):451-460. 40. Augustine JJ, Knauss TC, Schulak JA, Bodziak KA, Siegel C, Hricik DE. Comparative effects of sirolimus and mycophenolate mofetil on erythropoiesis in kidney transplant patients. Am J Transplant. 2004;4(12):2001-2006. 41. Ekberg H, Bernasconi C, Nรถldeke J, et al. Cyclosporine, tacrolimus and sirolimus retain their distinct toxicity profiles despite low doses in the Symphony study. Nephrol Dial Transplant. 2010;25(6):2004-2010. 42. Shameem R, Hamid MS, Wu S. Risk of anemia attributable to everolimus in patients with cancer: a meta-analysis of randomized controlled trials. Anticancer Res. 2015; 35(4):2333-2340. 43. Diekmann F, Rovira J, Diaz-Ricart M, et al. mTOR inhibition and erythropoiesis: microcytosis or anaemia? Nephrol Dial
Transplant. 2012;27(2):537-541. 44. Socolovsky M. Molecular insights into stress erythropoiesis. Curr Opin Hematol. 2007;14(3):215-224. 45. Sankaran VG, Weiss MJ. Anemia: progress in molecular mechanisms and therapies. Nat Med. 2015;21(3):221-230. 46. Zhang X, Camprecios G, Rimmele P, et al. FOXO3-mTOR metabolic cooperation in the regulation of erythroid cell maturation and homeostasis. Am J Hematol. 2014; 89(10):954-963. 47. Green RJ, Herget M. Outcomes of systemic/strategic team consultation: I. Overview and one-month results. Fam Process. 1989;28(1):37-58. 48. Kloosterman WP, Wienholds E, de Bruijn E, Kauppinen S, Plasterk RH. In situ detection of miRNAs in animal embryos using LNAmodified oligonucleotide probes. Nat Methods. 2006;3(1):27-29. 49. Diez-Roux G, Banfi S, Sultan M, et al. A high-resolution anatomical atlas of the transcriptome in the mouse embryo. PLoS Biol. 2011;9(1):e1000582. 50. Kim M, Tan YS, Cheng WC, Kingsbury TJ, Heimfeld S, Civin CI. MIR144 and MIR451 regulate human erythropoiesis via RAB14. Br J Haematol. 2015;168(4):583-597.
haematologica | 2018; 103(3)
ARTICLE
Bone Marrow Failure
Hypomorphic FANCA mutations correlate with mild mitochondrial and clinical phenotype in Fanconi anemia
Ferrata Storti Foundation
Roberta Bottega,1* Elena Nicchia,2* Enrico Cappelli,3 Silvia Ravera,4 Daniela De Rocco,1 Michela Faleschini,1 Fabio Corsolini,5 Filomena Pierri,3 Michaela Calvillo,3 Giovanna Russo,6 Gabriella Casazza,7 Ugo Ramenghi,8 Piero Farruggia,9 Carlo Dufour3 and Anna Savoia1,2
Institute for Maternal and Child Health – IRCCS “Burlo Garofolo”, Trieste; Department of Medical Sciences, University of Trieste; 3Clinical and Experimental Hematology Unit, “G. Gaslini” Children’s Hospital, Genoa; 4Department of Pharmacy (DIFAR), Biochemistry Lab, University of Genoa; 5U.O.S.D. Centro di Diagnostica Genetica e Biochimica delle Malattie Metaboliche, “G. Gaslini” Children’s Hospital, Genoa; 6Oncology Hematology Pediatric Unit, “Policlinico – Vittorio Emanuele”, University of Catania; 7Pediatric Onco-Hematology, Azienda Ospedaliera/ Universitaria Pisana, Pisa; 8Department of Pediatric and Public Health Sciences, University of Torino and 9Pediatric Onco-Hematology, ARNAS Civico Hospital, Palermo, Italy 1 2
Haematologica 2018 Volume 103(3):417-426
*RB and EN contributed equally to this study.
ABSTRACT
F
anconi anemia is a rare disease characterized by congenital malformations, aplastic anemia, and predisposition to cancer. Despite the consolidated role of the Fanconi anemia proteins in DNA repair, their involvement in mitochondrial function is emerging. The purpose of this work was to assess whether the mitochondrial phenotype, independent of genomic integrity, could correlate with patient phenotype. We evaluated mitochondrial and clinical features of 11 affected individuals homozygous or compound heterozygous for p.His913Pro and p.Arg951Gln/Trp, the two residues of FANCA that are more frequently affected in our cohort of patients. Although p.His913Pro and p.Arg951Gln proteins are stably expressed in cytoplasm, they are unable to migrate in the nucleus, preventing cells from repairing DNA. In these cells, the electron transfer between respiring complex I-III is reduced and the ATP/AMP ratio is impaired with defective ATP production and AMP accumulation. These activities are intermediate between those observed in wild-type and FANCA-/- cells, suggesting that the variants at residues His913 and Arg951 are hypomorphic mutations. Consistent with these findings, the clinical phenotype of most of the patients carrying these mutations is mild. These data further support the recent finding that the Fanconi anemia proteins play a role in mitochondria, and open up possibilities for genotype/phenotype studies based on novel mitochondrial criteria.
Correspondence: anna.savoia@burlo.trieste.it
Received: July 12, 2017. Accepted: December 14, 2017. Pre-published: December 21, 2017. doi:10.3324/haematol.2017.176131 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/417
Introduction Fanconi anemia (FA) is a rare autosomal or X-linked recessive disease characterized by congenital abnormalities, bone marrow failure, and predisposition to cancer (mainly leukemia and squamous cell carcinomas). FA is caused by mutations in at least 22 genes. The FA proteins co-operate in a pathway whose role in maintaining the genome integrity is well known. In the presence of DNA damage, eight of the FA proteins, including FANCA, assemble in a nuclear 'core complex' responsible for monoubiquitination of FANCD2, an FA player that localizes in foci where it interacts with other FA components for DNA repair.1 FA cells are consistently hypersensitive to interstrand cross-link inducing agents, such as diepoxybutane (DEB) or mitomycin C (MMC). In addition to their nuclear localization, the FA proteins are also localized in the cytoplasm, where they are likely to be involved in different processes that have yet to be defined, such as maintenance of mitochondrial aerobic metabolism, suppreshaematologica | 2018; 103(3)
©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
417
R. Bottega et al.
Table 1. Molecular and clinical data of Fanconi anemia individuals with biallelic or monoallelic p.His913Proand p.Arg951Gln alleles of FANCA.
Patient/ Status sex
Mutation
Present age (age at diagnosis)
Homozygous
c.2738A>C p.His913Pro
11 (10)
F2/F* Homozygous
c.2738A>C p.His913Pro
F3/M
Homozygous
c.2738A>C p.His913Pro
F4/F
Compound c.2738A>C heterozygous p.His913Pro c.3715_3729del p.Glu1239_Arg1243del Compound c.2738A>C heterozygous p.His913Pro c.894-?_1470+?del
F1/M
F5/M
F6/M
F7/M
F8/M
F9/M
Compound c.2738A>C heterozygous p.His913Pro c.2852G>A p.Arg951Glns Compound c.2852G>A heterozygous p.Arg951Gln c.284-?_426+?del p.Gly95Glufs*38 Compound c.2852G>A heterozygous p.Arg951Gln c.4261-19_4261-12del
Compound c.2852G>A heterozygous p.Arg951Gln c.3558dup p.Arg1187Glufs*28 F10/M Compound c.2852C>T heterozygous p.Arg951Gln c.2T>A p.? F11/F* Compound c.2851C>T heterozygous p.Arg951Trp c.1777-7_1779del
Congenital malformations
Malformation score
Bilateral thumb 11 hypoplasia, patency Moderate of the ductusarteriosus 24 Growth delay, strabismus, 2 (6) astigmatism, microphthalmia, Mild hypopigmentation (area), triangular face, micrognazia 26 Growth delay, 10 (16) strabismus, myopia, microphthalmia, Moderate widespread hyperpigmentation, triangular face, visual impairment, hypospadias 13 Growth delay, 9 (4) duodenal atresia, Moderate hyperpigmentation (area), cafè au lait spots 18 Growth delay, 22 (7) syndactyly, Severe triangular face, unilateral (right) renal hypoplasia, micropenis, hypospadias 19 Growth delay, 6 (15) astigmatism, microphthalmia, Moderate widespread hyperpigmentation, triangular face, micrognazia 22 Growth delay, 2 (11) cutaneous hyperpigmentation Mild
Dead at 9 (6)
Triangular face
1 Mild
Dead at 21 (10)
Growth delay
1 Mild
15 (6)
Growth delay, hyperpigmentation, triangular face, unilateral ectopic kidney
7 Moderate
20 (7)
Microphthalmia, triangular face, hyperpigmentation, hearing loss
Hematologic features
Haematologic score
Mild bi-lineage cytopenia; no transfusion support Cytopenia tri-lineage before HSCT (17-year-old); no transfusion before HSCT Leukopenia before HSCT (18-year-old); HSCT no transfusion before
3 Mild 7 Moderate
5 Moderate
Leukopenia and thrombocytopenia before HSCT (9-year-old); no transfusion before HSCT Cytopenia tri-lineage before HSCT (9-year-old); transfusion before HSCT Leukopenia at diagnosis; at present, blood count in the normal range
6 Moderate
Mild neutropenia
2 Mild
11 Severe
2 Mild
Thrombocytopenia 9 evolved in severe Severe tri-lineage cytopenia concomitantly with MDS/AML. Dead after two HSTC for relapse (9-year-old) Bi-lineage cytopenia. 7 Dead for CMV infection after Moderate HSCT (21-year-old)
Thrombocytopenia 8 evolved in tri-lineage Moderate cytopenia before HSCT (12-year-old); transfusions before HSCT 7 At diagnosis mild thrombocytopenia 7 Moderate evolved in tri-lineage Moderate left mild conductive cytopenia before HSCT (17-year-old); transfusions before HSCT
*Potential hematologic mosaicism. In F2 lymphoblasts a de novo compensatory mutation c.2737C>G/p.His913Ala is likely to restore the allele function. In F11 lymphoblast cell lines the c.1777-7_1779del was not detected, suggesting that a back mutation event occurred in these cells. For all of these patients, it was not possible to establish when and to what extent the revertant event occurred.
418
haematologica | 2018; 103(3)
Hypormorphic Fanconi anemia mutations
A
B
D
C
E
Figure 1. Functional studies of the p.His913Pro and p.Arg951Gln variants. (A) Complementation analysis determined as cell survival after mitomycin C (MMC) treatment of lymphoblast (LFB) cells from patients F3 (homozygous for p.His913Pro) and F7 (compound heterozygous for p.Arg951Gln and p.Gly95Glufs*38) transduced with retroviral vector expressing the wild-type FANCA cDNA. (B and C) Complementation of the G2 cell cycle arrest after melphalan exposure of F3 and F7 LFB. (D and E) Comparison of the cell survival and cell cycle analysis in LFB cells of patients carrying the p.His913Pro (F3, F4, F5 and F6) and the p.Arg951Gln (F7) mutations and in control cells: wild-type (+/+) LFB cells and LFB cells not expressing the protein (-/-) due to a homozygous large intragenic deletion (c.284-?_1826+?del) of FANCA. There is no significant difference between LFB-/- cells and LFB carrying the missense mutations.
sion of intracellular reactive oxygen species (ROS) levels, and protection from proinflammatory cytokine-induced apoptosis.2 Indeed, FA cells have structurally abnormal mitochondria, as they appear swollen with matrix rarefaction, altered cristae and reticulum fragmentation.3,4 All these features further affect the mitochondrial functions, suppressing cell respiration, perpetuating ROS production and switching respiration from oxidative phosphorylation (OXPHOS) to aerobic glycolysis.2-5 An important milestone in unraveling the role of the FA proteins in the cytoplasm is the discovery of their involvement in selective autophagy.6 They are fundamental for removal of damaged mitochondria probably modulating mitochondrial fission-fusion balance, explaining why mutations of the FA genes result in accumulation of morphological defective mitochondria and inbalance of the cellular redox status.6,7 The clinical expressivity in FA is extremely variable from severe to mild, sometimes without clinical signs. Except for mutations of the FANCD2 and BRCA2 genes, no strong association has been found between the clinical phenotype with the genotype, mainly because of the wide genetic heterogeneity and the limited number of patients who belong to rare complementation groups.8,9 Even considering FANCA, the most frequently mutated gene accounting for approximately two-thirds of the cases,10 the spectrum of mutation is extremly heterogeneous with private mutations and large intragenic deletions.11,12 We report functional studies of three missense FANCA mutations (p.His913Pro, p.Arg951Gln, and p.Arg951Gln), whose pathogenic effect in DNA repair is distinct from that in mitochondrial activity. The mutant proteins are haematologica | 2018; 103(3)
stably expressed but localized only in the cytoplasm, suggesting that they are functionally inactive at least for the DNA repair. They are instead functionally hypomorphic for the mitochondrial function. Consistent with the hypothesis of a residual activity of p.His913Pro, p.Arg951Gln, and p.Arg951Trp, the clinical phenotype of patients carrying these mutations is characterized by late onset of mild cytopenia, which tends to remain stable during follow up.
Methods Biological samples Eleven FA probands with positive chromosomal breakage test were included in this study. The institutional review board of the â&#x20AC;&#x153;G. Gasliniâ&#x20AC;? Hospital, Genoa, Italy, approved the study, and all the subjects or their legal guardians gave written informed consent to the investigation according to the Declaration of Helsinki. In 10 cases (F2-F11), mutations had been previously reported.13 Patient P1 was a novel case analyzed by next generation sequencing.14
Complementation assay Lymphoblast (LFB) cells were transduced with retroviral vectors expressing the cDNAs for FANCA, as previously reported.15,16 Mitomycin C (MMC) survival assay and cell cycle evaluation were performed as previously described.15
Western blot and immunofluorescence assay A full-length FANCA sequence was amplified and cloned into the pcDNA3.1-Flag tagged expression vector. The mutant FANCA cDNAs (p.His913Pro and p.Arg951Gln/Trp) were generated by 419
R. Bottega et al.
site-directed mutagenesis using specific primers (available upon request) and transfected, using calcium phosphate, in 293T cells treated with 2 mM hydroxyurea (HU) for 24 hours.17 Protein whole and fractionated cell extracts were prepared using M-PER™ Mammalian Protein Extraction Reagent and NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific), respectively. Primary antibodies were used as follows: antiFANCA (Savino et al.,18 1:500), anti-FANCD2 (Santa Cruz, sc20022, 1:500), add anti-HSP90 (Santa Cruz, sc-7947, 1:4000), antiORC2 (Abcam, ab68348; 1:500), anti-FLAG (OctA Probe H5, Santa Cruz, 1:500), and anti-α Tubulin (Santa Cruz, sc-5286, 1:2000). Immuno-reactivity was visualized using the Enhanced Chemiluminescent SuperSignalTm West Femto Maximum Sensitivity Substrate (Pierce). For immunofluorescence assay, anti-FLAG antibody was used
with anti-mouse FITC secondary antibody (F0479, DakoCytomation), while nuclei were stained with 1 ug/mL propidium iodide solution (Sigma Aldrich).
Biochemical assays FANCA pcDNA3.1 constructs were transfected in LFB cells not expressing the FANCA protein (FANCA-/-) using Lipofectamine (Thermo Fisher Scientific). Biochemical assays (ATP/AMP evaluation, electron transfer between complex I to complex III and Oxygen consumption measurements) were performed as described in Columbaro et al.19
Fo-F1 ATP synthase activity assay Evaluation of the Fo-F1 ATP synthase activity was performed as previously described.20 Briefly, 200,000 cells were incubated for 10
A
B
C
D
E
Figure 2. Expression of p.His913Pro and p.Arg951Gln FANCA proteins. (A) Western blot of total lysates from lymphoblast (LFB) cells of patients carrying the p.His913Pro (F3, F4, F5 and F6) and the p.Arg951Gln (F7 and F10) mutations, showing that the mutant FANCA proteins are expressed. The controls are wild-type (+/+) LFB cells and LFB cells not expressing the FANCA protein (-/-) due to a homozygous large intragenic deletion (c.284-?_1826+?del) of FANCA. (B) Western blot of cytoplasmic (C) and nuclear (N) fractionated cellular lysates after 2 mM hydroxyurea treatment (24 hours) of LFB from patients F3 and F7, showing that the endogenous p.His913Pro and p.Arg951Gln proteins do not translocate to the nucleus. Controls +/+ and -/-, as indicated in (A). (C) Western blot of cytoplasmic (C), nuclear (N) fractionated, or total (TL) cellular lysates from 293T cells transfected with the wild-type (wt) or the mutant (H913P and R951Q) forms of FANCA tagged with FLAG, confirming the exogenous p.His913Pro and p.Arg951Gln FANCA proteins are retained in the cytoplasm. (D) Immunofluorescence analyses on 293T cells transfected as indicated in (C). Nuclei are stained with propidium iodide (PI). (E) Western blot of different LFB cells exposed to 2 mM hydroxyurea (24 hours) showing no monoubiquitination of the FANCD2 protein. Control +/+, as indicated in (A).
420
haematologica | 2018; 103(3)
Hypormorphic Fanconi anemia mutations
minutes (min) in a proper medium and ATP synthesis was induced by the addition of 0.1 mM ADP. The reaction was monitored for 2 min, every 30 seconds (sec) in a luminometer (GloMax® 20/20n Luminometer, Promega Italia, Milan, Italy), by the luciferin/luciferase chemiluminescent method, with ATP standard solutions between 10-8 and 10-5 M (luciferin/luciferase ATP bioluminescence assay kit CLSII, Roche, Basel, Switzerland). Data were expressed as nmol ATP produced/min/106 cells. The oxidative phosphorylation efficiency (P/O ratio) was calculated as the ratio between the concentration of the produced ATP and the amount of consumed oxygen in the presence of respiring substrate and ADP.21
Clinical data Clinical information, phenotypic score and grade of cytopenia was defined as described in Svahn et al.8 The hematologic condition was evaluated by scores grouped in three categories, from 1 to 3, from > 3 to 7, and > 7, representing mild, moderate, and severe phenotypes, respectively. Regarding the malformation features, a score was assigned to every patient, dividing the population into three groups: mild (<6), moderate (6-15), and severe (>15).
A
Results Missense p.His913Pro and p.Arg951Gln mutants of FANCA are stably expressed with loss-of-function effect on DNA repair In our cohort of FA individuals, there are relatively recurrent missense mutations of the FANCA gene, such as c.2738A>C (p.His913Pro) accounting for 9 alleles from six different families (Table 1). Three patients were homozygous (F1, F2, and F3) and another 3 (F4, F5 and F6) were compound heterozygous for the mutation with the second allele being a small in-frame p.Glu1239_Arg1243del or large intragenic deletion, and another missense p.Arg951Gln mutation, respectively. All the six families were from Sicily, suggesting a founder effect of the c.2738A>C variant. Indeed, genotyping four polymorphic loci closed to FANCA, we found that the variant was associated with the same haplotype in all the families (data not shown). Bioinformatics tools showed low pathogenicity scores for p.His913Pro.13 Consistently, the CADD score (score 13.25) is lower than the threshold (score 15.00) commonly used to predict pathogenicity of missense variants.22
B
Figure 3. Hypomorphic effect of missense mutations on mitochondrial activity. (A) Intracellular concentrations of ATP and AMP, ATP/AMP ratio, and electron transfer between complex I and III were determined in lymphoblasts (LFB) from patients with mild (F6, F7), moderate (F3, F4) and severe (F6) hematologic scores. LFB from 3 Fanconi anemia (FA) patients not expressing FANCA (-/-) and 5 healthy individuals (+/+) have been used as controls. The three -/- cell lines are compound heterozygous (p.Arg18Profs*19/p.Asn1221Thrfs*26 and p.Ser175Leufs*5/p.Trp183*) and homozygous (Gly95Glufs*31) for nonsense or frameshift mutations. Each graph is representative of three experiments carried out in each cell line and data are expressed as mean±Standard Deviation (SD). t-test indicates a significant difference of P<0.05 (*) and P<0.01 (**) between the +/+ cells and the other samples, while # and ## indicate a significant difference of P<0.05 and P<0.01, respectively, between the -/- LFB cells and the other samples. (B) The same parameters reported in (A) were evaluated in an LFB -/- cell line transfected with the wild-type (wt) or mutant (H913P, R951Q and R951W) forms of FANCA tagged with FLAG, showing that the mitochondrial activity is intermediate between that of LFB +/+ and LFB -/cells transfected or untransfected (+/+ or -/- vectors) with the empty vector. Each graph is representative of three experiments carried out in each cell line and data are expressed as mean±Standard Deviation (SD). t-test indicates a significant difference of P<0.05 (* or #) and P<0.01 (** or ##) between LFB cells -/- transfected with empty vector or expressing the three missense mutant forms of FANCA and LFB cells +/+ (* and **) or LFB cells -/- (# and ##).
haematologica | 2018; 103(3)
421
R. Bottega et al.
To exclude the possibility that another FA gene or another unknown FANCA variant in linkage disequilibrium with c.2738A>C could be responsible for the disease, we further investigated the role of FANCA as a diseasecausing gene. First of all, complementation analysis of FANCA restored the MMC sensitivity and/or G2 arrest
induced by melphalan in LFB homozygous for p.His913Pro (F3) (Figure 1A and B). Then, to investigate the functional impact of p.His913Pro on protein stability and localization, we performed Western blot analysis using LFB from 4 affected individuals (F3-F6). In all the four cell lines, FANCA was not degraded and was
A
B
Figure 4. Oxymetric measures and ATP synthesis. (A) Amperometric traces of the oxygen consumption for LFB FANCA-/- cells transfected with wild-type (+/+), p.His913Pro, p.Arg951Gln, and p.Arg951Trp mutant forms of FANCA. (B) ATP synthesis rate in the same samples as in (A). The data in (A) and (B) are also depicted as histograms. Each bar graph is representative of three experiments and data are expressed as meanÂąStandard Deviation (SD). Anova test indicates a significant difference of P<0.05 (*) and P<0.01 (**) between the +/+ sample and the other samples, while ## indicates a significant difference of P<0.01 between LFB -/transfected with the p.His913Pro, p.Arg951Gln and p.Arg951Trp mutant forms of FANCA and LFB -/-.
422
haematologica | 2018; 103(3)
Hypormorphic Fanconi anemia mutations
expressed at similar level as in wild-type cells (FANCA+/+) (Figure 2A). The protein was instead undetectable, even at different exposures, in LFB (FANCA-/-) from an FA patient homozygous for a FANCA large intragenic deletion (exons 4 to 20) resulting in a frameshift mutation (p.Gly95Glyfs*31). Finally, in order to evaluate whether the mutant p.His913Pro protein could play a role in nuclear DNA repair, we analyzed fractionated cellular lysates. As expected, FANCA was detected in both cytoplasm and nucleus of the wild-type LFB cells. Instead, the protein was present only in the cytoplasm of cells from F3 (Figure 2B). Immunofluorescence and Western blot of fractionated cellular lysate from the 293T cells over-expressing the wild-type or the mutated FANCA tagged with FLAG confirmed that the mutant protein was retained in the cytoplasm (Figure 2C and D). Taken together, these data indicated that the p.His913Pro mutant form of FANCA is stably expressed but is unable to enter the nucleus and play its role in the DNA repair pathway. Of note, patient F6 is a compound heterozygous for the p.His913Pro and p.Arg951Gln mutations. The latter was identified in another 4 probands (F7, F8, F9, and F10) who are compound heterozygous for deleterious frameshift, splicing or start-loss mutation of FANCA (Table 1). Consistent with haplotype analysis (data not shown), the families carrying p.Arg951Gln come from different geographical areas, suggesting that the mutation originated from independent de novo events and that the presence of another disease-causing mutation in linkage disequilibrium was unlikely. In the LFB cells of F7, the sensitivity to MMC and the G2 arrest were restored by complementation analysis (Figure 1A and C). Like the p.His913Pro substitution, p.Arg951Gln did not affect FANCA stability, being the protein expressed in LFB of patients F7 and F10 (Figure 2A). Moreover, neither the endogenous nor the over-expressed mutant protein migrated into the nucleus (Figure 2B and C). Taken together, these data suggest that p.His913Pro and p.Arg951Gln are loss-of-function mutations for the role of FANCA in controlling the genomic integrity. Consistent with this hypothesis, FANCD2 was not monoubiquitinated in any of the six cell lines available for the analysis (Figure 2E). Moreover, in the same cells, the sensitivity to MMC and the percentage of cells blocked in the G2 phase are comparable with those of the FANCA-/- cells (Figure 1D and E). Finally, at position 951 an additional substitution (c.2851C>T/p.Arg951Trp) was identified in one allele of patient F11. However, the LFB cell line or other biological samples were not available for investigations to confirm that the mutation is associated with protein stability and lack of FANCD2 monoubiquitination, as reported by Karras et al.23
p.His913Pro and p.Arg951Gln are hypomorphic mutations for mitochondrial activity In order to evaluate whether FANCA could affect the mitochondrial function, we investigated the electron transport chain functionality and the cellular energy status in LFB from patients with mild (F6 and F7), moderate (F3 and F4), and severe (F5, the only LFB cell line available within this group) hematologic score. Consistent with the FA mitochondrial phenotype,2-4 the cells showed an altered ATP/AMP ratio with reduced ATP production and AMP accumulation, and impaired electron transfer between haematologica | 2018; 103(3)
complexes I and III (Figure 3A). Of note, the biochemical values were intermediate between those obtained in FANCA+/+ and FANCA-/- cells, suggesting that p.His913Pro and p.Arg951Gln are hypomorphic alleles for the mitochondrial activity. In order to demonstrate that these alleles are directly involved in determining a mild mitochondrial phenotype, we over-expressed the p.His913Pro and p.Arg951Gln, as well as p.Arg951Trp, FANCA proteins in FANCA-/- cells. Whereas the ATP and AMP concentrations, the ATP/AMP ratio, and the electron transport were complemented by transfection of the wild-type cDNA, the same phenotypes were only partially restored by expression of the three mutant cDNAs. Of note, the biochemical measurements are significantly different when FANCA-/- cells are compared with those over-expressing the mutant forms of FANCA (Figure 3B). The defect in the electron transfer between complexes I and III determines an impaired oxygen consumption and ATP synthesis in FANCA-/- cells, while in the samples expressing His913Pro, Arg951Gln and Arg951Trp, this activity is partially recovered, although with values lower than those obtained in cells expressing the wild-type FANCA protein (Figure 4A and B). Conversely, respiration was observed in all samples after stimulation with succinate, indicating that the electron transfer through complex II, III and IV pathways is not impaired (Figure 4A). This implies that the FA mutations may have a specific interaction with the complexes I, III and IV pathway. However, similarly to that reported in literature,21 the P/O ratio in the presence of pyruvate/malate is around 2.5 in all samples, suggesting that, despite the impaired mitochondrial function in FANCA-/- and in cells expressing the three mutant forms of FANCA, the residual oxygen consumption is devoted to ATP synthesis (Figure 4C).
Moderate clinical phenotype in patients with missense mutations affecting residues His913Pro and Arg951 The clinical data of the affected individuals carrying the p.His913Pro, p.Arg951Gln, and p.Arg951Trp mutations are reported in Table 1. Of the 11 patients, 8 were male and 3 females. Mean age at diagnosis was 8.9 years (range 4-16 years). Ten had mild or moderate somatic phenotype (range 1-11). Nine of these individuals had concordant mild or moderate hematologic scores ranging from 2 to 8. Three individuals (F1, F6, and F7) had mild cytopenia that did not require transfusions. Of note, F6 and F7 had a mild leukopenia/neutropenia at diagnosis; at present, blood count is in the normal range in F6 and neutropenia is stable in F7 after 4 and 11 years, respectively, from the first abnormal blood count. The remaining 6 probands underwent HSCT. The first abnormal blood count was a mild tri-lineage cytopenia in F2 and F9 or thrombocytopenia, as in F10 and F11. They all underwent HSCT for progression to aplastic anemia, which occurred 10-11 years after diagnosis in F2, F9, and F11. Even F3 and F4 did receive HSCT after two and five years, respectively, of stable moderate cytopenia because a matched healthy sibling brother was available. Except for F9, who died for HSCT complications due to cytomegalovirus infection, the other patients are all alive after a follow up ranging from 1 to 18 years (mean 9.5 years) and none have developed hematologic adverse events (myelodysplastic syndrome, acute myelogenous leukemia or cytogenetic alterations) and/or solid tumor. 423
R. Bottega et al.
Patients F5 and F8 were the 2 affected individuals with severe hematologic scores. In addition to having multiple congenital malformations (score 22), F5 was characterized by severe tri-lineage requiring blood transfusions. He successfully underwent HSCT two years after diagnosis. At diagnosis, F8 had no congenital malformation except for growth delay and mild thrombocytopenia. However, the mono-lineage cytopenia evolved in severe tri-lineage cytopenia concomitantly with myelodysplastic syndrome transformed in acute myelogenous leukemia. He died three years later for relapse after two HSCT.
Discussion Mutations of the FA genes are almost private, though common founder mutations have been reported in few populations, including Spanish Gypsies, the Afrikaner population of South Africa, Ashkenazi Jews, and people from some Italian geographical regions.24-27 This is the case of p.His913Pro, a mutation of the FANCA gene that is relatively frequent, even as a homozygous condition, in patients from Sicily. Indeed, the analysis of microsatellite markers in the six families revealed the presence of a common haplotype compatible with a founder effect. In contrast, p.Arg951Gln has been identified, never as a homozygous mutation, in individuals from different geographical areas, suggesting that it occurred as independent de novo mutational events. At the same residue, an additional rare pathogenic mutant (p.Arg951Trp) affects one allele of our cohort of patients.23 Since these alterations are missense mutations, their pathogenic effect on protein function has been ascertained by considering several aspects. Unlike p.Arg951Gln/Trp, p.His913Pro was classified as a variant of uncertain significance, as the predictive bioinformatic tools indicated that this amino acid substitution has no or mild effect on protein function.13 However, genetic studies support their pathogenic role, as they are present in FA patients and only rarely in controls (ExAC or other databases).13 Consistent with this hypothesis, we found that the mutant FANCA proteins (both endogenous and overexpressed) localize only in the cytoplasm, where they are stably expressed at similar levels to those observed in wild-type cells. As a consequence, the FA/BRCA pathway remains inactive because FANCD2 is not monoubiquitinated, as demonstrated by Western blot analysis. These data are consistent with those reported by Castella et al.,24 who showed that all the FANCA missense mutations are stably expressed. Like p.His913Pro and p.Arg951Gln, their products do not enter the nucleus, preventing FANCD2 being monoubiquitinated. On the contrary, the same authors demonstrated that null mutations, such as nonsense or frameshift alterations, are unstable and not detectable within cells.24 Therefore, our data confirm the hypothesis of a correlation between type of mutation and expression level of mutant FANCA, but as to why these mutant proteins are retained in the cytoplasm without any apparent function remains a subject of debate. In trying to unravel any role of the FA proteins in other cell compartments, we investigated the mitochondria OXPHOS function and the cellular energy status whose alteration may lead to an impaired mitophagy process.28 As demonstrated by biochemical assays, expression of the 424
p.His913Pro and p.Arg951Gln/Trp proteins are associated with a mild mitochondrial function impairment. In particular, the electron transport between complexes I and III, the ATP production, and the O2 consumption appear defective but not to the same extent as in FA cells homozygous for null mutations. Interestingly, the defective OXPHOS is limited to the pathway composed by complexes I, III and IV, considering that the oxygen consumption and the relative ATP synthesis is similar in control and FA samples when induced by succinate. This suggests that the potential function of FANCA, as well as other FA proteins, on mitochondria is confined to the OXPHOS led by complex I. However, the evaluation of P/O ratio shows that, in all samples, the oxygen consumption, even when very low, is finalized to the ATP synthesis, indicating that the mitochondria are always in a coupled status. In accordance with the recent discovery that the FA proteins are required for clearance of damaged mitochondria and to decrease the mitochondrial reactive oxygen species,6 our finding supports the role of FANCA in these organelles. Considering that small interfering RNA of the FA genes is associated with increased puncta, which is a marker for defects in clearance of damaged mitochondria,6 we can speculate that mutations of FANCA could negatively influence the efficiency of the mitophagy, causing an accumulation of damaged mitochondria. This may determine not only a decrement in the energy production, but also the increment of oxidative stress, which could induce damage of other mitochondria. However, we cannot exclude the possibility that the FANCA protein may have a more direct role in the modulation of the mitochondria structure and function, i.e. influencing the organization of inner mitochondrial membrane, whose integrity is essential for the proper functioning of the OXPHOS. Of interest, our data allow us to dissect the activity of FANCA in DNA repair from that in preserving the mitochondrial function, as has previously been hypothesized for FANCC.6 Indeed, one mutation of this gene (c.67delG), resulting in an N-terminus truncated but stable mutant form of FANCC, is unable to repair DNA but is able to restore the mitochondrial function. Therefore, we regard the p.His913Pro and p.Arg951Gln/Trp as variants with a hypomorphic effect of FANCA in maintaining the mitochondrial activity. Even mutations of the FANCD2 gene have been regarded as hypomorphic.29 However, their 'hypomorphic' effect is associated with the role of the protein in controlling genomic integrity. In FANCD2 individuals, at least one allele is always associated with expression of a low amount of FANCD2, which is detectable in both the non- and monoubiquitinated form, suggesting that residual activity of FANCD2 for the DNA repair processes is essential for life. On the contrary, biallelic null mutations of FANCA and FANCC are relatively common in FA and are always associated with lack of FANCD2 monoubiquitination.24 Considering that the c.67delG mutation of FANCC, which is able to restore the mitochondrial function,6 is associated with mild clinical phenotypes,9,30 we evaluated whether we could reach the same conclusion in our cohort of 11 patients. The mean age (8.9 years) at diagnosis was higher than that reported in the FA Italian cohort (6.8 years).8 Nine of them had concordant mild/moderate hematologic and somatic phenotypes. Regarding malformations, 4, 6, and one patients were classified as mild, haematologica | 2018; 103(3)
Hypormorphic Fanconi anemia mutations
moderate and severe, respectively, and percentages [0.36 vs. 0.43 (mild), 0.55 vs. 0.37 (moderate), and 0.09 vs. 0.20 (severe)] are not significantly different from those of the entire Italian cohort, although only in one case were the congenital malformations severe. For the hematologic features, 3 (0.30), 6 (0.50) and 2 (0.20) of patients received mild, moderate and severe scores, respectively, compared with 16% (mild), 27% (moderate), and 49% (severe) of the FA Italian population, suggesting that the missense mutations at residues His913 and Arg951 are associated with better hematologic prognosis.8 Two patients had severe hematologic score and underwent HSCT (together with another 6 affected individuals). However, HSCT per se should not be regarded as a marker of hematologic severity. Indeed, despite their moderate hematologic condition, 2 patients took advantage of having a suitable healthy matched sibling donor, which is known to provide the best outcome.31 Another 4 underwent familiar or matched unrelated donor HSCT even more than ten years after the first abnormal blood count, when cytopenia dropped from mild/moderate to severe. They are alive and well at 3-7 years after HSCT, except for F9 and F8 who died for cytomegalovirus infection after HSCT and relapse after two HSCT, respectively. Three probands (F2, F6, and F7) with mild hematologic score did not undergo HSCT and their hematologic condition appears to be stable even 11 years (F7) after diagnosis. Taken together, these data suggest that FA individuals carrying the p.His913Pro and p.Arg951Gln/Trp mutations at one or both the FANCA alleles have a relatively mild phenotype. This conclusion is consistent with data of Faivre et al.,9 showing that patients with altered FANCA protein have milder phenotypes than those with a complete loss of FANCA. However, the clinical phenotype depends on many factors, including different genetic and environmental factors, preventing us from confirming any clear correlations. Although FA affected individuals carrying the same disease-causing mutation (even among siblings) could have different outcomes, the 3 patients homozygous for p.His913Pro have mild/moderate clinical phenotype. Of note, our cohort numbers were limited, as the number of cases studied is relatively low. In order to
References 1. Longerich S, Li J, Xiong Y, Sung P, Kupfer GM. Stress and DNA repair biology of the Fanconi anemia pathway. Blood. 2014; 124(18):2812-2819. 2. Kumari U, Ya Jun W, Huat Bay B, Lyakhovich A. Evidence of mitochondrial dysfunction and impaired ROS detoxifying machinery in Fanconi anemia cells. Oncogene. 2014;33(2):165-172. 3. Cappelli E, Cuccarolo P, Stroppiana G, et al. Defects in mitochondrial energetic function compels Fanconi Anaemia cells to glycolytic metabolism. Biochim Biophys Acta. 2017;1863(6):1214-1221. 4. Ravera S, Vaccaro D, Cuccarolo P, et al. Mitochondrial respiratory chain Complex I defects in Fanconi anemia complementation group A. Biochimie. 2013;95(10):18281837. 5. Capanni C, Bruschi M, Columbaro M, et al. Changes in vimentin, lamin A/C and mitofilin induce aberrant cell organization
haematologica | 2018; 103(3)
6.
7.
8.
9.
assign statistical significance to the correlation, we should explore whether other FA mutations are hypomorphic for the mitochondrial activity and eventually correlate this with mild clinical outcome. Finally, another factor that could modulate the clinical phenotype in FA is the occurrence of hematopoietic mosaicism, in which cells have lost mitomycin C sensitivity due to back mutations or other genetic mechanisms, as it occurs in almost 20% of cases.32 In consideration of this, F2 and F11 are 2 individuals with potential hematopoietic mosaicism, since one of the two FA mutations was reverted to wild-type allele in their LFB cell lines.13 However, none of them were further investigated, preventing us from ascertaining whether and to what extent the mosaicism was present in the bone marrow and blood cells. In conclusion, at least for the mitochondrial activity, p.His913Pro and p.Arg951Gln/Trp are hypomorphic mutations, which are associated with a mild/moderate phenotype characterized by late onset of the disease and slow hematologic progression. Since the hematopoietic stem cells are relatively sensitive to their redox status,33 a residual activity of mitochondria in cells expressing these, and maybe other missense mutations, could help patients avoid experiencing worsening cytopenia and postponing their eligibility for HSCT. Therefore, we should explore whether other FA mutations with hypomorphic effect could explain the clinical variability in FA. Funding This study was supported by Telethon Foundation (grant GGP11076), Cariplo Foundation (2012-0529), Italian Ministry of Health (RF-2010-2309222), AIRFA (Italian Association for Research in Fanconi Anemia), ERG S.p.A., Cambiaso Risso Group, Rimorchiatori Riuniti S.p.A., and Saar Depositi Oleari Portuali S.p.A. We are grateful to the “Cell Line and DNA Biobank from Patients affected by Genetic Diseases” (“G. Gaslini” Institute) and Telethon Genetic Biobank Network (project n. GTB07001) for the sample providing. MF is supported by a fellowship (ID 19432) from AIRC (Italian Association of Cancer Research) and RB from "Umberto Veronesi" Foundation.
in fibroblasts from Fanconi anemia complementation group A (FA-A) patients. Biochimie. 2013;95(10):1838-1847. Sumpter R, Levine B. Novel functions of Fanconi anemia proteins in selective autophagy and inflammation. Oncotarget. 2016;7(32):50820-50821. Shyamsunder P, Esner M, Barvalia M, et al. Impaired mitophagy in Fanconi anemia is dependent on mitochondrial fission. Oncotarget. 2016;7(36):58065-58074. Svahn J, Bagnasco F, Cappelli E, et al. Somatic, hematologic phenotype, longterm outcome, and effect of hematopoietic stem cell transplantation. An analysis of 97 Fanconi anemia patients from the Italian national database on behalf of the Marrow Failure Study Group of the AIEOP (Italian Association of Pediatric HematologyOncology). Am J Hematol. 2016;91(7):666671. Faivre L, Guardiola P, Lewis C, et al. Association of complementation group and mutation type with clinical outcome in fan-
10.
11.
12.
13.
coni anemia. European Fanconi Anemia Research Group. Blood. 2000;96(13):40644070. Kee Y, D'Andrea AD. Molecular pathogenesis and clinical management of Fanconi anemia. J Clin Invest. 2012;122(11):37993806. Levran O, Diotti R, Pujara K, Batish SD, Hanenberg H, Auerbach AD. Spectrum of sequence variations in the FANCA gene: an International Fanconi Anemia Registry (IFAR) study. Hum Mutat. 2005;25(2):142149. Morgan NV, Tipping AJ, Joenje H, Mathew CG. High frequency of large intragenic deletions in the Fanconi anemia group A gene. Am J Hum Genet. 1999;65(5):13301341. De Rocco D, Bottega R, Cappelli E, et al. Molecular analysis of Fanconi anemia: the experience of the Bone Marrow Failure Study Group of the Italian Association of Pediatric Onco-Hematology. Haematologica. 2014;99(6):1022-1031.
425
R. Bottega et al. 14. Nicchia E, Greco C, De Rocco D, et al. Identification of point mutations and large intragenic deletions in Fanconi anemia using next-generation sequencing technology. Mol Genet Genomic Med. 2015; 3(6):500-512. 15. Antonio Casado J, Callén E, Jacome A, et al. A comprehensive strategy for the subtyping of patients with Fanconi anaemia: conclusions from the Spanish Fanconi Anemia Research Network. J Med Genet. 2007; 44(4):241-249. 16. Hanenberg H, Batish SD, Pollok KE, et al. Phenotypic correction of primary Fanconi anemia T cells with retroviral vectors as a diagnostic tool. Exp Hematol. 2002; 30(5):410-420. 17. Molina B, Marchetti F, Gómez L, et al. Hydroxyurea induces chromosomal damage in G2 and enhances the clastogenic effect of mitomycin C in Fanconi anemia cells. Environ Mol Mutagen. 2015; 56(5):457-467. 18. Savino M, Borriello A, D'Apolito M, et al. Spectrum of FANCA mutations in Italian Fanconi anemia patients: identification of six novel alleles and phenotypic characterization of the S858R variant. Hum Mutat. 2003;22(4):338-339. 19. Columbaro M, Ravera S, Capanni C, et al. Treatment of FANCA cells with resveratrol and N-acetylcysteine: a comparative study. PLoS One. 2014;9(7):e104857.
426
20. Ravera S, Dufour C, Cesaro S, et al. Evaluation of energy metabolism and calcium homeostasis in cells affected by Shwachman-Diamond syndrome. Sci Rep. 2016;6:25441. 21. Hinkle PC. P/O ratios of mitochondrial oxidative phosphorylation. Biochim Biophys Acta. 2005;1706(1-2):1-11. 22. Easton DF, Deffenbaugh AM, Pruss D, et al. A systematic genetic assessment of 1,433 sequence variants of unknown clinical significance in the BRCA1 and BRCA2 breast cancer-predisposition genes. Am J Hum Genet. 2007;81(5):873-883. 23. Karras GI, Yi S, Sahni N, et al. HSP90 Shapes the Consequences of Human Genetic Variation. Cell. 2017;168(5):856866.e812. 24. Castella M, Pujol R, Callén E, et al. Origin, functional role, and clinical impact of Fanconi anemia FANCA mutations. Blood. 2011;117(14):3759-3769. 25. Tipping AJ, Pearson T, Morgan NV, et al. Molecular and genealogical evidence for a founder effect in Fanconi anemia families of the Afrikaner population of South Africa. Proc Natl Acad Sci USA. 2001;98(10):5734-5739. 26. Kutler DI, Auerbach AD. Fanconi anemia in Ashkenazi Jews. Fam Cancer. 2004;3(34):241-248. 27. Savino M, Ianzano L, Strippoli P, et al. Mutations of the Fanconi anemia group A
28.
29.
30.
31.
32.
33.
gene (FAA) in Italian patients. Am J Hum Genet. 1997;61(6):1246-1253. MacVicar TD, Lane JD. Impaired OMA1dependent cleavage of OPA1 and reduced DRP1 fission activity combine to prevent mitophagy in cells that are dependent on oxidative phosphorylation. J Cell Sci. 2014; 127(Pt 10):2313-2325. Kalb R, Neveling K, Hoehn H, et al. Hypomorphic mutations in the gene encoding a key Fanconi anemia protein, FANCD2, sustain a significant group of FAD2 patients with severe phenotype. Am J Hum Genet. 2007;80(5):895-910. Yamashita T, Wu N, Kupfer G, et al. Clinical variability of Fanconi anemia (type C) results from expression of an amino terminal truncated Fanconi anemia complementation group C polypeptide with partial activity. Blood. 1996;87(10):4424-4432. MacMillan ML, DeFor TE, Young JA, et al. Alternative donor hematopoietic cell transplantation for Fanconi anemia. Blood. 2015;125(24):3798-3804. Lo Ten Foe JR, Kwee ML, Rooimans MA, et al. Somatic mosaicism in Fanconi anemia: molecular basis and clinical significance. Eur J Hum Genet. 1997;5(3):137-148. Du W, Adam Z, Rani R, Zhang X, Pang Q. Oxidative stress in Fanconi anemia hematopoiesis and disease progression. Antioxid Redox Signal. 2008;10(11):19091921.
haematologica | 2018; 103(3)
ARTICLE
Myelodysplastic Syndromes
Constitutional SAMD9L mutations cause familial myelodysplastic syndrome and transient monosomy 7
Victor B. Pastor,1,2* Sushree S. Sahoo,1,2,3* Jessica Boklan,4 Georg C. Schwabe,5 Ebru Saribeyoglu,5 Brigitte Strahm,1 Dirk Lebrecht,1 Matthias Voss,6 Yenan T. Bryceson,6 Miriam Erlacher,1,7 Gerhard Ehninger,8 Marena Niewisch,1 Brigitte Schlegelberger,9 Irith Baumann,10 John C. Achermann,11 Akiko Shimamura,12 Jochen Hochrein,13 Ulf Tedgård,14 Lars Nilsson,15 Henrik Hasle,16 Melanie Boerries,7,13 Hauke Busch,13,17 Charlotte M. Niemeyer1,7 and Marcin W. Wlodarski1,7
Department of Pediatrics and Adolescent Medicine, Division of Pediatric Hematology and Oncology, Medical Center, Faculty of Medicine, University of Freiburg, Germany; 2Faculty of Biology, University of Freiburg, Germany; 3Spemann Graduate School of Biology and Medicine, University of Freiburg, Germany; 4Center for Cancer and Blood Disorders, Phoenix Children's Hospital, AZ, USA; 5Children’s Hospital, Carl-Thiem-Klinikum Cottbus, Germany; 6 Department of Medicine, Huddinge, Hematology and Regenerative Medicine, Karolinska Institute, Stockholm, Sweden; 7German Cancer Consortium (DKTK), Freiburg, Germany and German Cancer Research Center (DKFZ), Heidelberg, Germany; 8Internal Medicine of Hematology/Medical Oncology, University Hospital, Dresden, Germany; 9Institute of Human Genetics, Hannover Medical School, Germany; 10Clinical Centre South West, Department of Pathology, Böblingen Clinics, Germany; 11Genetics & Genomic Medicine, UCL Great Ormond Street Institute of Child Health, University College London, UK; 12Boston Children's Hospital, Dana Farber Cancer Institute, and Harvard Medical School, MA, USA; 13Institute of Molecular Medicine and Cell Research, University of Freiburg, Germany; 14Department of Pediatric Oncology and Hematology, Skåne University Hospital, Lund, Sweden; 15 Department of Hematology, Oncology and Radiation Physics, Skåne University Hospital, Lund, Sweden; 16Department of Pediatrics, Aarhus University Hospital, Denmark and 17 Lübeck Institute of Experimental Dermatology, Germany
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):427-437
1
*VBP and SS contributed equally to this manuscript.
Correspondence: ABSTRACT
F
amilial myelodysplastic syndromes arise from haploinsufficiency of genes involved in hematopoiesis and are primarily associated with early-onset disease. Here we describe a familial syndrome in seven patients from four unrelated pedigrees presenting with myelodysplastic syndrome and loss of chromosome 7/7q. Their median age at diagnosis was 2.1 years (range, 1-42). All patients presented with thrombocytopenia with or without additional cytopenias and a hypocellular marrow without an increase of blasts. Genomic studies identified constitutional mutations (p.H880Q, p.R986H, p.R986C and p.V1512M) in the SAMD9L gene on 7q21, with decreased allele frequency in hematopoiesis. The non-random loss of mutated SAMD9L alleles was attained via monosomy 7, deletion 7q, UPD7q, or acquired truncating SAMD9L variants p.R1188X and p.S1317RfsX21. Incomplete penetrance was noted in 30% (3/10) of mutation carriers. Long-term observation revealed divergent outcomes with either progression to leukemia and/or accumulation of driver mutations (n=2), persistent monosomy 7 (n=4), and transient monosomy 7 followed by spontaneous recovery with SAMD9L-wildtype UPD7q (n=2). Dysmorphic features or neurological symptoms were absent in our patients, pointing to the notion that myelodysplasia with monosomy 7 can be a sole manifestation of SAMD9L disease. Collectively, our results define a new subtype of familial myelodysplastic syndrome and provide an explanation for the phenomenon of transient monosomy 7. Registered at: www.clinicaltrials.gov; #NCT00047268.
haematologica | 2018; 103(3)
marcin.wlodarski@uniklinik-freiburg.de
Received: September 12, 2017. Accepted: December 5, 2017. Pre-published: December 7, 2017.
doi:10.3324/haematol.2017.180778 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/427 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
427
V.B. Pastor et al.
Introduction Germline predisposition has been recognized as an underlying cause for the development of myelodysplastic syndromes (MDS) in children. Recently, it has also been gaining importance in the etiology of adult myeloid neoplasia, particularly in cases with a positive family history. Various genes are known to be associated with heritable forms of MDS and acute myeloid leukemia,1,2 e.g., GATA2,3 CEBPA,4 RUNX1,5 ANKRD26,6 ETV67 and DDX41,8 in addition to inherited bone marrow failure syndromes. Germline mutations in DDX41 can result in adult-onset myeloid neoplasia, while aberrations in RUNX1 and GATA2 are associated with myeloid neoplasia in younger individuals. We recently reported that GATA2 deficiency is the most common genetic cause of primary childhood MDS, accounting for 15% of all cases of advanced MDS, and 37% of MDS with monosomy 7 (MDS/-7).9 However, in the majority of cases of pediatric MDS, and also in a considerable number of cases of familial myeloid neoplasia, the presumed germline cause has not yet been discovered.10,11 Monosomy 7 is the most frequent cytogenetic lesion in children with MDS and, unlike in adults, it often occurs as the sole cytogenetic abnormality.12 Due to the rapid and progressive course of the disease, it is considered an urgent indication for hematopoietic stem cell transplantation.13 However, transient monosomy 7 has occasionally been documented in childhood MDS.14-16 Considering that complete (-7) and partial [del(7q)] deletion of chromosome 7 are common aberrations in MDS, extensive efforts have been undertaken to discern causative tumor suppressor genes located on chromosome 7. Asou and colleagues identified SAMD9 (Sterile Alpha Motif Domain-containing 9), its paralogue SAMD9L (SAMD9-like), and Miki/HEPACAM2 as commonly deleted genes within a 7q21 cluster in patients with myeloid neoplasia.17 Notably Samd9l-haploinsufficient mice were shown to develop myeloid malignancies characterized by different cytopenias and mimicking human disease with monosomy 7.18 In line with these findings, germline heterozygous gainof-function SAMD9L mutations p.H880Q, p.I891T, p.R986C, and p.C1196S were recently discovered in four pedigrees with variable degrees of neurological symptoms (ataxia, balance impairment, nystagmus, hyperreflexia, dysmetria, dysarthria) and hematologic abnormalities (single to tri-lineage cytopenias, MDS/-7). For most carriers, the clinical presentation was compatible with the diagnosis of ataxia-pancytopenia syndrome.19,20 Similarly, in two recent studies, we and others reported de novo gainof-function mutations in SAMD9 in a total of 18 patients with MIRAGE syndrome (myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes, and enteropathy), of whom four notably also developed MDS/-7.21,22 However, not all patients develop the full MIRAGE disease spectrum, as documented in one family with SAMD9-related MDS.23 The SAMD9 and SAMD9L genes share 62% sequence identity and apart from their putative role as myeloid tumor suppressors, their general function and their specific effect pertaining to hematopoiesis are not well-understood.18 In this study we aimed to identify the genetic cause in pedigrees with non-syndromic familial MDS. We discovered constitutional SAMD9L mutations associated with non-random patterns of clonal escape leading to loss of 428
the mutant allele. We further demonstrate in two cases that SAMD9Lâ&#x20AC;&#x201C;related disease can be associated with transient -7, occurring as a one-time clonal event followed by somatic correction of hematopoiesis achieved by UPD7q with double wild-type SAMD9L.
Methods Patients The diagnosis of MDS was established according to World Health Organization criteria.24,25 Patients 1 (P1), 2 (P2), 5 (P5), and 7 (P7) were enrolled in prospective study 98 of the European Working Group of MDS in Childhood (EWOG-MDS) (www.clinicaltrials.gov; #NCT00047268). Patient 6 (P6) was the father of P5 (family III). Family II (P3 and P4) was referred for evaluation of familial MDS from Phoenix Children´s Hospital, USA. The study had been approved by an institutional ethics committee (University of Freiburg, CPMP/ICH/135/95 and 430/16). Written informed consent to participation had been obtained from patients and parents.
Genomic studies and bioinformatics Exploratory whole exome sequencing was performed in bone marrow granulocytes in P1 and P2, as outlined in the Online Supplementary Material. Targeted deep sequencing for SAMD9/SAMD9L and genes related to MDS/inherited bone marrow failure syndromes was performed in other patients, except for P3 and P6 due to unavailability of material. All relevant variants were validated using Sanger sequencing. For germline confirmation, DNA was extracted from skin fibroblasts and/or hair follicles, and targets were amplified and sequenced as previously described.9 The degree of deleteriousness was calculated using the combined annotation-dependent depletion scoring system (CADD-score).26 The variants with CADD-scores higher than 20 were further evaluated for their role in hematologic disease or cancer, thereby focusing on the top 1% most deleterious variants in the human genome. In addition, pathogenicity calculations were performed using standard prediction tools. The evolutionary conservation across species and the physicochemical difference between amino acids were estimated by PhyloP, PhastCons and the Grantham score, respectively.27 Mutant clonal size was inferred from allelic frequencies and the total number of next-generation sequencing reads normalized to the ploidy level. Further details are provided in the Online Supplementary Methods.
Evaluation of variant allelic configuration Genomic DNA of P1 collected at the time of progress to chronic myelomonocytic leukemia (CMML) was amplified to obtain a 1333 bp region encompassing both SAMD9L mutations: p.V1512M (germline) and p.R1188X (acquired). Polymerase chain reaction products were TA-cloned and sequenced as previously reported.28 Sequences of 170 colonies were evaluated for the presence of SAMD9L mutations.
Cellular and functional studies Metaphase karyotyping and interphase fluorescence in situ hybridization were performed using bone marrow specimens according to standard procedures.12 Human colony-forming cell assays were performed in P1 (at CMML disease stage) and in P7 (at diagnosis) as previously described.29 Furthermore, to evaluate the effect of the patient-derived SAMD9L p.V1512M and p.R986C mutations on cellular proliferation, 293FT cells were dye-labeled and consequently transfected with wild-type or mutant teal fluorescent protein (TFP)-SAMD9L as previously described.20 The haematologica | 2018; 103(3)
SAMD9L-related familial MDS
transfected cells were tracked by flow cytometry for TFPSAMD9L expression and dye dilution as an indicator of cell division. SAMD9L variant p.T233N, recently reported as “disease-protective”,20 was used as a control.
Results Clinical phenotype of patients Patients P1-P6 (4 males and 2 females) belong to three unrelated families of German descent and were diagnosed with bona fide familial MDS after known inherited bone marrow failure syndromes had been excluded by targeted sequencing and functional tests (Figure 1, Table 1). Index patient P7 is the only child of a non-consanguineous
Swedish family. Detailed clinical descriptions of all the patients are given in the Online Supplementary Material. All affected individuals had normal measurements without dysmorphic stigmata at birth and at last follow-up (Table 1). Psychomotor development and neurocognitive function were normal and, in particular, no ataxia or movement disorders were diagnosed in the ten mutation carriers (7 patients and 3 silent carriers with SAMD9L mutations). Previous family histories were unremarkable for cytopenias, neurological disease, malignancies, or stillbirths, with the exception of the father of P7 who at the last follow-up presented with unclear ataxia. Prior non-invasive recurrent respiratory tract infections were noted in three of the seven patients (P1, P2, and P4) and endogenous eczema in two (P1 and P3). Moreover, P1 developed transient pancytope-
A
B
C
haematologica | 2018; 103(3)
Figure 1. Germline SAMD9L mutations in pedigrees with familial myelodysplastic syndrome. (A) Identification of four pedigrees with MDS and monosomy 7 harboring germline heterozygous SAMD9L mutations: p.V1512M (pedigree I), p.R986H (pedigree II) and p.R986C (pedigree III), p.H880Q (pedigree IV), and somatic mutations: p.R1188X (P1) and p.S1317RfsX21 (P7). Dotted symbols indicate healthy mutation carriers. Sanger sequencing on DNA extracted from hair follicles (HR) confirmed the germline status of mutations as visualized in electropherograms. Sequencing in P1 was performed on peripheral blood granulocytes (GR) revealing a minor mutational peak, corresponding to a variant allelic frequency of 8.3% by whole exome sequencing. In pedigree III, the mutation in P5 was confirmed in fibroblast (FB) DNA, while for P6 and remaining family members whole blood (WB) was analyzed. In pedigree IV other family members were not tested (n.t.). (B) SAMD9L and SAMD9 gene orientation on 7q22 in reverse strand direction (3‘-5‘). The SAMD9L protein is coded by one exon and contains two known functional sites: N-terminal sterile alpha motif (SAM) and nuclear localization sequence (NLS). Four germline and two somatic (*) mutations were identified in SAMD9L. Germline missense mutations are evolutionarily highly conserved. (C) TA cloning of the double mutated SAMD9L region of P1 revealed cis-configuration of mutations p.V1512M (germline) and p.R1188X (somatic) in ten out of 172 clones tested.
429
V.B. Pastor et al.
nia during infancy and suffered from self-limiting seizures during infancy with no structural brain abnormalities or neurological deficits identified. Peripheral blood findings at diagnosis included isolated thrombocytopenia in one (P2), thrombocytopenia with neutropenia in two (P1, P3) and pancytopenia in four patients (P4-7). Mean corpuscular volume was increased in five of the seven patients at diagnosis, and HbF was elevated in two out of three tested patients. MDS manifested at the median age of 2.1 (range, 1.0-42) years as hypocellular refractory cytopenia of childhood (RCC) or in the adult (P6) as refractory cytopenia with multilineage dysplasia (Table 1). Severe dysplasia with vacuolization was observed in two patients (P1, P2). A common cytogenetic feature in all patients was the complete or partial loss of chromosome 7 (Table 1).
The clinical course was remarkable for several patients. Three years after the initial diagnosis of RCC, P1 developed severe infections and hepatosplenomegaly, and for the first time required platelet transfusions, while his blood smear showed 21% blasts, compatible with the diagnosis of CMML (with no in vitro hypersensitivity to granulocyte-macrophage colony-stimulating factor), (Figure 2). Following hematopoietic stem cell transplantation, he developed late-onset acute graft-versus-host disease and died from cerebral hemorrhage. At the same time, his younger sister (P2) was diagnosed with RCC/-7; however, the parents decided against follow-up in the hematology clinic. Unexpectedly, her complete blood count normalized 3.7 years later and remained stable until the last follow-up 20 years after the initial diagnosis. She
Table 1. Clinical data of SAMD9L-mutated patients.
Patient # Gestational age; Dysmorphic Prior medical (UPN); measurements features, neurol. problems sex (percentile) symptoms
P1 (D084); 39w: 3550g male (P>50-75), 50cm (P20)
none
P2 (D154); 34w: 2670g female (P75-90), 49cm (P90)
none
P3 (US1); female
none
40w: 4080g (P90-97), 53cm (P50) P4 (US2); 41w: 3540g male (P25-50), 52cm (P40)
none
Timepoint; PB findings BM findings Cytogenetics MDS subtype (Plt, WBC, ANC: x109/L; Hb: g/dL) Cellularity Dysplasia Blasts (%) Metaphases FISH chr. 7
recurrent RTI 3.4 yrs; RCC Plt↓ (41), WBC↓ /ANC↓ and endogenous (4.3/0.34), MCV↑ eczema since infancy, transient pancytopenia 7 yrs; progress: Plt↓, WBC (mono: at 6 mo, Plt , ANC CMML 21%, blasts: 10%, at 2 yrs; self-limiting erythro-blasts: 26%) seizures at 3 yrs recurrent RTI since 2.0 yrs; RCC Plt (96), MCV age 1.5 yrs 5.7 yrs normal, MCV↑ 12 yrs normal, MCV↑ 13 yrs normal, MCV↑ 14, 15, 17, 18 yrs normal, MCV↑ 22 yrs normal, MCV↑ endogenous eczema 20 mo; RCC Plt↓ (88), ANC↓ (0.54), MCV ↑ recurrent RTI and endogenous eczema since infancy
12 mo; RCC 13 mo 15 mo 17 mo 18.5 mo
P5 (D637);40w: 3875g (P75), male 54cm (P75)
none
none
P6 (D637f); term: male normal P7 (SC054); term: female normal
none
none
none
pancytopenia and hypocellular
7.7 yrs; RCC
Plt↓ (5), ANC↓ (0.43), Hb↓ (8.3), HbF↑ (5.2%) ANC↓ (0.88), Hb↓ (9.8) normal normal, HbF↑ (11.3%) ANC (0.6)
Plt↓ (64), WBC↓/ANC (1.7/0.08), Hb↓ (9.8), MCV↑, HbF↑ (5.7%) 42 yrs; RCMD Plt↓ (72), WBC↓/ANC↓ (2.0/1.4), Hb↓ (5.8), MCV↑ 2.1 yrs; RCC Plt↓ (74), WBC↓/ANC↓ (4.0/0.8), Hb↓ (10.2) 2.3 yrs Plt↓ (90), Hb↓ (10.5) 5, 6, 7.5, 11, normal, HbF↑ 12, 18 yrs (1.2-2.8%)
↓
+++ vacuolization in E+M +++
<5
45,XY,-7 [6] / 46,XY [10]
60%
5
45,XY,-7 [5]
95%
<5
45,XX,-7‡
77%
n.p. N N N n.p. ↓
+++ vacuolization in M n.p. + N N n.p. +
n.p. <5 <5 <5 n.p. <5
n.p 46,XX n.p. 46,XX n.p. 45,XX,-7 [3] / 46,XX [18]
n.p n.p normal normal n.p. 16%
↓
++
<5
46,XY [20]
normal
N N N N
N N N N
7 <5 <5 <5
↓
++
↓
++
N
+
N
+ N
N
↓
46,XY [20] normal 46,XY [20] 5.5% n.p. 15% 45,XY,-7 [6] / 19% 46,XY,del(7) (q11.2q36) [4] / 46,XY [10] <5 45,XY,-7, der(18;21) n.p. (q10;q10),+21 [20/20]
<5 45,XY,der(1;7)(q10; n.p. p10)[11]/ 46,XY[5] <5 45,XX,-7 [4] / -7 46,XX [17] confirmed <5 46,XX [25] normal <5 normal normal
UPN: unique patient number; syndr.: syndromic; w: gestational week; RTI: respiratory tract infection (including otitis, bronchitis, pneumonia); yrs.: years of age; mo: months of age; Dx: diagnosis; RCC: refractory cytopenia of childhood; CMML: chronic myelomonocytic leukemia; BM: bone marrow; E: erythropoiesis, M: myelopoiesis; n.p.: not performed; PB: peripheral blood; Plt: platelets; MCV: mean corpuscular volume (according to age); WBC: white blood count; ANC: absolute neutrophil count; Dysplasia: +, mild; ++, moderate; +++, severe. N, normal; FISH: fluorescence in situ hybridization; *with occasional small dysplastic megakaryocytes. ‡ 51% of metaphases with monosomy 7, of those 17% additionally showed hypoploid metaphases with involvement of chromosomes 9, 14, 19, 21.
430
haematologica | 2018; 103(3)
SAMD9L-related familial MDS
achieved a spontaneous remission as indicated by normalization of bone marrow morphology/cellularity and cytogenetics (Figure 2, Table 1). A similar clinical picture was seen in P7 who was diagnosed with RCC with a -7 clone at the age of 2.1 years, and experienced rapid cytogenetic remission with normal marrow and blood counts until the last follow-up, 16 years after diagnosis (Table 1). Hematologic findings also normalized in P4, shortly after the initial manifestation; however, fluorescence in situ hybridization revealed chromosome 7 loss in bone marrow, which gradually worsened and culminated, after 3.5 months, in the emergence of two independent clones with -7 and del(7q) (Table 1). Due to high-risk cytogenetics and disease progression, five of the seven patients (P1, P3-6) underwent hematopoietic stem cell transplantation after myeloablative conditioning. At last follow-up, six of the seven patients were alive, four after transplantation, and two without therapy.
Constitutional and acquired SAMD9L mutations Exploratory whole exome sequencing performed in family I identified two shared candidate variants in P1 and P2 evaluated as highly conserved and deleterious by in silico prediction: SAMD9L (p.V1512M) and PTEN (p.Y188C) (Table 2, Online Supplementary Figure S1). Sequencing of DNA from hair follicles confirmed the constitutional nature of both novel mutations. The SAMD9L p.V1512M variant was inherited from the mother (Figure 1A) whereas PTEN p.Y188C was of paternal origin; both parents were asymptomatic and had normal complete blood counts at the time of testing. Finally, truncating acquired SAMD9L mutation p.R1188X (VAF 5.9%) was identified in P1 in hematopoiesis (Table 2). In pedigree II, targeted next-generation sequencing in P4 revealed SAMD9L p.R986H as the most plausible candidate constitutional mutation predicted to be highly conserved and deleterious (Table 2, Figure 1A,B). This mutation was found in four individuals in ExAC (out of 120976
alleles). Additional missense variants in JAK3 p.A877V (ExAC: 11 individuals, 121372 alleles), and FANCM p.L57F (ExAC: 195 individuals, 121190 alleles) had lower and moderate pathogenicity scores, respectively (Table 2). Chromosomal breakage studies on P4 were negative thus arguing against a pathogenic role of the heterozygous FANCM variant. Germline analysis revealed SAMD9L and JAK3 variants in P3, P4, and their father, while the FANCM variant was transmitted from the mother only to P4. Both parents were asymptomatic. In pedigree III, the SAMD9L p.R986C mutation was identified in P5 and the affected father, P6 (Figure 1A). This mutation has been reported in a family with ataxiapancytopenia phenotype, with one of three carriers developing MDS/-7 at the age of 18 months.20 The HLA-identical brother of P5 was thoroughly evaluated as a potential bone marrow donor. He was clinically healthy and had a normal complete blood count, but he did not qualify as a donor because of hypocellular bone marrow with mild dysplastic features. He was also a carrier of the p.R986C mutation. In P7 of pedigree IV, targeted next-generation sequencing identified two SAMD9L mutations (Table 2): missense p.H880Q with a variant allele frequency of 27% out of 8139 reads (likely constitutional; this mutation was reported in multiple individuals within a family with ataxia pancytopenia but no MDS phenotype) and nonsense p.S1317RfsX21 likely acquired in a subclone as inferred from the much lower variant allele frequency of 10% (5934 reads). In summary, inherited SAMD9L mutations p.V1512M, p.R986H, and p.R986C were identified in three families (each with 2 individuals diagnosed with MDS/-7 and 1 healthy carrier Figure 1A,B), and p.H880Q in P7 who presented with transient monosomy 7.
Acquired mutations in known oncogenes All patients with exception of P3 and P6 were evaluated for the presence of somatic mutations in leukemia-associated genes using whole exome sequencing or targeted
Figure 2. Bone marrow findings in P1 and P2 at different timepoints during the course of the disease. Hematoxylin and eosin staining of bone marrow (BM) at diagnosis of RCC in P1 showing dysplastic granulopoiesis with hypergranulation and a pseudoPelger cell (top left), myeloblast and dysplastic eosinophil (top right). BM at diagnosis in P2 (synchronous with monosomy 7) showing hypergranulation and vacuolization in myelocytes, and dysplastic erythropoiesis with double nuclei (bottom left). Normal BM morphology in P2, 15 years after initial BM confirming spontaneous phenotype reversion (bottom right).
haematologica | 2018; 103(3)
431
V.B. Pastor et al.
next-generation sequencing. In P1 previously reported leukemia driver mutations SETBP1 p.D868N, EZH2 p.V582M, and KRAS p.Q61P were identified as somatic (Table 2, Online Supplementary Figure S2). Similarly, in P5 a somatic RUNX1 mutation c.413_427+5dup20bp was detected (Table 2). This is a novel splice-donor site mutation not present in databases. No additional mutations in leukemia driver genes were observed in other affected cases.
Clonal escape mechanisms from SAMD9L mutations are not random In comparison with other constitutional variants, SAMD9L missense mutations showed significantly lower median allelic frequencies across all patients (53% versus 20%, P<0.05) (Table 2). This finding was corroborated by the consistent partial or complete loss of chromosome 7 (Figure 3, Table 1). In P1, -7 progressively expanded from 60% at the time that RCC was diagnosed to 95% at the time of progression to CMML. In addition, P1 and P7 harbored subclones with acquired stop-gain SAMD9L mutations located upstream of the constitutional missense substitutions (Figure 3). TA cloning confirmed the cis orienta-
tion of both mutations on the same chromosome in P1 (Figure 1C). Western blot of 293T cells transiently transfected with double-mutated (p.V1512M/p.R1188X) SAMD9L plasmid revealed stable expression of the truncated protein (data not shown). In P2 and P7, the initial -7 disappeared and was replaced by a large somatic clone with uniparental disomy (UPD) 7q that duplicated the paternal wild-type SAMD9L allele (Figures 3 and 4).
The natural history of SAMD9L-related myelodysplastic syndrome reveals divergent clinical outcomes We next studied in detail the disease course in patients who were left untreated for longer periods of time. In P4, the evolution from normal karyotype to independent clones with -7 and del(7q) was rapid and occurred within a few months (Table 1). Cytogenetic progress was associated with partial recovery of complete blood counts and normalization of bone marrow cellularity, pointing to the possibility of hematopoietic stem cell niche repopulation by -7/del(7q) retaining only the wild-type SAMD9L allele. P1 developed CMML 3.6 years after the initial diagnosis of MDS/-7. This patient carried somatic driver mutations representing major subclones co-existing on the -7 back-
Table 2. Overview of germline and somatic mutations.
Mutation type
Patient # (UPN)
Time -7% (FISH / point, age Metaphases)
P1 (D084)
7 yrs (CMML)
60% / [6/16]
17 yrs
0% / -
P2 (D154)
Genotype
SAMD9Lm
c.4534G>A
p.V1512M
Myeloid sample
VAF% (depth)
BM-GR
8.3% (277), 13.3%(2755) WES 57.5% (80) WES 21.5% (424), DS 19.3% (1261) WES 59.7% (62) Sanger DS 49.0% (13116) Sanger Sanger DS 43.0% (252) DS 48.1% (n.a.) DS 49.0% (401) Sanger Sanger Sanger DS7.5% (3422) Sanger Sanger DS 27% (8139)
GERMLINE
Family II
20 mo
P4 (US2)
15 mo
Father
Family III
Family IV
Mother P5 (D637) P6 (father) Brother P7 (SC054)
ACQUIRED
P1 (D084)
7.7 yrs 42 yrs 2.3 yrs 7 yrs (CMML)
-
High / Small
D/P/D/D(65%)
25.2
None
HR (Sanger)
High / Large High / Small
D/-/D/D(61%) D/P/D/D(65%)
25.8 25.2
None None
HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) HR (Sanger) FB (Sanger) -
High / Large High / Large High / Small High / Small Weak / Small High / Small Weak / Small Weak / Small High / Small Weak / Small Weak / Small High / Large High / Large High / Large Weak / Small
D/-/D/D(61%) D/-/D/D(61%) D/P/D/D(65%) D/D/D/D(87%) T/N/B/N(74%) D/D/D/D(87%) T/N/B/N(74%) D/D/B/D(52%) D/D/D/D(87%) T/N/B/N(74%) D/D/B/D(52%) D/N/B/D(87%) D/N/B/D(87%) D/N/B/D(87%) T/-/D/-(-)
25.8 25.8 25.2 26.5 23.3 26.5 23.3 17.9 26.5 23.3 17.9 21.7 21.7 21.7 23.7
None None None 4/0.003% 11/0.01% 4/0.003% 11/0.01% 195/0.32% 4/0.003% 11/0.01% 195/0.32% None None None None
Novel, stop-gain
35
-
Known driver mut. Known driver mut. Known driver mut. Novel, splice donor
28.2 26.7 34 n.a.
-
Novel, stop-gain 28.8
-
DS
PTENp SAMD9Lm
c.563A>G c.4534G>A
p.Y188C p.V1512M
BM-GR BM-GR
PTENp PTEN SAMD9L SAMD9Lp JAK3p 5.5% / [0/20] SAMD9Lp JAK3p FANCMm SAMD9L JAK3 FANCM - / [20/20] SAMD9Lp - / [11/16] SAMD9L SAMD9Lp - / [0/25] SAMD9L
c.563A>G c.563A>G c.4534G>A c.2957G>A c.2630C>T c.2957G>A c.2630C>T c.171G>C c.2957G>A c.2630C>T c.171G>C c.2956C>T c.2956C>T c.2956C>T c.2640C>A
p.Y188C p.Y188C p.V1512M p.R986H p.A877V p.R986H p.A877V p.L57F p.R986H p.A877V p.L57F p.R986C p.R986C p.R986C p.H880Q
BM-GR PB PB PB PB PB PB PB PB PB PB BM-GR PB PB BM-GR
60% / [6/16]
SAMD9L
c.3562C>T
p.R1188X
BM-GR
KRAS SETBP1 EZH2 RUNX1
Family I Father Mother P3 (US1)
WES
Germline Conservation/ Effect (SIFT / CADD- ExAC source PhysChem diff. MutTaster score browser / Polyphen 2 / n/ % in PredictSNP) population
16% / [3/21]
5.9% (202) 8.0% (884) 37.7% (106) WES 47.8% (355) WE S69.2% (130) Sanger FB (Sanger) WES DS
P5 (D637)
7.7yrs
- / [20/20]
P7 (SC054)
2.3yrs
- / [0/25]
c.182A>C p.Q61P BM-GR c.2602G>A p.D868N BM-GR c.1744G>A p.V582M BM-GR c.413_427 BM-GR +5dup20bp SAMD9L c.3951_ p.S1317RfsX21 BM-GR DS10% (5934) 3955delTAAAG* WES
-
Mut: mutation; UPN: unique patient number; FISH: fluorescence in situ hybridization; BM: bone marrow; PB: peripheral blood; GR: granulocytes; HR: hair follicles; FB: skin fibroblast; m.: maternal origin; p.: paternal origin; VAF: variant allelic frequency; WES: whole exome sequencing; DS: targeted deep sequencing; Sanger: identified by Sanger sequencing; n.a.: not available; +yrs/mo, years/months after diagnosis. Evolutionary conservation scores, Phylop and PhastCons; PhysChem diff., physicochemical difference between amino acids. In-silico prediction: SIFT: T-tolerated, Ddeleterious; Mutation Taster: D-disease causing, N-polymorphism, P-polymorphism automatic; PolyPhen2: D-probably damaging, B- benign; PredictSNP consensus classifier: D-deleterious, N-neutral (% accuracy). Combined annotation-dependent depletion (CADD-score) of 20 means that a variant is among the top 1% of deleterious variants in the human genome; CADD-20=1%, CADD-30=0.1%, CADD-40=0.01%, CADD-50=0.001%. * mutation classified as acquired based on low allelic frequency. Gene annotations: SAMD9L (NM_001303500.1), EHZ2 (NM_152998), SETBP1 (NM_015559), KRAS (NM_004985.4), FANCM (NM_001308134), JAK3 (NM_000215), PTEN (NM_000314.4), RUNX1 (001001890).
432
haematologica | 2018; 103(3)
SAMD9L-related familial MDS
ground (Table 2). An additional somatic SAMD9L mutation p.R1188X (co-occurring in cis with p.V1512M) was found in a minor clonal fraction of approximately 6%. In contrast, P2 and P7 showed an unexpected clinical course with spontaneous hematologic recovery, disappearing monosomy 7 and the presence of a large double wild-type UPD7q clone in the bone marrow (Figures 4 and 5). Both patients remained healthy, had normal follow-up bone marrow examinations with no signs of dysplasia (Table 1), and normal compete blood counts until their last follow-ups, 20 (P1) and 16 (P7) years after initial diagnosis.
SAMD9L mutations inhibit cell proliferation An inhibitory effect on cell proliferation reported for SAMD9L mutants overexpressed in 293FT cells in vitro was termed as gain-of-function. In contrast, ectopic expression of p.T233N was shown to mitigate cell proliferation to a lesser extent in comparison to wild-type SAMD9L, and was categorized as a disease-protective or loss-of-function variant.20 SAMD9L p.R986H was previously functionally studied and shown to be gain-of-function,20 while p.H880Q was shown to induce loss of heterozygosity by -7/del(7q) or UPD7q in an Epstein-Barr virus-transformed cell line in vitro.19 To determine the effect of the SAMD9L mutation identified in families I and III, we transiently transfected 293FT cells with vectors containing disease-associated mutations p.V1512M and p.R986C respectively, along with the loss-of-function variant p.T233N. Cell proliferation was assessed in dye dilution assays. Both SAMD9L p.V1512M and p.R986C decreased dye dilution in comparison to wild-type SAMD9L and p.T233N (Figure 6A,B), pointing to an
amplified growth restrictive effect of the disease-associated variants (Figure 6B).
Discussion In this study, we describe a familial MDS syndrome caused by heterozygous missense mutations in the SAMD9L gene located on chromosome 7q21. We present seven individuals from four unrelated pedigrees who developed MDS/-7 from the age of 1 to 42 years without any neurological involvement. Unlike other reported SAMD9L mutation carriers,19,20 the sole clinical manifestation in our cases was hematologic. Other novel findings outlined here are the description of the somatic mutational landscape likely contributing to the progression of MDS, the observation of transient monosomy 7, and finally the occurrence of non-random revertant mosaicism leading to complete hematologic recovery. The SAMD9L mutations p.R986H, p.R986C, and p.V1512M identified in this cohort affect evolutionarily highly conserved amino acid residues and are assessed as pathogenic by in silico prediction. The mutation p.H880Q (P7) shows a weak conservation score; however, this mutation had already been reported as causative for the ataxia-pancytopenia phenotype.19 We were not able to test SAMD9L genetics in P7, however the unclear ataxia that this patient had been evaluated for at his last visit points to a carrier status and indicates that there must be an overlap between sole hematologic and ataxia phenotypes in SAMD9L disease. Summarizing all SAMD9L mutations recently reported or identified in our cohort, a total of six
Figure 3. Mechanisms of clonal escape from SAMD9L germline mutations. Multiple mechanisms of clonal escape from damaging germline missense SAMD9L mutations are observed and lead to complete (monosomy 7) or partial (deletion 7q) loss of chromosome 7 with decreasing mutant SAMD9L allele (red circles), both situations can lead to MDS development; UPD7q and truncating somatic SAMD9L mutations (green circles), which have a benign outcome and contribute to normal hematopoiesis. Multiple clonal outcomes can occur in a single patient.
haematologica | 2018; 103(3)
433
V.B. Pastor et al.
Figure 4. Loss of mutated SAMD9L allele due to genomic deletion or mitotic recombination. Variant allelic frequency (VAF) scores for chromosome 7 in P1 and P2. single nucleotide polymorphisms and Indels detected using whole exome sequencing (~4000 variants with a VAF score >5% and <95%), show a complete loss of chromosome 7 in P1, as the VAF scores are either low or high. P2, unlike P1 demonstrates a partial loss of the chromosome 7 after position 7q11.22 towards the q terminal site. The read depth of the single nucleotide polymorphisms for P2 was maintained throughout for chromosome 7 with no loss thus confirming that loss of heterozygosity is due to UPD and not -7q. Whole exome sequencing VAF values are marked by a star within the graph. VAF: variant allelic frequency; UPD: uniparental isodisomy. Blue line: centromere; red line: SAMD9L gene position; yellow dotted line: start of UPD. For P7, targeted nextgeneration sequencing identified 14 informative (heterozygous) polymorphisms located on chromosome 7q with an average depth of 1036 reads (Online Supplementary Table S1). Single nucleotide polymorphisms are represented in a VAF graph depicting the skewing of heterozytosity towards one allelle occurring after position g.66098482 (rs3764903).
germline mutations can be discerned (p.H880Q, p.I981T, p.R986H, p.R986C, p.C1196S, and p.V1512M). Of note, all these mutations cluster exclusively to the C-terminal half of the protein. Further, upon comparing reported mutations in the paralogue gene SAMD9 (p.R982H/C)22 with that of the present study in SAMD9L (p.R986H/C), we identified a potential mutational hotspot affecting highly conserved regions in both SAMD9L (p.984-989: GVRIIH) and SAMD9 (p980-985: GVRIIH) proteins. The reported constitutional variants in both SAMD9/SAMD9L were classified as gain-of-function based on the observation of decreased cell proliferation in a 293T cell line.20-22 Similarly, we observed growth deficiency in 293T cells harboring SAMD9L p.V1512M and p.R986C. Based on these findings, one cautious speculation hints at a gain-of-function effect that is toxic to cells. This is supported by the discovery of an acquired stopgain SAMD9L mutation in P1 and P7 that likely â&#x20AC;&#x153;eliminatesâ&#x20AC;? 434
Figure 5. Clonal evolution and spontaneous reversion due to UPD7q. Clonal evolution model in P2 (D154) depicting disease history during an observation period of 20 years. At diagnosis, initial bone marrow harbored monosomy 7 (77% by fluorescence in situ hybridization and 51% by metaphase karyotyping). Blood counts normalized 3.7 years later and since then P2 maintained normal complete blood counts until last follow-up at the age of 22 years. From the age of 12 years, repeated yearly bone marrow examinations revealed normocellular hematopoiesis with no dysplasia and normal cytogenetics. Bone marrow collected at the age of 17 years (*) was subjected to whole exome and targeted deep sequencing. Germline heterozygous SAMD9L mutation p.V1512M was detected at a variant allelic frequency (VAF) of ~20%, corresponding to a clonal size of ~40% in a diploid chromosome 7 background. Concurrently, a spontaneous genetic correction of the SAMD9L locus occurred resulting from uniparental isodisomy (UPD)7q of paternal origin. This self-corrected clone occurred either initially (dotted line) or after termination of monosomy 7 and contributed to normal hematopoiesis. Abbreviations: Dx, diagnosis; pat, paternal origin; mat, maternal origin; UPD; uniparental isodisomy, LFU; last follow-up.
germline missense mutations. In the cases studied here, complete or partial deletion of chromosome 7 and also UPD7q was non-random and each time resulted in loss of the germline-mutated SAMD9L gene copy. Additional studies, which are essential to further define the effect of SAMD9L variants, might be challenging due to the growth inhibitory effect of the alterations. It also remains to be determined whether SAMD9L missense mutations lead to increased protein stability, alter protein structure, enhance an unknown functional domain, or exert a completely neomorphic effect. We describe three silent mutation carriers from separate families demonstrating no previous relevant medical history. Despite normal complete blood counts and mean corpuscular volume, the brother of P5 had a hypocellular marrow with mild dysplasia, evidently attributed to the identical pathogenic SAMD9L mutation. This finding emphasizes the need for thorough hematologic workup, including marrow studies, in potential sibling donors especially when they lack a genetic marker for familial disease. The intrafamiliar heterogeneity regarding the hematologic presentation remains elusive; one can speculate that other yet unknown genetic or epigenetic mechanisms might act as modifiers. Thus far there is only limited knowledge about the regulation and cellular functions of SAMD9L. It has been postulated that both SAMD9L and the adjacent paralogous SAMD9 gene share functional redundancy, shaped by a long-term, possibly virus-induced selective pressure.30 haematologica | 2018; 103(3)
SAMD9L-related familial MDS
A
B
Figure 6. Functional evaluation of SAMD9L mutations. (A,B) The effect of SAMD9L mutations on cell proliferation was assessed by dye dilution assays. 293FT cells were transiently transfected with TFP-SAMD9L wild type (WT), the disease-associated mutations p.R986C and p.V1512M, and the protective variant p.T233N previously reported by Tesi et al.20 (A) Histograms depict the dye levels in transfected cells. Dye levels were monitored in TFP-transfected cells (filled gray histograms) and compared to cells expressing uniformly intermediate levels of TFP-SAMD9L wild-type (blue histograms) or variants (red/orange lines), as indicated. A single representative experiment is shown. (B) Cumulative summary of three independent experiments on inhibition of cell proliferation associated with indicated TFP-SAMD9L mutations. Values (mean Âą SD) are calculated based on a scale defined by 0 (dye levels in TFP-transfected cells) and -1 (dye levels in cells transfected with TFP-SAMD9L wild-type). Unpaired t-test, two tailed: *P<0.05; ** P<0.005; ***P<0.001.
Both genes can be upregulated by type I and II interferons20,31-33 and by this might suppress inflammatory pathways and exert anti-viral properties.31,34 Although there was no evidence for a defined immunodeficiency in our cohort, three of the seven patients experienced recurrent respiratory infections with or without cytopenias before they developed RCC/-7. Further evidence supports that SAMD9/SAMD9L genes act as tumor suppressors as their inactivation is associated with increased cellular proliferation, i.e. in normophosphatemic familiar tumoral calcinosis (SAMD9), in hepatitis B virus-associated hepatocellular carcinoma (SAMD9L), in MDS/acute myeloid leukemia with microdeletion in 7q21 (both genes), and in Samd9lhaploinsufficient mice. Based on the observations in this murine model, the cytokine-receptor complexes cannot be properly degraded due to impaired endosomal function in SAMD9L-haploinsufficient cells, which results in constitutive intracellular signaling with prolonged cell survival.18 Moreover, Samd9l+/- and Samd9l-/- mice develop MDS with normo/hypercellular bone marrow, drawing a parallel to human SAMD9L-related MDS in which the initial marrow hypocellularity associated with the â&#x20AC;&#x153;toxicâ&#x20AC;? mutation is restored upon the loss of the mutant SAMD9L allele. This loss is accomplished through -7, del(7q) or UPD7q, and leads to a proliferative advantage with clonal expansion. In our cohort, although initially all patients presented with hypocellular marrow, longer observation periods in P1, P2, P4, and P7 revealed successive normalization of marrow cell content associated with increasing -7/del(7q) clone or the appearance of UPD7q. In P1, the leukemic progression was further aggravated by the accumulation of typical MDS driver mutations in SETBP1, KRAS and EZH2 within the -7 clone. P5 also demonstrated an acquired splice site mutahaematologica | 2018; 103(3)
tion in RUNX1, a known leukemic driver gene. Notably, typical adult MDS driver mutations (i.e., TET2, DNMT3A, and IDH1/2) were not encountered in our cohort. This is in line with previously published findings discussing SETBP1, RAS pathway mutations and RUNX1 (identified in our SAMD9L-mutated patients) as recurrent drivers of pediatric MDS.10 Building on our observations we propose the following mechanism of MDS evolution in SAMD9L disease: the bone marrow attempts to circumvent the toxicity of the constitutional SAMD9L mutation and selects for fitter, yet premalignant 7/del(7q) clones (with only one wild-type SAMD9L copy), or benign clones with truncated SAMD9L (Figure 3). Over time the resulting haploinsufficiency of tumor supressor genes on 7q (e.g. EZH2 or CUX1) in all patients likely provides the first step towards progression. Finally, additional driver somatic mutations might be encountered in some but not all patients. Somatic revertant mosaicism has been reported in inherited bone marrow failure syndromes with hypocellular bone marrow, including telomeropathy with germline mutations in TERT,35 and Fanconi anemia in which mosaicism in blood occurs at rates of up to 140 times higher than in the general population.36,37 However, in general, revertant mosaicism is a rare facet to clonal hematopoiesis because spontaneous correction of the pathogenic allele is a random event. In our study, we report two patients (P2 and P7) who presented with RCC/-7 at young age and demonstrated complete hematopoietic remission with normal cytogenetics throughout an observation period ranging from 16 to 20 years. The clinical picture of these patients fits the previously described transient monosomy 7 syndrome; to our knowledge eight patients with primary MDS with transient -7 or del(7q) have been reported in the literature.14-16,38-40 Their ages at diagnosis ranged from 8 435
V.B. Pastor et al.
months to 3 years, and spontaneous remission was achieved within 1-20 months from diagnosis. It seems that the -7 clones in our patients either vanished spontaneously or were outcompeted by “fitter” UPD7q-corrected clones with a diploid copy of the wild-type SAMD9L allele. Based on these observations, a watch-and-wait strategy might be proposed for younger patients with RCC/-7 who have no additional somatic driver mutations and are clinically stable. However, prolonged “watchful waiting” poses the risk of progression as witnessed in P1, who developed CMML and acquired oncogenic mutations 3.6 years after he was diagnosed with RCC. In conclusion, our observations establish the molecular basis of a distinct subtype of familial MDS and point to the notion that MDS with chromosome 7 loss can be the sole and common manifestation of SAMD9L-related disease. The negative mutational effect leads to escape and outgrowth of clones carrying -7/del(7q) with only wildtype SAMD9L allele, which might spontaneously dissapear or persist and provide the first step towards disease progression. Finally, this is the first description of longterm revertant mosaicism due to non-random UPD7q in SAMD9L disease, and a plausible explanation for transient monosomy 7 syndrome. Acknowledgments We are grateful to Sophia Hollander, Christina Jäger, Yahaira Pastor, Alexandra Fischer, Wilfried Truckenmüller (Freiburg),
References 1. Churpek JE, Godley LA. How I diagnose and manage individuals at risk for inherited myeloid malignancies. Blood. 2016;128 (14):1800-1813. 2. Feurstein S, Drazer MW, Godley LA. Genetic predisposition to leukemia and other hematologic malignancies. Semin Oncol. 2016;43(5):598-608. 3. Hahn CN, Chong CE, Carmichael CL, et al. Heritable GATA2 mutations associated with familial myelodysplastic syndrome and acute myeloid leukemia. Nat Genet. 2011; 43(10):1012-1017. 4. Smith ML, Cavenagh JD, Lister TA, Fitzgibbon J. Mutation of CEBPA in familial acute myeloid leukemia. N Engl J Med. 2004;351(23):2403-2407. 5. Song WJ, Sullivan MG, Legare RD, et al. Haploinsufficiency of CBFA2 causes familial thrombocytopenia with propensity to develop acute myelogenous leukaemia. Nat Genet. 1999;23(2):166-175. 6. Pippucci T, Savoia A, Perrotta S, et al. Mutations in the 5' UTR of ANKRD26, the ankirin repeat domain 26 gene, cause an autosomal-dominant form of inherited thrombocytopenia, THC2. Am J Hum Genet. 2011;88(1):115-120. 7. Zhang MY, Churpek JE, Keel SB, et al. Germline ETV6 mutations in familial thrombocytopenia and hematologic malignancy. Nat Genet. 2015;47(2):180-185. 8. Polprasert C, Schulze I, Sekeres MA, et al. Inherited and somatic defects in DDX41 in myeloid neoplasms. Cancer Cell. 2015;27 (5):658-670. 9. Wlodarski MW, Hirabayashi S, Pastor V, et
436
10.
11.
12.
13.
14.
15.
16.
17.
Tamara Szattler (Stockholm) and Bart Przychodzien (USA) for their excellent laboratory assistance and data management, and to Dr. Anett Schmidt, Dr. Dagmar Möbius, and Dr. Elisabeth Holfeld (Cottbus) for clinical care. The authors are grateful to the Genomics Core Facility at the German Cancer Research Center/DKFZ, Heidelberg, Germany for performing whole exome sequencing and Claritas Genomics (Cambridge, MA, USA) for targeted sequencing in P4. Funding Deutsche Krebshilfe (Max Eder grant #109005) to MWW, BMBF (DKTK German Cancer Consortium, topic molecular diagnostics of pediatric malignancies) to CMN and MWW, BMBF (e:Med FKZ 01ZX1409B) to MB, the German Science Foundation (DFG, SFB 850, to MB and EXE306 to HB), the Research Council, Swedish Cancer Society, and Foundation for Strategic Research to YTB. MWW and ME are past trainees of EHA-ASH translational research training in hematology. MV is an EMBO Long-term Fellow (ALTF 206-2015 co-funded by the European Commission (LTFCOFUND2013, GA-2013609409). JCA is a Wellcome Trust Senior Research Fellow in Clinical Science (grant 098513/Z/12/Z) with support from the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital for Children, NHS Foundation Trust and UCL, and Great Ormond Street Children's Charity. This study was supported in part by the Excellence Initiative of the German Research Foundation (GSC4, Spemann Graduate School).
al. Prevalence, clinical characteristics, and prognosis of GATA2-related myelodysplastic syndromes in children and adolescents. Blood. 2016;127(11):1387-1397. Pastor V, Hirabayashi S, Karow A, et al. Mutational landscape in children with myelodysplastic syndromes is distinct from adults: specific somatic drivers and novel germline variants. Leukemia. 2017;31(3): 759-762. Churpek JE, Pyrtel K, Kanchi KL, et al. Genomic analysis of germ line and somatic variants in familial myelodysplasia/acute myeloid leukemia. Blood. 2015;126(22): 2484-2490. Gohring G, Michalova K, Beverloo HB, et al. Complex karyotype newly defined: the strongest prognostic factor in advanced childhood myelodysplastic syndrome. Blood. 2010;116(19):3766-3769. Kardos G, Baumann I, Passmore SJ, et al. Refractory anemia in childhood: a retrospective analysis of 67 patients with particular reference to monosomy 7. Blood. 2003;102(6):1997-2003. Mantadakis E, Shannon KM, Singer DA, et al. Transient monosomy 7 - A case series in children and review of the literature. Cancer. 1999;85(12):2655-2661. Parker TM, Klaassen RJ, Johnston DL. Spontaneous remission of myelodysplastic syndrome with monosomy 7 in a young boy. Cancer Genet Cytogenet. 2008;182(2): 122-125. Leung EW, Woodman RC, Roland B, Abdelhaleem M, Freedman MH, Dror Y. Transient myelodysplastic syndrome associated with isochromosome 7q abnormality. Pediatr Hematol Oncol. 2003;20(7):539-545. Asou H, Matsui H, Ozaki Y, et al.
18.
19.
20.
21.
22.
23.
24.
25.
Identification of a common microdeletion cluster in 7q21.3 subband among patients with myeloid leukemia and myelodysplastic syndrome. Biochem Biophys Res Commun. 2009;383(2):245-251. Nagamachi A, Matsui H, Asou H, et al. Haploinsufficiency of SAMD9L, an endosome fusion facilitator, causes myeloid malignancies in mice mimicking human diseases with monosomy 7. Cancer Cell. 2013;24(3):305-317. Chen DH, Below JE, Shimamura A, et al. Ataxia-pancytopenia syndrome is caused by missense mutations in SAMD9L. Am J Hum Genet. 2016;98(6):1146-1158. Tesi B, Davidsson J, Voss M, et al. Gain-offunction SAMD9L mutations cause a syndrome of cytopenia, immunodeficiency, MDS and neurological symptoms. Blood. 2017;129(16):2266-2279. Narumi S, Amano N, Ishii T, et al. SAMD9 mutations cause a novel multisystem disorder, MIRAGE syndrome, and are associated with loss of chromosome 7. Nat Genet. 2016;48(7):792-797. Buonocore F, Kuhnen P, Suntharalingham JP, et al. Somatic mutations and progressive monosomy modify SAMD9-related phenotypes in humans. J Clin Invest. 2017;127(5): 1700-1713. Schwartz JR, Wang S, Ma J, et al. Germline SAMD9 mutation in siblings with monosomy 7 and myelodysplastic syndrome. Leukemia. 2017;31(8):1827-1830. Baumann I, Niemeyer CM BJ, Shannon K. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon: IARC Press; 2008, 2008:104-107. Vardiman JW, Harris NL, Brunning RD. The World Health Organization (WHO) classifi-
haematologica | 2018; 103(3)
SAMD9L-related familial MDS
26.
27. 28.
29.
30.
cation of the myeloid neoplasms. Blood. 2002;100(7):2292-2302. Kircher M, Witten DM, Jain P, O'Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat Genet. 2014;46(3):310-315. Grantham R. Amino acid difference formula to help explain protein evolution. Science. 1974;185(4154):862-864. Wlodarski MW, O'Keefe C, Howe EC, et al. Pathologic clonal cytotoxic T-cell responses: nonrandom nature of the T-cell-receptor restriction in large granular lymphocyte leukemia. Blood. 2005;106(8):2769-2780. Vraetz T, Emanuel PD, Niemeyer CM. In vitro regulation of colony stimulating factormediated hematopoiesis in healthy individuals and patients with different types of myeloproliferative disease. Methods Mol Biol. 2003;215:293-309. Lemos de Matos A, Liu J, McFadden G, Esteves PJ. Evolution and divergence of the mammalian SAMD9/SAMD9L gene family. BMC Evol Biol. 2013;13:121.
haematologica | 2018; 103(3)
31. Chefetz I, Ben Amitai D, Browning S, et al. Normophosphatemic familial tumoral calcinosis is caused by deleterious mutations in SAMD9, encoding a TNF-alpha responsive protein. J Invest Dermatol. 2008;128(6): 1423-1429. 32. Tanaka M, Shimbo T, Kikuchi Y, Matsuda M, Kaneda Y. Sterile alpha motif containing domain 9 is involved in death signaling of malignant glioma treated with inactivated Sendai virus particle (HVJ-E) or type I interferon. Int J Cancer. 2010;126(8):19821991. 33. Hershkovitz D, Gross Y, Nahum S, et al. Functional characterization of SAMD9, a protein deficient in normophosphatemic familial tumoral calcinosis. J Invest Dermatol. 2011;131(3):662-669. 34. Liu J, Wennier S, Zhang L, McFadden G. M062 is a host range factor essential for myxoma virus pathogenesis and functions as an antagonist of host SAMD9 in human cells. J Virol. 2011;85(7):3270-3282. 35. Jongmans MCJ, Verwiel ETP, Heijdra Y, et al. Revertant somatic mosaicism by mitotic
36.
37.
38.
39.
40.
recombination in dyskeratosis congenita. Am J Hum Genet. 2012;90(3):426-433. Waisfisz Q, Morgan NV, Savino M, et al. Spontaneous functional correction of homozygous fanconi anaemia alleles reveals novel mechanistic basis for reverse mosaicism. Nat Genet. 1999;22(4):379-383. Reina-Castillon J, Pujol R, Lopez-Sanchez M, et al. Detectable clonal mosaicism in blood as a biomarker of cancer risk in Fanconi anemia. Blood Advances. 2017;1(5): 319-329. Nagasawa M, Tomizawa D, Tsuji Y, et al. Pancytopenia presenting with monosomy 7 which disappeared after immunosuppressive therapy. Leuk Res. 2004;28(3):315-319. Benaim E, Hvizdala EV, Papenhausen P, Moscinski LC. Spontaneous remission in monosomy 7 myelodysplastic syndrome. Br J Haematol. 1995;89(4):947-948. Scheurlen W, Borkhardt A, Ritterbach J, Huppertz HI. Spontaneous hematological remission in a boy with myelodysplastic syndrome and monosomy 7. Leukemia. 1994;8(8):1435-1438.
437
ARTICLE
Myeloproliferative Neoplasms
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):438-446
Benefits and pitfalls of pegylated interferon-α2a therapy in patients with myeloproliferative neoplasm-associated myelofibrosis: a French Intergroup of Myeloproliferative neoplasms (FIM) study
Jean-Christophe Ianotto,1* Aurélie Chauveau,2* Françoise Boyer-Perrard,3 Emmanuel Gyan,4 Kamel Laribi,5 Pascale Cony-Makhoul,6 Jean-Loup Demory,7 Benoit de Renzis,8 Christine Dosquet,9 Jerome Rey,10 Lydia Roy,11 Brigitte Dupriez,12 Laurent Knoops,13 Laurence Legros,14 Mohamed Malou,15 Pascal Hutin,16 Dana Ranta,17 Omar Benbrahim,18 Valérie Ugo,19 Eric Lippert2** and Jean-Jacques Kiladjian20**
Service d’Hématologie Clinique, Institut de Cancéro-hématologie, CHRU Brest, France; Laboratoire d’Hématologie, CHRU de Brest and INSERM U1078, Université de Bretagne Occidentale, Brest, France; 3Service des Maladies du Sang, CHU d’Angers, France; 4 Hématologie et Thérapie Cellulaire, CRU de Cancérologie H.S. Kaplan, Tours, France; 5 Service d’Hématologie, CH Le Mans, France; 6Service d’Hématologie, CH AnnecyGenevois, France; 7Service d’Hématologie, Hôpital St Vincent de Paul, Lille, France; 8 Service d’Hématologie, CHU Clermont-Ferrand, France; 9APHP, Saint Louis Hospital, Cell Biology Department, Paris, France; 10Département d’Hématologie, Institut Paoli-Calmette, Marseille, France; 11Service d’Hématologie, Hôpital de Créteil, France; 12Service d’Hématologie Clinique, CHU de Lens, France; 13Cliniques Universitaires Saint-Luc and Université Catholique de Louvain, Brussels, Belgium; 14Service d’Hématologie, CHU de Nice, France; 15Service d’Oncologie et D’Hématologie, Hôpital de Morlaix, France; 16Service de Médecine Interne et de Maladies Infectieuses, Hôpital Laennec, Quimper, France; 17 Département d’Hématologie, Hôpital Universitaire de Nancy, Vandœuvre-lès-Nancy, France; 18Service d’Hématologie, Hôpital La Source, Orléans, France; 19Laboratoire d’Hématologie, CHU d’Angers, France and 20Centre d’Investigation Clinique, Hôpital SaintLouis, APHP, Université Paris Diderot, Inserm, Paris, France 1 2
*JCI and AC contributed equally to this manuscript.
Correspondence:
**EL and JJK contributed equally to this manuscript.
jean-jacques.kiladjian@aphp.fr ABSTRACT Received: September 22, 2017. Accepted: December 6, 2017. Pre-published: December 7, 2017.
doi:10.3324/haematol.2017.181297 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/438 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
438
W
e have previously described the safety and efficacy of pegylated interferon-α2a therapy in a cohort of 62 patients with myeloproliferative neoplasm-associated myelofibrosis followed in centers affiliated to the French Intergroup of Myeloproliferative neoplasms. In this study, we report their long-term outcomes and correlations with mutational patterns of driver and nondriver mutations analyzed by targeted next generation sequencing. The median age at diagnosis was 66 years old, the median follow-up since starting pegylated interferon was 58 months. At the time of analysis, 30 (48.4%) patients were alive including 16 still being treated with pegylated interferon. The median survival of patients with intermediate and high-risk prognostic Lille and dynamic International Prognostic Scoring System scores treated with pegylated interferon was increased in comparison to that of historical cohorts. In addition, overall survival was significantly correlated with the duration of pegylated interferon therapy (70 versus 30 months after 2 years of treatment, P<10-12). JAK2V617F allele burden was decreased by more than 50% in 58.8% of patients and two patients even achieved complete molecular response. Next-generation sequencing analyses performed in 49 patients showed that 28 (57.1%) of them carried non-driver mutations. The presence of at least one additional mutation was associated with a reduction of both overall and leukemia-free survival. These findings in a large series of patients with myelofibrosis suggest that pegylated haematologica | 2018; 103(3)
Pegylated interferon-α in myelofibrosis
interferon therapy may provide a survival benefit for patients with intermediate- or high-risk Lille and dynamic International Prognostic Scoring System scores. It also reduced the JAK2V617F allele burden in most patients. These results further support the use of pegylated interferon in selected patients with myelofibrosis. (Clinicaltrials.gov #NCT02910258 and #NCT02897297).
Introduction Primary myelofibrosis is a Philadelphia chromosomenegative myeloproliferative neoplasm characterized by splenomegaly, constitutional symptoms and cytopenia and/or proliferative features in peripheral blood. Secondary myelofibrosis may develop from either polycythemia vera or essential thrombocythemia.1,2 The main causes of death of patients with myelofibrosis include disease progression leading to cachexia or infection, and acceleration and transformation of their disease into acute myeloid leukemia.3 The JAK2V617F mutation can be found in about half of myelofibrosis patients, while 20-30% carry a calreticulin (CALR) mutation and 5-10% a mutation in the thrombopoietin receptor gene MPL. These three driver mutations influence the clinical presentation and outcome: for example, CALR-mutated myelofibrosis patients are predominantly male, have higher platelet and lower leukocyte and red cell counts, and longer survival than those with the JAK2V617F mutation.4,5 In addition to these three driver mutations, other mutations are frequently found in myelofibrosis patients, mainly in genes involved in epigenetic regulation or the splicing machinery. Some of these mutations have been associated with poorer survival and Vannucchi et al. have defined five “high molecular risk” genes: ASXL1, EZH2, SRSF2, and IDH1/2. Mutations in any of these genes dramatically decreased overall and event-free survival of the affected patients and the presence of more than one additional mutation conferred an even worse outcome.6,7 Several strategies have been used to alleviate the proliferative aspects of myelofibrosis (e.g., hydroxyurea, pipobroman, 6-mercaptopurine) or the cytopenic ones (e.g., thalidomide and its derivatives, androgens, recombinant erythropoietin), as well as to manage splenomegaly (e.g., hydroxyurea, radiotherapy, splenectomy). The efficacy of these approaches is generally modest, especially with regards to cytopenia and does not clearly modify disease evolution.2 Ruxolitinib, a non-specific JAK1/JAK2 inhibitor approved for the treatment of symptomatic myelofibrosis patients, was recently shown to be very effective in reducing the inflammatory component of these diseases with significant improvement of pruritus, fever, weight loss and splenomegaly. Although still debated, results of the COMFORT I and II studies also suggest that ruxolitinib may increase overall survival of high-risk myelofibrosis patients compared to that of patients treated in the placebo or “best available therapy” arms.8-12 To date, however, allogeneic stem cell transplantation remains the only curative option for these patients.13-16 We have previously reported on the feasibility and hematologic results of myelofibrosis treatment with pegylated interferon-α2a in a prospective observational study conducted by the French Intergroup of Myeloproliferative neoplasms (FIM).17,18 Herein, we report the long-term outcomes of this large cohort of patients, focusing on survival haematologica | 2018; 103(3)
and incidence of acute leukemia. In addition, next-generation sequencing data enabled us to assess the impact of interferon treatment on the prognosis associated with mutational patterns, and with the presence of non-driver mutations.
Methods Patients’ recruitment Between December 2006 and April 2011 we prospectively recruited 62 patients treated with pegylated interferon-α2a in 17 centers affiliated to the FIM group. The inclusion criteria and methodology have been described elsewhere.18 This study was approved by the local Institutional Review Board and registered in ClinicalTrials.gov (NCT02910258). All participants gave written informed consent. The patients treated in the CHRU of Brest were also registered in the OBENE observatory (NCT02897297). Pegylated interferon-α2a was initiated by physicians in accordance with local and national guidelines. During the period of this study, ruxolitinib was only available through clinical trials (approval for use in myelofibrosis in France was obtained in August 2012).
Molecular analyses Samples from all the patients were characterized for the three driver mutations. Genomic DNA was extracted from blood neutrophils or total leukocytes using the Flexigene DNA kit (Qiagen, Germany) according to the manufacturer’s recommendation. JAK2V617F was quantified by real-time quantitative polymerase chain reaction analysis according to previously described methods.19 MPLW515K/L mutations were screened for using the MPLW515L/K MutaScreen Kit (Qiagen) according to manufacturer’s instructions. Quantitative polymerase chain reactions were performed on ABI7500 instruments (Applied Biosystems). CALR exon 9 mutations were screened for by fragment analysis according to published methods.4 Polymerase chain reaction products were analyzed on an ABI3130 instrument (Applied Biosystems).
Next-generation sequencing Targeted next-generation sequencing was performed in 49 samples collected at the time of starting pegylated interferon-α2a treatment (34 JAK2V617F-positive, 12 CALR-positive, 3 triple-negative). The next-generation sequencing panel included 26 genes (ASXL1, BCOR, CBL, CSF3R, DNMT3A, ETNK1, ETV6, EZH2, IDH1, IDH2, JAK2, KRAS, MPL, NRAS, PDGFRA, RUNX1, SETBP1, SF3B1, SH2B3, SRSF2, STAG2, TET2, TP53, U2AF1, ULK1, and ZRSR2) and the sequencing was performed using AmpliseqTM (Thermo Fisher Scientific, Foster City, CA, USA) custom design. Library preparation and sequencing using PGMTM (Thermo Fisher Scientific) were performed according to the manufacturer’s instructions. Mutations were detected using the Variant Caller v4.2 plugin from Torrent Suite Software and IonReporter v5.2 (Life Technologies). For mutation calling, arbitrary filters were fixed with variant allele frequencies >2% and depth >50X. False positive variants were dropped after BAM analysis on Alamut® 439
J-C. Ianotto et al.
(Interactive Biosoftware). Only exonic non-synonymous mutations were analyzed.
Statistical analyses The Student t-test, chi-squared test and Kaplan-Meier curves were applied using the R-project (3.1.2 version, BiostaGV website, hosted by the Institute for Statistics and Mathematics of Wirtschaftuniversität Wien, Austria). Results were considered statistically significant if the P value was less than 0.05, and each value was expressed plus or minus the standard deviation. Overall survival was defined as the period between diagnosis of myelofibrosis or, when indicated, initiation of pegylated interferon-α2a treatment and last visit or death. Leukemia-free survival was defined as survival without transformation to acute leukemia. Univariate analyses were performed based on either an exact Fisher test or Wilcoxon rank sum test. Variables that were found to be associated with the outcome at the 10% level were then introduced into a multivariate logistic model.
7.4 years from the diagnosis of myelofibrosis whereas the median leukemia-free survival had not been reached (Figure 1A,B). The 5-year actuarial survival rate for the whole cohort was 69.4% from diagnosis and 54.8% from the first prescription of pegylated interferon-α2a. The duration of pegylated interferon-α2a therapy had a significant impact on overall survival: the median overall survival was 30 months in patients who received less than 2 years of treatment compared to 70 months for patients who received the drug for more than 2 years (P<0.0001). As expected, the Lille and DIPSS scores differentiated patients treated with pegylated interferon-α2a in terms of overall survival (Figure 1C,D). However, the median survival observed in this cohort was clearly longer than that
Table 1. Characteristics of the patients and their disease.
Variable
Results Patients’ characteristics Sixty-two patients with primary myelofibrosis (n=29, 46.8%) or secondary myelofibrosis (n=33, 53.2%) were included in this study. The median age of these patients at the time their myelofibrosis was diagnosed was 66 years old (range, 33-81) and the mean interval between the diagnosis of myelofibrosis and the beginning of pegylated interferon-α2a treatment was 19.1 months. Forty-two patients (68%) had been previously treated and 40 of them (95%) had received hydroxyurea. The patients’ characteristics are summarized in Table 1. At the time of starting treatment with pegylated interferon-α2a, the median age of the patients with primary myelofibrosis was 64 years, whereas that of the patients with secondary myelofibrosis was 70.5 years old. The majority of the patients were over 65 years old (35/62, 56.5%). The male/female sex ratio was 1.38. Splenomegaly was present in 43 patients (69.4%), constitutional symptoms were documented in 28 (45.2%). More than two-thirds of the patients (44/62, 71%) were in a proliferative phase (leukocytosis and/or thrombocytosis). However, 36 patients were anemic (58.1%) and 13 were transfusion-dependent (36%). Most of the patients were classified in the “Intermediate-2” risk category according to International Prognostic Scoring System (IPSS) or Dynamic IPSS (DIPSS) scores. The mutational status for driver mutations was identified for all patients: 42 (68%) had JAK2V617F, 14 had CALR exon 9 mutation (9 with type 1, 3 with type 2, and 2 with other mutations), one had MPLW515, four were triple-negative (3NEG) and one had coexisting JAK2 and MPL mutations. Karyotype was available for 37 (59.7%) patients, and was normal in 70.3% (Table 1).
Survival and leukemic transformation The median follow-up after starting treatment with pegylated interferon-α2a was 58 months (range, 9-107), whereas the median follow-up after having been diagnosed with myelofibrosis was 69.6 months (range, 10178). The median duration of pegylated interferon-α2a treatment was 39 months (range, 6-107). At the time of the analysis, 30 patients (48.4%) were still alive. The median overall survival of the cohort was 440
Value [range]
Myelofibrosis subtype, n. Primary myelofibrosis Post-PV myelofibrosis Post-ET myelofibrosis Median age at the beginning of interferon (years) All the patients Primary myelofibrosis Secondary myelofibrosis Patients ≥ 65 years, n. (%) Mean time between diagnosis of myelofibrosis and start of interferon therapy (months) Risk category, % Lille score (Low / Intermediate / High) IPSS score (Low / Int1 / Int2 / High) DIPSS score (Low / Int1 / Int2 / High) Male / female, n. Previous therapy, n. (%) Number of patients Median number of treatments per patient Hydroxyurea Pipobroman Anagrelide 6 Mercaptopurine Driver mutations, n. (%) JAK2V617F CALR MPL MPL/JAK2 Triple negative Karyotype, n. (%) Normal Deletion 20q Anomaly 9 Trisomy 21 Monosomy Y Clinical parameters, n. (%) Splenomegaly Constitutional symptoms Biological parameters Median white blood cell count (109/L) Median hemoglobin (g/L) Median platelet count (109/L)
29 19 14 67 [33-81] 64 [35-81] 70.5 [33-79] 35 (56,5) 19.1
50 / 40.3 / 9.7 14.7 / 27.9 / 36.1 / 21.3 16.1 / 37.1 / 41.9 / 4.9 36/26 42 (68) 1.6 40 (95) 16 (38) 10 (24) 7 (17) 62 42 (67.7) 14 (22.6) 1 (1.6) 1 (1.6) 4 (6.4) 37 (59.7) 26 (70.3) 4 (10.8) 3 (8.1) 2 (5.4) 2 (5.4) 43 (69.4) 28 (45.2) 10.5 [1.3-78.3] 103 [74-160] 378 [23-1396]
PV: polycythemia vera; ET: essential thrombocythemia; IPSS: International Prognostic Scoring System; DIPSS: Dynamic International Prognostic Scoring System; Int1: intermediate-1; Int2: intermediate-2.
haematologica | 2018; 103(3)
Pegylated interferon-α in myelofibrosis
reported in the reference cohorts used to establish the prognostic scores, especially in higher risk categories: according to the Lille score (8.9 versus 7.75 years for low risk, 5.42 versus 2.17 years for intermediate rsik and 4.46 versus 1.08 years for high-risk) and to the DIPSS score (6.9 versus 4 years for intermediate-2 and 4.58 versus 1.5 years for high-risk patients).20,21 For patients with IPSS intermediate-2 or high-risk, the 5-year actuarial survival rate was 60% from diagnosis and 48.6% from the first prescription of interferon. We did not observe any differences between patients with primary and secondary myelofibrosis with regards to median overall survival (7.4 versus 7 years, P=0.82) or leukemia-free survival (not reached for both, P=0.95). The type of driver mutation had a statistically significant impact on survival: the median overall survival was 13.5 years for CALR-mutated patients compared to 7 years for JAK2-mutated patients (P<0.0001). Causes of death were documented in 29/32 patients (90.6%): eight had a secondary malignancy including transformation to AML in seven and secondary cancer in one, seven died of complications of myelofibrosis/cytopenia, five transplanted patients had fatal graft-versus-host disease (GvHD), five had cardiovascular events and four died of infections. Overall, the disease evolved to acute myeloid leukemia in eight patients (13%), only one of whom was alive at the time of the analysis. Transformation to AML occurred during pegylated interferon-α2a treatment in three
patients at a median time of 1.2 years after initiation of pegylated interferon-α2a therapy (1.3 years since the diagnosis of myelofibrosis). In five patients, the transformation to acute myeloid leukemia occurred after discontinuation of pegylated interferon-α2a, at a median of 4.2 years after initiation of interferon and 6.8 years since the diagnosis of myelofibrosis (the median duration of interferon treatment in these 5 patients was 2.1 years).
A
B
C
D
Discontinuation of pegylated interferon-α2a At the time of analysis, 16 patients (25.8% of the entire cohort, 53.3% of the living patients) were still being treated with pegylated interferon-α2a. Forty-five patients (72.6%) had discontinued interferon treatment: 25 (55.6%) due to resistance and 20 (44.4%) due to intolerance. Resistance was defined by myelofibrosis progression (n=19), transformation to acute myeloid leukemia (n=3) or failure of the disease to improve (n=3). Intolerance included occurrence of new cytopenia (n=8), psychiatric complications (n=6), fatigue (n=2), cutaneous porphyria (n=1), type 2 diabetes mellitus (n=1) or other (n=2). The median duration of pegylated interferon-α2a treatment was 20 months in patients with resistance compared to 12 months in those with intolerance. Patients developing intolerance to pegylated interferon-α2a had longer median overall survival and leukemia-free survival than patients with resistance (P=10-5 and P=0.048, respectively) (Figure 2A,B). Of the 45 patients who stopped interferon treatment, 15
Figure 1. Survival of the whole study cohort. (A) Overall and (B) leukemia-free survival of the whole cohort and survivals according to the prognostic (C) Lille and (D) DIPSS scores.
haematologica | 2018; 103(3)
441
J-C. Ianotto et al.
A
B
Figure 2. Survival according to treatment status. Kaplan-Meier estimated (A) overall and (B) leukemia-free survival differentiating patients who were still being treated with pegylated-interferon from patients who had stopped interferon because of intolerance or resistance.
(33.3%) were given ruxolitinib, seven (15.6%) underwent allogeneic stem cell transplantation (ASCT) and 23 patients (51.1%) were treated with different drugs or received no further medicine (Figure 3). The median survival after cessation of pegylated interferon-α2a was 17 months (range, 3-62). The median survival of patients who received ruxolitinib was 22 months compared to 14 months for those who did not (P=0.12) or 10 months for patients who underwent ASCT (P=0.003). Nineteen patients (30.6%) received an erythropoietinstimulating agent during interferon therapy; we did not observe that these patients, compared to those not given such agents, had increased resistance to pegylated interferon-α2a, greater occurrence of acute myeloid leukemia or a difference in overall or leukemia-free survival.
Evolution of the allele burden of driver mutations The median mutant allele burdens at the time of starting pegylated interferon-α2a treatment, studied in 31 JAK2mutated and eight CALR-mutated patients were 66.8% (range, 8.9-98.3) and 41.2% (range, 32-46.1), respectively. The JAK2V617F allele burden was quantified serially in 27/31 patients. In this group, the median allele burden prior to pegylated interferon-α2a treatment was 57.3%; the burden remained stable during the first year and then decreased to 47.1% at 24 months and 29% at 36 months. We observed a decrease of mutant allele burden with a more than 10% reduction in 17/27 patients (63%), more than 20% in 15/27 patients (55.6%), and more than 50% in 10/27 patients (37%). Four patients (15%) achieved a reduction of more than 95%, including two patients who had complete molecular responses (below the detection threshold of 0.1% in our assay). Among the other patients, 7/27 (26%) had a stable allele burden (±10%) and three patients had a more than 10% increase in allele burden (Figure 4). We did not observe any difference of outcome (death or acute myeloid leukemia evolution) between patients whose JAK2V617F allele burden did or did not decrease. Sequential quantification of mutant CALR allele burden was available in only four patients. The median value remained essentially stable, altering from 42.4% to 46.8%. Only one patient experienced a reduction of mutant CALR allele burden, which decreased by 33%. 442
Figure 3. Patients’ treatment. ASCT: allogeneic stem cell transplantation; disc: discontinuation (of Peg-Ifn); dur: duration; FU: follow-up; m: months; n: number; Peg-Ifn: pegylated-interferon.
Impact of non-driver mutations Of the 49 patients analyzed with targeted next-generation sequencing, 28 (57.1%) carried at least one additional mutation different from the driver mutation. Overall 44 mutations were identified (1.6 per patient) in 16 different genes (Figure 5). Of these mutations, 47% affected epigenetic regulators, 21% signaling and 16% splicing or other categories (Figure 5). The most frequent mutations involved ASXL1 and TET2 genes (7 cases each). The number of patients harboring non-driver mutations was similar between JAK2-mutated (21/34, 61.8%) and CALRmutated (6/12, 50%) patients (P=0.51). Additional mutations were found in 68% (23/34) of patients who discontinued pegylated interferon-α2a treatment (9/15, 60% for intolerance and 14/19, 74% for resistance) compared to only 33% (5/15) of patients who remained on pegylated interferon-α2a treatment (P=0.02). Patients with at least one non-driver mutation had shorter overall survival than those with only driver mutations (6.1 years versus not reached, P=0.06) (Figure 6A). The same was true for leukemia-free survival (not reached in both groups, P=0.026) (Figure 6B). In detail, leukemiafree survival was significantly different between patients haematologica | 2018; 103(3)
Pegylated interferon-Îą in myelofibrosis
carrying no (median not reached), one (median not reached) or several additional mutations (median 6.7 years) (P=0.026). A similar trend was observed for overall survival (not reached, 7 years and 6 years, respectively), but the difference did not reach statistical significance. Nine patients (18% of the tested patients, 32% of those with non-driver mutations) carried at least one of the mutations belonging to the high molecular risk (HMR) group: six patients had one mutation and three had two or more mutations. Carrying a mutation in these HMR genes was associated with reduced overall and leukemia-free survival but, surprisingly, HMR mutations did not have a stronger impact than any other additional mutations (Figure 6C,D). HMR mutations were found in five (24%) JAK2-mutated patients, three (50%) CALR-mutated patients, and one triple-negative patient (P=0.32). Mutations in ASXL1 were identified in three (14%) JAK2â&#x20AC;&#x201C;mutated and three (50%) CALR-mutated patients (P=0.1). The presence of ASXL1 mutations had no impact on either the leukemiafree survival or the overall survival of CALR-positive patients, and it was not relevant whether the mutation pattern was CALR-positive/ASXL1-negative or CALRnegative/ASXL1-positive.
year actuarial survival rate was 69.4% from diagnosis for the whole cohort, and 60% for patients with intermediate-2 or high risk according to the IPSS. These findings suggest a positive impact of interferon therapy on both overall and leukemia-free survivals in addition to the high rate of clinical and hematologic responses that we previously reported.17,18 By comparison, two prospective trials (COMFORT-I and II) have reported the results of the use of ruxolitinib in myelofibrosis patients with IPSS intermediate-2 or high-risk score.8,9 In a recent actualization after 5 years of the COMFORT-II study,22 Harrison et al. observed an overall survival of 59.4% (median not reached). However, the study was not originally designed to assess survival, and these results have been debated.12,23,24 At the time of the analysis, 16 patients (25.8%) were still being treated with pegylated interferon-Îą2a whereas 45 (72.6%) had stopped treatment, 25 (55.6%) due to resistance and 20 (44.4%) due to intolerance. The overall survival of patients who continued their therapy was
Discussion This study reports long-term outcomes of the largest cohort of interferon-treated myelofibrosis patients to our knowledge. The first finding is an unexpectedly long median overall survival of 89 months after myelofibrosis diagnosis in this population of patients, of whom the majority had intermediate or high-risk disease according to the DIPSS (84%) and Lille (50%) scoring systems. For these categories, the observed overall survival in this study was clearly longer than that expected according to the DIPSS (6.9 versus 4 years for intermediate-2 risk patients and 4.58 versus 1.5 years for high-risk patients) and the Lille (5.42 versus 2.17 years for intermediate-risk and 4.46 versus 1.08 years for high-risk) score categories.20,21 The 5-
A
Figure 4. Variations of the JAK2V617F allele burden during the follow-up. Relative variation of the JAK2V617F allele burden for each of the 27 patients for whom sequential testing was done.
B
Figure 5. Non-driver mutations identified by next-generation sequencing among 49 tested patients. The black color indicates high molecular risk (HMR) mutations. (A) Number of patients with each mutation; (B) number of additional mutations identified per patient. The percentages correspond to the proportion of HMR mutations among additional mutations.
haematologica | 2018; 103(3)
443
J-C. Ianotto et al. A
B
C
D
Figure 6. Survival according to non-driver mutation status. (A) Overall survival and (B) leukemia-free survival according to the presence of at least one mutation. (C) Overall survival and (B) leukemia-free survival according to the presence of one of the high molecular risk mutations. High molecular risk is defined by the presence of one of the five following mutations: ASXL1, SRSF2, EZH2 or IDH1/2.
longer than that of patients who stopped (P<10-6). The patients who were identified as resistant to pegylated interferon-Îą2a had the worst survival, suggesting that resistance to interferon is a marker of aggressive disease. However, these results should be interpreted in the light of subsequent treatment that clearly affected survival. For example, patients who received ruxolitinib after discontinuing pegylated interferon-Îą2a had a median survival of 22 months compared to 14 months for patients treated with other therapies. All seven patients who underwent ASCT died within a median of 10 months, mainly from GvHD (5/7 patients). Although numbers are small, this is in stark contrast with previous studies of ASCT in myelofibrosis patients reporting 5-year overall survival rates between 41 and 55%.25,26 There is no available study of the impact of interferon on the outcome of ASCT in Philadelphia chromosome-negative myeloproliferative neoplasms. However, in chronic myeloid leukemia, Pigneux and colleagues showed that interferon therapy increased the incidence of GvHD (65 versus 38%, P=0.01) and decreased disease-free and overall survival rates at 5 years (33 versus 41%, P=0.005% and 41 versus 55%, P=0.002, respectively).27 Collectively, our results suggest that interferon therapy should not be initiated in patients with myelofibrosis who have a high probability of undergoing ASCT within a few months. Interferon has been shown to decrease the mutant allele burden of JAK2 or CALR driver mutations in both poly444
cythemia vera and essential thrombocythemia. Accordingly, we observed a greater than 10% reduction of the JAK2V617F allele burden in 63% of the patients and a greater than 95% reduction in four (15%). This molecular response in myelofibrosis patients is, however, less than the 89.6% JAK2V617F allele burden reduction and 24% complete molecular responses reported by Kiladjian et al. in patients with polycythemia vera.28,29 Very good molecular responses were also recently reported by Masarova et al. after long-term follow-up of pegylated interferon-Îą2a therapy in patients with polycythemia vera or essential thrombocythemia.30 Among 63 JAK2V617F-positive patients, they observed a molecular response rate of 63% (including 16% complete molecular responses), with a reduction of the median mutant allele burden from 41 to 12%. The reduction of the JAK2V617F allele burden did not have any impact on overall survival or leukemia-free survival in our study. Silver et al. also recently reported such an absence of correlation between molecular response and clinical outcomes in a series of 30 myelofibrosis patients treated with interferon.31 Such a level of response seems unique to interferon since in the COMFORT-II study, ruxolitinib achieved a greater than 20% reduction of JAK2V617F burden in only 30% of the patients (compared to 55.6% in our series), and none reached a complete molecular response.22 We were also able to evaluate the CALR molecular response in a few patients and found that only one patient had a significant decrease in CALR mutant allele burden. haematologica | 2018; 103(3)
Pegylated interferon-α in myelofibrosis
This is in contrast with the results obtained by Verger et al. in patients with essential thrombocythemia in whom a reduction of the median mutant CALR allele burden from 41 to 26% was observed in a cohort of 31 patients.32 Besides the classical driver mutations, we identified at least one additional mutation in 28/49 patients with a mean of 1.6 additional mutations per patient. The presence of at least one mutation significantly reduced leukemiafree survival and the presence of more than one mutation was associated with a decrease of both overall survival and leukemia-free survival. Such a negative impact of additional mutations on survival is in line with previous findings in myelofibrosis patients.7 Vannucchi et al. have reported that mutations in ASXL1, EZH2, SRSF2 or IDH1/2 carried a stronger adverse prognostic impact, defining a high molecular risk profile.6 In our study, the presence of these HMR mutations affected outcome, but not more than other nonHMR mutations did. ASXL1 mutations alone did not significantly affect survival, but this may be due to the limited number of patients carrying this mutation in our series. However, our data could indicate that the higher risk associated with mutations affecting ASXL1, EZH2, SRSF2 or IDH1/2 could be in part reduced by interferon therapy. Another possible explanation for the discrepancy between our results regarding HMR mutations and those published by Vannucchi and colleagues is that our cohort of patients included a higher proportion with secondary myelofibrosis. Indeed, Rotunno et al. reported that only SRSF2 mutations affected survival in secondary myelofibrosis.33 Lastly, additional mutations were more frequently found in patients intolerant of or resistant to interferon, a finding in agreement with the results published by Silver et al. indicating that additional mutations are more frequent in patients who could not remain on pegylated interferon therapy.31 They also reported that a higher number of mutations, including HMR mutations, is associated with poorer response to interferon. The limitations of our study include the absence of eval-
References 1. Mesa RA, Verstovsek S, Cervantes F, et al. Primary myelofibrosis (PMF), post polycythemia vera myelofibrosis (post-PV MF), post essential thrombocythemia myelofibrosis (post-ET MF), blast phase PMF (PMFBP): consensus on terminology by the International Working Group for Myelofibrosis Research and Treatment (IWG-MRT). Leuk Res. 2007;31(6):737-740. 2. Tefferi A. Primary myelofibrosis: 2017 update on diagnosis, risk-stratification, and management. Am J Hematol. 2016;91(12): 1262-1271. 3. Cervantes F, Tassies D, Salgado C, Rovira M, Pereira A, Rozman C. Acute transformation in nonleukemic chronic myeloproliferative disorders: actuarial probability and main characteristics in a series of 218 patients. Acta Haematol. 1991;85(3):124-127. 4. Klampfl T, Gisslinger H, Harutyunyan AS, et al. Somatic mutations of calreticulin in myeloproliferative neoplasms. N Engl J Med. 2013;369(25):2379-2390. 5. Nangalia J, Massie CE, Baxter EJ, et al. Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. N
haematologica | 2018; 103(3)
uation of symptoms and measurements of cytokine levels. Such studies were not possible due to the lack of a validated specific tool in French for symptom assessment in myelofibrosis at the time the study was initiated, and to the absence of stored plasma or serum given the observational nature of the study. Other important information would have been gained from sequential evaluation of bone marrow biopsies since it has been shown that interferon therapy may reduce fibrosis in selected cases.34 Although all patients had a biopsy for the diagnosis of their disease, investigators did not perform new biopsies after interferon therapy, so the impact of the treatment on this aspect of the disease could not be studied in this cohort of patients. In conclusion, while we have previously reported the clinical and hematologic efficacy of pegylated interferonα2a treatment in myelofibrosis patients, this long-term analysis suggests that interferon therapy may also improve overall survival and leukemia-free survival. In contrast, interferon therapy before ASCT could increase the risk of GvHD and should probably be avoided in this context. Intolerance of or resistance to interferon identifies a group of patients with a dismal outcome, as does the presence of additional mutations. These results indicate that even in the ruxolitinib era, the place of pegylated interferon-α2a should be discussed in patients with myelofibrosis, the optimal target population possibly being high-risk myelofibrosis patients without the prospect of ASCT and with proliferative disease. Acknowledgments The authors would like to thank all the clinical research team members who participated in data collection. We also thank the “Ligue contre le cancer” for their continuous support of research into myeloproliferative neoplasms at Brest Hospital. This study is part of the “CTIM3” project supported by a grant from the French Cancer Institute (INCa), TRANSLA13-140 and of the FIMBANK project (INCa BCB 2013).
Engl J Med. 2013;369(25):2391-2405. 6. Vannucchi AM, Lasho TL, Guglielmelli P, et al. Mutations and prognosis in primary myelofibrosis. Leukemia. 2013;27(9):18611869. 7. Guglielmelli P, Lasho TL, Rotunno G, et al. The number of prognostically detrimental mutations and prognosis in primary myelofibrosis: an international study of 797 patients. Leukemia. 2014;28(9):1804-1810. 8. Harrison C, Kiladjian JJ, Al-Ali HK, et al. JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N Engl J Med. 2012;366(9):787-798. 9. 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. 10. Vannucchi AM, Kantarjian HM, Kiladjian JJ, et al; COMFORT Investigators. A pooled analysis of overall survival in COMFORT-I and COMFORT-II, 2 randomized phase III trials of ruxolitinib for the treatment of myelofibrosis. Haematologica. 2015;100(9): 1139-1145. 11. Passamonti F, Vannucchi AM, Cervantes F, et al. Ruxolitinib and survival improvement in patients with myelofibrosis. Leukemia. 2015;29(3):739-740.
12. Cervantes F, Pereira A. Does ruxolitinib prolong the survival of patients with myelofibrosis? Blood. 2016;129(7):832-837. 13. Alchalby H, Kröger N. Allogeneic stem cell transplant vs. Janus kinase inhibition in the treatment of primary myelofibrosis or myelofibrosis after essential thrombocythemia or polycythemia vera. Clin Lymphoma Myeloma Leuk. 2014;14 (Suppl):S36-41. 14. Kröger NM, Deeg JH, Olavarria E, et al. Indication and management of allogeneic stem cell transplantation in primary myelofibrosis: a consensus process by an EBMT/ELN International Working Group. Leukemia. 2015;29(11):2126-2133. 15. Kröger N, Giorgino T, Scott BL, et al. Impact of allogeneic stem cell transplantation on survival of patients less than 65 years of age with primary myelofibrosis. Blood. 2015;125(21):3347-3350. 16. Robin M, Porcher R, Wolschke C, et al. Outcome after transplantation according to reduced-intensity conditioning regimen in patients undergoing transplantation for myelofibrosis. Biol Blood Marrow Transplant. 2016;22(7):1206-1211. 17. Ianotto JC, Kiladjian JJ, Demory JL, et al. PEG-IFN-alpha-2a therapy in patients with
445
J-C. Ianotto et al.
18.
19.
20.
21.
22.
23.
446
myelofibrosis: a study of the French Groupe d'Etudes des Myelofibroses (GEM) and France Intergroupe des syndromes Myéloprolifératifs (FIM). Br J Haematol. 2009;146(2):223-225. Ianotto JC, Boyer-Perrard F, Gyan E, et al. Efficacy and safety of pegylated-interferon α-2a in myelofibrosis: a study by the FIM and GEM French cooperative groups. Br J Haematol. 2013;162(6):783-791. Lippert E, Boissinot M, Kralovics R, et al. The JAK2-V617F mutation is frequently present at diagnosis in patients with essential thrombocythemia and polycythemia vera. Blood. 2006;108(6):1865–1867. Dupriez B, Morel P, Demory JL, et al. Prognostic factors in agnostic myeloid metaplasia: a report on 195 cases with a new scoring system. Blood. 1996;88(3):1013-1018. Passamonti F, Cervantes F, Vannucchi AM, et al. A dynamic prognostic model to predict survival in primary myelofibrosis: a study by the IWG-MRT (International Working Group for Myeloproliferative Neoplasms Research and Treatment). Blood. 2010;115(9):1703-1708. Harrison CN, Vannucchi AM, Kiladjian JJ, et al. Long-term findings from COMFORT-II, a phase 3 study of ruxolitinib vs best available therapy for myelofibrosis. Leukemia. 2016;30(8):1701-1707. Barosi G, Zhang MJ, Gale PR. Does ruxoli-
24.
25.
26.
27.
28.
29.
tinib improve survival of persons with MPN-associated myelofibrosis? Should-it? Leukemia. 2014;28(11): 2267-2270. Martí-Carvajal AJ, Anand V, Solà I. Janus kinase-1 and Janus kinase-2 inhibitors for treating myelofibrosis. Cochrane Database Syst Rev. 2015;(4):CD010298. Scott BL, Gooley TA, Sorror ML, et al. The Dynamic International Prognostic Scoring System for myelofibrosis predicts outcomes after hematopoietic cell transplantation. Blood. 2012; 119(11): 2657-2664. Gupta V, Malone AK, Hari PN, et al. Reduced-intensity hematopoietic cell transplantation for patients with primary myelofibrosis: a cohort analysis from the center for international Blood and Marrow Transplant Research. Biol Blood Marrow Transplant. 2014;20(1):89-97. Pigneux A, Tanguy ML, Michallet M, et al; Société Française de Greffe de Moelle. Prior treatment with alpha interferon does not adversely affect the outcome of allogeneic transplantation for chronic myeloid leukaemia. Br J Haematol. 2002;116(1):193201. Kiladjian JJ, Cassinat B, Turlure P, et al. High molecular response rate of polycythemia vera patients treated with pegylated interferon alpha-2a. Blood. 2006;108(6):2037-2040. Kiladjian JJ, Cassinat B, Chevret S, et al. Pegylated interferon-alfa-2a induces com-
30.
31.
32.
33.
34.
plete hematologic and molecular responses with low toxicity in polycythemia vera. Blood. 2008;112(8):3065-3072. Masarova L, Patel KP, Newberry KJ, et al. Pegylated interferon alfa-2a in patients with essential thrombocythaemia or polycythaemia vera: a post-hoc, median 83 months follow-up of an open-label, phase 2 trial. Lancet Haematol. 2017;4(4):e165-e175. Silver RT, Barel AC, Lascu E, et al. The effect of initial molecular profile on response to recombinant interferon-α (rIFNα) treatment in early myelofibrosis. Cancer. 2017;123(14): 2680–2687. Verger E, Cassinat B, Chauveau A, et al. Clinical and molecular response to interferon-α therapy in essential thrombocythemia patients with CALR mutations. Blood. 2015;126(24):2585-2591. Rotunno G, Pacilli A, Artusi V, et al. Epidemiology and clinical relevance of mutations in post polycythemia vera and post essential thrombocythemia myelofibrosis: a study on 359 patients of the AGIMM group. Am J Hematol. 2016;91(7): 681-686. Pizzi M, Silver RT, Barel A, Orazi A. Recombinant interferon-α in myelofibrosis reduces bone marrow fibrosis, improves its morphology and is associated with clinical response. Mod Pathol. 2015;28(10):13151323.
haematologica | 2018; 103(3)
ARTICLE
Chronic Myeloid Leukemia
CD36 defines primitive chronic myeloid leukemia cells less responsive to imatinib but vulnerable to antibody-based therapeutic targeting
Ferrata Storti Foundation
Niklas Landberg,1 Sofia von Palffy,1 Maria Askmyr,1 Henrik Lilljebjörn,1 Carl Sandén,1 Marianne Rissler,1 Satu Mustjoki,2 Henrik Hjorth-Hansen,3 Johan Richter,4 Helena Ågerstam,1 Marcus Järås1 and Thoas Fioretos1
Division of Clinical Genetics, Department of Laboratory Medicine, Lund University, Sweden; 2Hematology Research Unit Helsinki, Department of Clinical Chemistry and Hematology, University of Helsinki, and Helsinki University Hospital Comprehensive Cancer Center, Finland; 3Department of Hematology, St Olavs Hospital, Trondheim, Norway and 4Department of Hematology, Oncology and Radiation Physics, Skåne University Hospital, Lund, Sweden 1
Haematologica 2018 Volume 103(3):447-455
ABSTRACT
T
yrosine kinase inhibitors (TKIs) are highly effective for the treatment of chronic myeloid leukemia (CML), but very few patients are cured. The major drawbacks regarding TKIs are their low efficacy in eradicating the leukemic stem cells responsible for disease maintenance and relapse upon drug cessation. Herein, we performed ribonucleic acid sequencing of flow-sorted primitive (CD34+CD38low) and progenitor (CD34+CD38+) chronic phase CML cells, and identified transcriptional upregulation of 32 cell surface molecules relative to corresponding normal bone marrow cells. Focusing on novel markers with increased expression on primitive CML cells, we confirmed upregulation of the scavenger receptor CD36 and the leptin receptor by flow cytometry. We also delineate a subpopulation of primitive CML cells expressing CD36 that is less sensitive to imatinib treatment. Using CD36 targeting antibodies, we show that the CD36 positive cells can be targeted and killed by antibody-dependent cellular cytotoxicity. In summary, CD36 defines a subpopulation of primitive CML cells with decreased imatinib sensitivity that can be effectively targeted and killed using an anti-CD36 antibody.
Introduction Chronic myeloid leukemia (CML) arises when a reciprocal t(9;22) translocation, generating the BCR/ABL1 fusion gene, occurs in a hematopoietic stem cell (HSC).1,2 Currently, the disease is often controlled by daily administered tyrosine kinase inhibitors (TKIs) and patients rarely progress into an accelerated phase or blast crisis.3 However, BCR/ABL1 transcripts are still detectable during treatment, even in the majority of patients with complete clinical and cytogenetic responses.4 Among TKItreated patients with undetectable minimal residual disease (MRD), 40-60% lose their molecular remission after TKI cessation.5 This is generally believed to be caused by CML stem cells, which are partially resistant to TKI treatment.6-8 Even patients with undetectable residual disease have been shown to harbor primitive CML cells.9 These primitive CML cells reside within the CD34+CD38low population, and have been shown by us and others to express both IL1RAP and CD26.10-14 However, the exact immunophenotype of these primitive CML cells is not clearly defined, and the identification of additional cell surface molecules on primitive CML cells may translate into new therapeutic opportunities. Herein, we performed ribonucleic acid (RNA) sequencing of CML CD34+CD38low cells, and identified CD36 and the leptin receptor (LEPR) as being specifically upregulated on primitive CML cells compared to corresponding normal bone marrow (NBM) cells. We further demonstrate that the CD36 expressing subpopulation of primitive CML cells is less sensitive to imatinib treatment, and that CD36 antibodies can induce selective killing of CML cells by antibody-dependent cellular cytotoxicity haematologica | 2018; 103(3)
Correspondence: niklas.landberg@med.lu.se or thoas.fioretos@med.lu.se Received: April 10, 2017. Accepted: December 18, 2017. Pre-published: December 28, 2017. doi:10.3324/haematol.2017.169946 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/447 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
447
N. Landberg et al.
(ADCC), thus providing a putative new therapeutic opportunity for targeting imatinib-resistant CML stem cells.
Methods Patient samples and CD34 enrichment Bone marrow (BM) and peripheral blood (PB) from TKI-naive chronic phase CML patients (n=34; Online Supplementary Table S1) were obtained after written informed consent and in accordance with the Declaration of Helsinki. Ten of these patients were included in the NordCML006 study (clinicaltrials.gov identifier: 00852566) and 15 in the ongoing BFORE study (clinicaltrials.gov identifier: 02130557).15,16 Mononuclear cells were isolated using lymphoprep (GE Healthcare Bio-Sciences AB, Sweden) and CD34 enrichment was performed using magnetic beads (Miltenyl Biotec, Germany) according to manufacturer’s instructions. The study was conducted with the approval of a regional ethics committee in Lund (Dnr 2011/289).
Flow cytometric analyses and FACS sorting of primary cells Analyses of cell surface protein expression and fluorescenceactivated cell sorting (FACS) was performed on a LSR Fortessa or a FACS Aria II (BD Bioscience, USA). The antibodies and viability dyes used are listed in Online Supplementary Table S2. Isotype controls were used at corresponding concentrations. Two or more CML samples were analyzed for each cell surface marker. Prior to RNA extraction, carried out according to the manufacturer’s instructions (Thermo Fisher Inc, USA), viable, single CD34+CD38low cells (5% lowest CD38 expressing cells of the CD34+ cells) from CML and NBM were sorted into a PicoPure RNA Isolation Kit Extraction Buffer (Thermo Fisher Scientific Inc).
RNA sequencing To analyze gene expression, complementary (c)DNA was amplified using the SMARTer Ultra Low Input RNA Kit for Sequencing (Takara Bio Europe, France). Sequencing libraries were prepared from the amplified cDNA using the Nextera Library DNA Preparation Kit (Illumina, USA). Paired 2x151 base pair (bp) RNA sequencing was performed on a NextSeq 500 (Illumina). The reads were aligned to human reference genome hg19 using TopHat 2.0.7.17 Gene expression values were calculated as fragments per kilobase of transcript per million reads (fpkm) using Cufflinks 2.2.0.18 A total of ten diagnostic CML samples and four NBM controls were analyzed. RNA sequencing data have been deposited at the European Genome-phenome Archive (EGA) under the accession code EGAS00001002421. Qlucore Omics Explorer (v 3.1 Qlucore AB, Sweden) was used to identify differentially expressed genes.
Cell cycle status and cell culture with imatinib To distinguish cells in G0/G1 phase from cells in S/G2/M phase in primary CML patient samples, deep red anthraquinone 5 (DRAQ5; BioStatus, UK) was added after staining for CD34, CD38, IL1RAP and CD36. Cells were incubated at room temperature for 20 minutes and subsequently analyzed using a LSR Fortessa (BD Bioscence). To determine sensitivity to imatinib treatment, 2000 CML cells per well were FACS sorted into 96-well plates according to CD34+CD38lowIL1RAP+CD36+ and CD34+CD38lowIL1RAP+CD36- phenotypes (Online Supplementary Figure S1) and challenged with imatinib at 5μM or dimethyl sulfoxide (DMSO) at a corresponding concentration for 72 hours. Viable cells were evaluated using CountBright Absolute Counting Beads (Thermo Fisher Inc) and 4',6-diamidino-2-phenylindole (DAPI) on a LSR Fortessa (BD Bioscience) after three days in culture. 448
ADCC assay For ADCC assays we used polyclonal antibodies targeting CD36 produced in rabbit (Innovagen, Sweden) and unspecific polyclonal rabbit immunoglobulin G (IgG) isotypes (Abcam) as control. Target cells subjected to ADCC were KU812 cells, CD34+enriched CML samples and CD34+-enriched NBM samples. Target cells were labeled with PKH26 (Sigma-Aldrich, USA), plated at 10,000 cells/well and incubated for 30 minutes at room temperature with antibodies of concentrations between 0.001-10μg/ml. One hundred thousand natural killer (NK) cells harvested from healthy donors and isolated using magnetic beads (Milteny Biotec) were added to each well, and the ADCC effect was analyzed the following morning using DAPI as a viability marker. Specific ADCC-induced cell death was calculated according to the following formula: (Percentage DAPI+ cellsantibody – Percentage DAPI+ cellsno antibody) / (0.01 x Percentage DAPI– cellsno antibody).
Statistical Analyses Prism 6 (GraphPad Software, USA) was used for statistical analyses. When possible, the Mann-Whitney U-test was used to determine statistically significant differences between groups, in other settings the Student’s t-test was used. The Spearman’s rank test was used to determine correlations.
Results RNA sequencing of CD34+CD38low CML cells identifies a distinct gene expression profile In order to identify cell surface markers that are upregulated on primitive CML cells, we performed RNA sequencing of sorted CD34+CD38+ progenitor cells and more primitive CD34+CD38low CML cells from ten newly diagnosed patients in chronic phase (Figure 1A). Corresponding healthy BM cells, sorted using the same strategy, were used as controls (n=4). The CML CD34+CD38low population had a distinct gene expression profile when compared to normal CD34+CD38low and CML CD34+CD38+ cells, visualized using unsupervised principal component analysis with a variance threshold of 0.27 retaining 1005 genes (Figure 1B). Using a previously curated list of 1418 genes encoding cell surface proteins, a two-group comparison revealed a statistically significant leukemia-specific upregulation of 32 genes (fold change >3, Q-value < 0.05) in the primitive CML cell compartment (Figure 1C,D). Correspondingly, 24 cell surface-associated genes were found to be significantly downregulated in the same cells (fold change <0.33, Qvalue < 0.05, Online Supplementary Figure S2).
Validation of cell surface protein expression of upregulated genes To assess if the upregulated genes identified by RNA sequencing corresponded to an increased protein expression at the cell surface, we performed flow cytometric analyses of 16 putative targets using commercially available antibodies (Table 1 and Online Supplementary Table S2). We confirmed the previously reported leukemia-specific upregulation of the following four markers: IL1RAP, IL2RA (CD25), DPP4 (CD26) and NCAM1 (CD56) within the CD34+CD38low compartment (Table1).11-13,19 Seven additional proteins, including CD36, LEPR (CD295), TFRC (CD71), ITGB3 (CD61), CD7, FCGR2A (CD32), and GP6 were found to be expressed on primary CD34+CD38low CML cells by flow cytometry (Table 1). Most of these cell surface prohaematologica | 2018; 103(3)
CD36 defines CML cells less sensitive to imatinib
teins were also expressed on the CML cell lines KU812 and BV173, whereas fewer displayed expression on K562 and LAMA84 cells. The remaining five candidate markers (IL12RBI, HMMR, ECE1, TNFRSF18, and TYRO3) could not be detected on primary CML cells, possibly due to suboptimal antibodies or absent or very low cell surface protein expression (Table 1). Finally, CD93, known to be expressed on acute myeloid leukemia (AML) cells and recently reported to be upregulated in CML,20,21 was also expressed on primitive CML cells, but did not show a significant upregu-
lation in our gene expression data relative to corresponding NBM cells (Table 1).
CD36 and LEPR are selectively expressed on CD34+CD38low CML cells compared to corresponding NBM cells Of the newly identified cell surface proteins, CD36, LEPR, ITGB3 and TFRC showed the highest expression on CD34+CD38low CML cells, and were therefore evaluated for expression in NBM. Whereas CD36 and LEPR could not be
A
B
C
D
Figure 1. RNA sequencing of sorted primary CML cells from bone marrow of ten newly diagnosed chronic phase patients. (A) Schematic illustration of cell populations analyzed by RNA sequencing. BM aspirates were used to isolate mononuclear cells and then enrich for CD34 expressing cells with subsequent FACS sorting. The cells displaying the lowest 5% according to CD38 expression were defined as the primitive population and the highest 80% as the more mature progenitor population. (B) Unsupervised principal component analysis of BM from CML patients (n=10) and healthy donors (n=4) sorted into hematopoietic and leukemic progenitor and primitive cell populations. (C) Heat map showing overexpression of cell surface genes in the primitive CML population (CD34+CD38low) as compared to healthy HSCs as determined by RNA sequencing. (D) Schematic figure of total number of transcribed genes detected, number of cell surface associated genes used to filter the results, number of upregulated genes in the primitive CML cell population, and genes available for validation on protein level. HPC: hematopoietic progenitor cells; HSC: hematopoietic stem cells; CML: chronic myeloid leukemia; RNA-seq: ribonucleic acid sequencing.
haematologica | 2018; 103(3)
449
N. Landberg et al. Table 1. Cell surface expression of 17 candidate markers in CML cells as determined by flow cytometry.
Gene Symbol CD36 LEPR TFRC ITGB3 CD7 FCGR2A GP6 IL12RB1 HMMR ECE1 TNFRSF18 TYRO3 IL1RAP* IL2RA* DPP4* NCAM1* CD93**
CD
CD34 CD38
CD36 CD295 CD71 CD61 CD7 CD32 nc CD212 CD168 nc CD357 nc nc CD25 CD26 CD56 CD93
++ + ++ ++ + +/+ ++ ++ ++ + ++
+
low
Primary CMLa CD34+CD38+ ++ + ++ ++ ++ + + ++ +/+/-
KU812 ++ ++ ++ ++ + ++ +/++ ++ ++ ++
K562 ++ ++ ++ + + ++ + +
CML cell lines BV173
LAMA84
+/+ ++ ++ +/+ ++ +/-
+ ++ ++ ++ + ++ +/+ + ++ ++ + -
a Two or more chronic myeloid leukemia (CML) patient samples were analyzed for each cell surface marker. ++: strong expression; +: moderate expression; +/−: expression on some cells or weak expression; −: no expression; CD: cluster of differentiation; nc: not clustered; *: previously described to be upregulated on CD34+CD38low CML cells; **: did not pass the thresholds used to identify differentially expressed cell surface markers in the RNA sequencing analysis.
detected or exhibited low expression on the cell surface of CD34+CD38low NBM cells, ITGB3 and TFRC showed a clearly detectable expression, albeit lower than in the corresponding CML cells (Figure 2A). Further, LEPR was not expressed on more mature CD34+CD38+ cells, whereas CD36 showed some expression in these cells (Online Supplementary Figure S3A). When examining specific subpopulations of normal hematopoietic cells, CD36 displayed a higher expression in more mature subsets (megakaryocyte-erythroid progenitors [MEP], granulocytemachrophage progenitors [GMP], and common myeloid progenitors [CMP]) as compared to more primitive subsets [HSCs, multipotent progenitors [MPP], and lymphoidprimed multipotent progenitors [LMPP]; Online Supplementary Figure S3B). Notably, the expression of CD36 separated the CD34+CD38low CML cells into distinct CD36 positive and negative populations. The overexpression of CD36 in CD34+CD38low cells in relation to corresponding cells from healthy BM samples was confirmed in an independent cohort of 16 CML patients (P=0.0089, Figure 2B). Staining for LEPR resulted in a weak signal, but extended analyses of ten primary CML samples with a high leukemic burden in the CD34+CD38low compartment (as defined by 50-100% IL1RAP expression, mean 86% IL1RAP+ cells), confirmed a significantly higher LEPR expression compared with corresponding healthy cells (P=0.002), devoid of LEPR (Figure 2C). Moreover, all CML samples displayed higher mean fluorescence intensity (MFI) for LEPR compared with paired isotype control stained samples, whereas the NBM samples did not (Online Supplementary Figure S4). Given the specific expression of LEPR in CML cells, we investigated if CML cells would respond to leptin.22 However, when assessing cell growth in vitro and colony forming capacity upon stimulation with leptin, no effects were observed (Online Supplementary Figure S5A-S5D). We conclude that CD36 450
and LEPR are specifically upregulated on the surface of primitive CML cells.
CD36 expression separates CD34+CD38lowIL1RAP+ CML cells into two distinct populations We previously demonstrated that IL1RAP expression can be used to identify BCR/ABL1 positive cells within the CD34+CD38low compartment of BM cells from CML patients, with all cells in the IL1RAP positive fraction being BCR/ABL1 positive.11,13 Because CD36 was found to be expressed on a subpopulation of the CD34+CD38low CML cells, we investigated the co-expression of CD36 and IL1RAP. Although a significant correlation between CD36 and IL1RAP expression was observed (r=0.679, P=0.0048, Figure 3A), CD36 was distinctly expressed on a subset of the CD34+CD38lowIL1RAP+ cells (Figure 3B; Online Supplementary Figure S6). To investigate the co-expression of IL1RAP and CD36 in relation to the BCR/ABL1 status of the cells, we sorted cells based on IL1RAP and CD36 expression within the CD34+CD38low cell fraction from three CML patients. By fluorescence in situ hybridization (FISH) analyses, we found that on average 98% of CD34+CD38lowIL1RAP+CD36+ cells and 98% of CD34+CD38lowIL1RAP+CD36– cells were BCR/ABL1 positive. By contrast, only 3% of the CD34+CD38lowIL1RAP–CD36– cells were BCR/ABL1 positive (Figure 3C,D). Hence, CD36 divides the CD34+CD38lowIL1RAP+ compartment into a CD36 positive and a CD36 negative population that are both predominantly BCR/ABL1 positive.
Primitive CML cells expressing CD36 are less sensitive to imatinib treatment To delineate the difference between the CD36 positive and negative cell populations of primitive CML cells, we sorted CD34+CD38lowIL1RAP+CD36+ and CD34+ CD38low haematologica | 2018; 103(3)
CD36 defines CML cells less sensitive to imatinib
A
B
C
Figure 2. Differentially expressed proteins on primitive CML cells. (A) Representative histograms showing the specific expression of CD36 and LEPR but not ITGB3 and TFRC on leukemic CD34+CD38low cells compared to corresponding cells in NBM. (B) Dot plot showing consistent overexpression of CD36 in the CD34+CD38low cell fraction in CML samples (n=16) compared to NBM samples (n=5). (C) Overexpression of LEPR in the CD34+CD38low cell fraction in a cohort of ten CML patients with high IL1RAP expression as compared to NBM samples (n=4). **P<0.01. CML: chronic myeloid leukemia; NBM: normal bone marrow; MFI: mean fluorescence intensity; LEPR: leptin receptor.
IL1RAP+CD36â&#x2C6;&#x2019; CML cells. Both populations are enriched for BCR/ABL1 positive cells, given that IL1RAP expression marks these cells.11,13 The two cell populations exhibited similar growth and survival after three days in cell culture (n=3, P=0.48, Figure 3E). Interestingly, the CD36 expressing cell population showed significantly decreased imatinib sensitivity as compared to cells lacking CD36 (n=3, P=0.019, Figure 3F). By contrast, the two cell populations exhibited a similar sensitivity to nilotinib, suggesting that the decreased in vitro sensitivity of CD36 expressing cells to imatinib can be overcome by the second generation TKI nilotinib (Online Supplementary Figure S7). In order to determine if the difference in response to imatinib was caused by differences in quiescence, cell cycle status was evaluated in three CML patient samples. No difference in cell cycle status was observed between CD36 positive and negative cells within the CD34+CD38lowIL1RAP+ population, with the haematologica | 2018; 103(3)
majority of cells being in G0 or G1 phase (Figure 3G,H). As anticipated, more mature CD34+CD38+ cells were cycling to a higher degree (Online Supplementary Figure S8). These findings suggest that within the CD34+CD38lowIL1RAP+ compartment, CD36 defines cells that are quiescent and less sensitive to imatinib treatment. Morever, within the primitive CD34+CD38lowIL1RAP+ fraction, CD36 positive and CD36 negative cells showed similar expression of other putative CML stem cell markers, such as CD25 and CD26 (Online Supplementary Figure S9).
CD36 expression declines during TKI treatment In order to investigate whether TKI treatment affects CD36 expression, we first cultured primary CML cells in vitro. However, CD36 expression rapidly decreased during in vitro culture even without the presence of TKI (Online Supplementary Figure S10A-S10C). Subsequently, in a more 451
N. Landberg et al.
direct assay, we measured CD36 expression in BM cells from three CML patients treated for three months with imatinib, bosutinib or dasatinib, respectively. All three samples showed a substantial reduction of CD36 expression in the CD34+CD38low compartment as compared to matched diagnostic samples (Figure 4A). To assess the BCR/ABL1 sta-
tus of the cells during treatment, only patient #11 treated with imatinib had a sufficient number of cells to allow for FACS sorting and subsequent FISH analyses. The CD34+CD38lowCD36+ cells contained 44% BCR/ABL1 positive cells, whereas CD34+CD38lowCD36- cells only contained 6% BCR/ABL1 positive cells (Figure 4B,C). This
A
B
C
D
E
F
G
H
Figure 3. A subgroup of primitive CML cells less sensitive to imatinib express CD36 (A) Linear regression and Spearman’s rank correlation show significant correlation between IL1RAP and CD36 expression in primitive CML cells, Y=0.76X + 2.4; r=0.68, P=0.0048. (B) Contour plot of co-expression of IL1RAP and CD36 in a representative CML sample (CML #5). (C) FISH on sorted cells from three CML patients showed a mean of 98% BCR/ABL1 positive cells within CD34+CD38lowIL1RAP+CD36+ cells and 98% BCR/ABL1 positive cells within CD34+CD38lowIL1RAP+CD36– cells. In the CD34+CD38lowIL1RAP–CD36– cell fraction a mean of 3% were BCR/ABL1 positive; mean based on cells from two CML patients, the third patient had no cells with a CD34+CD38lowIL1RAP–CD36– phenotype. (D) FISH showing a BCR/ABL1 positive (upper panel) and negative (lower panel) cell. (E) CD34+CD38lowIL1RAP+ CML cells FACS sorted according to CD36 expression does not appear to differ in cell growth and survival in vitro. The mean of three CML samples is shown; error bars depict standard deviation. (F) CD34+CD38lowIL1RAP+ CML cells FACS sorted according to CD36 expression and treated with imatinib at a concentration of 5μM show that CD36 expressing cells are more resistant to imatinib treatment in vitro. The mean of three CML samples is shown; error bars depict standard deviation. (G) Representative histograms from cell cycle analysis using DRAQ5 to determine DNA content show a majority of both CD36+ and CD36– cells in G0/G1 phase within the CD34+CD38lowIL1RAP+ population. (H) Data on cell cycle status from three CML patient samples are summarized showing mean and standard deviation. *P<0.05. ns; not significant.
452
haematologica | 2018; 103(3)
CD36 defines CML cells less sensitive to imatinib
patient, with the highest CD36 expression after three months of TKI treatment, was subsequently the only one of the three patients that failed to achieve major molecular response (MMR) after 12 months of treatment, a definition of optimal response, according to the 2013 European LeukemiaNet Guidelines (Online Supplementary Table S1).3
CD36 targeting antibodies induce specific killing of CML cells Having found that CD36 marks cells that are less sensitive to imatinib treatment, we next explored CD36 as a potential target for an antibody-based therapy. To this end, we generated a polyclonal antibody targeting CD36. In
A A
B
C B
C
Figure 4. CD36 expression is reduced after TKI treatment. (A) BM aspirates from three CML patients treated with imatinib, bosutinib or dasatinib for three months showed a substantial reduction of CD36 expression in the CD34+CD38low compartment as compared to diagnosis. (B) FISH for BCR/ABL1 content on sorted CD34+CD38lowCD36+ and CD34+CD38lowCD36â&#x20AC;&#x201C; cells from patient #11 after 3 months imatinib treatment show a higher BCR/ABL1 content in CD36 expressing cells. (C) FISH showing a BCR/ABL1 positive (upper panel) and negative (lower panel) cell. CML: chronic myeloid leukemia.
haematologica | 2018; 103(3)
Figure 5, Polyclonal antibodies targeting CD36 induce specific killing of CML cells. (A) The CML cell line KU812 with high CD36 expression can be specifically killed by a polyclonal rabbit anti-CD36 antibody in a dose-dependent manner by ADCC when incubated with human NK cells. (B) Primary CD34+ CML cells can be specifically killed by a polyclonal rabbit anti-CD36 antibody in a dose-dependent manner by ADCC. The mean of three CML samples is shown; error bars depict standard deviation. (C) NBM samples from healthy donors show a minute ADCC-induced cell death only at the highest tested concentration of antibody. The mean of two NBM samples is shown; error bars depict standard deviation. CML: chronic myeloid leukemia. ADCC: antibody-dependent cellular cytotoxicity; NBM: normal bone marrow.
453
N. Landberg et al.
ADCC assays, we found specific and dose-dependent cell killing of CD36 expressing KU812 cells (Figure 5A). By contrast, the polyclonal CD36 antibody was not associated with direct toxic effects in the absence of human NK effector cells (Online Supplementary Figure S11). Notably, we also found that CD36 can be used as a target for ADCC-mediated cell killing of primary CD34+ CML cells (n=3, Figure 5B), whereas minimal cell death was observed in corresponding NBM cells (n=2, Figure 5C). We conclude that CD36 targeting antibodies can direct human NK cells to specifically eliminate CML cells by ADCC.
Discussion CML is propagated by leukemic stem cells which are partially insensitive to TKI-based therapies, and therefore believed to be responsible for disease relapse upon withdrawal of treatment.6-9,23 The identification of cell surface markers upregulated on primitive CML cells may provide important biological insights, allow for their prospective isolation and characterization, and provide novel means for therapeutic targeting of the treatment resistant cells. Herein, we used RNA sequencing of CD34+CD38low CML cells to identify genes encoding cell surface proteins upregulated on primitive CML cells and with low or absent expression on corresponding normal cells. In total, we identified upregulation of 32 candidate cell surface markers, of which 16 were further evaluated for protein expression by flow cytometry. Previous studies aimed at identifying differentially expressed genes encoding cell surface markers on candidate CML stem cells mainly used microarray-based approaches.11,12,24,25 Of the identified markers, some have previously been described as being upregulated at the transcriptional level, but apart from IL2RA (CD25), DPP4 (CD26) and IL1RAP, their biological roles have not been functionally studied in the context of chronic phase CML. We focused in particular on novel cell surface molecules present on primitive CML cells and with low to absent expression on corresponding normal cells, since this may reveal new therapeutic markers on primitive CML cells that can be selectively targeted.26 A similar approach recently allowed us to identify IL1RAP as a therapeutic target on CML stem cells.11,27 Of the 32 upregulated transcripts, we validated the cell surface protein expression of four previously well-studied molecules: IL1RAP, IL2RA, DPP4, and NCAM1.11-13,19,28 In addition, we identified seven novel cell surface markers expressed on primitive CML cells: CD36, LEPR (CD295), TFRC (CD71), ITGB3 (CD61), CD7, FCGR2A (CD32), and GP6. Of these markers, ITGB3 has been shown to be upregulated and functionally important for the growth and homing of AML cells,29 and TFRC has been described to be expressed at various levels in AML.30 However, we found that only CD36 and LEPR were specifically upregulated on primitive CML cells when compared to corresponding cells from healthy BM. The CD36 molecule is a heavily glycosylated transmembrane protein and scavenger receptor expressed in adipose tissue as well as on thrombocytes, monocytes and macrophages with a role in phagocytosis.31 It has been suggested to have a role in the formation of atherosclerotic plaques and the associated inflammation.32 Interestingly, the expression of CD36 was recently shown to demark a metabolically distinct and treatment refractory subgroup of 454
leukemic stem cells in AML and blast crisis CML.33 Moreover, CD36 has been shown to be essential in the metastatic spread of several forms of cancer.34 CD36 antibodies were found to block metastasis in xenograft models of oral carcinoma, malignant melanoma and breast cancer, possibly by interfering with the metabolic use of fatty acids.34 However, the expression of CD36 and its putative therapeutic importance in chronic phase CML has not been addressed thus far. We found that CD36 is upregulated on CD34+CD38low CML cells. Interestingly, we also discovered that the CD36 expressing subpopulation within the CD34+CD38lowIL1RAP+ CML fraction was less sensitive to imatinib treatment, and that these cells could be specifically killed by ADCC using CD36 antibodies. IL1RAP targeting has been shown to induce killing of CML cells in a similar fashion,11,27 but herein we describe that CD36 antibodies specifically target a cell population within the IL1RAP expressing population that is less sensitive to imatinib. This could provide new means with which to target the cells responsible for relapse after imatinib cessation. However, whether targeting IL1RAP or CD36 would be the best approach remains unclear; it is possible that it could be beneficial to target both markers for a potentially additive effect. Many CML patients are treated with imatinib, and we therefore sought to determine the effect of imatinib on CD36 expression. However, CD36 expression is distinctly reduced during in vitro culture even in the absence of imatinib. Instead, a more direct approach was used, and we show that CD36 expression within the primitive CD34+CD38low population is drastically decreased during the first three months of therapy. It remains of interest, but is unclear whether the repeated measure of CD36 expression could act as a surrogate for response during imatinib treatment. In addition to CD36, we also found LEPR to be upregulated on primitive CML cells. LEPR is the receptor for the well-studied peptide hormone leptin, which is mainly produced by adipocytes and is known to be involved in the regulation of bodyweight, BM microenvironment, normal hematopoiesis, and proliferation of AML cells.35-41 We observed no growth promoting effects in vitro on primitive CML cells following stimulation by the ligand leptin, but the mechanism of action in CML could be different. Indeed, we note with interest that both CD36 and LEPR are involved in adipose tissue homeostasis, and a potential interplay between the adipocyte containing BM microenvironment and CML stem cells could provide important growth or survival signals for the neoplastic stem cells. In conclusion, we herein identify upregulation of several novel cell surface markers, including CD36 and LEPR, on primitive CML cells that may provide novel ways to study and target CML stem cells. In addition, we define CD36 as a marker of cells within the primitive CML cell population in chronic phase CML with decreased sensitivity to imatinib that are vulnerable to antibody-based therapeutic targeting. Acknowledgments The authors would like to thank the Nordic CML Study Group for their help in obtaining and distributing the CML samples between the Nordic countries, in particular we wish to thank Kimmo Porkka, Jesper Stentoft, Bjørn Tore Gjertsen, Jeroen Janssen, Kourosh Lotfi, Leif Stenke, and Ulla StrÜmberg. The NordCML006 and the BFORE studies were supported by haematologica | 2018; 103(3)
CD36 defines CML cells less sensitive to imatinib
research funding from Bristol-Myers Squibb and Pfizer, respectively. In addition, we wish to thank Linda Magnusson for critical experimental assistance. Funding This work was supported by the Swedish Cancer Society, the
References 1. Ren R. Mechanisms of BCR–ABL in the pathogenesis of chronic myelogenous leukaemia. Nat Rev Cancer. 2005;5(3):172183. 2. Holyoake TL, Vetrie D. The chronic myeloid leukemia stem cell: stemming the tide of persistence. Blood. 2017; 129(12):1595-1606 3. Baccarani M, Deininger MW, Rosti G, et al. European LeukemiaNet recommendations for the management of chronic myeloid leukemia: 2013. Blood. 2013; 122(6):872884. 4. Mahon FX, Etienne G. Deep molecular response in chronic myeloid leukemia: the new goal of therapy? Clin Cancer Res. 2014 Jan 16;20(2):310-322. 5. Saussele S, Richter J, Hochhaus A, Mahon FX. The concept of treatment-free remission in chronic myeloid leukemia. Leukemia. 2016;30(8):1638-1647. 6. Jiang X, Zhao Y, Smith C, et al. Chronic myeloid leukemia stem cells possess multiple unique features of resistance to BCRABL targeted therapies. Leukemia. 2007; 21(5):926-935. 7. Corbin AS, Agarwal A, Loriaux M, Cortes J, Deininger MW, Druker BJ. Human chronic myeloid leukemia stem cells are insensitive to imatinib despite inhibition of BCR-ABL activity. J Clin Invest. 2011; 121(1):396-409. 8. Hamilton A, Helgason GV, Schemionek M, et al. Chronic myeloid leukemia stem cells are not dependent on Bcr-Abl kinase activity for their survival. Blood. 2012; 119(6):15011510. 9. Chomel JC, Bonnet ML, Sorel N, et al. Leukemic stem cell persistence in chronic myeloid leukemia patients with sustained undetectable molecular residual disease. Blood. 2011;118(13):3657-3660. 10. Eisterer W, Jiang X, Christ O, et al. Different subsets of primary chronic myeloid leukemia stem cells engraft immunodeficient mice and produce a model of the human disease. Leukemia. 2005;19(3):435441. 11. Jaras M, Johnels P, Hansen N, et al. Isolation and killing of candidate chronic myeloid leukemia stem cells by antibody targeting of IL-1 receptor accessory protein. Proc Natl Acad Sci USA. 2010; 107(37):16280-16285. 12. Herrmann H, Sadovnik I, Cerny-Reiterer S, et al. Dipeptidylpeptidase IV (CD26) defines leukemic stem cells (LSC) in chronic myeloid leukemia. Blood. 2014; 123(25):3951-3962. 13. Landberg N, Hansen N, Askmyr M, et al. IL1RAP expression as a measure of leukemic stem cell burden at diagnosis of chronic myeloid leukemia predicts therapy outcome. Leukemia. 2015;30(1):253-257. 14. Warfvinge R, Geironson Ulfsson L,
haematologica | 2018; 103(3)
15.
16.
17.
18.
19.
20.
21.
22. 23. 24.
25.
26.
27.
Swedish Children’s Cancer Foundation, the Medical Faculty of Lund University, the Swedish Research Council, BioCARE, and the ISREC Foundation by a joint grant to Swiss Cancer Center, Lausanne, CREATE Health Cancer Center, and the Medical Faculty of Lund University from the Biltema Foundation. SM is supported by the Finnish Cancer Institute.
Sommarin MNE, et al. Single-cell molecular analysis defines therapy response and immunophenotype of stem cell subpopulations in CML. Blood. 2017;129(17):23842394. Mustjoki S, Richter J, Barbany G, et al. Impact of malignant stem cell burden on therapy outcome in newly diagnosed chronic myeloid leukemia patients. Leukemia. 2013;27(7):1520-1526. Hjorth-Hansen H, Stenke L, Söderlund S, et al. Dasatinib induces fast and deep responses in newly diagnosed chronic myeloid leukaemia patients in chronic phase: clinical results from a randomised phase-2 study (NordCML006). Eur J Haematol. 2014;94(3):243-250. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. 2013;14(4):R36. Trapnell C, Williams BA, Pertea G, et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 2010;28(5):511-515. Janssen JJWM, Deenik W, Smolders KGM, et al. Residual normal stem cells can be detected in newly diagnosed chronic myeloid leukemia patients by a new flow cytometric approach and predict for optimal response to imatinib. Leukemia. 2011;26(5):977-984. Iwasaki M, Liedtke M, Gentles AJ, Cleary ML. CD93 marks a non-quiescent human leukemia stem cell population and is required for development of MLLrearranged acute myeloid leukemia. Cell Stem Cell. 2015;17(4):412-421. Kinstrie R, Horne GA, Morrison H, et al. CD93 is a novel biomarker of leukemia stem cells in chronic myeloid leukemia. Blood. 2015;126(23):49. Tartaglia LA. The Leptin Receptor. J Biol Chem. 1997;272(10):6093-6096. Dick JE. Stem cell concepts renew cancer research. Blood. 2008;112(13):4793–4807. Diaz-Blanco E, Bruns I, Neumann F, et al. Molecular signature of CD34+ hematopoietic stem and progenitor cells of patients with CML in chronic phase. Leukemia. 2007;21(3):494-504. Gerber JM, Gucwa JL, Esopi D, et al. Genome-wide comparison of the transcriptomes of highly enriched normal and chronic myeloid leukemia stem and progenitor cell populations. Oncotarget. 2013; 4(5):715728. Majeti R. Monoclonal antibody therapy directed against human acute myeloid leukemia stem cells. Oncogene. 2010; 30(9):1009-1019. Ågerstam H, Hansen N, Palffy von S, et al. IL1RAP antibodies block IL-1–induced expansion of candidate CML stem cells and
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
mediate cell killing in xenograft models. Blood. 2016;128(23):2683–2693. Sadovnik I, Hoelbl-Kovacic A, Herrmann H, et al. Identification of CD25 as dependent growth-regulator of leukemic stem cells. Clin Cancer Res. 2016;22(8):2051-2061. Miller PG, Al-Shahrour F, Hartwell KA, et al. In vivo RNAi screening identifies a leukemia-specific dependence on integrin beta 3 signaling. Cancer Cell. 2013; 24(1):4558. Liu Q, Wang M, Hu Y, et al. Significance of CD71 expression by flow cytometry in diagnosis of acute leukemia. Leuk Lymphoma. 2014;55(4):892-898. Silverstein RL, Febbraio M. CD36, a scavenger receptor involved in immunity, metabolism, angiogenesis, and behavior. Sci Signal. 2009;2(72):re3. Harb D, Bujold K, Febbraio M, Sirois MG, Ong H, Marleau S. The role of the scavenger receptor CD36 in regulating mononuclear phagocyte trafficking to atherosclerotic lesions and vascular inflammation. Cardiovasc Res. 2009;83(1):42-51. Ye H, Adane B, Khan N, et al. Leukemic stem cells evade chemotherapy by metabolic adaptation to an adipose tissue niche. Cell Stem Cell. 2016;19(1):23-37. Pascual G, Avgustinova A, Mejetta S, et al. Targeting metastasis-initiating cells through the fatty acid receptor CD36. Nature. 2016;541(7635):41-45. Mantzoros CS, Magkos F, Brinkoetter M, et al. Leptin in human physiology and pathophysiology. Am J Physiol Endocrinol Metab. 2011;301(4):E567-84. Chen H, Charlat O, Tartaglia LA, et al. Evidence that the diabetes gene encodes the leptin receptor: identification of a mutation in the leptin receptor gene in db/db mice. Cell. 1996;84(3):491-495. Bennett BD, Solar GP, Yuan JQ, Mathias J, Thomas GR, Matthews W. A role for leptin and its cognate receptor in hematopoiesis. Curr Biol. 1996;6(9):1170-1180. Feldmana DE, Chena C, Punjb V, Hidekazu Tsukamotoc DE, Machida K. Pluripotency factor-mediated expression of the leptin receptor (OB-R) links obesity to oncogenesis through tumor-initiating stem cells. Proc Natl Acad Sci USA. 2012;109(3):829-834. Yue R, Zhou BO, Shimada IS, Zhao Z, Morrison SJ. Leptin receptor promotes adipogenesis and reduces osteogenesis by regulating mesenchymal stromal cells in adult bone marrow. Cell Stem Cell. 2016; 18(6):782-796. Uddin S, Mohammad RM. Role of leptin and leptin receptors in hematological malignancies. Leuk Lymphoma. 2016; 57(1):1016. Ozturk K, Avcu F, Ural AU. Aberrant expressions of leptin and adiponectin receptor isoforms in chronic myeloid leukemia patients. Cytokine. 2012; 57(1):61-67.
455
ARTICLE
Acute Myeloid Leukemia
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):456-465
A 29-gene and cytogenetic score for the prediction of resistance to induction treatment in acute myeloid leukemia Tobias Herold,1,2,3 Vindi Jurinovic,4 Aarif M. N. Batcha,2,3,4 Stefanos A. Bamopoulos,1 Maja Rothenberg-Thurley,1 Bianka Ksienzyk,1 Luise Hartmann,1,2,3 Philipp A. Greif,1,2,3 Julia Phillippou-Massier,5 Stefan Krebs,5 Helmut Blum,5 Susanne Amler,3 Stephanie Schneider,1 Nikola Konstandin,1 Maria Cristina Sauerland,6 Dennis Görlich,6 Wolfgang E. Berdel,7 Bernhard J. Wörmann,8 Johanna Tischer,1 Marion Subklewe,1 Stefan K. Bohlander,9 Jan Braess,10 Wolfgang Hiddemann,1,2,3 Klaus H. Metzeler,1,2,3 Ulrich Mansmann2,3,4* and Karsten Spiekermann1,2,3*
Department of Internal Medicine III, University of Munich, Germany; 2German Cancer Consortium (DKTK), Partner Site Munich, Munich, Germany; 3German Cancer Research Center (DKFZ), Heidelberg, Germany; 4Institute for Medical Informatics, Biometry and Epidemiology, University of Munich, Germany; 5Laboratory for Functional Genome Analysis (LAFUGA), Gene Center, Ludwig-Maximilians-Universität (LMU) München, Germany; 6Institute of Biostatistics and Clinical Research, University of Munich, Germany; 7Department of Medicine, Hematology and Oncology, University of Münster, Germany; 8German Society of Hematology and Oncology, Berlin, Germany; 9Department of Molecular Medicine and Pathology, University of Auckland, Auckland, New Zealand and 10Department of Oncology and Hematology, Hospital Barmherzige Brüder, Regensburg, Germany 1
*UM and KS contributed equally to this work.
ABSTRACT
P
Correspondence: tobias.herold@med.uni-muenchen.de
Received: August 15, 2017. Accepted: December 7, 2017. Pre-published: December 14, 2017. doi:10.3324/haematol.2017.178442 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/2/xxx ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
456
rimary therapy resistance is a major problem in acute myeloid leukemia treatment. We set out to develop a powerful and robust predictor for therapy resistance for intensively treated adult patients. We used two large gene expression data sets (n=856) to develop a predictor of therapy resistance, which was validated in an independent cohort analyzed by RNA sequencing (n=250). In addition to gene expression markers, standard clinical and laboratory variables as well as the mutation status of 68 genes were considered during construction of the model. The final predictor (PS29MRC) consisted of 29 gene expression markers and a cytogenetic risk classification. A continuous predictor is calculated as a weighted linear sum of the individual variables. In addition, a cut off was defined to divide patients into a high-risk and a lowrisk group for resistant disease. PS29MRC was highly significant in the validation set, both as a continuous score (OR=2.39, P=8.63·10-9, AUC=0.76) and as a dichotomous classifier (OR=8.03, P=4.29·10-9); accuracy was 77%. In multivariable models, only TP53 mutation, age and PS29MRC (continuous: OR=1.75, P=0.0011; dichotomous: OR=4.44, P=0.00021) were left as significant variables. PS29MRC dominated all models when compared with currently used predictors, and also predicted overall survival independently of established markers. When integrated into the European LeukemiaNet (ELN) 2017 genetic risk stratification, four groups (median survival of 8, 18, 41 months, and not reached) could be defined (P=4.01·10-10). PS29MRC will make it possible to design trials which stratify induction treatment according to the probability of response, and refines the ELN 2017 classification.
Introduction Approximately 20-30% of younger adult patients with acute myeloid leukemia (AML) and up to 50% of older adults are refractory to induction treatment.1 There are several definitions of treatment failure due to resistant disease (RD) or primary refractory AML.1-5 One of the earliest consensus definitions classified RD as the persistence of leukemic blasts in either the peripheral blood or the bone marrow in a patient alive at least seven days following treatment, excluding patients with haematologica | 2018; 103(3)
Prediction of primary resistant AML
death in aplasia or death due to indeterminate cause.4 More recent recommendations define primary refractory disease as a failure to achieve complete remission (CR) or CR with incomplete hematologic regeneration (CRi) after two courses of induction treatment.2 Regardless of the definition, RD is associated with extremely poor survival. Patients with RD can currently not be identified with high specificity before the start of treatment and therapeutic resistance remains one of the main problems in AML therapy.3 It is difficult to quantify predictive ability. Usually, area under receiver-operating characteristic curve (AUC) is used to describe predictive ability, where a value of 1 indicates perfect prediction and 0.5 indicates no prediction.3,6 AUC values of 0.6-0.7, 0.7-0.8 and 0.8-0.9 are considered as poor, fair and good, respectively.3,6 An AUC of more than 0.9 would be desirable. Several tools have been developed to predict therapeutic response in AML. The AML-score by Krug et al., based on standard clinical and laboratory variables including genetics, was developed to predict CR or early death of older patients (≥60 years) treated with intensive chemotherapy.7 The score reached a “poor” prognostic ability of AUC=0.68 in the validation set.7 A study using comparable variables to identify RD analyzing 4601 patients of all age groups from the MRC/NCRI, HOVON/SAKK, SWOG and MD Anderson Cancer Center achieved an AUC prediction of AUC=0.78 by bootstrap adjusted validation in the training sets.3 The inclusion of extensive genetic testing data from the initial diagnostic workup in this classifier was not able to significantly improve the ability to predict primary resistance to treatment in younger patients.6,8 These ‘maximal’ models yielded AUCs of 0.77-0.80 but were not validated in independent data sets.6 We hypothesized that we could improve the prediction of RD by combining standard clinical and laboratory variables, mutation data and gene expression data of large homogeneously treated patient cohorts to design a new classifier.
Methods Patients In this study, we used three independent data sets, hereafter referred to as training set 1, 2 and validation set. All patients included in the analysis received cytarabine- and anthracyclinebased induction treatment. All patients included in the German AML Cooperative Group (AMLCG) trials were scheduled to receive at least one high-dose cytarabine-containing course as part of their double induction treatment before they were considered resistant. Training set 1 consisted of 407 patients randomized and treated in the multicenter phase III AMLCG-1999 trial (clinicaltrials.gov identifier 00266136) between 1999 and 2005.9,10 The patients are part of a previously published gene expression data set (GSE37642) and samples were analyzed on Affymetrix arrays.11,12 Patient selection was based on the availability of information on response to induction treatment. All patients with a t(15;17), myelodysplastic syndrome (MDS), or an overall survival (OS) of less than 16 days were excluded. Training set 2 consisted of samples from 462 AML patients treated in various trials of the Haemato-Oncology Foundation for Adults in the Netherlands (HOVON). These samples were analyzed by Affymetrix arrays and clinical and gene expression haematologica | 2018; 103(3)
data are publicly available (GSE14468).13,14 Thirteen patients had to be excluded due to early death (OS <16 days) or missing follow-up data. Finally, the validation set consisted of all patients with available material treated in the AMLCG-2008 study (clinicaltrials.gov identifier 01382147), a randomized, multicenter phase III trial (n=210).15 These patients were analyzed by RNAseq. Because of the high response rate in the AMLCG-2008 trial (CR and CRi: 289 of 387, 75%) and the low rate of resistant patients (48 of 387, 12%), we decided to include an additional 40 patients with RD in the validation set to increase the statistical power; these patients were treated in the AMLG-1999 trial. We selected these patients by including all patients with RD of the AMLCG-1999 trial that were not part of training set 1 and who had sufficient material for analysis. Subsequently, only patients matching the control treatment arm of the AMLCG-2008 trial were selected for RNAseq and included in the validation set (n=40). Cytogenetic data were missing in 8 cases from the validation set (not done: n=3; no cells dividing: n=5); since cytogenetic information is required to calculate most risk scores, these patients were excluded from subsequent analysis. The gene expression data are publicly available through the Gene Expression Omnibus Web site (GSE106291). Details regarding the treatment regimens are described in the Online Supplementary Appendix. A detailed flow chart describing the patient cohorts and selection process is shown in Online Supplementary Figure S1. All study protocols were in accordance with the Declaration of Helsinki and were approved by the institutional review boards of the participating centers. All patients provided written informed consent for inclusion on the clinical trial and in the genetic analyses.
Molecular workup Cytogenetic analyses in the AMLCG trials were performed centrally, and risk groups were defined according to the 2010 UK Medical Research Council (MRC) and the European LeukemiaNet (ELN) 2017 genetic risk classification (ELN2017). Patients were characterized for NPM1 and CEBPA mutations, FLT3 internal tandem duplications (FLT3-ITD), and KMT2A (formerly MLL) partial tandem duplications (KMT2A-PTD) using standard methods described recently.16 Targeted amplicon sequencing of 68 recurrently mutated genes as published recently was used for genetic characterization in training set 1 and the validation set.17 RNAseq libraries were prepared using the Sense mRNA Seq Library Prep Kit V2 (Lexogen, n=238) and the TruSeq RNA Library Preparation V2 Kit (Illumina, n=12). Between 500-1000 ng total RNA [RNA integrity number (RIN) >7] were used as input material. All sequencing was paired end and performed using polyadenylatedselected and, in case of the Lexogen libraries, stranded RNA sequencing. Processing details and sequencing metrics are provided in the Online Supplementary Appendix. Samples were sequenced on a HiSeq 1500 instrument (Illumina) as 100 bp reads to a targeted depth of 20 million mappable paired reads per sample according to the “Standards, Guidelines and Best Practices for RNA-Seq v.1.0 (June 2011)”18 recommendations of the ENCODE Consortium. Samples were aligned with STAR 2.4.019 to the human hg19 reference genome and analyzed by DESeq2.20 Details regarding the workflow are provided in the Online Supplementary Appendix.
Development of the predictive classifier The aim of the study was to develop a predictor that accurately identifies patients with RD. To achieve this goal, we used clinical markers, cytogenetics (defined according to the MRC), mutational analysis of 68 recurrently mutated genes in AML and gene expression markers to construct a predictive model. All gene expression variables were scaled to a mean value of 0 and variance equal to 1. 457
T. Herold et al.
After pre-selection of variables in training set 1 and 2, penalized logistic regression (Lasso) in training set 1 was used to develop a predictive classifier that was able to identify patients with RD. Details regarding the development of the classifier are given in the Online Supplementary Appendix and Online Supplementary Figure S2, or were published previously.21 The predictor was validated in a fully independent cohort of patients where gene expression was analyzed by a different method, namely RNA-Seq. The final predictor and cut off were developed before the validation data set became available.
Cut-off development Our aim was to provide a meaningful cut off to guide clinical decision making. In the case of AML, this means that patients should not be excluded from induction chemotherapy unless there is a very high likelihood that it will be ineffective. Therefore, from the clinical perspective, a high specificity of the predictor is desirable. The cut off was, therefore, designed to have a specificity of 0.9 in training set 1. However, by adjusting the cut off to achieve high specificity, the sensitivity of the predictor decreases.
Statistical analysis All statistical analyses were performed using the R 3.3.1 software package (R Foundation for Statistical Computing, Vienna, Austria). Definitions of response to treatment are shown in Figure 1A. We used a slightly modified definition of RD as recommended by Cheson et al.4 In the original definition, patients had to survive at least seven days post induction treatment and have persistent AML in blood or bone marrow after induction treatment. To address the difference in treatment length between various induction regimens [HAM (5 days), â&#x20AC;&#x153;7+3â&#x20AC;? [7 days], TAD (9 days) or sHAM (11 days)], only patients surviving at least 16 days after the start of treatment, independently of the treatment arm, were considered for the development of the predictor in the training sets. Day 16 was selected because the first induction response testing in AMLCG trials was scheduled at this time point. Only patients who started study treatment and those with a definite induction result (CR/CRi or RD) were included in training set 1 and in the analysis of RD in the validation set. Patients with death in aplasia or of indeterminate cause were excluded from training set 1 (but not from the validation set). This subgroup of patients (n=15) in
A
B
Figure 1. Definitions of response and study design. (A) Figure showing the details of the response definition. (B) Flow chart showing the study design and distribution of patients. *Patients analyzed by targeted sequencing for 68 genes recurrently mutated in acute myeloid leukemia. RD: resistant disease; AML: acute myeloid leukemia; PB: peripheral blood; PS29MRC: Predictive Score 29 MRC; HOVON: Haemato-Oncology Foundation for Aults in the Netherlands.
458
haematologica | 2018; 103(3)
Prediction of primary resistant AML
set 2 were significantly younger and had a higher response rate than patients in the other two sets. Due to the addition of resistant patients, the validation set included significantly fewer patients with favorable cytogenetics. The median follow up was more than eight years in the training sets and 4.2 years in the validation set.
combination with patients with RD (non-responder; see Figure 1A) was analyzed separately in the validation set. Overall survival was defined as time from study entry until death from any cause. Patients alive were censored at the time of their last follow up. The prognostic impact of the classifier was evaluated by the Kaplan-Meier method and the log-rank test. Because resistant AML can only be cured by stem cell transplantation (SCT), all survival analyses were censored for SCT if not otherwise indicated. The χ2-test was applied for categorical variables and the Wilcoxon test for continuous variables for statistical comparisons. For three-way comparison of numerical variables, we used the Kruskal-Wallis Test. Multivariable logistic models for resistance to induction treatment and multivariable Cox models for OS were used to adjust for potential confounders. P≤0.05 was considered statistically significant.
Validation of the final predictor The final score (Predictive Score 29 MRC, PS29MRC) consists of 29 gene expression markers and the cytogenetic risk groups defined according to the MRC classification,22 and is calculated as a weighted linear sum of the individual predictors (Figure 2). The score in the validation set ranged from -2.75 to +3.72. In univariate analysis, PS29MRC as a continuous variable (PS29MRCcont) was a highly significant predictor of RD in the validation set with an odds ratio (OR) of 2.39 (95%CI: 1.80, 3.26; P=8.63·10-9, AUC=0.76) (Figure 3A). Similar results were seen with PS29MRC as a dichotomous variable (PS29MRCdic) applying the pre-defined cut off (Online Supplementary Appendix) defined in training set 1 (OR 8.03, 95%CI: 4.07, 16.46; P=4.29·10-9). We subsequently designed multivariable models including all vari-
Results Patients A flow chart of the study is given in Figure 1B. A total of 1106 patients were included in the analysis. Patients' characteristics are shown in Table 1. Patients in training
Table 1. Patients’ characteristics.
Training Set 1§
Training Set 2
Validation Set$
407 354 (87) 57 (18-85) 199 (49) 52 (13) 253 (63) 94 (24) 21.2 (0.4-666) 9.0 (3.5-15.4) 53.0 (1-1760) 452 (76-4613) 80 (10-100) 112 (28) 300 (74) 107 (26) 188 (71) 32 [283] 8.6
449 NA 46 (15-77) 227 (51) 70 (16) 298 (66) 81 (18) NA NA NA NA NA NA 372 (83) NA* 189 (59) 38 [289] 8.6
250 208 (83) 58 (18-74) 120 (48) 14 (6) 188 (75) 48 (19) 17.0 (0.5-406) 9.0 (4.5-16.0) 61.5 (1-997) 408 (107-6601) 73 (6-100) 40 (16) 164 (66) 71 (28) 67 (56) 36 [150] 4.2
Number of patients de novo AML, n (%) Age, median (range) Male sex, n (%) MRC favorable, n (%) MRC intermediate, n (%) MRC unfavorable, n (%) White-cell count (x 10 9/L), median (range) Hemoglobin (g/dL), median (range) Platelet count (x 10 9/L), median (range) LDH (U/L), median (range) Bone marrow blasts, % median (range) ECOG performance status > 1, n (%) CR/CRi, n (%) AML with resistant disease, n (%) Cumulative relapse, n (%) 5-year survival in %, [n of deaths] Median follow up (years)
P 0.22 <0.001 0.79 <0.001
0.14 0.57 0.063 0.13 0.0016 <0.001 <0.001 0.33 0.0031 0.096
The data set was restricted to patients with definitive induction results (CR/CRi or resistant disease). $ The data set includes additional 40 patients with resistant disease from the AMLCG-1999 trial. * In training set 2 only the information of responder (CR/CRi) and non-responder (n=77, 17%) was available. n: number; AML: acute myeloid leukemia; LDH: lactate dehydrogenase; ECOG: Eastern Cooperative Oncology Group; CR: complete remission; CRi: incomplete hematologic regeneration. §
Table 2. Univariate and multivariable analysis of the prediction of resistant disease in the validation set.
Variable PS29MRCdic Age continuous NPM1mut RUNX1mut TP53mut
Multivariable analysis, n=235 OR [95%-CI]
P
Univariate analysis OR [95%-CI]
P
4.44 [2.00; 10.16] 1.06 [1.03; 1.10] 0.48 [0.19; 1.142] 1.05 [0.50; 2.44] 7.16 [1.76; 38.61]
0.00030 0.00012 0.094 0.90 0.010
8.03 [4.07; 16.46] 1.07 [1.04; 1.10] 0.23 [0.11; 0.46] 2.13 [1.07; 4.21] 12.03 [3.72; 53.84]
4.29·10-9 3.87·10-6 6.62·10-5 0.029 0.00016
OR: Odds Ratio; CI: Confidence Interval.
haematologica | 2018; 103(3)
459
T. Herold et al.
ables with P≤0.05 in training set 1 and the validation set for the prediction of RD. The results of PS29MRCdic are shown in Table 2. Only PS29MRCdic, age and TP53 mutations remained significant in the model. Comparable results were seen with PS29MRC as a continuous variable (Online Supplementary Table S1). When we included all non-responders in the analysis (Figure 1A), PS29MRCcont (OR 2.34, 95%CI: 1.80, 3.13; P=1.48·10-9, AUC=0.75) and PS29MRCdic (OR 8.04, 95%CI: 4.20, 16.06; P=9.27·10-10) remained highly significant. In the multivariable model, only PS29MRCdic, age and TP53 mutations remained significant (PS29MRCdic: OR 5.08, 95%CI: 2.39; 11.19; P=3.41·10-5; age: OR 1.06, 95%CI: 1.03, 1.09; P=0.00013; TP53: OR 6.13, 95%CI: 1.56, 31.90; P=0.016). Comparable results were seen when only patients treated in the AMLCG 2008 trial were considered (Online Supplementary Table S2). To exclude the possibility that patients that achieve a CR after two courses of induction treatment are misclassified as resistant, we also tested if PS29MRC was able to correctly forecast RD at day 60. In univariate, as well as in multivariable analysis, PS29MRC showed comparable results (Online Supplementary Table S3).
Characterization of the dichotomized classifier PS29MRCdic was designed in the training set to identify resistant AML patients with high specificity. By applying the pre-defined cut off, we were able to reach a specificity of 148 of 164 (90%) and sensitivity of 33 of 71 (46%) in the independent validation set. The performance of the classifier in the training sets and the validation set is shown in Online Supplementary Figure S3. The accuracy of PS29MRCdic was 77% in the validation set. When we included all patients with death in aplasia or
of indeterminate cause in the analysis (Figure 1A), the sensitivity of PS29MRCdic in predicting non-response to induction treatment was 40 of 86 (47%) and the accuracy 75%. In the cytogenetic subgroups favorable (n=14; resistant: n=0; CR/CRi: n=13), intermediate (n=188; resistant: n=42; CR/CRi: n=136) and adverse (n=48; resistant: n=29; CR/CRi: n=15), the classifier showed an accuracy of 100%, 78% and 66%, respectively (Online Supplementary Figure S4). When we applied the ELN2017 genetic risk stratification, the accuracy of PS29MRCdic in the subgroups favorable, intermediate and unfavorable was 89%, 74% and 70%, respectively (Figure 3B). Since not achieving a CR/CRi is highly correlated with OS, we analyzed the performance of PS29MRCdic to serve as a prognostic tool. The classifier was a highly significant predictor of survival in univariate (HR 2.81, 95%CI: 1.98, 3.99; P=7.73·10-9) and multivariable (HR 2.15, 95%CI: 1.39, 3.31; P=0.00052) models, including all variables significant (P≤0.05) in training set 1 and the validation set (Online Supplementary Table S4). When we integrated PS29MRCdic in the ELN2017 genetic risk stratification, four risk groups with a median OS of 8 months (95%CI: 5-10) in the PS29MRCdic high-risk group, 16 (95%CI: 8-41) months in the ELN2017 unfavorable group, and 'not reached' in the intermediate and favorable risk groups could be defined (Figure 4A-C). Approximately 50% of all ELN2017 unfavorable and approximately 12% of all ELN2017 intermediate-risk patients were classified as high risk according to PS29MRCdic. The probability of survival in the four risk groups at 24 months was 12%, 38%, 57% and 76%, respectively. Comparable results were seen when OS was not censored for SCT (Online Supplementary Figure S5).
Figure 2. Signature and weights. Variables included in the predictive score PS29MRC. The final score is calculated as the weighted sum of these values (MRC high risk/low risk as 1 or -1, respectively). The final classifier consisted of 29 gene expression markers and the favorable and unfavorable cytogenetic MRC groups. Variables in red are associated with resistant disease; variables in blue are predictive for a response to induction treatment.
460
haematologica | 2018; 103(3)
Prediction of primary resistant AML
We also classified the heterogeneously treated AML patients included in the TCGA analysis23 with PS29MRCdic (available gene expression samples n=183). The predictor was highly predictive for OS in the independent data set (Online Supplementary Figure S6).
Performance of the classifier in genetic subgroups Mutation profiles defining genetic subgroups in AML have been recently defined.17,24 We used these subgroups to further characterize the predictive potential of PS29MRCdic. A detailed picture of the analysis is shown in Figure 5A and Online Supplementary Figure S7A. The classifier reached very high accuracy in the genetic subgroups defined by Metzeler et al.17 as core binding factor alterations (CBF), KMT2A rearranged AML, AML with biallelic CEBPA mutations, and AML with NPM1 mutations. Moderate predictive accuracy was achieved in the high-risk subgroups defined by TP53 and RUNX1 alterations. Here, PS29MRCdic was not able to significantly predict OS (Figure 5B and D). In the large subgroup of patients without classifiable genetic alterations, PS29MRCdic was able to significantly predict survival (P=0.027) (Figure 5E).
The results of the classification of genetic subgroups according to Papaemmanuil et al.24 show comparable results with a moderate predictive accuracy of PS29MRC in high-risk subgroups defined by TP53 mutations, chromosomal aneuploidy or both, AML with mutated chromatin, RNA-splicing genes or both, and AML with driver mutations but no class-defining lesions. In all other subgroups, even though some had small sample sizes, PS29MRC reached very high predictive accuracy. In contrast to the results seen with the classification according to Metzeler et al.,17 significantly, OS could only be predicted in the subgroup of patients with mutated chromatin, RNA-splicing genes, or both (Online Supplementary Figure S7B).
Performance of the classifier in comparison to currently used models In pairwise comparisons to published, predictive classifiers like the model by Walter et al.3 (integrating information on age, performance status, white blood cell count, platelet count, bone marrow blasts, sex, type of AML, cytogenetics and NPM1 and FLT3-ITD status), or the modified molecular version of this score,6 PS29MRCdic
A
B
Figure 3. Receiver operating characteristic curve (ROC) of predictive score PS29MRC as a continuous variable (PS29MRCcont) and barplots showing the predictive performance of the PS29MRC as a dichotomous variable (PS29MRCdic) in the validation set. (A) ROC curve showing the performance of PS29MRCcont and other predictive scores in the validation set at varying thresholds. Area under receiver-operating characteristic curve (AUC): PS29MRCcont: 0.76; Walter-Score: 0.71; Retrained response LSC17: 0.61. (B) Bar plots showing the performance of PS29MRCdic in subgroups defined by the European LeukemiaNet (ELN) 2017 genetic risk classification (ELN2017). CR: complete remission.
haematologica | 2018; 103(3)
461
T. Herold et al.
reached superior predictive power (Table 3). When we restricted the analysis to patients aged 60 years or over (n=118), only PS29MRCdic was left as significant variable (HR=2.04, 95%CI: 1.35, 3.21; P=0.0012). Recently, a highly significant prognostic tool based on “stemness” gene expression markers (LSC17) was published.25 A modified version of this classifier (retrained response LSC17) was developed to predict resistant disease. In multivariable testing, PS29MRCdic outperformed these predictors that are solely based on gene expression variables (Table 3). The extremely high OR is the result of a very low variance of LSC17 (range -0.40 to 0.41). Comparable results were achieved for PS29MRC as continuous variable (Online Supplementary Table S5 and Online Supplementary Figure S3A).
rently established predictive markers and the intermediate cytogenetic risk group. The development process of PS29MRC included extensive evaluation of mutational data from recurrently mutated genes in AML. Interestingly, as shown previously by Walter et al.,6 we were not able to improve the predictive ability of our model by including information on the mutational status of these AML associated genes in the classifier.6 The gene expression markers included in our signature can only be surrogates of cellular pathways predicting resistant disease or directly responsible for the refractory phenotype. MIR155HG, the host gene of miR-155, is one of the most important markers in our signature. High expression of the microRNA miR-155 has already been shown to be associated with an aggressive phenotype in AML with normal karyotype.26 mir-155 is regulated by NFκB and could be repressed by the inhibitor MLN4924 (Pevonedistat) which has already entered phase II clinical trials (clinicaltrials.gov identifier: 02610777).27 Allogeneic SCT is currently the only curative option for patients with RD.2 However, to our knowledge there is currently no accepted standard treatment that can be offered to patients with a very high probability of RD before undergoing SCT.28 Approaches including low-dose chemotherapy or hypomethylating agents are possible options.29 Of note, the rate of SCT was lower in AMLCG patients that were resistant to induction treatment than for patients with an indication for SCT in the post-remission phase due to unfavorable cytogenetic markers (data not shown). One explanation could be the usually poor physical condition of resistant AML patients due to prolonged cytopenia after induction treatment and refractory disease. However, so far no randomized trial has demonstrated that avoiding intensive induction treatment and selecting other approaches would result in a higher SCT rate. The extremely poor prognosis of patients with RD shows an urgent need for alternative treatment approaches since a
Discussion We developed a powerful predictor for primary therapy resistance in AML. PS29MRC was validated in a fully independent patient cohort using a different technique to measure gene expression. The predictor also strongly associated with survival, which emphasizes the importance of the initial response to therapy. In our analysis, the predictive power for survival was limited to the intermediate and unfavorable ELN2017 genetic risk groups, possibly due to the low rate of resistant patients in the favorable genetic risk subgroup. PS29MRC identified a highrisk patient group of approximately 20% of all intensively treated AML patients with a median survival of only eight months, and a survival probability of only 12% at 24 months. Therefore, regarding the toxicity and side effects of induction treatment, it appears questionable as to whether this treatment can be still considered “standard” for this patient subgroup. PS29MRC showed limited effectiveness in high-risk groups defined by, for example, TP53 alterations, but was stronger in patients without cur-
Table 3. Univariate and multivariable analysis of the prediction of resistant disease of PS29MRCdic and alternative models in the validation set.
Variable PS29MRCdic AML-score by Walter et al.6
Variable PS29MRCdic Molecular Version of the AML-score by Walter et al.6
Variable PS29MRCdic LSC17
Variable PS29MRCdic Retrained response LSC17
Multivariable analysis, n=225§ OR [95%-CI] P 8.77 [4.27; 18.84] 1.21 [1.05; 1.39]
-9
8.15·10 0.0089
Multivariable analysis, n=225§ OR [95%-CI] P 8.82 [4.28; 18.98] 1.21 [0.96; 1.53]
-9
8.54·10 0.11
Multivariable analysis, n=235 OR [95%-CI] P 6.10 [2.99; 12.87] 11.51 [1.44; 99.78]
-6
1.10·10 0.023
Multivariable analysis, n=235 OR [95%-CI] P 7.44 [3.65; 15.78] 1.22 [0.66; 2.24]
-8
6.81·10 0.51
Univariate analysis* OR [95%-CI]
P
9.98 [4.92; 21.25] 1.28 [1.12; 1.48]
5.95·10-10 0.00068
Univariate analysis* OR [95%-CI]
P
9.98 [4.92; 21.25] 1.37 [1.11; 1.69]
5.95·10-10 0.0032
Univariate analysis OR [95%-CI]
P
8.03 [4.07; 16.46] 51.36 [7.80; 388.15]
4.29·10-9 7.29·10-5
Univariate analysis OR [95%-CI]
P
8.03 [4.07; 16.46] 1.97 [1.16; 3.39]
4.29·10-9 0.013
§ N=10 patients had to be excluded due to missing variables to calculate the AML-score by Walter et al.6 * To allow a fair comparison, univariate analyses were performed on the subset of patients with available information on all compared variables. OR: Odds Ratio; CI: Confidence Interval.
462
haematologica | 2018; 103(3)
Prediction of primary resistant AML
substantial number of patients do not benefit from induction treatment. PS29MRC would offer an accurate tool to design and implement such trials. We were able to demonstrate that the information on treatment response is included in the AML bulk itself at initial diagnosis. However, even though our study included a large amount of data in order to construct a better predictor, we still were only just able to reach a fair AUC in an independent validation set which is higher, but still in the range, of recent publications.3,6,7 Similar to the work of Walter et al.,6 we seem to have reached an obstacle that could not be overcome even by the addition of more information. Since all patients were considered eligible for intensive treatment, there seem to be additional, currently unknown variables that influence the response to induction treatment. It is tempting to speculate which other variables in addition to the disease itself affect response. Maybe clonal heterogeneity, individual drug metabolism or co-medications and interactions are more important than currently assumed. For example, CYP2E1 expression levels, which are associated with response to treatment in our study, influence cytarabine metabolism,30 and CYP2E1 expression levels
might be influenced by smoking.31 Furthermore, maybe the inclusion of more patients could help to increase the predictive ability. A recent publication analyzing 1540 AML patients and 231 predictor variables suggested that large knowledge banks of matched genomic-clinical data can support clinical decision making.32 Considering, for example, the work by Walter et al.6 and our study, we would strongly recommend including gene expression markers into these approaches because of their predictive potential. Gene expression data sets published by TCGA,23 HOVON,13,14 AMLSG33 and AMLCG,11,12 as well as the LEUCEGENE Project,34 already summarize more than 2000 AML patients that could be used to improve our prognostic and predictive abilities to personalize AML treatment. The implementation of a gene expression-based classifier in routine clincial practice is difficult because gene expression analysis is currently not included in the recommended molecular work up of newly diagnosed AML.35 However, advances in next generation sequencing result in more cost effective and robust methods to measure gene expression, such as NanoString or RNAseq.36,37 It is highly probable that these techniques will be available in future
A
B
C
haematologica | 2018; 103(3)
Figure 4. Refinement of the European LeukemiaNet (ELN) 2017 genetic risk classification (ELN2017) by predictive score PS29MRC. (A) Pie charts showing the distribution of patients according to ELN2017 and refined risk criteria. (B) Kaplan-Meier estimates of acute myeloid leukemia (AML) patients in the validation set according to ELN2017 and the refined ELN2017 classification. (C) Scheme of reclassification of the three ELN2017 risk groups into four groups by integrating PS29MRC as a dichotomous variable (PS29MRCdic) (high risk) with the ELN2017 risk classification.
463
T. Herold et al.
A
B
D
C
E
clinical settings and this will help to improve risk classification of patients. In summary, failure of induction treatment is one of the remaining challenges in the treatment of AML. From a clinical perspective, risk stratification before the start of treatment would be desirable. Patients with a high risk of RD could avoid the side effects of intensive induction and might be assigned to novel, experimental treatment strategies. We were able to validate a predictor that reached outstanding specificity and accuracy in an independent data set. PS29MRC could be instrumental in helping design clinical trials that overcome the current paradigm of intensive induction as standard treatment in all eligible patients. Acknowledgments This work is in memory of Thomas Büchner, who died while helping the authors with this project. The authors thank all participants of the AMLCG trials and recruiting centers. 464
Figure 5. Predictive ability of predictive score PS29MRC in genetic subgroups of acute myeloid leukemia (AML). (A) Bar plots showing the predictive ability of PS29MRC as a dichotomous variable (PS29MRCdic) in various genetic subgroups. The yaxis shows the absolute number of patients included. Patients in blue were predicted to respond to treatment (PS29MRCdic, low risk). Patients in red were predicted as resistant disease (PS29MRCdic, high risk). The accuracy is given in percentage. (B-E) Overall survival of AML patients in selected genetic subgroups. Kaplan-Meier estimates of AML patients classified according to PS29MRCdic as low risk and high risk.
Funding This work was supported by the Wilhelm-Sander-Stiftung (grant 2013.086.1) to TH, UM and KS; by funding from Deutsche Forschungsgemeinschaft (DFG Collaborative Research Centre SFB 1243) to PAG, HB, MS, WH, KHM and KS; by the “Verein zur Förderung von Wissenschaft und Forschung an der Medizinischen Fakultät der LMU München“ to TH. SKB is supported by Leukaemia & Blood Cancer New Zealand and the family of Marijanna Kumerich. Contribution: TH, UM and KS conceived and designed the experiments. TH, MR-T, BK, LH and KHM performed experiments. TH, VJ, AMNB, SAB, MR-T and KHM analyzed data. VJ, AMNB, SAB and UM provided bioinformatics support. JP-M, SK and HB managed the Genome Analyzer IIx platform and the RNA sequencing. MR-T, BK, LH, PAG, SS, NK, JT, MS, JB, SKB and KS characterized patient samples; MCS, DG, JB, SA, WEB, BJW and WH coordinated the AMLCG clinical trials. UM and KS supervised the project. TH, VJ, SAB, KHM and SKB wrote the manuscript. haematologica | 2018; 103(3)
Prediction of primary resistant AML
References 1. Dohner H, Estey EH, 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. 2. Dรถhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017; 129(4):424-447. 3. Walter RB, Othus M, Burnett AK, et al. Resistance prediction in AML: analysis of 4601 patients from MRC/NCRI, HOVON/SAKK, SWOG and MD Anderson Cancer Center. Leukemia. 2015; 29(2):312-320. 4. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol. 2003;21(24):4642-4649. 5. Ferguson P, Hills RK, Grech A, et al. An operational definition of primary refractory acute myeloid leukemia allowing early identification of patients who may benefit from allogeneic stem cell transplantation. Haematologica. 2016;101(11):13511358. 6. Walter RB, Othus M, Paietta EM, et al. Effect of genetic profiling on prediction of therapeutic resistance and survival in adult acute myeloid leukemia. Leukemia. 2015; 29(10):2104-2107. 7. Krug U, Rollig C, Koschmieder A, et al. Complete remission and early death after intensive chemotherapy in patients aged 60 years or older with acute myeloid leukaemia: a web-based application for prediction of outcomes. Lancet. 2010; 376(9757):2000-2008. 8. Dohner H, Weisdorf DJ, Bloomfield CD. Acute Myeloid Leukemia. N Engl J Med. 2015;373(12):1136-1152. 9. Buchner T, Krug UO, Peter Gale R, et al. Age, not therapy intensity, determines outcomes of adults with acute myeloid leukemia. Leukemia. 2016;30(8):1781-1784. 10. Buchner T, Berdel WE, Schoch C, et al. Double induction containing either two courses or one course of high-dose cytarabine plus mitoxantrone and postremission therapy by either autologous stem-cell transplantation or by prolonged maintenance for acute myeloid leukemia. J Clin Oncol. 2006;24(16):2480-2489. 11. Herold T, Metzeler KH, Vosberg S, et al. Isolated trisomy 13 defines a homogeneous AML subgroup with high frequency of mutations in spliceosome genes and poor prognosis. Blood. 2014;124(8):1304-1311. 12. Li Z, Herold T, He C, et al. Identification of a 24-gene prognostic signature that
haematologica | 2018; 103(3)
13.
14.
15.
16.
17.
18.
19. 20.
21.
22.
23.
improves the European LeukemiaNet risk classification of acute myeloid leukemia: an international collaborative study. J Clin Oncol. 2013;31(9):1172-1181. Wouters BJ, Lowenberg B, ErpelinckVerschueren CA, van Putten WL, Valk PJ, 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. 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. Braess J, Kreuzer K-A, Spiekermann K, et al. High Efficacy and Significantly Shortened Neutropenia Of Dose-Dense S-HAM As Compared To Standard Double Induction: First Results Of a Prospective Randomized Trial (AML-CG 2008). Blood. 2013;122(21):619-619. Greif PA, Dufour A, Konstandin NP, et al. GATA2 zinc finger 1 mutations associated with biallelic CEBPA mutations define a unique genetic entity of acute myeloid leukemia. Blood. 2012;120(2):395-403. Metzeler KH, Herold T, RothenbergThurley M, et al. Spectrum and prognostic relevance of driver gene mutations in acute myeloid leukemia. Blood. 2016;128(5):686698. ENCODE Consortium. Standards, Guidelines and Best Practices for RNA-Seq v.1.0. June 2011. [Available from: https://genome.ucsc.edu/encode/protocols/dataStandards/ENCODE_RNAseq_St andards_V1.0.pdf. Last accessed: 31st January 2018]. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15-21. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. Herold T, Jurinovic V, Metzeler KH, et al. An eight-gene expression signature for the prediction of survival and time to treatment in chronic lymphocytic leukemia. Leukemia. 2011;25(10):1639-1645. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood. 2010;116(3):354-365. TCGA. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013; 368(22): 2059-2074.
24. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med. 2016;374(23):2209-2221. 25. Ng SW, Mitchell A, Kennedy JA, et al. A 17gene stemness score for rapid determination of risk in acute leukaemia. Nature. 2016;540(7633):433-437. 26. Marcucci G, Maharry KS, Metzeler KH, et al. Clinical role of microRNAs in cytogenetically normal acute myeloid leukemia: miR155 upregulation independently identifies high-risk patients. J Clin Oncol. 2013; 31(17):2086-2093. 27. Khalife J, Radomska HS, Santhanam R, et al. Pharmacological targeting of miR-155 via the NEDD8-activating enzyme inhibitor MLN4924 (Pevonedistat) in FLT3ITD acute myeloid leukemia. Leukemia. 2015;29(10):1981-1992. 28. Thol F, Schlenk RF, Heuser M, Ganser A. How I treat refractory and early relapsed acute myeloid leukemia. Blood. 2015; 126(3):319-327. 29. Welch JS, Petti AA, Miller CA, et al. TP53 and Decitabine in Acute Myeloid Leukemia and Myelodysplastic Syndromes. N Engl J Med. 2016;375(21):2023-2036. 30. Iacobucci I, Lonetti A, Candoni A, et al. Profiling of drug-metabolizing enzymes/transporters in CD33+ acute myeloid leukemia patients treated with Gemtuzumab-Ozogamicin and Fludarabine, Cytarabine and Idarubicin. Pharmacogenomics J. 2013;13(4):335-341. 31. Zevin S, Benowitz NL. Drug interactions with tobacco smoking. An update. Clin Pharmacokinet. 1999;36(6):425-438. 32. Gerstung M, Papaemmanuil E, Martincorena I, et al. Precision oncology for acute myeloid leukemia using a knowledge bank approach. Nat Genet. 2017;49(3):332340. 33. Gaidzik VI, Paschka P, Spath D, et al. TET2 mutations in acute myeloid leukemia (AML): results from a comprehensive genetic and clinical analysis of the AML study group. J Clin Oncol. 2012; 30(12):1350-1357. 34. Pabst C, Bergeron A, Lavallee VP, et al. GPR56 identifies primary human acute myeloid leukemia cells with high repopulating potential in vivo. Blood. 2016; 127(16):2018-2027. 35. Dohner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017; 129(4):424-447. 36. Wang M, Lindberg J, Klevebring D, et al. Validation of risk stratification models in acute myeloid leukemia using sequencingbased molecular profiling. Leukemia. 2017;31(10):2029-2036. 37. Ng SW, Mitchell A, Kennedy JA, et al. A 17gene stemness score for rapid determination of risk in acute leukaemia. Nature. 2016;540(7633):433-437.
465
ARTICLE
Non-Hodgkin Lymphoma
Ferrata Storti Foundation
Chemotherapy-induced differential cell cycle arrest in B-cell lymphomas affects their sensitivity to Wee1 inhibition Xiaoguang Wang,1 Zhangguo Chen,1 Ameet K. Mishra,1 Alexa Silva,1 Wenhua Ren,2 Zenggang Pan3 and Jing H. Wang1
Department of Immunology and Microbiology; 2Department of Medicine Division of Pulmonary Sciences and Critical Care Medicine and 3Department of Pathology, University of Colorado, Anschutz Medical Campus, Aurora, CO, USA 1
Haematologica 2018 Volume 103(3):466-476
ABSTRACT
C
Correspondence: jing.wang@ucdenver.edu
Received: July 6, 2017. Accepted: November 30, 2017. Pre-published: December 7, 2017.
hemotherapeutic agents, e.g., cytarabine and doxorubicin, cause DNA damage. However, it remains unknown whether such agents differentially regulate cell cycle arrest in distinct types of Bcell lymphomas, and whether this phenotype can be exploited for developing new therapies. We treated various types of B cells, including primary and B lymphoma cells, with cytarabine or doxorubicin, and determined DNA damage responses, cell cycle regulation and sensitivity to a Wee1 inhibitor. We found that cyclin A2/B1 upregulation appears to be an intrinsic programmed response to DNA damage; however, different types of B cells arrest in distinct phases of the cell cycle. The Wee1 inhibitor significantly enhanced the apoptosis of G2 phase-arrested B-cell lymphomas by inducing premature entry into mitosis and mitotic catastrophe, whereas it did not affect G1/S-phase-arrested lymphomas. Cytarabine-induced G1-arrest can be converted to G2-arrest by doxorubicin treatment in certain B-cell lymphomas, which correlates with newly acquired sensitivity to the Wee1 inhibitor. Consequently, the Wee1 inhibitor together with cytarabine or doxorubicin inhibited tumor growth in vitro and in vivo more effectively, providing a potential new therapy for treating B-cell lymphomas. We propose that the differential cell cycle arrest can be exploited to enhance the chemosensitivity of B-cell lymphomas.
Introduction doi:10.3324/haematol.2017.175992 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/466 Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
466
Cytarabine, known as Ara-C, rapidly converts to cytosine arabinoside triphosphate, which can be incorporated into DNA during the process of DNA synthesis, and eventually causes DNA damage, probably by stalling replication forks and generating DNA double-stranded breaks. Given that cancer cells proliferate rapidly, Ara-C can kill cancer cells by interfering with their DNA synthesis during the S phase of the cell cycle. Ara-C has been the backbone of induction chemotherapy for acute myeloid leukemia and acute lymphocytic leukemia for several decades.1,2 For non-Hodgkin lymphomas, Ara-C is used as an upfront therapy for mantle cell lymphoma and Burkitt lymphoma, and as part of some salvage regimens when nonHodgkin lymphomas relapse. However, it remains incompletely understood how Ara-C treatment regulates DNA damage responses in primary B cells and B-cell lymphomas. The current treatment of B-cell non-Hodgkin lymphomas typically includes RCHOP, a combination of anti-CD20 (rituximab), three chemotherapy agents (cyclophosphamide, doxorubicin, vincristine), and one steroid (prednisone).3,4 This regimen has increased the rates of complete response for both young and elderly patients with diffuse large B-cell lymphoma.5,6 Both cyclophosphamide and doxorubicin are also DNA-damaging agents, although their functional mechanisms are different from those of Ara-C. Doxorubicin is commonly used to treat cancers, including breast cancer, bladder cancer, lymphoma and acute lymphoblastic leukemia.7 haematologica | 2018; 103(3)
G2-arrested B lymphoma is prone to Wee1 inhibition
Doxorubicin can stabilize the complex of topoisomerase II and broken DNA strands, thereby preventing the broken DNA double helix from being resealed and causing stalled DNA replication. Furthermore, the formation of doxorubicin-DNA adducts could activate DNA damage responses independent of topoisomerase II.8 When cells experience DNA damage, the cell cycle can be arrested in the G1, S or G2 phase for DNA repair.9 If the DNA damage is beyond recovery or the level of double-stranded breaks exceeds the repair capacity, cells never enter mitosis but die or undergo senescence.9 It does, however, remain poorly understood how doxorubicin treatment regulates cell cycle arrest and cell death in B-cell lymphomas. Cell cycle checkpoints are critical to control the progression of the cell cycle of DNA-damaged cells. The active complex of CDK1 and cyclinB1 controls entry into the mitotic (M) phase, and the expression of CDK1 is constitutive. Tyr15 phosphorylation mediated by Wee1 and Myt1 would inactivate CDK1, thus inhibiting mitotic entry. CyclinB1 expression increases at late S phase and reaches the peak at late G2 phase. CyclinB1 down-regulation would arrest cells at G2 phase, thus reducing mitotic entry.10,11 Further study proved that cyclinB1 is rate limiting but not essential for mitotic entry and progression.12 Abrogation of the G2/M checkpoint, for instance, by reducing the phosphorylation level of CDK1, enhances premature mitotic entry upon DNA damage, leading to increased cell death via mitotic catastrophe.9,13 Previous studies have shown that combined treatment with genotoxic drugs and Wee1 inhibitor efficiently controls leukemia progression.14-16 It remains unclear whether Wee1 inhibitor enhances the M phase entry of cell cycle-arrested B-cell lymphomas and, if so, whether G1, S or G2 phasearrested lymphomas are sensitive to Wee1 inhibitor. In the current study, we employed primary mouse B cells, and various mouse and human B-cell lymphoma lines to test how B cells respond to Ara-C or doxorubicin treatment and to elucidate the relationships among DNA damage, cell cycle arrest and the cell death pathway. Our data suggest that cyclinB1/A2 upregulation is an intrinsically programmed DNA damage response. We show that different types of B cells exhibit differential cell cycle arrest upon Ara-C or doxorubicin treatment. Overall, our studies may reveal new mechanistic insights into DNA damage responses and cell cycle regulation, identify biomarkers to predict chemosensitivity and facilitate the development of novel therapies for B-cell lymphomas and beyond.
Methods Cell culture and SOMAscan assay CH12 lymphoma cells were cultured as described previously.17 G1XP lymphomas were generated, established and cultured as described previously.18 Ramos, OCI-LY1, OCI-LY3, OCI-LY7 and DHL-16 were gifts from Dr. Wing C. Chan (University of Nebraska, NE, USA) and were cultured in 10% fetal bovine serum lymphocyte medium. Lymphoma cells were cultured at 0.5×106/mL, and treated with Ara-C (Cat. SY004943, Accela, San Diego, CA, USA), MK1775 (Cat. 2373, Biovision, Milpitas, CA, USA) or doxorubicin (Cat. 159101, MP Biomedicals) at indicated concentrations for 6 h or 24 h. Splenic B cells were isolated from wildtype (wt) naïve mice using a negative selection kit (Stem Cell Technologies, Canada), cultured with anti-CD40 and interleukin4 as described previously,19 and collected 4 days after culture for haematologica | 2018; 103(3)
Ara-C treatment or for western blot and flow cytometry analysis. Details of the SOMAscan assay and computational analysis are provided in the Online Supplementary Materials and Methods.
Western blot, flow cytometry, and knockdown of cyclins Primary antibodies used in the western blots are listed in Online Supplementary Table S3. Secondary horseradish peroxidase-conjugated anti-mouse, anti-goat and anti-rabbit antibodies were from Jackson ImmunoResearch (West Grove, PA, USA) and developed by ECL™ western blotting detection reagents (GE Healthcare, Little Chalfont, UK) according to the instructions provided with the kit. Details of the flow cytometry analysis including cell cycle analysis and cyclin knockdown are given in the Online Supplementary Materials and Methods.
Fluorescence microscopy Cells were collected and placed on poly-L-lysine-treated cover slips for 30 min, fixed by 4% paraformaldehyde for 1 h at room temperature and permeabilized by 0.1% Triton X-100 for 30 min. After blocking with 2% bovine serum albumin for 30 min, cells were covered by Vecta shield mounting medium (Cat.H-1200, Vector Laboratories, Burlingame, CA, USA) and slides. Images were acquired with an Eclipse TE2000 (Nikon).
In vivo treatment of the transplant G1XP lymphoma model G1XP lymphomas were generated by crossing Cγ1Cre knockin, Xrcc4, and Trp53 conditional knockout mice, as described previously.18 Animal work was approved by the Institutional Animal Care and Use Committee of University of Colorado Anschutz Medical Campus (Aurora, CO, USA). Details of the treatment are provided in the Online Supplementary Materials and Methods.
Results Ara-C treatment induced apoptosis via caspase3 activation and DNA fragmentation Ara-C is analogous to deoxycytidine and is thought to be an S phase-specific agent.20,21 How Ara-C causes the death of B cells remains incompletely understood. We treated mouse B-cell lymphoma, CH12 cells,22 with Ara-C, and found that Ara-C increased the percentage of annexin-V-positive CH12 cells and caused DNA fragmentation (Online Supplementary Figure S1A-C). Since the occurrence of DNA fragmentation is related to caspase3-induced apoptosis, we examined whether Ara-C treatment affected caspase3 activation. Our data showed that cleaved caspase3 was increased in a dose-dependent manner upon Ara-C treatment (Online Supplementary Figure S1D,E). Recent studies found that chemotherapeutic agents can cause necroptosis, a regulated form of necrosis or inflammatory cell death.23 We, therefore, examined the expression of RIP3 (RIPK3) and CaMKII, essential components of the necroptosis pathway, upon Ara-C treatment.24-26 We found that Ara-C did not increase RIP3 or CaMKII in various B-cell lymphomas (Online Supplementary Figure S1F). We, therefore, conclude that Ara-C causes apoptosis by activating caspase3 in B-cell lymphomas and does not induce necroptosis.
Ara-C treatment upregulated cyclinB1 and cyclinA2 in various types of B cells To elucidate how primary B cells or B-cell lymphomas respond to DNA damage, we employed cutting-edge 467
X. Wang et al.
aptamer-based multiplexed proteomic technology,27 SOMAscan, to identify the differentially expressed proteins in untreated versus Ara-C-treated B cells. Due to the breakthrough of SOMAmer,27 the SOMAscan™ proteomic assay enables us to quantify 1310 proteins across approximately eight logs of concentration as shown by previous studies.28,29 Among 1310 proteins tested, we found that there were only three proteins differentially expressed between untreated and Ara-C-treated primary B cells from wt or Xrcc4/p53 double conditional knockout mice.18 These were cyclinB1/cdk1, cyclinA2/cdk2 and importin B1 (IMB1), all of which were significantly upregulated (fold change ≥2, P<0.05) in Ara-C-treated primary activated B cells (Figure 1A and Online Supplementary Tables S1 and S2). We present the top ten upregulated/downregulated proteins in Ara-C-treated wt or double conditional knock-out primary B cells (Online Supplementary Figure S2A,B). We recently established a unique mouse model by specifically deleting a NHEJ gene, Xrcc4, and Trp53, in ger-
minal center B cells, which results in the spontaneous development of G1XP lymphomas.18 Cell lines were established from G1XP lymphomas. In line with our data from primary B cells, we found that cyclinB1 and cyclinA2 were also upregulated by Ara-C in our newly developed G1XP lymphomas and in CH12 lymphomas (Figure 1A). We performed western blotting to validate the results of the SOMAscan assay. Consistently, we found that cyclinB1 and cyclinA2 were indeed upregulated in Ara-Ctreated wt or p53 conditional knock-out primary B cells, and in CH12 or G1XP lymphoma cells (Figure 1B). Notably, we found that CDK1 phosphorylation (pCDK1) at Tyr15 was also enhanced by Ara-C treatment in primary B cells or B lymphoma cells; in contrast, total protein expression of CDK1 and CDK2 was not increased (Figure 1B). When CDK1 is phosphorylated at Tyr15 by Wee1 kinase, it is inactivated and blocks M phase entry. Thus, these data suggest that although cyclinB1 is upregulated, the cyclinB1/CDK1 complex stays inactive and
A
B
C
D
468
Figure 1. Upregulation of cyclinB1 and cyclinA2 in different types of B cells upon AraC treatment. (A) SOMAscan analysis of untreated or Ara-C-treated B cells. CH12 lymphoma (n=3), G1XP lymphoma (n=3), wt (n=3) or Xrcc4/p53 conditional knockout (n=3) primary B cells were treated with 10 μM Ara-C for 16 h. Cell lysates were prepared and subject to SOMAscan assay. Data are presented as mean±s.e.m. (B) Increased expression of cyclinB1, cyclinA2 and pCDK1 in different types of B cells upon Ara-C treatment. CH12 or G1XP lymphomas, wt or p53-/- primary B cells were treated with 1 μM Ara-C for 24 h. CyclinB1, CDK1, pCDK1, cyclinA2 and CDK2 were detected by western blot and β-actin as the loading control. (C) Ramos lymphomas were treated and analyzed as described in (B). (D) Upregulation of cyclinB1, cyclinA2 and pCDK1 in different types of B-cell lymphomas upon doxorubicin (DOX) treatment. CH12, Ramos and Ly1 lymphomas were treated with 100 nM doxorubicin for 24 h and analyzed as described in (B). Data are representative results of three independent experiments.
haematologica | 2018; 103(3)
G2-arrested B lymphoma is prone to Wee1 inhibition
probably would not promote M phase entry (see below). To generalize our findings, we treated several human Bcell lymphoma lines, including Ramos, OCI-LY1, OCILY3, OCI-LY7 and DHL-16, and found that Ara-C treatment upregulated cyclinB1, cyclinA2, and pCDK1 expression in all of them (Figure 1C, Online Supplementary Figure S3A). Additionally, we examined other players involved in DNA damage responses and found that Ara-C treatment activated Chk1 and Chk2 in most of the lymphoma lines examined. However, we did not detect obvious differences in the levels of CDC25A or pCDC25C (Online Supplementary Figure S3B). Next, we examined the kinetics of cyclinB1/A2 upregulation. CyclinB1 and cyclinA2 were upregulated after 8 h of Ara-C treatment (data not shown), and significant induction occurred after 20 h or 24 h (Online Supplementary Figure S3C). Overall, our results showed a time-dependent effect of Ara-C treatment on upregulating cyclinB1/A2. Lastly, we showed that another DNA-damaging agent, doxorubicin, can also upregulate cyclinB1/A2 in several B-
cell lymphoma lines (Figure 1D). We conclude that cyclinB1 and cyclinA2 upregulation appears to be an intrinsically programmed DNA damage response, which occurs in both activated primary B cells and various B-cell lymphomas.
Ara-C induces differential cell cycle arrest in different types of B cells Since cyclins control cell cycle progression, we examined how Ara-C affected cell cycling in mouse primary B cells and B-cell lymphomas in a time- and dose-dependent manner. Based on the Ara-C doses employed in previous studies,30 we chose to test the effects of treatment with 1 μM and 10 μM Ara-C. Six hours of Ara-C treatment did not affect cell cycle progression significantly in wt primary B cells regardless of Ara-C dosage (Figure 2A, top panel). In contrast, after 24 h of treatment, wt primary B cells predominantly arrested in the S phase in the presence of 10 μM Ara-C, whereas, 1 μM Ara-C treatment caused a modest increase in the percentage of S and G2 phase-arrested
A
B
C
haematologica | 2018; 103(3)
Figure 2. Different types of B cells are arrested in distinct phases of the cell cycle upon Ara-C treatment. (A) Primary B cells were arrested in the S phase of the cell cycle upon Ara-C treatment. Anti-CD40/IL-4 activated wt primary B cells were treated with 1 μM or 10 μM Ara-C for 6 or 24 h. (B) Mouse B-cell lymphomas were arrested in the G2 phase of the cell cycle upon Ara-C treatment. CH12 or G1XP lymphoma cells were treated with 1 μM Ara-C for 6 or 24 h. (C) Ramos lymphoma cells were treated with 1 μM Ara-C for 24 or 36 h. Cells were collected at indicated time points and fixed by 70% ethanol. After propidium iodide (PI) staining, cell cycles were determined by FACS (FL2A). Data are representative results of three to five independent experiments.
469
X. Wang et al.
cells (Figure 2A, bottom panel). Taken together, our data show that Ara-C induces mouse primary B cells preferentially arrested in the S phase of the cell cycle. Contrary to our findings in primary B cells, in CH12 cells we found that Ara-C treatment caused arrest in the G1 phase at an early time point (6 h: untreated 38.2% versus Ara-C 63.5%) and in the G2/M phase at a later time point (24 h: untreated 10.5% versus Ara-C 42.7%) (Figure 2B, top panel). To further corroborate our findings, we determined the cell cycle progression of G1XP lymphomas upon Ara-C treatment. Consistently, our data showed that Ara-C also arrested G1XP lymphomas in the G1 phase at an early time point and in the G2/M phase at the later time point (Figure 2B, bottom panel). Ara-C treatment resulted in apoptosis regardless of G1 or G2 phase of the cell cycle (Online Supplementary Figure S4A). All cells of the sub-G1 phase were positive for cleaved caspase3, indicating their apoptotic phenotypes (Online Supplementary Figure S4B). Nevertheless, the ratio among G1/S/G2 remained the same with or without sub-G1 phase included. Since we were more interested in living cells, sub-G1 phase cells were not included in the cell cycle analysis. To distinguish whether Ara-C induces G2 or M phase arrest, we determined the phosphorylation level of histone3 (pH3), which serves as a marker for M phase entry since chromosome condensation requires H3 phosphorylation. We found that pH3 level was reduced upon Ara-C treatment, demonstrating that Ara-C-treated lymphomas were arrested in the G2 phase but failed to progress into the M
phase (Online Supplementary Figure S5). These observations are consistent with our findings that the level of pCDK1 (Tyr15) was increased upon Ara-C treatment (Figure 1B), which would inactivate CDK1 and block M phase entry. In sharp contrast to our findings in CH12 and G1XP, we found that Ramos cells, a human Burkitt lymphoma line, were preferentially arrested in the G1 phase of the cell cycle (Figure 2C). Thus, Ara-C induces differential cell cycle arrest in different types of B cells, namely, S phase arrest in primary mouse B cells, G2 phase arrest in CH12 and G1XP lymphoma cells and G1 arrest in Ramos cells.
p53 deficiency does not affect Ara-C-induced cell cycle arrest or M phase blockage The TP53 gene is the most frequently mutated gene in human cancers.31 p53 is capable of inducing cell cycle arrest, apoptosis, or senescence, modulating DNA repair or metabolism, and serving as the guardian of the genome.32-34 Thus, we tested whether p53 deficiency might affect cell cycle progression upon Ara-C treatment. We treated wt or p53 conditional knock-out primary B cells with Ara-C (1 μM or 10 μM). Our data showed that primary B cells were arrested in the S phase after 24 h of Ara-C treatment regardless of p53 genotype (Figure 3A). Next, we sought to determine whether p53 deficiency affects M phase entry. The percentage of the pH3-positive population is relatively low in primary B cells given that they do not proliferate as fast as lymphomas (Figure 3B, Untreated). In order to increase the percentage of M phase cells, we used col-
A
B
Figure 3. p53 deficiency does not affect the cell cycle changes induced by Ara-C treatment in primary B cells. Wt and p53 conditional knockout primary B cells were treated with 1 μM or 10 μM Ara-C for 24 h. (A) Cell cycles were determined by propidium iodide (PI) staining and (B) mitotic entry was determined by pH3 and PI staining. Data are representative results of three independent experiments.
470
haematologica | 2018; 103(3)
G2-arrested B lymphoma is prone to Wee1 inhibition
cemid, a microtubule-depolymerizing drug that blocks M phase progression (Figure 3B, Untreated versus Untreated plus colcemid). However, we did not detect a significant difference in cell cycle arrest between wt and p53 conditional knock-out primary B cells (Figure 3B). Ara-C treatment blocked M phase entry regardless of p53 genotype (Figure 3B). We, therefore, conclude that p53 deficiency does not affect cell cycle progression of primary B cells upon Ara-C treatment. This observation is consistent with our findings with G1XP lymphomas that are deficient in both Xrcc4 and Trp53, yet, we found that G1XP lymphomas were arrested in the G2 phase, similar to p53-proficient CH12 cells. These data suggest that Ara-C-induced G2 phase arrest is also independent of p53.
Increased cyclinA2 and cyclinB1 were not required for Ara-C-induced G2 phase arrest To address whether increased cyclinA2 and cyclinB1 were required for Ara-C-induced G2 phase arrest,
A
C
haematologica | 2018; 103(3)
Wee1 inhibitor sensitized G2 phase-arrested lymphomas to Ara-C treatment Ara-C treatment increased pCDK1 (Figure 1B), indicating that pCDK1 may contribute to Ara-C-induced inhibitory effects on mitotic entry (Online Supplementary Figure S5). Since Wee1 kinase can phosphorylate CDK1, we employed Wee1 inhibitor to reduce pCDK1 in G1XP
B
D
E
cyclinA2, cyclinB1 or both were knocked down in CH12 cells by transient transfection of corresponding short hairpin RNA. Western blot confirmed the reduced expression of cyclinA2, cyclinB1, or both in knocked down CH12 cells compared with scrambled controls (Online Supplementary Figure S6A). However, the reduction of cyclinA2, cyclinB1, or both had no obvious effects on AraC-induced G2 phase arrest (Online Supplementary Figure S6B). Additionally, the cleaved caspase3 was not affected by reduced expression of cyclinA2, cyclinB1 or both (Online Supplementary Figure S6A).
F
Figure 4. Wee1 inhibitor (MK1775) enhances premature mitotic entry, mitotic catastrophe and apoptosis upon Ara-C treatment. (A, B) Wee1 inhibitor enhances premature mitotic entry of G2 phase-arrested B-cell lymphomas upon Ara-C treatment. G1XP lymphoma cells were either untreated or treated with 1 μM Ara-C, 100 nM MK1775, or both for 24 h. Cells were collected and fixed by 70% ethanol. After propidium iodide (PI) and pH3 staining, cell cycles were determined by FACS (FL1-H/FL2-A). Statistical significance was calculated with one-way ANOVA, Tukey multiple comparison test, **P≤0.01 in (B). (C) Increased apoptosis of B-cell lymphomas upon combined treatment with Ara-C and MK1775. G1XP lymphoma cells were treated as described in (A). pCDK1 and caspase3 were detected by western blot and βactin was the loading control. (D) Increased mitotic catastrophe of B-cell lymphoma upon combined treatment with Ara-C and MK1775. G1XP lymphoma cells were treated as described in (A). Cells undergoing mitotic catastrophe (white arrow) were detected by fluorescent microscopy with DAPI. (E) Increased cell death of Ara-C-induced G2 phase-arrested lymphomas upon MK1775 treatment. CH12 and G1XP lymphoma cells were treated as described in (A). Statistical significance was calculated with one-way ANOVA, Tukey multiple comparison test, *P≤0.05. (F) G1/S phase-arrested lymphomas upon Ara-C treatment are not sensitive to MK1775 treatment. Ramos, Ly1, Ly7 and DHL-16 lymphomas cells were either untreated or treated with 1 μM Ara-C, 100 nM MK1775, or both for 24 h. Cell numbers were counted and are presented as the percentage of the untreated group. Data are representative results of three independent experiments.
471
X. Wang et al.
lymphomas and tested whether mitotic entry was affected. Cell cycle analysis showed that Wee1 inhibitor enhanced the mitotic entry of Ara-C-induced G2 phasearrested lymphomas (Figure 4A,B). Consistently, Wee1 inhibitor reduced the level of pCDK1 (Figure 4C). Furthermore, combined treatment of Ara-C and Wee1 inhibitor induced a higher level of cleaved caspase3 (Figure 4C), promoted more cells to undergo mitotic catastrophe (Figure 4D), and caused more cell death (Figure 4E). Ara-C treatment resulted in G1 or S phase arrest in other lymphoma lines including DHL-16, Ly1, Ramos, and Ly7 (Online Supplementary Figure S7). Intriguingly, Wee1 inhibitor exhibited no effects on the level of mitotic entry (Online Supplementary Figure S7), cleaved caspase3 (Online Supplementary Figure S8) or cell death (Figure 4F) of the G1 or S phase-arrested lymphomas. Overall, our data suggested that Wee1 inhibitor appears to preferentially promote
472
Newly acquired sensitivity to Wee1 inhibitor in doxorubicin-induced G2-phase-arrested lymphomas To generalize our findings that Wee1 inhibitor may preferentially affect G2-phase-arrested lymphomas, we treated various lymphoma lines with doxorubicin and found that doxorubicin induced G2 phase-arrest in CH12, Ramos and Ly7 cells but G1 phase-arrest in Ly1 and DHL16 cells (Figure 5A). Consistently, we showed that combined treatment with doxorubicin and Wee1 inhibitor resulted in more cell death in G2 phase-arrested lymphoma lines compared with doxorubicin alone (Figure 5B). In contrast, the combined treatment had no effects on the survival of non-G2 phase-arrested lymphoma lines (Figure 5C). In line with our findings in Ara-C-treated lymphomas (Figure 4), we found that Wee1 inhibitor also
Figure 5. Newly acquired sensitivity to Wee1 inhibitor in doxorubicin-induced G2 phase-arrested lymphoma cells. (A) Cell cycle arrest of various lymphoma lines upon doxorubicin treatment. CH12, Ramos, Ly7, Ly1 and DHL-16 were either untreated or treated with 100 nM doxorubicin, 100 nM MK1775, or both for 24 h. Cells were collected and fixed by 70% ethanol. After propidium iodide (PI) staining, cell cycles were determined by flow cytometry (FL2-A). (B) Increased cell death of doxorubicin-induced G2 phase-arrested lymphomas upon MK1775 treatment. CH12, Ramos and Ly7 lymphomas cells were treated as described in (A). Statistical significance was calculated with one-way analysis of variance, Tukey multiple comparison test, *Pâ&#x2030;¤0.05. (C) Non-G2 phase arrested lymphomas, upon doxorubicin treatment, are not sensitive to MK1775 treatment. Ly1 and DHL-16 lymphoma cells were treated as described in (A). Cell numbers were counted and presented as the percentage of the untreated group. Data are representative results of three independent experiments.
A
B
mitotic entry, mitotic catastrophe and cell death in G2 phase-arrested lymphomas.
C
haematologica | 2018; 103(3)
G2-arrested B lymphoma is prone to Wee1 inhibition
enhanced the mitotic entry of doxorubicin-induced G2 phase-arrested lymphomas (Figure 6A,B), whereas it did not affect the non-G2 phase-arrested lymphomas (Figure 6B). In summary, our data showed that Ramos and Ly7 lymphomas were arrested in G1 or S phase upon Ara-C treatment (Online Supplementary Figure S7); however, doxorubicin treatment promoted G2 phase arrest in these two lymphoma lines, which correlated with their newly acquired sensitivity to Wee1 inhibitor.
Combined Ara-C and Wee1 inhibitor profoundly suppressed tumor growth in vivo To test the therapeutic effects of combined Ara-C and Wee1 inhibitor treatment, we established an in vivo transplant model in which G1XP lymphoma cells injected into syngeneic recipient mice develop into secondary B-cell lymphomas (Figure 7A,B). These secondary G1XP lymphomas were initially sensitive to Ara-C treatment, but became unresponsive to continuous Ara-C treatment and relapsed ∼20 days after the initial Ara-C treatment (Figure 7A,B). Wee1 inhibitor alone had no obvious effects on tumor growth or recipient survival compared with the vehicle control (Figure 7A,B). In contrast, combined treatment of Wee1 inhibitor and Ara-C significantly prolonged recipients’ survival (Figure 7A) and effectively suppressed tumor growth (Figure 7B). Of note, some of the recipient mice in the Ara-C/MK1775 group died early from unknown causes other than lymphoma. At day 34, tumor size was significantly reduced in the combined treatment group compared with that in the group treated with AraC alone (Figure 7C); animals in the groups treated with the vehicle control or Wee1 inhibitor were terminated before this time point due to tumor size exceeding the limit of institutional guidelines. In addition, combined treatment with Ara-C and Wee1 inhibitor attenuated the side effects of Ara-C treatment alone since the weight of recipients was significantly greater in the combined treatment group (Figure 7D). We conclude that, when combined with other chemotherapeutic agents, Wee1 inhibitor may provide a novel therapeutic intervention for B-cell lymphomas that can be arrested in the G2 phase of the cell cycle.
teins included in the SOMAscan assay, only three were differentially expressed between untreated and Ara-Ctreated primary B cells, including cyclinB1/A2. However, we failed to identify the functional consequence of cyclinB1/A2 upregulation since double knockdown did not perturb the cell cycle or cause more apoptosis. Perhaps upregulation of these cyclins is indeed a stereotyped universal response to DNA damage, but is only functionally significant in certain types of DNA damage, such as irradiation.35 Another possibility is that other cyclins may play a compensatory role when cyclins A2 and B1 are knocked down. Our preliminary RNA-seq data did not show the upregulation of cyclinB1/A2 transcripts in the Ara-C-treated group. Hence, we predict that Ara-C treatment may somehow regulate the protein level of cyclinB1/A2, for example, via post-translational modulation of cyclinB1/A2 that leads to high and stabilized protein levels being maintained. Future studies are needed to elucidate the mechanisms that upregulate these cyclins upon Ara-C treatment, independently from blocking the cell cycle. Previous studies showed that DNA damage often elicits innate immune responses, such as upregulation of NKG2D ligands or interferon responses, in macrophages36
A
B
Discussion Althrough Ara-C and doxorubicin have been used in clinics for several decades, it remains incompletely understood how primary B cells or B-cell lymphomas respond to such DNA-damaging agents. In the current study, we employed Ara-C and doxorubicin to treat primary mouse B cells and various B-cell lymphoma lines and present four novel findings: (i) upregulation of cyclinA2 and cyclinB1 appears to be an intrinsically programmed DNA damage response; (ii) Ara-C or doxorubicin induces differential cell cycle arrest in different types of B cells; (iii) Wee1 inhibitor sensitizes G2 phase-arrested B lymphoma cells to Ara-C treatment by inducing premature mitotic entry and mitotic catastrophe. Furthermore, Ara-C-induced G1 phasearrest can be converted to G2 phase-arrest by doxorubicin treatment in certain B-cell lymphomas (e.g., Ramos cells), which correlates with newly acquired sensitivity to Wee1 inhibitor, and (iv) combined treatment with Ara-C and Wee1 inhibitor profoundly suppressed the tumor growth of transplanted G1XP lymphomas in vivo. We present unexpected findings that, among 1310 prohaematologica | 2018; 103(3)
Figure 6. Wee1 inhibitor (MK1775) enhances premature mitotic entry of G2 phase-arrested lymphomas upon treatment with doxorubicin. CH12, Ramos, Ly1, Ly7 and DHL-16 lymphoma cells were treated as described in Figure 5. Cells were collected and fixed by 70% ethanol. After propidium iodide (PI) and pH3 staining, cell cycles were determined by flow cytometry (FL1-H/FL2-A). (A) Representative FACS data of Ly7 lymphoma cells are shown. (B) Statistical analysis of premature mitotic entry in various lymphoma lines. Statistical significance was calculated with ANOVA. Tukey multiple comparison test, *P≤0.05. Data are representative results of three independent experiments.
473
X. Wang et al.
or B-cell lymphomas.37 However, we did not detect upregulation of any factors related to interferon responses in Ara-C-damaged primary B cells. This observation is consistent with previous findings in 3’ repair exonuclease 1 (Trex 1) conditional knockout mice.38 TREX1 is an exonuclease that degrades cytosolic DNA and RNA and one of the unknown substrates of TREX1 can trigger cytoplasmic DNA sensor cyclic GMP–AMP synthase (cGAS). Loss of Trex1 in dendritic cells was sufficient to cause the release of interferon and systemic autoimmunity, whereas, deletion of Trex1 in B cells via CD19Cre did not produce any detectable interferon responses.38 Taken together, these data suggest that primary B cells appear to tolerate DNA damage with a higher threshold than other types of cells. This phenomenon is likely attributed to the physiological process of programmed generation of double-stranded breaks in primary B cells, namely, class switch recombination and somatic hypermutation. Double-stranded breaks are the essential intermediates of class switch recombination,39 and recent studies revealed that somatic hypermutation can result in double-stranded break formation in activated B cells.40 Hence, it may not be surprising that
A
C
most non-Hodgkin lymphomas (80-90%) derive from B cells (B-cell non-Hodgkin lymphoma).41 This higher incidence of B-cell lymphomas is probably attributable to Bcell-specific DNA recombination induced by activationinduced deaminase.39 While primary B cells do not appear to trigger any immune responses upon DNA damage, this scenario would help to protect DNA-damaged B cells from being attacked by innate immune cells or T cells. Since Ara-C functions as an S phase-specific chemotherapeutic drug, it would presumably arrest cells in the S phase. We found, however, that different types of B-cell lymphomas were arrested in distinct phases of the cell cycle by Ara-C treatment. The differential cell cycle arrests might depend on the specificity of chemotherapy agents (e.g., Ara-C versus doxorubicin), the severity of induced DNA damage and the genetic or epigenetic profiles of different types of B-cell lymphomas. Intriguingly, we found that the cell cycle arrest pattern in certain B-cell lymphomas can be shifted with different chemotherapeutic agents. For instance, Ramos and Ly7 were arrested by Ara-C in the G1 or S phase; however, they could be preferentially arrested in the G2 phase by doxorubicin
B
D
Figure 7. Combined treatment with Wee1 inhibitor (MK1775) and Ara-C eradicates transplanted G1XP lymphoma and attenuates side effects. (A) Kaplan-Meier survival curve of recipient mice inoculated with G1XP lymphomas and treated with vehicle (n=18), Ara-C (n=13), MK1775 (n=13) or both (n=13). Data are combined from three independent experiments. When tumor size reached 4000 mm3 or other humane end points were met (e.g., necrotic tumors), mice were euthanized in accordance with institutional guidelines. (B) Combined treatment with Ara-C and MK1775 profoundly suppressed the growth of G1XP lymphomas. Recipients were treated daily with vehicle control or MK1775 from day 21 until termination, or with Ara-C alone from day 21 to day 46, or Ara-C/MK1775 from day 21 to day 28 after tumor inoculation. (C) Combined treatment remarkably reduced the tumor size. The tumor size was monitored for the group treated with Ara-C alone (n=26) vs. the Ara-C/MK1775 (n=26) treated group for about 1 week (from day 32 to day 38 after tumor inoculation), whereas the vehicle control and MK1775 groups were already terminated. Data are shown for theday 34 time point. (D) The weight of recipient mice was increased in the combined treatment group. The weight of recipient mice was monitored similarly as described in (C) for the group treated with Ara-C alone (n=13) vs. the Ara-C/MK1775 (n=13) group. Data are shown for day 34 time point. Statistical significance was calculated with a T comparison test, ***P≤0.001.
474
haematologica | 2018; 103(3)
G2-arrested B lymphoma is prone to Wee1 inhibition
treatment, which concomitantly occurred with acquired sensitivity to Wee1 inhibitor. We propose that the plasticity of G2 phase-arrest could be exploited for the design of personalized B-cell lymphoma therapy: (i) the cell cycle arrest pattern could be determined upon treatment with different chemotherapeutic agents for each individual patient; (ii) G2 phase-arrest inducing agents could be identified, which could then be combined with Wee1 inhibitor to achieve more effective treatment for the corresponding patient. Combined treatment with Wee1 inhibitor and Ara-C has been tested for acute myeloid leukemia and T-cell acute lymphocytic leukemia,14,15 and Wee1 inhibition was also tested in different types of cancers including medulloblastoma and hepatocellular carcinoma;42-45 however, the underlying mechanisms remain incompletely understood. Mitotic entry is restricted by the phosphorylation of CDK1, and inhibition of pCDK1 could abrogate the G2/M checkpoint and propel G2-phase cells to enter the M phase.44,46,47 This premature M phase entry would elicit mitotic catastrophe or apoptosis, which has been suggested to act as an effective means of killing cancer cells.9,13 Previous studies found that Wee1 inhibition promoted G1 or S phase-arrested cells to undergo premature mitotic
References 1. Wang WS, Tzeng CH, Chiou TJ, et al. Highdose cytarabine and mitoxantrone as salvage therapy for refractory non-Hodgkin's lymphoma. Jpn J Clin Oncol. 1997;27(3):154157. 2. Momparler RL. Optimization of cytarabine (ARA-C) therapy for acute myeloid leukemia. Exp Hematol Oncol. 2013;2:20. 3. Vose JM, Link BK, Grossbard ML, et al. Phase II study of rituximab in combination with chop chemotherapy in patients with previously untreated, aggressive nonHodgkin's lymphoma. J Clin Oncol. 2001;19(2):389-397. 4. Czuczman MS, Grillo-Lopez AJ, White CA, et al. Treatment of patients with low-grade B-cell lymphoma with the combination of chimeric anti-CD20 monoclonal antibody and CHOP chemotherapy. J Clin Oncol. 1999;17(1):268-276. 5. 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. 6. Pfreundschuh M, Trumper L, Osterborg A, et al. CHOP-like chemotherapy plus rituximab versus CHOP-like chemotherapy alone in young patients with good-prognosis diffuse large-B-cell lymphoma: a randomised controlled trial by the MabThera International Trial (MInT) Group. Lancet Oncol. 2006;7(5):379-391. 7. Tacar O, Sriamornsak P, Dass CR. Doxorubicin: an update on anticancer molecular action, toxicity and novel drug delivery systems. J Pharm Pharmacol. 2013;65(2):157-170. 8. Yang F, Teves SS, Kemp CJ, et al. Doxorubicin, DNA torsion, and chromatin dynamics. Biochim Biophys Acta. 2014;1845(1):84-89. 9. Vitale I, Galluzzi L, Castedo M, et al. Mitotic
haematologica | 2018; 103(3)
10.
11.
12.
13.
14.
15.
16.
17.
18.
entry.16,48 However, our data suggest that only G2 phasearrested B cell lymphomas were sensitive to combined treatment with Wee1 inhibitor and Ara-C or doxorubicin. In line with in vitro data, our in vivo transplant G1XP model demonstrated a potential of combined Ara-C and Wee1 inhibitor treatment in eradicating B-cell lymphomas. Prior studies showed that Ara-C treatment activated Chk1,49 and Chk1 inhibitors synergize with Ara-C in suppressing acute myeloid leukemia.49,50 Consistently, we found that Ara-C treatment activated Chk1 and Chk2 in most of the B-cell lymphoma lines examined. Overall, these data suggest that cell cycle modulators may enhance the sensitivity of cancer cells to chemotherapeutic agents (e.g., Ara-C or doxorubicin), and that the combinatorial therapy may be more effective. Of note, we observed that a small percentage of recipients died early in the Ara-C/MK1775 group in the absence of lymphoma recurrence, suggesting that the toxic effects of combined Ara-C/MK1775 treatment need to be tested in future studies. Acknowledgments We thank Michael Rice, Stephanie Cung, and Brittany C. Waschke for their technical help. We apologize to those whose work is not cited here due to length restrictions.
catastrophe: a mechanism for avoiding genomic instability. Nat Rev Mol Cell Biol. 2011;12(6):385-392. Yuan J, Kramer A, Matthess Y, et al. Stable gene silencing of cyclin B1 in tumor cells increases susceptibility to taxol and leads to growth arrest in vivo. Oncogene. 2006;25(12):1753-1762. Yuan JP, Yan RL, Kramer A, et al. Cyclin B1 depletion inhibits proliferation and induces apoptosis in human tumor cells. Oncogene. 2004;23(34):5843-5852. Soni DV, Sramkoski RM, Lam M, et al. Cyclin B1 is rate limiting but not essential for mitotic entry and progression in mammalian somatic cells. Cell Cycle. 2008;7(9):1285-1300. Denisenko TV, Sorokina IV, Gogvadze V, et al. Mitotic catastrophe and cancer drug resistance: a link that must to be broken. Drug Resist Updat. 2016;24:1-12. Ford JB, Baturin D, Burleson TM, et al. AZD1775 sensitizes T cell acute lymphoblastic leukemia cells to cytarabine by promoting apoptosis over DNA repair. Oncotarget. 2015;6(29):28001-28010. Van Linden AA, Baturin D, Ford JB, et al. Inhibition of Wee1 sensitizes cancer cells to antimetabolite chemotherapeutics in vitro and in vivo, independent of p53 functionality. Mol Cancer Ther. 2013;12(12):2675-2684. Zhou L, Zhang Y, Chen S, et al. A regimen combining the Wee1 inhibitor AZD1775 with HDAC inhibitors targets human acute myeloid leukemia cells harboring various genetic mutations. Leukemia. 2015;29(4): 807-818. Chen Z, Getahun A, Chen X, et al. Imbalanced PTEN and PI3K signaling impairs class switch recombination. J Immunol. 2015;195(11):5461-5471. Chen Z, Elos MT, Viboolsittiseri SS, et al. Combined deletion of Xrcc4 and Trp53 in mouse germinal center B cells leads to novel B cell lymphomas with clonal heterogeneity.
J Hematol Oncol. 2016;9:2. 19. Chen ZG, Ranganath S, Viboolsittiseri SS, et al. AID-initiated DNA lesions are differentially processed in distinct B cell populations. J Immunol. 2014;193(11):5545-5556. 20. Sampath D, Cortes J, Estrov Z, et al. Pharmacodynamics of cytarabine alone and in combination with 7-hydroxystaurosporine (UCN-01) in AML blasts in vitro and during a clinical trial. Blood. 2006;107(6):2517-2524. 21. Mesa RA, Loegering D, Powell HL, et al. Heat shock protein 90 inhibition sensitizes acute myelogenous leukemia cells to cytarabine. Blood. 2005;106(1):318-327. 22. Nakamura M, Kondo S, Sugai M, et al. High frequency class switching of an IgM+ B lymphoma clone CH12F3 to IgA+ cells. Int Immunol. 1996;8(2):193-201. 23. Linkermann A, Green DR. Necroptosis. N Engl J Med. 2014;370(5):455-465. 24. Dillon CP, Weinlich R, Rodriguez DA, et al. RIPK1 blocks early postnatal lethality mediated by caspase-8 and RIPK3. Cell. 2014;157(5):1189-1202. 25. Orozco S, Yatim N, Werner MR, et al. RIPK1 both positively and negatively regulates RIPK3 oligomerization and necroptosis. Cell Death Differ. 2014;21(10):1511-1521. 26. Vanden Berghe T, Linkermann A, JouanLanhouet S, et al. Regulated necrosis: the expanding network of non-apoptotic cell death pathways. Nat Rev Mol Cell Biol. 2014;15(2):135-147. 27. Gold L, Ayers D, Bertino J, et al. Aptamerbased multiplexed proteomic technology for biomarker discovery. PLoS One. 2010;5(12): e15004. 28. Guiraud S, Edwards B, Squire SE, et al. Identification of serum protein biomarkers for utrophin based DMD therapy. Sci Rep. 2017;7:43697. 29. Coenen-Stass AM, McClorey G, Manzano R, et al. Identification of novel, therapyresponsive protein biomarkers in a mouse
475
X. Wang et al.
30.
31.
32. 33. 34.
35. 36.
476
model of Duchenne muscular dystrophy by aptamer-based serum proteomics. Sci Rep. 2015;5:17014. Ersvaer E, Brenner AK, Vetas K, et al. Effects of cytarabine on activation of human T cells - cytarabine has concentration-dependent effects that are modulated both by valproic acid and all-trans retinoic acid. BMC Pharmacol Toxicol. 2015;16:12. Surget S, Khoury MP, Bourdon JC. Uncovering the role of p53 splice variants in human malignancy: a clinical perspective. Onco Targets Ther. 2013;7:57-68. Lane DP. Cancer. p53, guardian of the genome. Nature. 1992;358(6381):15-16. Carr AM, Green MH, Lehmann AR. Checkpoint policing by p53. Nature. 1992;359(6395):486-487. Bieging KT, Mello SS, Attardi LD. Unravelling mechanisms of p53-mediated tumour suppression. Nat Rev Cancer. 2014;14(5):359-370. Porter LA, Singh G, Lee JM. Abundance of cyclin B1 regulates gamma-radiation-induced apoptosis. Blood. 2000;95(8):2645-2650. Hartlova A, Erttmann SF, Raffi FA, et al. DNA damage primes the type I interferon system via the cytosolic DNA sensor STING to promote anti-microbial innate immunity.
Immunity. 2015;42(2):332-343. 37. Lam AR, Le Bert N, Ho SS, et al. RAE1 ligands for the NKG2D receptor are regulated by STING-dependent DNA sensor pathways in lymphoma. Cancer Res. 2014;74(8):2193-2203. 38. Peschke K, Achleitner M, Frenzel K, et al. Loss of trex1 in dendritic cells is sufficient to trigger systemic autoimmunity. J Immunol. 2016;197(6):2157-2166. 39. Chen Z, Wang JH. Generation and repair of AID-initiated DNA lesions in B lymphocytes. Front Med. 2014;8(2):201-216. 40. Yeap LS, Hwang JK, Du Z, et al. Sequenceintrinsic mechanisms that target AID mutational outcomes on antibody genes. Cell. 2015;163(5):1124-1137. 41. Kuppers R. Mechanisms of B-cell lymphoma pathogenesis. Nat Rev Cancer. 2005;5(4): 251-262. 42. Harris PS, Venkataraman S, Alimova I, et al. Integrated genomic analysis identifies the mitotic checkpoint kinase WEE1 as a novel therapeutic target in medulloblastoma. Mol Cancer. 2014;13:72. 43. Kogiso T, Nagahara H, Hashimoto E, et al. Efficient induction of apoptosis by Wee1 kinase inhibition in hepatocellular carcinoma cells. Plos One. 2014;9(6): e100495.
44. Mueller S, Haas-Kogan DA. WEE1 kinase as a target for cancer therapy. J Clin Oncol. 2015;33(30):3485-3487. 45. Pappano WN, Zhang Q, Tucker LA, et al. Genetic inhibition of the atypical kinase Wee1 selectively drives apoptosis of p53 inactive tumor cells. BMC Cancer. 2014;14:430. 46. Elledge SJ. Cell cycle checkpoints: preventing an identity crisis. Science. 1996;274 (5293):1664-1672. 47. van Vugt MA, Medema RH. Getting in and out of mitosis with Polo-like kinase-1. Oncogene. 2005;24(17):2844-2859. 48. Aarts M, Sharpe R, Garcia-Murillas I, et al. Forced mitotic entry of S-phase cells as a therapeutic strategy induced by inhibition of WEE1. Cancer Discov. 2012;2(6):524539. 49. Schenk EL, Koh BD, Flatten KS, et al. Effects of selective checkpoint kinase 1 inhibition on cytarabine cytotoxicity in acute myelogenous leukemia cells in vitro. Clin Cancer Res. 2012;18(19):5364-5373. 50. Karp JE, Thomas BM, Greer JM, et al. Phase I and pharmacologic trial of cytosine arabinoside with the selective checkpoint 1 inhibitor Sch 900776 in refractory acute leukemias. Clin Cancer Res. 2012;18(24): 6723-6731.
haematologica | 2018; 103(3)
ARTICLE
Non-Hodgkin Lymphoma
Blood cytokine concentrations in pediatric patients with anaplastic lymphoma kinasepositive anaplastic large cell lymphoma
Ferrata Storti Foundation
Fabian Knörr,1* Christine Damm-Welk,1* Stephanie Ruf,1 Vijay Kumar Singh,1 Martin Zimmermann,2 Alfred Reiter1 and Wilhelm Woessmann1 NHL-BFM Study Center, Department of Pediatric Hematology and Oncology, Justus-Liebig University, Giessen and 2Department of Pediatric Hematology and Oncology, Children’s Hospital, Hannover Medical School, Germany
1
*FK and CDW contributed equally to this work.
Haematologica 2018 Volume 103(3):477-485
ABSTRACT
P
atients with anaplastic lymphoma kinase-positive anaplastic large cell lymphoma often present with B-symptoms or hemophagocytosis and generate an anti-tumor immune response. Specific serum cytokine levels or profiles may reflect the tumor burden, non-specific immune stimulation by the tumor or differences in the strength of the patients’ anti-lymphoma immunity. We systematically correlated pretreatment concentrations of 25 cytokines with clinical and biological characteristics in a well-characterized cohort of 119 uniformly treated pediatric patients with anaplastic large cell lymphoma. Fifteen patients with anaplastic large cell lymphoma in remission and 11 patients with low-stage B-cell lymphoma served as controls. Concentrations of interleukin-9, interleukin-10, interleukin-17a, hepatocyte growth factor, soluble interleukin-2 receptor, and soluble CD30 were significantly higher in initial sera of patients than in the sera of subjects from both control groups, indicating an anaplastic large cell lymphoma-type cytokine signature. The levels of interleukin-6, interferon-γ, interferon γ-induced protein, and soluble interleukin-2 receptor correlated with the stage, initial general condition, minimal disseminated disease, anaplastic lymphoma kinase-antibody titers, and the risk of relapse among patients with anaplastic lymphoma kinase-positive anaplastic large cell lymphoma. Only interleukin-6 showed an independent prognostic value in multivariate analyses. Pre-treatment cytokine profiles in patients with anaplastic large cell lymphoma reflect a tumor signature as well as tumor burden and also differences in the strength of the patients’ immune response.
Correspondence: Christine.Damm-Welk@paediat.med.unigiessen.de Received: August 16, 2017. Accepted: December 7, 2017. Pre-published: December 14, 2017.
doi:10.3324/haematol.2017.177972 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/477
Introduction Patients with anaplastic lymphoma kinase (ALK)-positive anaplastic large cell lymphoma (ALCL) often present with B symptoms or a macrophage activation syndrome indicating an inflammatory or immune reaction.1 In addition, ALK-positive ALCL elicits a specific host immune response, as indicated by the production of antiALK-autoantibodies,2 and a cellular immune response against ALK.3-6 In some aspects, the immune response in ALCL is comparable to that occurring in patients with Hodgkin lymphoma, in whom elevated serum cytokine levels have been described at the time of diagnosis.7,8 Elevated levels of serum cytokines as immune mediators, such as interleukin (IL)6, IL-8, IL-10, IL-17a, and IL-22, have also been shown in small series of ALK-positive ALCL patients.9,10 In vitro production of IL-6, IL-8 and interferon (IFN)-γ by an ALKpositive ALCL cell line HSC-M1 has been reported.11 Soluble CD30 (sCD30) and the soluble IL-2 receptor (sIL-2R) can be shed from ALCL cells and their levels were haematologica | 2018; 103(3)
©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
477
F. Knörr et al.
found to be elevated in seven ALK-positive ALCL patients.12 sCD30 levels correlated with an inferior survival in a series of adult ALCL patients with unknown ALK-status.13,14 IL-9 has been described as part of an autocrine signaling pathway including JAK3 with a possible role in the pathogenesis of ALK-positive ALCL.15 IL-22 expressed in three ALCL cell lines contributes to STAT3 activation and tumorigenicity of ALK-positive ALCL.16 While above-mentioned reports highlight ALCL as a cytokine-active lymphoma with hints towards an ALCLtype cytokine signature, the type and pattern of cytokine expression among ALK-positive ALCL patients has not been analyzed systematically. Pretreatment serum cytokine levels may reflect the biological activity of the tumor as well as host immune characteristics. Correlations of initial cytokine concentrations and patterns with patients’ characteristics, antibody titers against ALK as a surrogate measure of an autologous immune response against ALK, as well as outcomes might allow the definition of cytokine profiles that are associated with disease activity, tumor burden and the patients’ specific immune response. We, therefore, investigated whether pretreatment cytokine concentrations in patients with ALK-positive ALCL correlate with biological and clinical characteristics and ALK-antibody titers in a uniformly treated, large cohort of 119 children and adolescents with ALK-positive ALCL.
Methods Eligibility NPM-ALK-positive ALCL patients treated in the BerlinFrankfurt-Muenster group study NHL-BFM95 or German patients enrolled in the European intergroup trial ALCL99 between August 1998 and December 2008 were potentially eligible for inclusion in this study after giving written informed consent to participation. Both studies were approved by the institutional ethics committee of the primary investigator of the NHL-BFM study group (AR). Patients with completely resected stage I disease were excluded because their treatment was different. Patients were included if there were pretreatment serum or plasma samples available and had detectable anti-ALK antibody titer levels. Eligibility was confirmed by demonstration of NPM-ALK positivity of the tumor either by NPM-ALK polymerase chain reaction, two-color fluorescence in situ hybridization for t(2;5) or nuclear and cytoplasmic staining for ALK.
Patients The inclusion criteria were fulfilled by 119 patients. Staging procedures included bone marrow aspiration cytology and a spinal tap. Bone marrow involvement was defined by cytologically detectable ALCL cells, irrespective of their number. The patients’ treatment consisted of a cytoreductive prephase followed by six chemotherapy courses, as described elsewhere.17 As control, serum or plasma samples from 15 of those patients in remission without concurrent infection taken before the start of the sixth chemotherapy course were analyzed. In addition, serum samples taken at the time of diagnosis from 11 age-matched patients with Burkitt lymphoma from risk groups R1 and R2 (stage I – III, lactate dehydrogenase below 500 U/L) included in the B-NHL BFM 04 study served as a second control group. Methods and the patients’ results regarding ALK-antibody titers and minimal disseminated disease at diagnosis were described and published previously.18-20 478
Measurement of cytokine levels Blood samples were centrifuged and supernatants were immediately frozen and stored at -80°C until analysis. Samples were assessed for the levels of following soluble mediators: IL1β, IL-2, sIL-2R, IL-4, IL-5, IL-6, IL-8, IL-9, IL-10, IL-12p70, IL13, IL-17a, IL-22, IL-23, tumor necrosis factor-α (TNF-α), IFNγ, monokine induced by γ-interferon (MIG), interferon γinduced protein 10 (IP-10), hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF), monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein (MIP)-1α, MIP-1β, granulocyte colony-stimulating factor (G-CSF) and soluble CD30 (sCD30). The measurements were performed using FlowCytomix kits (eBioscience, Frankfurt, Germany) according to the manufacturer’s instructions. Samples were processed on a FACS Calibur (BD, Heidelberg, Germany) and data were analyzed using the FlowCytomix Software (version 2.4, eBioscience, Frankfurt, Germany).
Statistical methods Statistical calculations were performed using the R statistical package (R Foundation for Statistical Computing, Vienna, Austria). Cytokine levels are reported as median values and were compared between different groups according to diagnosis, clinical and biological characteristics using Mann-Whitney U and Kruskal-Wallis tests. The level of statistical significance was 0.05. Event-free survival was defined as the time from diagnosis to relapse, secondary tumor or death from any cause. Estimates of overall and event-free survival were performed using the Kaplan-Meier method. Differences between groups were compared by log-rank test. A multivariate analysis was performed using the proportional hazards method described by Cox on cytokines showing significant differences in univariate analysis and known risk factors20 with forward selection keeping only significant variables (P<0.05) in the model.
Results Patients’ characteristics The median age of the 119 ALCL patients at diagnosis was 12.0 years (range, 0.3 – 17.8) and 58% (69 patients) were male. The median follow-up was 6.6 years. The 3year event-free survival rate of the 119 ALCL patients was 66.4 ± 4.3% and their 3-year overall survival rate was 86.5 ± 3.1%. Detailed clinical data of the patients and controls are shown in Online Supplementary Table S1.
Pretreatment cytokine levels in patients with anaplastic lymphoma kinase-positive anaplastic large cell lymphoma Concentrations of IL-9, IL-10, IL-17a, HGF, sIL-2R, and sCD30 were significantly higher in ALCL patients at the time of diagnosis than in the patients in either control group (patients with B-cell non-Hodgkin lymphoma and ALCL patients in remission) (Figure 1). The median concentrations of all cytokines are shown in Online Supplementary Table S2. In 28 ALCL patients, measurements of sIL-2R were above the upper detection limit of 221 869.99 pg/mL. For the analyses, these samples were attributed a value of 221 869.99 pg/mL. IP-10 could not be measured in one patient. Correlations between cytokine concentrations are shown in Online Supplementary Table S3. haematologica | 2018; 103(3)
Blood cytokine concentrations in ALK-positive ALCL
Correlation of cytokine levels with clinical and laboratory characteristics Correlations between cytokine concentrations and various clinical and laboratory characteristics are summarized in Table 1.
Age and sex There was no significant difference in the median cytokine concentrations between girls and boys. Concentrations of IL17a (P=0.018) and MIP-1α (214.4 pg/mL versus 106.5 pg/mL, P=0.02) were higher in patients aged 0 – 9 years (41/119) than in patients over 9 years old. Levels of IL-2 (P=0.02) and IL-8 (32.1 pg/mL versus 79.3 pg/mL, P=0.038) were higher in patients older than 9 years (78/119).
Stage and organ involvement Patients with Murphy stage III or IV (88/116) had significantly higher concentrations of IL-6 (P=0.002), IL-10 (P=0.02), IFN-γ (P=0.013), IP-10 (P=0.009), MIG (P=0.001), VEGF (P=0.048), HGF (P=0.017), sCD30 (P<0.001), and sIL-2R (P<0.001) compared to patients with a lower stage (Online Supplementary Figure S1). Bone marrow infiltration, defined as blasts detectable in bone marrow smears, was detected in 16 of the 119 (13.4%) ALCL patients. These patients had significantly higher concentrations of IL-6 (30.1 pg/mL versus 0 pg/mL, P=0.011), IL-10 (281.2 pg/mL versus 0 pg/mL, P=0.003), IL-13 (7/16 versus 18/103 elevated, P=0.013), TNF-α (8/16 versus 20/103 elevated, P=0.005), IFN-γ (8.2 pg/mL versus
Figure 1. Cytokine levels in patients with anaplastic lymphoma kinase-positive anaplastic large cell lymphoma and controls. Logarithmic representation of cytokine levels in pg/mL for 119 ALCL patients at the time of diagnosis, 15 ALCL patients in remission and 11 patients with B-cell non-Hodgkin lymphoma (B-NHL).
haematologica | 2018; 103(3)
479
480 0.0
0.0
0.0 0.0
0.0
P
n.s.
0.013 0.0
0.0 0.0
0.0
P
n.s.
n.s. 0.0
0.0 0.0
0.0
P
n.s.
0.043
n.s.
n.s.
MIP1-β 96.3
40.0
0.0
IL-8
G-CSF
1113.5 887.5
VEGF
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.018
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.0
n.s.
79.3 0.038
535.5
96.7
106.5 0.020
138.9
336.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.2
0.0
0.0
0.0
904.5
1032.5 n.s.
2751.1 2488.7 n.s.
6.7
32.1
450.9
81.2
214.4
137.3
276.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
5.1
7
0.0
0.0
719.2
2000.9
0.0
40.0
481.7
86.3
134.5
118.3
276.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0 n.s.
n.s.
n.s.
n.s.
0.024
0.0
100.3
744.8
104.9
150.4
314.9
404.5
0.0
0.0
0.0
0.0
n.s.
n.s.
n.s.
n.s.
n.s.
0.025
n.s.
0.040
n.s.
n.s.
n.s.
65.9 <0.001
0.0
0.0
280.0 <0.001
6.0
74.1 <0.001
0.0
0.0
38.0
50.9
450.1
90.2
179.2
115.5
272.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
8.2
0.0
0.0
0.0
1254.1 1639.7 0.006
n.s.
n.s.
n.s.
0.035
n.s.
n.s.
n.s.
n.s.
0.008
0.046
0.04
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
1213.6
n.s.
3086.7 0.002
0.0
45.0
533.5
89.2
116.7
187.2
342.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
37.4 <0.001
4.5
14.9 <0.001
0.0
0.0
0.0
41.6
530.9
85.7
158.3
153.7
276.6
0.0
0.0
0.0
0.0
69.3
0.0
0.0
10.3
5.1
14.9
0.0
0.0
1278.0
81300.3 79788.9
873.1
2184.7 3636.6
0.0
50.9
487.5
94.8
126.1
136.9
341.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.2
0.0
0.0
0.0
n.s.
n.s.
0.003
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
<0.001
n.s.
n.s.
n.s.
n.s.
0.010
n.s.
0.0
0.0
P
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
1213.6
2811.8
0.0
64.9
511.8
96.4
148.0
251.7
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.003
377.2 <0.001
0.0
0.0
0.0
0.0
0.0
0.0
0.0
13.2
5.1
11.0
0.0
0.0
53811.2 125200.9 0.008
816.0
2675.9
0.0
38.7
501.4
84.2
136.9
117.7
230.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.2
0.0
0.0
0.0
30878.7 130878.5 <0.001
0.0
0.0
LDH ≤ 240 > 240 N = 62 N = 57
0.0
0.0
P
n.s.
n.s. 0.0
0.0 0.0
0.0
BM involvement Negative Positive N= 103 N = 16
1239.0
3111.9
0.0
43.6
551.9
85.4
134.4
183.8
333.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
6.6
2.2
15.2
0.0
0.0
0.0
49.5
499.6
90.7
148.0
118.3
276.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
4.5
0.0
0.0
0.0
n.s.
1085.4
0.006 2678.2
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.002
n.s.
n.s.
0.044
n.s.
<0.001
n.s.
n.s.
732.3
4789.9
0.0
49.3
548.7
75.7
98.9
343.9
441.0
8.2
3.7
0.0
0.0
26.2
0.0
0.0
236.5
2.6
30.1
0.0
0.0
46643.1 123590.3 <0.001 84067.9 75992.3
790.2
2159.4
1.1
64.8
460.9
98.6
139.3
118.0
278.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
16.5
0.0
0.0
0.0
25329.4 125883.3 <0.001 82290.2 190649.2
0.0
0.0
CRP ≤ 4 mg/L > 4 mg/L N = 55 N = 64 P
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.002
0.006
0.002
0.005
n.s.
n.s.
0.044
0.013
n.s.
0.003
n.s.
0.011
n.s.
n.s.
0.01
n.s.
n.s. 0.0
0.0
P
n.s.
n.s.
0.0
0.0
0.0
0.0
n.s.
n.s.
Histology n. common common N = 37 N= 50 P
0.0
n.s.
0.013
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.020
n.s.
0.002
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.009
1146.1 0.048
2955.0 0.017
0.0
43.6
535.5
93.6
125.6
183.8
344.2 <0.001
0.0
0.0
0.0
0.0
0.0
0.0
0.0
6.2
2.2
13.7
0.0
0.0
0.0
n.s. n.s.
n.s. n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
0.6
n.s.
94.8 n.s.
466.4 n.s.
94.1 n.s.
132.8 n.s.
193.2 n.s.
339.3 n.s.
0.0
0.0
0.0
0.0
0.0 0.018
0.0
0.0
0.0
32.0 n.s.
0.0 0.047
0.0
933.2 1031.2 n.s.
2685.2 3021.0 n.s.
0.0
40.6
583.7
96.3
201.7
130.3
296.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
13.2
0.0
11771.9 107003.4 <0.001 125200.9 78485.7 n.s.
691.6
1677.0
48.1
45.5
367.4
83.5
214.2
102.9
187.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
8.2
0.0
0.0
0.0
17857.6 118455.5 <0.001 85557.6 89737.8 n.s.
0.0
0.0
Stage I + II III + IV N= 28 N= 88
Median values of cytokine levels in pg/mL according to clinical parameters of 119 ALCL patients at the time of diagnosis. P values are given when significant differences were found. General condition was grouped as: 1 – no or slight impairment; 2 – moderate or strong impairment; 3 – bedridden and severely ill. Bone marrow (BM) involvement was considered positive if blasts were seen on smears, irrespective of cell numbers. IP-10 could not be measured in one patient. LDH: lactate dehydrogenase, CRP: C-reactive protein.
35553.0 121979.7 0.002
803.9
3146.6 10545.8 <0.001 2013.3
0.0
68.9
557.3
95.1
135.8
131.9
338.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
13.2
8.0
17.1
0.0
0.0
sCD30 63697.1105707.4 n.s. 79788.9 81300.3 n.s. 63367.9 140824.1 75364.2 n.s.
n.s.
2751.1 2488.7 n.s.
n.s.
n.s.
n.s.
HGF
0.0
71.1
514.0 501.4
MCP1
83.5
n.s.
130.3 139.0
MIP1-α 164.0 121.2
n.s.
n.s.
IP10
0.0
n.s.
0.0
0.0
n.s.
n.s.
286.1 330.6
0.0
TNF
0.0
0.0
MIG
0.0
IL-23
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
IFN
0.0
IL-22
0.0
0.0
0.0
0.0
IL-13
IL-17A
0.0
0.0
0.0
IL-10
IL-12p70 0.0
0.0
6.0
0.0
4.5
IL-6
IL-9
0.0
0.0
0.0
0.020
0.0
IL-5
0.0
P
n.s. n.s.
0.0
0.0
Leukocytosis ≤ 10/nL > 10/nL N = 72 N =47
n.s.
n.s.
0.0
B symptoms no yes N = 56 N = 63
0.0
0.0
P
n.s.
General condition 1 2 3 N = 81 N = 24 N= 14
IL-4
0.0
IL-2
0.0
Age ≤9 >9 N = 41 N = 78
sIL-2R 94163.8 91444.0 n.s. 82570.5 102780.3 n.s. 63245.2 121650.4 221870.0 <0.001 28242.1 145441.8 <0.001 95534.4 85983.0
0.0
IL-1
Sex Male Female N = 69 N = 50
Table 1. Cytokine concentrations and clinical characteristics.
F. Knörr et al.
haematologica | 2018; 103(3)
Blood cytokine concentrations in ALK-positive ALCL
0 pg/mL, P=0.002), IP-10 (343.9 pg/mL versus 118.3 pg/mL, P=0.002), MIG (441.0 pg/mL versus 276.4 pg/mL, P=0.006), and sIL-2R (190649.2 pg/mL versus 82290.3 pg/mL, P=0.01) compared to patients without bone marrow involvement. The central nervous system was involved in two of the 116 patients who had a lumbar puncture. Both patients had high concentrations of IL-6 (170.9 pg/mL, 513.5 pg/mL) and IL-17a (2623.3 pg/mL, 1468.3 pg/mL). Correlations of cytokine concentrations with other organ involvement (skin, mediastinum, liver, lung and spleen) are shown in Online Supplementary Table S4. Patients in a poorer general condition (24/119) at the time of diagnosis (4 or 5 on a subjective scale from 1 to 5) had significantly higher concentrations of IL-1β (P=0.013), IL-6 (P<0.001), IL-10 (P<0.001), IL-12p70 (P=0.024), IL-17a (P<0.001), HGF (P<0.001), sIL-2R (P<0.001), VEGF (P=0.006), IFN-γ (P=0.039), and IP-10 (P=0.025) (Online Supplementary Figure S2).
B symptoms, leukocytosis, C-reactive protein, and lactate dehydrogenase levels In comparison to patients without B symptoms, patients with B symptoms (63/119) had significantly higher concentrations of IL-6 (P=0.001), IL-10 (P<0.001), IL 12p70 (P=0.024), IFN-γ (P=0.04), MIG (P=0.046), HGF (P=0.002), sIL-2R (P<0.001), sCD30 (P=0.002), and IP-10
(P=0.008), and lower median concentrations of G-CSF (P=0.035). Likewise, patients with CRP values above 4 mg/L (64/119) had significantly higher concentrations of sIL-2R (P<0.001), IL-6 (P<0.001), IL-10 (P=0.044), IL-17a (P=0.002), HGF (P=0.006), and sCD30 (P<0.001) than patients with a lower CRP. Patients with lactate dehydrogenase values above 240 U/L (57/119) had significantly higher concentrations levels of sIL-2R (P<0.001), IP-10 (P=0.003), MIG (P<0.001), and sCD30 (P=0.008). Patients with leukocytosis (white blood cells: >10 x 109/L, 47/119) had significantly higher concentrations of IL-1β (P=0.043), IL-17a (P<0.001), IL-6 (P=0.01), and HGF (P=0.003).
Histological subtype Histological subtype was analyzed in 87 ALCL patients. The 37 patients with non-common histological subtype had significantly higher concentrations of IL-6 (13.2 pg/mL versus 0 pg/mL, P=0.047) and IL-17a (P=0.018) as compared to patients with a common histology.
Correlations of cytokine levels with biological characteristics Minimal disseminated disease Minimal disseminated disease, defined as a positive polymerase chain reaction for NPM-ALK transcripts in bone marrow or peripheral blood, was detected at diagno-
Figure 2. Cytokine levels correlated with minimal disseminated disease. Logarithmic representation of cytokine levels in pg/mL of ALK-positive ALCL patients at time of diagnosis according to the detection of minimal disseminated disease (positive n=55, negative n=46). G-CSF levels were higher in patients without minimal disseminated disease. Only cytokines for which significant differences were found are shown.
haematologica | 2018; 103(3)
481
F. Knörr et al.
sis in 55 of 101 evaluable patients. The presence of minimal disseminated disease, an independent prognostic factor in ALK-positive ALCL, was associated with elevated concentrations of IL-6 (20.7 pg/mL versus 0 pg/mL, P=0.001), IL-10 (37.4 pg/mL versus 0 pg/mL, P<0.001), IL17a (P=0.031), MCP-1 (588.8 pg/mL versus 383.8 pg/mL, P=0.008), HGF (3877.2 pg/mL versus 2101.1 pg/mL, P=0.002), IP-10 (272.94 pg/mL versus 108.3 pg/mL, P<0.001), sCD30 (137 449.9 pg/mL versus 51 781.1 pg/mL, P=0.001), and sIL-2R (159 428.4 pg/mL versus 30 087.6 pg/mL, P<0.001) (Figure 2). G-CSF (48.1 pg/mL versus 0 pg/mL, P=0.041) and MIP-1α concentrations (214.2 pg/mL versus 106.6 pg/mL, P=0.042) were significantly higher in patients without minimal disseminated disease. Using quantitative polymerase chain reaction, 29 of 90 patients were found to have a normalized copy number >10, where the number of copies of NPM-ALK is normalized to 10 000 copies of ABL. Concentrations of sIL-2R (P<0.001), IL-6 (P=0.004), IL-10 (P<0.001), IFN-γ (P<0.001), MIG (P=0.041) and IP-10 (P=0.003) were higher among these patients than among patients with lower normalized copy numbers.
Anti-anaplastic lymphoma kinase-antibody titers Anti-ALK antibody titers inversely correlate with the risk of relapse and may serve as a surrogate for the strength of the ALK-specific immune response.19,20 The
patients were grouped according to the strength of the antibody titer into those with low (≤1/750), intermediate (>1/750 – <1/60750) and high (≥1/60750) titers. There were 34, 53, and 32 patients in the low, intermediate and high titer groups, respectively. Patients with a low titer had significantly higher median concentrations of sIL-2R (P=0.013), IL-6 (P<0.001), IFN-γ (P=0.022), and IP-10 (P=0.021), but lower concentrations of IL-23 (P=0.008) compared to patients with an intermediate or high titer (Figure 3).
Correlation of cytokine levels with outcomes In univariate analysis, patients with concentrations above the median of IL-6, IL-10, IL-17a, IFN-γ, MCP-1, HGF, IP-10 and sIL-2R had a significantly lower 3-year event-free survival rate compared to patients with levels below the median (Table 2). The greatest difference in event-free survival rates was found between patients with IL-6 concentrations above the detection threshold and patients with no detectable IL-6 [event-free survival: 85.7% (95% confidence interval: 77.5 - 94.8) versus 44.6% (95% confidence interval: 33.4 - 59.8), P(log-rank)<0.001]. In a stepwise Cox regression analysis, including known risk factors and all cytokines for which findings were significant in the univariate analysis, only IL-6 retained an independent prognostic value with a hazard ratio of 2.9 ± 0.4 (Table 2).
Figure 3. Cytokine levels correlated with antianaplastic lymphoma kinase antibody titer. Median cytokine levels are shown for patients with low (≤1/750, n=34), intermediate (>1/750 – <1/60750, n=53), and high (≥1/60750, n=32) anti-ALK antibody titers. Logarithmic representation, values in pg/mL. Only cytokines for which significant differences were found are shown.
482
haematologica | 2018; 103(3)
Blood cytokine concentrations in ALK-positive ALCL
Discussion The aim of this study was to describe pretreatment serum cytokine concentrations and correlate them with clinical and biological characteristics among pediatric patients with NPM-ALK-positive ALCL. Although blood was collected at 51 different trial sites, serum or plasma was used depending on availability, the time from blood collection to freezing varied by several hours and the storage period differed considerably, the correlations found between cytokines and clinical characteristics such as IL-6 and the patients’ general condition (Table 1) are indicative of reliable measurements.
Elevated pretreatment cytokine levels in patients with ALCL compared to either post-treatment samples or other non-Hodgkin-lymphoma patients have been reported in two series of patients. Savan et al. found higher levels of IL-22 and IL-8 in nine of 11 untreated ALCL patients compared to post-treatment controls.9 Mellgren et al. recorded higher levels of IL-6, IL-10, MIP-1α, and sIL-2R in six pediatric ALCL patients at time of diagnosis compared to the levels in children with other non-Hodgkin lymphoma.10 Our systematic analysis of serum cytokine levels confirmed and extended these findings in a large group of children with untreated NPM-ALK-positive ALCL. IL-9, IL-10, IL-17a, HGF, sIL-2R, and sCD30 levels form a kind
Table 2. Cytokines in univariate and multivariate analyses.
Univariate Patients Minimal disseminated disease Positive Negative AntiALK antibody titer ≤ 1/750 > 1/750 Clinical risk (HR) Standard risk High risk B symptoms Positive Negative Histology Non-common Common sIL-2R > median ≤ median IL-6 > 0 pg/mL = 0 pg/mL IL-10 > 0 pg/mL = 0 pg/mL IL-17a > 0 pg/mL = 0 pg/mL IFN-γ > 0 pg/mL = 0 pg/mL IP-10 > median ≤ median MCP1 > median ≤ median HGF > median ≤ median sCD30 > median ≤ median
101 55 46 119 34 85 119 42 77 119 63 56 87 37 50 119 59 60 119 56 63 119 53 66 119 43 76 119 29 90 118 59 59 119 59 60 119 59 60 119 60 59
Events
Multivariate
HR
P
HR
P
6.0 ± 0.4
<0.001
6.6 ± 0.5
< 0.001
3.7 ± 0.3
< 0.001
3.6 ± 0.4
< 0.001
6.5 ± 0.5
< 0.001
4.5 ± 0.5
0.005
2.3 ± 0.3
0.01
2.2 ± 0.4
0.038
2.7 ± 0.3
0.002
5.0 ± 0.4
< 0.001
2.9 ± 0.4
0.007
2.9 ± 0.3
0.001
2.5 ± 0.3
0.004
3.5 ± 0.3
< 0.001
2.3 ± 0.3
0.008
2.6 ± 0.3
0.003
2.5 ± 0.3
0.004
2.0 ± 0.3
0.028
31 6 21 21 5 37 29 13 17 13 28 14 32 10 27 15 22 20 18 24 27 15 29 13 28 14 27 15
Stepwise regression was used to test whether cytokines have additional prognostic value having taken into account the known risk factors minimal disseminated disease, antiALK-antibody titer and clinical risk group.37 HR, hazard ratio.
haematologica | 2018; 103(3)
483
F. Knörr et al.
of cytokine-signature for ALK-positive ALCL when compared with those of both remission samples and samples from age-matched children with low-stage B-cell nonHodgkin lymphoma as separate controls. The concentrations of sIL-2R and sCD30 were expectedly higher in ALCL patients than in controls since ALCL cells, by definition, express CD30 and show strong staining for CD25, the α-subunit of the IL-2 receptor.12-14,21 Both molecules can be shed by the tumor cells.12,22 The detection of IL-9 would be in accordance with the previously described autocrine IL-9/JAK3 signaling in ALCL.15 ALCL cells have been described to resemble a Th17 phenotype and to produce IL-17.9,23 Cumulatively, these data suggest that these elevated serum cytokines might be produced by the lymphoma. Within the cohort of patients with ALK-positive ALCL, high levels of IL-6, IFN-γ, IP-10, and sIL-2R correlated with high stage, initial poor general condition, minimal disseminated disease, low ALK-antibody titers, and lower event-free survival at 3 years. The concentrations of sIL-2R and IL-6 correlate with the extent of disease, relapse risk and survival in different tumor types including Hodgkin lymphoma and peripheral T-cell lymphoma.24-27 The levels of sIL-2R, sCD30 and IL-6 have been described as independent prognostic markers in both Hodgkin lymphoma patients and patients with peripheral T-cell lymphoma.8,27,28 Several strong independent biological prognostic parameters are available in patients with ALK-positive ALCL.29 It is not, therefore, unexpected that only IL-6 retained an independent prognostic value for event-free survival in our cohort of ALKpositive ALCL patients in a multivariate analysis including the established risk factors, minimal disseminated disease and anti-ALK antibody titers.18-20 sIL-2R was described as a marker of disease activity in a cohort of nine ALK-negative and ALK-positive ALCL patients evaluated at different time points.12 As for sIL-2R, higher levels of sCD30 were associated with higher stage, presence of minimal disseminated disease and other clinical characteristics in our cohort. These findings support a role of sCD30 and sIL-2R as markers of tumor burden.25 The cumulative observations that IL-23 levels correlated directly with the anti-ALK antibody titers in our study and that this cytokine has been shown to be produced by activated dendritic cells,30 is involved in Th17 effector functions31 and has a role in autoimmunity32 might suggest that
References 1. Greer JP, Kinney MC, Collins RD, et al. Clinical features of 31 patients with Ki-1 anaplastic large-cell lymphoma. J Clin Oncol. 1991;9(4):539-547. 2. Pulford K, Falini B, Banham AH, et al. Immune response to the ALK oncogenic tyrosine kinase in patients with anaplastic large-cell lymphoma. Blood. 2000;96(4): 1605-1607. 3. Passoni L, Scardino A, Bertazzoli C, et al. ALK as a novel lymphoma-associated tumor antigen: identification of 2 HLAA2.1-restricted CD8+ T-cell epitopes. Blood. 2002;99(6):2100-2106.
484
IL-23 could support the production of autoantibodies. Elevated concentrations of IL-10 were correlated with minimal disseminated disease positivity, disease stage and significantly lower event-free survival at 3 years in univariate analyses and could hint toward an immune evasion of the tumor. ALK-positive ALCL express PD-L1,33 involved in suppression of the immune response, and IL10-secretion in ALK-positive ALCL is induced via STAT3 signaling.34 Elevated concentrations of IL-10 may reflect immune evasion of the tumor and suppression of cytotoxic T-cell functions. We also investigated whether a Th-subset-specific serum cytokine pattern could be identified in ALCL patients. Although some patients showed a pattern of elevated IFN-γ, IP-10 and MIG (these latter two both produced upon stimulation with IFN-γ35) and levels of IL-17 and IL-23 might hint towards the activation of Th17 cells, the majority of ALCL patients did not show a conclusive pattern. The concept of a certain Th response linked to a disease has been questioned by the discovery of a plethora of newly described subsets and the plasticity of those cell types.36 In addition, a multitude of host factors, tumor dissemination and individual tumor characteristics could influence the cytokine expression pattern. In summary, our findings suggest that expression of IL9, IL-10, IL-17a, HGF, sIL-2R, and sCD30 form a cytokine signature typical of ALK-positive ALCL. The levels of IL6, IFN-γ, IP-10, and sIL-2R correlated with lymphoma dissemination, other poor prognostic factors and the risk of relapse among pediatric patients with ALK-positive ALCL. Our data underline the role of immune mediators in explaining part of the typical clinical presentation of ALCL patients with B symptoms and further signs of systemic inflammation. IL-6, as a classical cytokine marker of inflammation, was also an independent prognostic parameter. More work is needed to elucidate the role of the cellular immune response to ALK-positive ALCL and to understand the role of mediators in the tumor microenvironment in patients. Acknowledgments This work was supported by a grant from the Deutsche José Carreras Leukämie-Stiftung (DJCLS08/09) to WW. FK, CDW and WW were additionally supported by Forschungshilfe Peiper. We wish to thank J. Schieferstein and S. Schwalm for their expert technical assistance.
4. Ait-Tahar K, Cerundolo V, Banham AH, et al. B and CTL responses to the ALK protein in patients with ALK-positive ALCL. Int J Cancer. 2006;118(3):688-695. 5. Ait-Tahar K, Barnardo MC, Pulford K. CD4 T-helper responses to the anaplastic lymphoma kinase (ALK) protein in patients with ALK-positive anaplastic large-cell lymphoma. Cancer Res. 2007;67(5):18981901. 6. Singh VK, Werner S, Hackstein H, et al. Analysis of nucleophosmin-anaplastic lymphoma kinase (NPM-ALK)-reactive CD8(+) T cell responses in children with NPMALK(+) anaplastic large cell lymphoma. Clin Exp Immunol. 2016;186(1):96-105. 7. Skinnider BF, Mak TW. The role of cytokines in classical Hodgkin lymphoma.
Blood. 2002;99(12):4283-4297. 8. Marri PR, Hodge LS, Maurer MJ, et al. Prognostic significance of pretreatment serum cytokines in classical Hodgkin lymphoma. Clin Cancer Res. 2013;19(24):68126819. 9. Savan R, McFarland AP, Reynolds DA, et al. A novel role for IL-22R1 as a driver of inflammation. Blood. 2011;117(2):575-584. 10. Mellgren K, Hedegaard CJ, Schmiegelow K, Muller K. Plasma cytokine profiles at diagnosis in pediatric patients with nonHodgkin lymphoma. J Pediatr Hematol Oncol. 2012;34(4):271-275. 11. Al-Hashmi I, Decoteau J, Gruss HJ, et al. Establishment of a cytokine-producing anaplastic large-cell lymphoma cell line containing the t(2;5) translocation: potential role
haematologica | 2018; 103(3)
Blood cytokine concentrations in ALK-positive ALCL
12.
13.
14.
15.
16.
17.
18.
19.
of cytokines in clinical manifestations. Leuk Lymphoma. 2001;40(5-6):599-611. Janik JE, Morris JC, Pittaluga S, et al. Elevated serum-soluble interleukin-2 receptor levels in patients with anaplastic large cell lymphoma. Blood. 2004;104(10):33553357. Nadali G, Vinante F, Stein H, et al. Serum levels of the soluble form of CD30 molecule as a tumor marker in CD30+ anaplastic large-cell lymphoma. J Clin Oncol. 1995;13(6):1355-1360. Zinzani PL, Pileri S, Bendandi M, et al. Clinical implications of serum levels of soluble CD30 in 70 adult anaplastic large-cell lymphoma patients. J Clin Oncol. 1998;16(4):1532-1537. Qiu L, Lai R, Lin Q, et al. Autocrine release of interleukin-9 promotes Jak3-dependent survival of ALK+ anaplastic large-cell lymphoma cells. Blood. 2006;108(7):2407-2415. Bard JD, Gelebart P, Anand M, Amin HM, Lai R. Aberrant expression of IL-22 receptor 1 and autocrine IL-22 stimulation contribute to tumorigenicity in ALK+ anaplastic large cell lymphoma. Leukemia. 2008;22(8):1595-1603. Seidemann K, Tiemann M, Schrappe M, et al. Short-pulse B-non-Hodgkin lymphomatype chemotherapy is efficacious treatment for pediatric anaplastic large cell lymphoma: a report of the Berlin-FrankfurtMunster Group Trial NHL-BFM 90. Blood. 2001;97(12):3699-3706. Damm-Welk C, Busch K, Burkhardt B, et al. Prognostic significance of circulating tumor cells in bone marrow or peripheral blood as detected by qualitative and quantitative PCR in pediatric NPM-ALK-positive anaplastic large-cell lymphoma. Blood. 2007;110(2):670-677. Ait-Tahar K, Damm-Welk C, Burkhardt B, et al. Correlation of the autoantibody response to the ALK oncoantigen in pediatric anaplastic lymphoma kinase-positive anaplastic large cell lymphoma with tumor dissemination and relapse risk. Blood.
haematologica | 2018; 103(3)
2010;115(16):3314-3319. 20. Mussolin L, Damm-Welk C, Pillon M, et al. Use of minimal disseminated disease and immunity to NPM-ALK antigen to stratify ALK-positive ALCL patients with different prognosis. Leukemia. 2013;27(2):416-422. 21. Josimovic-Alasevic O, Durkop H, Schwarting R, Backe E, Stein H, Diamantstein T. Ki-1 (CD30) antigen is released by Ki-1-positive tumor cells in vitro and in vivo. I. Partial characterization of soluble Ki-1 antigen and detection of the antigen in cell culture supernatants and in serum by an enzyme-linked immunosorbent assay. Eur J Immunol. 1989;19(1):157-162. 22. Miles RR, Cairo MS, Satwani P, et al. Immunophenotypic identification of possible therapeutic targets in paediatric nonHodgkin lymphomas: a Children's Oncology Group report. Br J Haematol. 2007;138 (4):506-512. 23. Matsuyama H, Suzuki HI, Nishimori H, et al. miR-135b mediates NPM-ALK-driven oncogenicity and renders IL-17-producing immunophenotype to anaplastic large cell lymphoma. Blood. 2011;118(26):68816892. 24. Rutkowski P, Kaminska J, Kowalska M, Ruka W, Steffen J. Cytokine serum levels in soft tissue sarcoma patients: correlations with clinico-pathological features and prognosis. Int J Cancer. 2002;100(4):463-471. 25. Bien E, Balcerska A. Serum soluble interleukin 2 receptor alpha in human cancer of adults and children: a review. Biomarkers. 2008;13(1):1-26. 26. Lippitz BE. Cytokine patterns in patients with cancer: a systematic review. Lancet Oncol. 2013;14(6):e218-228. 27. Gupta M, Stenson M, O'Byrne M, et al. Comprehensive serum cytokine analysis identifies IL-1RA and soluble IL-2Ralpha as predictors of event-free survival in T-cell lymphoma. Ann Oncol. 2016;27(1):165172. 28. Visco C, Nadali G, Vassilakopoulos TP, et al. Very high levels of soluble CD30 recog-
29.
30.
31.
32. 33.
34.
35.
36.
37.
nize the patients with classical Hodgkin's lymphoma retaining a very poor prognosis. Eur J Haematol. 2006;77(5):387-394. Damm-Welk C, Pillon M, Woessmann W, Mussolin L. Prognostic factors in paediatric anaplastic large cell lymphoma: role of ALK. Front Biosci (Schol Ed). 2015;7:205-216. Oppmann B, Lesley R, Blom B, et al. Novel p19 protein engages IL-12p40 to form a cytokine, IL-23, with biological activities similar as well as distinct from IL-12. Immunity. 2000;13(5):715-725. Aggarwal S, Ghilardi N, Xie MH, de Sauvage FJ, Gurney AL. Interleukin-23 promotes a distinct CD4 T cell activation state characterized by the production of interleukin-17. J Biol Chem. 2003;278(3):19101914. Croxford AL, Mair F, Becher B. IL-23: one cytokine in control of autoimmunity. Eur J Immunol. 2012;42(9):2263-2273. Andorsky DJ, Yamada RE, Said J, Pinkus GS, Betting DJ, Timmerman JM. Programmed death ligand 1 is expressed by non-Hodgkin lymphomas and inhibits the activity of tumor-associated T cells. Clin Cancer Res. 2011;17(13):4232-4244. Kasprzycka M, Marzec M, Liu X, Zhang Q, Wasik MA. Nucleophosmin/anaplastic lymphoma kinase (NPM/ALK) oncoprotein induces the T regulatory cell phenotype by activating STAT3. Proc Natl Acad Sci USA. 2006;103(26):9964-9969. Farber JM. A macrophage mRNA selectively induced by gamma-interferon encodes a member of the platelet factor 4 family of cytokines. Proc Natl Acad Sci USA. 1990;87(14):5238-5242. Nakayamada S, Takahashi H, Kanno Y, O'Shea JJ. Helper T cell diversity and plasticity. Curr Opin Immunol. 2012;24(3):297302. Le Deley MC, Reiter A, Williams D, et al. Prognostic factors in childhood anaplastic large cell lymphoma: results of a large European intergroup study. Blood. 2008;111(3):1560-1566.
485
ARTICLE
Non-Hodgkin Lymphoma
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):486-496
Clinicopathological characteristics of T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies: a single-center case series of 25 patients and a review of the literature
Marieke L. Nijland,1# Lianne Koens,2# Steven T. Pals,3 Ineke J.M. ten Berge,1 Frederike J. Bemelman1 and Marie JosĂŠ Kersten4*
Renal Transplant Unit, Department of Nephrology, Academic Medical Center; Department of Pathology, Academic Medical Center; 3Department of Pathology and Lymphoma and Myeloma Center Amsterdam (LYMMCARE), Academic Medical Center and 4Department of Hematology and Lymphoma and Myeloma Center Amsterdam (LYMMCARE), Academic Medical Center, Amsterdam, the Netherlands 1 2
#
MLN and LK contributed equally to this work.
ABSTRACT
A
Correspondence: m.j.kersten@amc.uva.nl
Received: May 3, 2017. Accepted: December 13, 2017. Pre-published: December 21, 2017. doi:10.3324/haematol.2017.169987 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/486
lthough it is known that B-cell lymphomas occur more frequently in immunocompromised patients, thus far such an association has not been clearly established for T-cell lymphomas. Of the 251 patients who were diagnosed with a T-cell non-Hodgkin lymphoma in our center between 1999 and 2014, at least 25 were identified in immunocompromised patients. Herein, we retrospectively analyzed the clinical and pathological characteristics of these 25 cases. In addition, we searched the literature and present an overview of 605 previously published cases. The actual number of patients with B-cell chronic lymphocytic leukemia and patients on immunosuppressive drugs for inflammatory bowel disease or rheumatoid arthritis in the total cohort of 251 patients diagnosed with T-cell non-Hodgkin lymphoma was much higher than the number of patients expected to have these diseases in this cohort, based on their prevalence in the general population. This, together with the large number of additional cases found in the literature, suggest that the risk of developing T-cell non-Hodgkin lymphoma is increased in immunocompromised patients. Compared to T-cell nonHodgkin lymphoma in the general population, these lymphomas are more often located extranodally, present at a younger age and appear to have a poor outcome. The observations made in the study herein should raise awareness of the possible development of T-cell non-Hodgkin lymphoma in immunodeficient patients, and challenge the prolonged use of immunosuppressive drugs in patients who are in clinical remission of their autoimmune disease.
Introduction Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
486
It has long been recognized that patients with either a primary or acquired immunodeficiency are at increased risk for the development of malignant lymphomas.1,2 Hematopoietic stem cell and solid organ transplant recipients, for example, can develop post-transplant lymphoproliferative disease (PTLD);3 patients infected with the human immunodeficiency virus (HIV), patients with primary immunodeficiencies and patients treated for inflammatory bowel disease (IBD) with immunosuppressive drugs all have an increased risk for developing lymphoma.4-8 Moreover, this complication is seen in patients with autoimmune diseases like rheumatoid arthritis (RA), primary SjĂśgren syndrome and systemic lupus erythematosus. However, it is not clear whether the lymphomas in these patients are triggered by chronic inflammation caused by the disease itself or by the (immunosuppressive) therapies used.9-13 In all the groups studied, the reported lymphomas are predominantly of B-cell origin.3,5,11 haematologica | 2018; 103(3)
T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies
Table 1. Patient characteristics.
Primary disorder AI diseases
Sex
Age Disorder at Dx defined CD
Location lymphoma
Histologic subtype
Staging*
Prior IS drugs or chemotherapy
Period given
Interval start IST- Dx (m)
Kidney
Precursor T-LBL
IE
Prednisolone Azathioprine Adalimumab Azathioprine Azathioprine Prednisolone Azathioprine Prednisolone Azathioprine Infliximab Adalimumab Azathioprine Methotrexate Etanercept Methotrexate Prednisolone Prednisolone NR Prednisolone
NR >2y 1-2 y NR >2y NR >2y NR >2y 1-2 y >2y >2y >2y 1-2 y 6-12 m >2y 1-2 y
NR NR NR NR 44 NR 210 NR NR NR NR 173 NR 18 6 60 13
>2y
NR
Chlorambucil NR Chlorambucil Prednisolon Cyclophosphamide Rituximab PI3K-inhibitor Cyclophosphamide Prednisolone CT, multiple courses
NR
92
NR NR NR NR 1-2 y 6-12 m NR NR
NR NR NR NR 21 10 NR 143
Cyclophosphamide Rituximab Fludarabine Peginterferon Chlorambucil MMF Prednisolone Tacrolimus Chlorambucil Cyclophosphamide Prednisolone
3-6 m 3-6 m 3-6 m 6-12 m >2y <3 m 3-6 m 3-6 m NR NR NR
13 13 13 NR 165 5 5 5 NR NR NR
Prednisolone MS Azathioprine Tacrolimus Tacrolimus Prednisolone Azathioprine
>2y >2y >2y >2y >2y >2y >2y
367 367 367 367 95 315 315
1 male
23
2 female 3 male 4 male
20 CD and PSC Skin Primary C-CD30+ T-LPD 46 CD Liver HSTCL 21 CD Spleen, BM, liver PTCL, NOS
IA IV IV
5 male 6 male
39 43
III IV
7 female 8 female 9 male 10 female 11 male 12 male 13 male 14 male
36 40 52 75 70 54 43 46
CD CD
LN Precursor T-LBL Small bowel, PTCL, NOS kidneys, pancreas, heart
CD LN, pelvic cavity, skin ALCL, ALKRA Nasopharynx ENKL, nasal type RA Skin Mycosis Fungoides RA LN, spleen, liver AITL PMR LN, spleen, bone AITL M. SjĂśgren Skin Mycosis Fungoides Sarcoidosis LN, liver AITL PSC LN ALCL, ALK+ (small cell variant) 81 B-CLL PB T-PLL 62 B-CLL Skin Primary C-CD30+ T-LPD 78 B-CLL, AIHA LN, PTCL, NOS and MN spleen, BM
II II IA IV IV IA III II
4 male
59
PTCL, NOS
IV
5 male
75
Mantle cell lymphoma
ALCL, ALK-
IV
6 male Organ 1 male transplantation
75 70 MN
B-CLL Kidney Tx,
2 male
71
Kidney Tx
3 female 4 male
46 78
Liver Tx LN, bone marrow EBV+ PTCL, NOS Kidney tx Small bowel T-NHL, unclassifiable
1 male
26
Hematologic malignancy
HIV
1 male 2 male 3 male
Multiple BM, spleen, liver myeloma and ASCT
HIV
LN, lung
Spleen PTCL, NOS Medulla Clonal T-cell oblongata, population in liquor, pons CNS lymphoma at imaging
Small bowel
LN, lung
ALCL, ALK-
ALCL, ALK-
IV IA IV
III IV
IIIE
IV IV
II
None
*For cutaneous lymphomas the tumor-node-metastasis-blood (TNMB) system was used. All other lymphomas were staged by the Ann Arbor staging system. AI: autoimmune; AIHA: autoimmune hemolytic anemia; AITL: angioimmunoblastic T-cell lymphoma; ALCL: T-cell anaplastic large cell lymphoma; ALK: anaplastic lymphoma kinase; ASCT: autologous stem cell transplant; B-CLL: B-cell chronic lymphatic leukemia; BM: bone marrow; CT: chemotherapy; CD: Crohnâ&#x20AC;&#x2122;s disease; Dx: diagnosis; EBV: Epstein-Barr virus; ENKL: extranodal NK/T-cell lymphoma; HIV: human immunodeficiency virus; HSTCL: hepatosplenic T-cell lymphoma; IQR: interquartile range; IS(T): immunosuppressive (therapy); LN: lymph nodes; MMF: mycophenolate mofetil; MS: mycophenolate sodium; MN: membranous nephropathy; NR: not recorded/retraceable; PB: peripheral blood; PMR: polymyalgia rheumatica; Primary C-CD30+ T-LPD: primary cutaneous CD30+ T-cell lymphoproliferative disease; PSC: primary sclerosing cholangitis; PTCL, NOS: peripheral T-cell lymphoma, not otherwise specified; RA: rheumatoid arthritis; T-LBL: T-cell lymphoblastic lymphoma; T-PLL: T-cell prolymphocytic leukemia; T-NHL: T-cell non-Hodgkin lymphoma; Tx: transplantation.
haematologica | 2018; 103(3)
487
M.L. Nijland et al.
Much less is known about the development of T-cell non-Hodgkin lymphomas (T-NHL) in patients with immunodeficiencies or autoimmune diseases. For HIV patients and solid organ transplant recipients some large
case series and reviews on T-NHL have been published.1420 In IBD patients, the development of a specific and rare subtype of T-NHL, hepatosplenic T-NHL (HSTCL), has been associated with the use of thiopurines, either alone
Table 2. Patient and lymphoma characteristics, ordered by primary disorder.
Disorder
AI disease
Hematologic malignancy Solid organ Tx
HIV
Total
Number of cases
Total: 14 CD: 7 (50%) RA: 3 (21.4%) Sarcoidosis: 1 (7.1%) PR: 1 (7.1 %) M. Sjogren: 1 (7.1%) PSC: 1 (7.1%) 43 (33-53)
Total: 6 B-CLL: 4 (66.7%) MCL: 1 (16.7%) MM: 1 (16.7%)
Total: 4 Kidney: 3 (75%) Liver: 1 (25%)
Total: 1
25
75 (61-79)
71 (52-76)
26
52 (40-73)
10 (71.4%) 4 (28.6%) IST: 13 (92.9%) NR: 1 (7.1%)
6 (100%) CT: 5 (83.3%) NR:1 (16.7%)
3 (75%) 1 (25%) IS: 4 (100%)
1 (100%) No IS: 1(100%)
44 (13-173)
92 (17-154)
205 (27.5-354)
-
20 (80%) 5 (20%) IST: 17 (68%) CT: 5 (20%) No IST: 1 (4%) NR: 2 (8 76 (14,3-171)
3-6 m:6-12 m:1 (7.1%) 1-2 y: 3 (21.4%) >2 y: 8 (57.1%) NR: 2 (14.3%) CTCL: 3 (21.4%) AITL: 3 (21.4%) PTCL, NOS: 2 (14.3%) Prec T-LBL: 2 (14.3%) ALCL, ALK+ : 1 (7.1%) ENKL: 1 (7.1%) HSTCL: 1 (7.1%) ALCL, ALK-: 1 (7.1%)
3-6 m: 6-12 m: 2 (33.3%) 1-2 y: 1 (16.7%) >2 y: 1 (16.7%) NR: 2 (33.3%) PTCL, NOS: 3 (50%) CTCL: 1 (16.7%) ALCL, ALK-: 1 (16.7%) T-PLL : 1 (16.7%
3-6 m: 1 (25%) 6-12 m: 1-2 y: >2 y: 3 (75%) NR: PTCL, NOS: 1 (25%) ALCL, ALK-: 1 (100 %) T-NHL, unclass: 1 (25 %) ALCL, ALK-: 1 (25%) Clonal T-cell population in liquor: 1 (25%)
LN: 5 (35.7%) EN: 13 (92.9%) Exclusive EN: 9 (64.2%)
LN: 3 (50%) EN: 5 (83.3%) Exclusive EN: 3 (50%)
LN: 1 (25%) EN: 4 (100%) Exclusive EN: 3 (75%)
Median age Dx (y [IQR]) Sex Male Female IS therapy or chemotherapy
Median interval start drugs â&#x20AC;&#x201C; Dx (m [IQR]) Duration of IST or CT (treatment longest given)
Histologic subtype
Location
LN: EN: 1 (100%) Exclusive EN: 1 (100%)
3-6 m: 1 (4%) 6-12 m: 3 (12%) 1-2 y: 4 (16%) >2 y: 12 (48%) NR: 4 (16%) PTCL, NOS: 6 (24%) ALCL, ALK-: 4 (16%) AITL:3 (12%) CTCL: 4 (16%) Prec T-LBL: 2 (8%) ALCL, ALK+ : 1 (4%) ENKL: 1 (4%) HSTCL: 1 (4%) T-PLL: 1 (4%) Clonal T cell population in liquor: 1 (4%) T-NHL, unclass: 1 (4%) LN: 9 (36%) EN: 23 (92%) Exclusive EN: 16 (64%)
continued in the next page
488
haematologica | 2018; 103(3)
T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies
or in combination with tumor necrosis factor (TNF)-α inhibitors.21,22 Even less is known about the development of T-NHL in patients with other immunodeficiencies, as only case reports and some small case series have been published. Herein, we present a relatively large series of 25 immunodeficient patients in whom T-NHL was diagnosed in a single referral center in the period 1999 to 2014. In this cohort study, we describe the clinical characteristics of these cases and correlate them to the pathological features of T-NHL. Furthermore, we present a review of the literature on T-NHL in immunocompromised patients. To the best of our knowledge, this is the largest series of T-NHL in patients with varying causes of immunodeficiency reported so far.
Methods Histopathological material and reports from patients treated in or referred to the Academic Medical Center in Amsterdam are stored prospectively in a database. This database was queried for samples on which a T-cell receptor (TCR) gene rearrangement analysis was performed between 1999 and 2014. In our center, analysis of TCR gene rearrangement on tumor tissue is standard practice in the workup if T-NHL is suspected. In cases where the diagnosis of T-NHL was confirmed by histology, molecular testing and clinical features, the corresponding clinical data were searched for the presence of immunodeficiency prior to the diagnosis of TNHL. For all cases, the biopsies were reviewed and immunohistochemical stains for CD2, CD3, CD4, CD5, CD8, granzyme B, PD1, CD30, ALK1, CD21, CD20, TdT, CD56 and in situ hybridization for Epstein-Barr virus (EBV)-encoded coded ribonucleic acid (RNA; EBER) were analyzed. If these were not performed in the routine diagnostic work-up (especially in older cases), they were additionally performed for this purpose. The lymphomas were (re)classified according to the World Health Organization (WHO) 2008 classification of lymphoid malignancies.23 Clinical informa-
tion for all patients was collected in an anonymized database. The survival status and the cause of death were determined at the cutoff date of April 1, 2015. For staging, the Ann Arbor system was used for all patients, with the exception of those with cutaneous lymphomas, for whom the Mycosis Fungoides Cooperative Group (MFCG) tumor-node-metastasis (TNM) staging system of cutaneous T-NHL was used.24 Statistical analyses of the data were performed using SPSS (version 23.0 for Windows). A search in PubMed was performed to find additional cases of T-NHL in patients with varying causes of immunodeficiency using the following search terms: “Immunocompromised”, “immunodeficiency”, “decreased immunity”, “reduced immunity”, “HIV”, “autoimmune disease”, “rheumatoid arthritis”, “IBD”, “inflammatory bowel disease”, “Crohn”, “ulcerative colitis”, “hematologic malignancy”, “Hodgkin”, “Waldenström”, “B-cell lymphoma”, “Bcell lymphoma”, and “leukemia”. These terms were combined using the “AND”-function with the search terms: “T-NHL”, “T-cell lymphoma”, “peripheral T-NHL”, “peripheral T-cell lymphoma” or “PTCL”. When available, Medical Subject Headings (MeSH) terms were used. In order to be included in the review herein, the cases of patients were required to have a pre-existing immunodeficiency due to HIV, immunosuppressive therapy, hematologic malignancies or a primary immunodeficiency before they developed a T-NHL. Reference lists of selected articles were used to identify additional articles. Articles published in a language other than English were excluded. Articles using earlier published cases were checked and duplicate cases were eliminated. Since more comprehensive reviews describing case series and previously published cases are available in the literature for T-NHL occurring in solid organ transplant recipients or in HIV patients, only these papers were included.
Results Patients’ characteristics A total of 251 T-NHL cases were found in our histopathological database. Forty-two cases, for which no
continued from the previuos page
Extranodal sites
Stage* I/II III/IV Median survival (m)
BM/Bone: 4 Skin: 4 Liver: 4 Nasopharynx: 1 Bowel: 1 Lung: 1 Heart: 1 Pancreas: 1 PCO:1
BM/Bone: 2 Skin: 1 Liver: 1 PB: 1 Spleen: 2 Lung: 1
7 (50%) 7 (50%) 62
1 (16.7%) 5 (83.3%) 7.2
BM/Bone: 1 Bowel: 2 Heart: 1 CZS: 1
4 (100%) 8
Lung: 1
BM/Bone: 7 (28%) Skin: 5 (20%) Liver: 5 (20%) Spleen: 3 (12%) Bowel: 3 (12%) Lung:3 (12%) Heart: 2 (8%) Pancreas: 2 (8%) PB:1 (4%) PCO: 1 (4%) CZS: 1 (4%)
1 (100%) 96
9 (36%) 16 (64%) 11.3
*For cutaneous lymphomas the TNMB system was used. All other lymphomas were staged by the Ann Arbor staging system. AI: autoimmune; AITL: angioimmunoblastic T-cell lymphoma; ALCL: anaplastic large cell lymphoma; B-CLL: B-cell chronic lymphatic leukemia; BM: bone marrow; CT: chemotherapy; CD: Crohn’s disease; Dx: diagnosis; CTCL: primary cutaneous T-cell lymphoma; EN: extranodal; ENKL: extranodal NK/T cell lymphoma; HIV: human immunodeficiency virus; HSTCL: hepatosplenic T-cell lymphoma; IS(T): immunosuppressive (therapy); LN: lymphnodes involved; MCL: mantle cell lymphoma; MM: multiple myeloma; NR: not recorded/retraceable; PB: peripheral blood; PCO: pelvic cavity organs; PR: polymyalgia rheumatica; Prec T-LBL: precursor T-cell lymphoblastic lymphoma; Prim. C-CD30+ T-LPD: primary cutaneous CD30+ T-cell lymphoproliferative disease; PSC: primary sclerosing cholangitis; PTCL, NOS: peripheral T-cell lymphoma, not otherwise specified; RA: rheumatoid arthritis; T-NHL: T-cell non-Hodgkin lymphoma; T-PLL; T-cell prolymphocytic leukemia; Tx: transplantation.
haematologica | 2018; 103(3)
489
M.L. Nijland et al.
clinical data were available, were excluded (see flowchart, Figure 1). Twenty-five of the remaining 209 cases (12%) were identified in patients with an immunodeficiency. Table 1 and Table 2 give an overview of the clinical characteristics, ordered by the underlying disorder in the latter. The majority of cases of T-NHL were found in patients with an auto-immune disease (56%). Other underlying disorders were hematologic malignancies (24%), solid organ transplantation (16%) and HIV infection (4%). Previously published case reports and case series are summarized in Online Supplementary Table S1, also ordered by the underlying disorder. For comparison, in our PubMed search 605 cases of immunodeficiency-related T-NHL were identified, of which 201 occurred in patients with HIV infection (33%), 197 in transplant recipients (33%), 143 in patients with underlying autoimmune diseases (24%), 55 following previously treated hematologic malignancies (9%) and nine in patients with a primary immunodeficiency (1%). In our series, T-NHL occurred at a median age of 52 years (interquartile range 39.5-73), which is nine to ten years younger than that observed in studies on T-NHL in the general population.25,26 Patients with Crohnâ&#x20AC;&#x2122;s disease and HIV were younger than patients with hematologic malignancies and solid organ transplantations. This is con-
sistent with the literature on HIV and PTLD, in which the mean/median ages reported were 38-39 and 43.5-57.5 respectively.14,15,17-20 In line with these studies, we found a male predominance (80%).14,15,17-20,26,27 Most patients had been treated with either immunosuppressive therapy or chemotherapy prior to the diagnosis of lymphoma (88%). For patients for whom information was available regarding the exact start date of the therapy with immunosuppressive or chemotherapeutic agents, the median interval between the start of the immunosuppressive treatment and the time of diagnosis was 76 months (interquartile range 14.3171; N=16). In all patients for whom the duration of the drug use could be deduced from the clinical records, 60% had used one or more immunosuppressive or chemotherapeutic agents for at least two years (N=20). In all 25 patients, most (60%) had used prednisolone for varying periods. Azathioprine was used by 32%, including 86% of the patients with Crohnâ&#x20AC;&#x2122;s disease; those for whom it was documented (7 out of 8) used this drug for longer than two years. Drugs somewhat less frequently used were chlorambucil and cyclophosphamide (both 16%), mostly by patients with hematologic malignancies, and tacrolimus (12%) by solid organ transplant recipients. A few patients had been treated with adalimumab, infliximab, rituximab,
Figure 1. Cases included. Flowchart of inclusion of cases of T-NHL in patients with immunodeficiencies due to an underlying disorder or immunosuppressive drugs in the period 1999-2014. AI: autoimmune; Hem. malignancy: hematologic malignancy; HIV: human immunodeficiency virus; SOT: solid organ transplantation; T-NHL: T-cell non-Hodgkin lymphoma.
490
haematologica | 2018; 103(3)
T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies
mycophenolate mofetil or sodium, a phosphatidylinositol 3-kinase (PI3K) inhibitor or pegylated interferon (PEGINF). In the cases reported in the literature, the use of thiopurines was also widespread (63% of all patients with autoimmune diseases and 86% of patients with IBD) (Table 3). TNF-α inhibitors like adalimumab, infliximab and etanercept were, in contrast to our series, the most frequently used drugs in patients with autoimmune diseases; 86% was treated a TNF-α inhibitor, often in combination with thiopurines (63% of this group). In the case reports concerning hematologic malignancies most patients were treated with a chemotherapeutic regime containing multiple agents (56%) and a few with chlorambucil only (6%) (Table 3).
Lymphoma characteristics An overview of the histological characteristics of the lymphomas of our patients and those reported in the literature is provided in Table 2 and Table 4, respectively. TNHL were morphologically and immunophenotypically highly variable. In our series, peripheral T-NHL (PTCLNOS), was seen most frequently (24%), followed by anaplastic lymphoma kinase (ALK)-negative T-cell anaplastic large cell lymphomas (ALCL) (16%). There were three cases (12%) of angioimmunoblastic T-NHL (AITL), limited to the group of patients with autoimmune diseases. We saw only one case of HSTCL. This distribution was comparable to that in the general European population.26 However, between the subtype distribution in the reported cases and that in the general population there were some differences. Most notable were the high frequency of primary cutaneous T-NHL in the immunocompromised patients (29% vs. 1.7%), a more frequent occurrence of HSTCL (12% vs. 1.4%), and the relative lack of AITL cases (1% vs. 18.5%) in this group.26 There was no predominance of either CD4+ or CD8+ lymphomas. Of the 19 cases in which CD4/CD8 staining had been carried out, seven (36.8%) were CD8+ , six (31.6%) were CD4+ , three (15.8%) were CD4+ CD8+ , and three cases (15.8%) were CD4–CD8–. EBER was performed in 23 lymphoma cases and the majority was negative (91.3%). EBER was positive in two cases (8.7%), one of which concerned extranodal natural killer (NK)/T-cell lymphoma and the other involving PTCL-NOS with EBVpositive T cells and concurrent EBV-positive B-cell blasts.
In two other cases, both AITL, the malignant T cells were negative, but the B-cell compartment was EBV-positive. Figure 2 shows an example of a PTCL-NOS in a patient with B-cell chronic lymphocytic leukemia (B-CLL) as the underlying disorder. Extranodal involvement was observed in the vast majority of patients (92%), and 16 patients (64%) showed an exclusively extranodal localization of the lymphoma. The most commonly involved organs were the bone marrow or bone, skin, liver, spleen, small bowel and lung. The heart, pancreas, peripheral blood, pelvic cavity organs and central nervous system were affected in some cases (Table 2). This rate of extranodal involvement is higher than the 65-72% which has been reported in the general, immunocompetent population,25,28 and is consistent with more frequently occurring extranodal localizations of B-cell lymphomas in patients with either primary or acquired immunodeficiencies.3,17,29-31 The majority of patients had Ann Arbor stage III/IV disease at presentation (64%). The four patients staged according to the MFCG TNM staging system of cutaneous T-cell lymphomas had stage I or II disease.
Treatment and outcome As shown in Table 5, in our series lymphoma treatment was very heterogeneous due to the different histological subtypes and clinical stages of disease. After a median follow-up of six years, 15 patients had died. The median overall survival (OS) was 11.3 months (table 2) and the 1- and 5-year survival rates were 47% (95% confidence interval (CI) 27-67%) and 31% (95% CI 7-55%), respectively. This is somewhat lower than the 5year survival rate reported in patients with T-NHL in the general population (38-49%),25,28,32 and consistent with a worse survival of HIV and post-transplant patients with TNHL.14,15,17-19 Since worse outcomes have been reported for certain histological subtypes as compared to others,25,26,28 outcomes were calculated separately for the total number of patients with PTCL-NOS, AITL and HSTCL (n=10). The median OS in this group was 7.5 months with a 5year survival rate of 20% (95% CI -6-46%), which is indeed lower than the median OS of 60 months and the 5year survival rate of 41% (95% CI 8-74%) in patients with the remaining histological subtypes. Twelve of the 15 deceased patients died within four months following diagnosis. The main causes of death, in
Table 3. Use of drugs in cases reported in the literature.
AID IBD Other AID Total HM CLL Other HM Total Other ID
Nr pts
TNF-αi
64 79 143 Nr pts 44 11 55 None 9
60 63 123 CT 20 9 29
Thiop TNF-αi + thiop 55 35 90 CA only 3 3
48 30 78 Melfalan 1 1
MTX 2 11 13 None 13 1 14
CsA 6 6 NR 8 8
Other 5 5
AID: autoimmune disease; CA: chlorambucil; CLL: chronic lymphocytic leukemia; CsA: cyclosporine A; CT: chemotherapy, multiple agents; HM: hematologic malignancies; IBD: inflammatory bowel disease; ID: immunodeficiencies ; MTX: methotrexate; Nr pts: number of patients; Thiop: thiopurines; TNF-αi: TNF-α inhibitors.
haematologica | 2018; 103(3)
491
M.L. Nijland et al. Table 4. Distribution of histologic subtypes in cases reported in the literature. ALCL ALK+ ALK â&#x20AC;&#x201C; ALK NR AITL ATLL CTCL EATL ENKL HSTCL IVL NKL NKTCL Prec T-LBL PTCL, NOS SPTCL T-LGL Other T-NHL Total
HIV
Post tx
AID
HM
Other ID
Total
52 2 27 23 2 11 48
16 4
12 2 3 7 3
3 2 1
1
34
39
4
3 47
1
1 1
5 5
10 1
2
34 143
55
9
84 10 32 42 6 17 173 1 13 73 1 6 14 8 146 9 9 45 605
1 1
10 76
201
12 6 48 1 8 23 1 6 4 8 53 3 9 11 197
1
1
AID: autoimmune disease; AITL: angioimmunoblastic T-cell lymphoma; ALCL: anaplastic large cell lymphoma, systemic; ALK: anaplastic lymphoma kinase; ATLL: adult T-cell leukemia/lymphoma; CTCL: primary cutaneous T-cell lymphoma; EATL: enteropathy-associated T-cell lymphoma; ENKL: extranodal NK/T-cell lymphoma; HM: hematologic malignancies; HSTCL: hepatosplenic T-cell lymphoma; ID: immunodeficiency; IVL: intravascular lymphoma; NKL: natural killer cell lymphoma(/leukemia); NKTCL: natural killer/T-cell lymphoma; NR: Not recorded or not known; PTCL, NOS: peripheral T-cell lymphoma, not otherwise specified; Prec T-LBL: precursor T-cell lymphoblastic lymphoma; SPTCL: subcutaneous panniculitis-like T-cell lymphoma; T-LGL: T-cell large granular lymphocytic leukemia; T-NHL: T-cell non-Hodgkin lymphoma; tx: transplantation.
those cases for which this information was available, were lymphoma progression (N=6) or sepsis (N=2). One patient developed two other malignancies and the exact cause of death remained unknown.
Discussion In the general population, T-cell neoplasms are uncommon, representing about 5 to 10% of all NHL in Western countries.26 While B-cell lymphomas are known to occur more frequently in transplant recipients and in patients with IBD, HIV or autoimmune diseases,3,5,6,11 for T-NHL this data has not been elucidated as of yet. Since T-NHL are tumors of the immune system, akin to B-cell lymphomas, it is likely that the risk of developing a T-NHL is also increased in patients with an impaired immune system. However, a higher incidence of T-NHL has only been reported for a few specific patient groups, including HIV patients,14,15 solid organ transplant recipients19 and patients with a history of coeliac disease, psoriasis or eczema.33,34 Furthermore, a higher incidence of the rare hepatosplenic T-NHL has been documented in IBD patients on thiopurines.21 The present study of 25 patients is the largest case series of T-NHL in patients with varying causes of acquired immunodeficiencies published thus far. These 25 patients, diagnosed at our center between 1999 and 2014, constitute 12% of the total number of cases of T-NHL in which clinical data were available. Three of these 25 patients developed AITL, which may be accompanied by autoim492
mune features such as arthritis and synovitis, and can therefore resemble RA.35 The autoimmune diseases in these three patients were RA, polymyalgia rheumatica and sarcoidosis, which had been diagnosed 30 years, five years and three years, respectively, prior to the diagnosis of AITL. These lengthy intervals signify that it is most unlikely that they were manifestations of the AITL itself, since arthritis occurring more than six months in advance of the diagnosis is extremely uncommon.35 To determine whether T-NHL occurs more often in immunocompromised patients than in the general population, the prevalence of patients with an immunodeficiency within the cohort of patients with T-NHL should be compared to the prevalence of patients with an immunodeficiency in the general population. Since no data are available on the overall prevalence of immunodeficiencies in the general population, we compared the frequencies of specific underlying disorders in the Dutch context. We did this for IBD, B-CLL and RA, the three most consistent underlying disorders which we witnessed, apart from solid organ transplantation for which an association with the occurrence of T-NHL is already known.17,19,30 Based on a prevalence of IBD in 432 patients per 100,000 inhabitants in 2010,36 we would expect that 0.90 of the 209 TNHL would occur in patients with IBD, compared to the seven we actually found. For B-CLL, a prevalence of 5,061 patients (average number in 2013/2014)37 out of a total population of 16,829,28938 leads to an expectation of 0.063 patients with B-CLL in our cohort, within which we identified an actual number of four patients. A similar calculation for RA, which had a prevalence of 116,000 patients in haematologica | 2018; 103(3)
T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies
Figure 2 . The hematoxylin and eosin stain shows large atypical cells in a background of small monomorphic lymphocytes. (A) The large atypical cells are positive for CD3 (B) and show loss of expression of CD5. (C) The small lymphocytes in the background are B-cells (CD20) with co-expression of CD5 (D), consistent with residual B-CLL in the background of this T-cell lymphoma.
201139 out of a population of 16,574,98938 leads to an expected number of 1.46 compared to an actual number of 3 patients. The numbers of expected cases for IBD and RA are probably overestimated, since the prevalence of IBD and RA in the general population that we used for these calculations also included all non-treated patients, who were excluded in our case series. Regarding RA, it is known that only about 10% of the patients with this disease who are registered are treated by a rheumatologist or other medical specialist, which generally corresponds with the use of immunosuppressive drugs.39 Using this information, when we recalculated the expected number of patients with RA in our cohort, we found a number of only 0.146 compared to the actual number of three. The probability of observing a certain number of cases, given an expected number of cases, can be calculated via the Poisson distribution. By using this method, we found a probability of 4.34E-05 that at least seven cases of IBD would occur given an expected number of 0.90. The probability of at least four cases of B-CLL compared to an expected number of 0.063 is 6.24E-07. For RA the probability of three cases occurring is 0.00047 when the expected number is 0.146. These calculations, even when bearing in mind the referral bias of our center, suggest that for B-CLL patients and those on immunosuppressive drugs for the treatment of IBD and RA, the risk of developing a T-NHL is higher than in the general population. haematologica | 2018; 103(3)
Taking into account the additional 596 cases we found in the literature of T-NHL in patients with impaired immunity due to HIV, hematologic malignancies and immunosuppressive drugs, this observation might be extrapolated to all patients with secondary immunodeficiencies. Whether patients with primary immunodeficiencies are also at risk for developing T-NHL is less clear, since we found only nine such cases in the literature. In this particular group of patients polyclonal T-cell proliferations are seen more often than overt T-cell lymphomas.8 The pathogenesis of T-NHL in immunodeficient patients is unclear. In our series the majority of lymphomas were EBV-negative, suggesting that, in contrast to B-cell lymphoproliferative disorders in immunocompromised patients, EBV does not play a role in this setting. Infection by other viruses, such as human T-lymphotropic virus type 1 (HTLV1) and human herpes virus (HHV)-6, possibly plays a role in development, as is suggested for Tcell PTLD.40-42 Another possible mechanism is that immune dysfunction contributes to lymphomagenesis by diminished immunosurveillance when malignant mutations arise in lymphoid cells due to other causes, for instance chronic antigenic stimulation or environmental factors, including mutagenic effects of chemotherapeutic or immunosuppressive drugs.42,43 Furthermore, a common biological basis of the first and second malignancy in hematologic malignancies has been suggested, ie., due to 493
M.L. Nijland et al.
malignant transformation of a stem cell with the capacity to differentiate in either a B- or a T-NHL or a shared genetic predisposition for the two lymphomas.44,45 An example of this is the ten-eleven translocation 2 (TET2) mutation.
TET2 is a dioxygenase that plays an important role in hematopoietic stem cells and progenitor cells by catalyzing multiple steps of 5-methylcytosine oxidation. TET2 mutations are commonly found in myeloid cancers, in
Table 5. Treatment and outcome.
Histological Subtype
First-line therapy
PTCL, NOS
None, confirmation diagnosis after death
PTCL, NOS PTCL, NOS
None, palliative treatment None, diagnosis at autopsy
PTCL, NOS PTCL, NOS PTCL, NOS ALCL ALKALCL ALKALCL ALK-
Cytarabine, idarubicine PR DHAP-VIM-DHAP PD CHOEP, autologous SCT and HD-MTX CR CHOP, alemtuzumab CR None, palliative treatment Death CHOP and intrathecal MTX CR, relapse DHAP CR CHOP PD DHAP; third line: brentuximab; fourth line: Cyclophosphamide, fludarabine, TBI, followed by allo-CB SCT Reversed CHOP
Death
Death
1
ALCL ALK+
Cyclofosfamide, doxorubicine, prednisolone
CR, relapse
Death
41
AITL
CHOP
PR
Death
3
AITL
CHOP
SD
AITL MF
Prednisolone PUVA, topical steroids, MTX, prednisolone PUVA, topical steroids No treatment, self-limiting RT Multiple chemotherapeutic agents (HOVON-70-protocol) Prednisolone, vincristin, daunorubicin, PEG-asparaginase and intrathecal MTX CHOP NR, treatment in other center Prednisolone RT None, palliative treatment
NR PD
ALCL ALK-
MF Prim. C-CD30+ T-LPD Prim. C-CD30+ T-LPD Precursor T-LBL Precursor T-LBL
ENKL, nasal type HSTCL T-PLL Liquor/CNS T-NHL, unclass.
R
Second-line therapy R
Death Death
OC
S (m)
Cause of death
Death
0
Sepsis (mycobacterial infection) Rapid progression MOF and ARDS, high lymphoma load Refractory disease 19
<1 0
Alive <1 Alive CR
4 Alive 111
Rapid progression 102 Alive
21
DHAP-VIM-DHAP + alemtuzumab
SD
Death
9
Sepsis caused by bowel perforation due to lymphoma NR, besides lymphoma also gallbladder cancer and HCC Chemotherapy complications NR
RT
PR
Death Death
4 61
NR NR
Alive Alive
65 74
Alive
69
CR CR
Alive Alive
76 8
CR NR NR
Alive Death Death Death Death
49 3 4 1 1
DHAP; third line: VIM
SD
CR CR CR, relapse RT
CR
NR NR Rapid progression Rapid progression
AITL: angioimmunoblastic T-cell lymphoma; ALCL: T-cell anaplastic large cell lymphoma; ALK: anaplastic lymphoma kinase; ARDS: acute respiratory distress syndrome; CB SCT: cord blood stem cell transplantation; CHOEP: cyclophosphamide, doxorubicin, vincristine, etoposide and prednisolone; CHOP: cyclophosphamide, doxorubicin, vincristine and prednisolone; CR: complete remission; DHAP: dexamethasone, high dose cytarabine and cisplatinum; ENKL: extranodal NK/T-cell lymphoma; HSTCL: hepatosplenic T-cell lymphoma; LTF: lost to follow-up; MF: mycosis fungoides; MOF: multi-organ failure; MTX: methotrexate; NR: not recorded or not known; OC: outcome; PD: progressive disease; PR partial remission; Prim. C-CD30+ T-LPD: primary cutaneous CD30+ T-cell lymphoproliferative disease; PTCL, NOS: peripheral T-cell lymphoma, not otherwise specified; PUVA: psoralen plus ultraviolet A; R: response; RT: radiotherapy; S: survival; SD: stable disease; T-LBL: T-cell lymphoblastic lymphoma; T-PLL: T-cell prolymphocytic leukemia; TBI: total body irradiation; VIM: ifosfamide, mitoxantrone and etoposide.
494
haematologica | 2018; 103(3)
T-cell non-Hodgkin lymphoma arising in patients with immunodeficiencies
about 11.9 % of all T-NHL, particularly in AITL and PTCL, in NOS, and in about 2% of B-NHL. The same TET2 mutations, for example, have been found in patients with AML/MDS secondary to a previous lymphoma, suggesting a shared genetic origin.46,47 Moreover, it has been observed that TET2 loss in mice leads to hypermutagenicity in hematopoietic stem cells and progenitor cells, resulting in an increased risk of various hematologic malignancies.48 Considering that most patients had been treated with immunosuppressive drugs or chemotherapy for intervals of longer than two years, it seems that prolonged treatment with these types of drugs increases the risk of malignant lymphoma development. An association between the use of thiopurines alone or combined with TNF-Îą inhibitors and the development of HSTCL in IBD patients has been reported previously.15 The majority of patients with IBD in our series and in the reported cases had been using azathioprine. It is possible, therefore, that the use of thiopurines also contributes to the development of other T-NHL. Obviously, inherent to its retrospective design, our study has some limitations. The clinical data were not complete for all patients, thus making it difficult to assess the temporal relationship between the underlying disorders or drug use and the development of T-NHL for some. In addition, the group was too small to run subgroup analyses. Moreover, since the overall prevalence of immunodeficiencies in the general population is not known, we could only compare the prevalence of specific underlying disorders in our cohort of patients with T-NHL to those in the general population.
References 1. Gatti RA, Good RA. Occurrence of malignancy in immunodeficiency diseases. A literature review. Cancer. 1971;28(1):89-98. 2. Penn I, Hammond W, Brettschneider L, Starzl TE. Malignant lymphomas in transplantation patients. Transplant Proc. 1969;1(1):106-112. 3. Parker A, Bowles K, Bradley JA, et al. Diagnosis of post-transplant lymphoproliferative disorder in solid organ transplant recipients - BCSH and BTS Guidelines. Br J Haematol. 2010;149(5):675-692. 4. Beral V, Peterman T, Berkelman R, Jaffe H. AIDS-associated non-Hodgkin lymphoma. Lancet. 1991;337(8745):805-809. 5. Levine AM, Seneviratne L, Espina BM, et al. Evolving characteristics of AIDS-related lymphoma. Blood. 2000;96(13):4084-4090. 6. Bewtra M. Lymphoma in inflammatory bowel disease and treatment decisions. Am J Gastroenterol. 2012;107(7):964-970. 7. Beaugerie L, Brousse N, Bouvier AM, et al. Lymphoproliferative disorders in patients receiving thiopurines for inflammatory bowel disease: a prospective observational cohort study. Lancet. 2009;374(9701):16171625. 8. Gratzinger D, Jaffe ES, Chadburn A, et al. Primary/congenital immunodeficiency: 2015 SH/EAHP workshop report-part 5. Am J Clin Pathol. 2017;147(2):204-216. 9. Simon TA, Thompson A, Gandhi KK, Hochberg MC, Suissa S. Incidence of
haematologica | 2018; 103(3)
10.
11.
12.
13.
14.
15.
16.
Conclusion The 25 cases presented herein, together with the 596 cases found in the literature of T-NHL in patients with varying causes of immunodeficiencies suggest that patients with a secondary immunodeficiency are at increased risk for the development of T-NHL. Prolonged treatment with immunosuppressive or chemotherapeutic drugs seems to contribute to the risk. T-NHL in immunodeficient patients are histologically very heterogeneous. The distribution of subtypes resembles that present in the general population, with the exception of primary cutaneous T-NHL and HSTCL, both of which have been reported more often in patients with an immunodeficiency, and AITL, which has been reported less frequently in this group. Overall, the prognosis seems worse compared to T-NHL of similar subtypes in the general population. TNHL occur predominantly in men, and in immunodeficient patients they tend to be more often located extranodally, which is in line with B-cell lymphomas in this group of patients. In our series of immunodeficient patients, the lymphomas occurred on average nine to ten years earlier than T-NHL in the general population. The observations in the study herein should raise awareness of the possible development of T-NHL in immunodeficient patients and challenge the prolonged use of immunosuppressive drugs in patients who are in clinical remission of their autoimmune disease. Funding The study was supported by Lymph&Co and by a grant from the Egbers Foundation.
malignancy in adult patients with rheumatoid arthritis: a meta-analysis. Arthritis Res Ther. 2015;17:212. Wolfe F, Michaud K. Lymphoma in rheumatoid arthritis: the effect of methotrexate and anti-tumor necrosis factor therapy in 18,572 patients. Arthritis Rheum. 2004;50(6):1740-1751. Zintzaras E, Voulgarelis M, Moutsopoulos HM. The risk of lymphoma development in autoimmune diseases: a meta-analysis. Arch Intern Med. 2005;165(20):2337-2344. Askling J, Fored CM, Baecklund E, et al. Haematopoietic malignancies in rheumatoid arthritis: lymphoma risk and characteristics after exposure to tumour necrosis factor antagonists. Ann Rheum Dis. 2005;64(10):1414-1420. Baecklund E, Iliadou A, Askling J, et al. Association of chronic inflammation, not its treatment, with increased lymphoma risk in rheumatoid arthritis. Arthritis Rheum. 2006;54(3):692-701. Castillo JJ, Beltran BE, Bibas M, et al. Prognostic factors in patients with HIVassociated peripheral T-cell lymphoma: a multicenter study. Am J Hematol. 2011;86(3):256-261. Gilardin L, Copie-Bergman C, Galicier L, et al. Peripheral T-cell lymphoma in HIVinfected patients: a study of 17 cases in the combination antiretroviral therapy era. Br J Haematol. 2013;161(6):843-851. Biggar RJ, Engels EA, Frisch M, Goedert JJ. Risk of T-cell lymphomas in persons with
17.
18.
19.
20.
21.
22.
AIDS. J Acquir Immune Defic Syndr. 2001;26(4):371-376. Herreman A, Dierickx D, Morscio J, et al. Clinicopathological characteristics of posttransplant lymphoproliferative disorders of T-cell origin: single-center series of nine cases and meta-analysis of 147 reported cases. Leuk Lymphoma. 2013;54(10):21902199. Perez K, Castillo J, Dezube BJ, Pantanowitz L. Human immunodeficiency virus-associated anaplastic large cell lymphoma. Leuk Lymphoma. 2010;51(3):430-438. Swerdlow SH. T-cell and NK-cell posttransplantation lymphoproliferative disorders. Am J Clin Pathol. 2007;127(6):887895. Seckin D, Barete S, Euvrard S, et al. Primary cutaneous posttransplant lymphoproliferative disorders in solid organ transplant recipients: a multicenter European case series. Am J Transplant. 2013;13(8):21462153. Deepak P, Sifuentes H, Sherid M, Stobaugh D, Sadozai Y, Ehrenpreis ED. T-cell nonHodgkin's lymphomas reported to the FDA AERS with tumor necrosis factoralpha (TNF-alpha) inhibitors: results of the REFURBISH study. Am J Gastroenterol. 2013;108(1):99-105. Kotlyar DS, Osterman MT, Diamond RH, et al. A systematic review of factors that contribute to hepatosplenic T-cell lymphoma in patients with inflammatory bowel disease. Clin Gastroenterol Hepatol.
495
M.L. Nijland et al. 2011;9(1):36-41 e1. 23. Swerdlow SH, Campo E, Harris NL, et al. WHO classification of tumours of haematopoietic and lymphoid tissues (4th ed). Lyon, France: International Agency for Research on Cancer. 2008. 24. Olsen E, Vonderheid E, Pimpinelli N, et al. Revisions to the staging and classification of mycosis fungoides and Sezary syndrome: a proposal of the International Society for Cutaneous Lymphomas (ISCL) and the cutaneous lymphoma task force of the European Organization of Research and Treatment of Cancer (EORTC). Blood. 2007;110(6):1713-1722. 25. Lopez-Guillermo A, Cid J, Salar A, et al. Peripheral T-cell lymphomas: initial features, natural history, and prognostic factors in a series of 174 patients diagnosed according to the R.E.A.L. Classification. Ann Oncol. 1998;9(8):849-855. 26. Vose J, Armitage J, Weisenburger D. International peripheral T-cell and natural killer/T-cell lymphoma study: pathology findings and clinical outcomes. J Clin Oncol. 2008;26(25):4124-4130. 27. Jaffe ES, Harris NL, Stein H, Isaacson PG. Classification of lymphoid neoplasms: the microscope as a tool for disease discovery. Blood. 2008;112(12):4384-4399. 28. Arrowsmith ER, Macon WR, Kinney MC, et al. Peripheral T-cell lymphomas: clinical features and prognostic factors of 92 cases defined by the revised European American lymphoma classification. Leuk Lymphoma. 2003;44(2):241-249. 29. Ferreri AJ. Risk of CNS dissemination in extranodal lymphomas. Lancet Oncol. 2014;15(4):e159-169. 30. Zucca E, Roggero E, Bertoni F, Cavalli F.
496
31.
32.
33.
34.
35.
36.
37. 38.
Primary extranodal non-Hodgkin's lymphomas. Part 1: Gastrointestinal, cutaneous and genitourinary lymphomas. Ann Oncol. 1997;8(8):727-737. Tran H, Nourse J, Hall S, Green M, Griffiths L, Gandhi MK. Immunodeficiency-associated lymphomas. Blood Rev. 2008;22(5):261-281. Gisselbrecht C, Gaulard P, Lepage E, et al. Prognostic significance of T-cell phenotype in aggressive non-Hodgkin's lymphomas. Groupe d'Etudes des Lymphomes de l'Adulte (GELA). Blood. 1998;92(1):76-82. Ekstrom Smedby K, Vajdic CM, Falster M, et al. Autoimmune disorders and risk of non-Hodgkin lymphoma subtypes: a pooled analysis within the InterLymph Consortium. Blood. 2008;111(8):40294038. Wang SS, Flowers CR, Kadin ME, et al. Medical history, lifestyle, family history, and occupational risk factors for peripheral T-cell lymphomas: the InterLymph NonHodgkin Lymphoma Subtypes Project. J Natl Cancer Inst Monogr. 2014;2014(48):66-75. Tsochatzis E, Vassilopoulos D, Deutsch M, Filiotou A, Tasidou A, Archimandritis AJ. Angioimmunoblastic T-cell lymphomaassociated arthritis: case report and literature review. J Clin Rheumatol. 2005;11(6):326-328. De Groof J, Rossen N, van Rhijn B, et al. P644 Epidemiology and characteristics of inflammatory bowel disease in a large population-based cohort in the Netherlands. Poster presentations European Crohn's and Colitis Organisation, 2015. The Netherlands Cancer Registry. Statistics Netherlands (CBS).
39. National Institute of Public Health and the Environment (RIVM). 40. Hoshida Y, Li T, Dong Z, et al. Lymphoproliferative disorders in renal transplant patients in Japan. Int J Cancer. 2001;91(6):869-875. 41. Lin WC, Moore JO, Mann KP, Traweek ST, Smith C. Post transplant CD8+ gammadelta T-cell lymphoma associated with human herpes virus-6 infection. Leuk Lymphoma. 1999;33(3-4):377-384. 42. Fisher SG, Fisher RI. The epidemiology of non-Hodgkin's lymphoma. Oncogene. 2004;23(38):6524-6534. 43. Barzilai A, Trau H, David M, et al. Mycosis fungoides associated with B-cell malignancies. Br J Dermatol. 2006;155(2):379-386. 44. Campidelli C, Sabattini E, Piccioli M, et al. Simultaneous occurrence of peripheral Tcell lymphoma unspecified and B-cell small lymphocytic lymphoma. Report of 2 cases. Hum Pathol. 2007;38(5):787-792. 45. Herro E, Dicaudo DJ, Davis MD, Weaver AL, Swanson DL. Review of contemporaneous mycosis fungoides and B-cell malignancy at Mayo Clinic. J Am Acad Dermatol. 2009;61(2):271-275. 46. Ko M, An J, Pastor WA, Koralov SB, Rajewsky K, Rao A. TET proteins and 5methylcytosine oxidation in hematological cancers. Immunol Rev. 2015;263(1):6-21. 47. Solary E, Bernard OA, Tefferi A, Fuks F, Vainchenker W. The Ten-Eleven Translocation-2 (TET2) gene in hematopoiesis and hematopoietic diseases. Leukemia. 2014;28(3):485-496. 48. Pan F, Wingo TS, Zhao Z, et al. Tet2 loss leads to hypermutagenicity in haematopoietic stem/progenitor cells. Nat Commun. 2017;8:15102.
haematologica | 2018; 103(3)
ARTICLE
Chronic Lymphocytic Leukemia
In vitro and in vivo evidence for uncoupling of B-cell receptor internalization and signaling in chronic lymphocytic leukemia
Ferrata Storti Foundation
Eve M. Coulter,1 Andrea Pepper,2 Silvia Mele,3 Najeem’deen Folarin,4 William Townsend,1 Kirsty Cuthill,4 Elizabeth H. Phillips,1 Piers E. M. Patten1,4 and Stephen Devereux4
School of Cancer and Pharmaceutical Sciences, Faculty of Life Sciences and Medicine, King’s College London; 2Brighton and Sussex Medical School, Medical research Building, University of Sussex, Brighton; 3St John’s Institute of Dermatology, Department of Genetics and Molecular Medicine, King's College London and 4Department of Haematological Medicine, Kings College Hospital, London, UK 1
Haematologica 2018 Volume 103(3):497-505
ABSTRACT
B
-cell receptor activation, occurring within lymph nodes, plays a key role in the pathogenesis of chronic lymphocytic leukemia and is linked to prognosis. As well as activation of downstream signaling, receptor ligation triggers internalization, transit to acidified endosomes and degradation of ligand-receptor complexes. Herein, we investigated the relationship between these two processes in normal and leukemic B cells. We found that leukemic B cells, particularly anergic cases lacking the capacity to initiate downstream signaling, internalize and accumulate ligand in acidified endosomes more efficiently than normal B cells. Furthermore, ligation of either surface CD79B, a B-cell receptor component required for downstream signaling, or surface Immunoglobulin M (IgM) by cognate agonistic antibody, showed that the two molecules internalize independently of each other in leukemic but not normal B cells. Since association with surface CD79B is required for surface retention of IgM, this suggests that uncoupling of B-cell receptor internalization from signaling may be due to the dissociation of these two molecules in leukemic cells. A comparison of lymph node with peripheral blood cells from chronic lymphocytic leukemia patients showed that, despite recent B-cell receptor activation, lymph node B cells expressed higher levels of surface IgM. This surprising finding suggests that the Bcell receptors of lymph node- and peripheral blood-derived leukemic cells might be functionally distinct. Finally, long-term therapy with the Bruton’s tyrosine kinase inhibitors ibrutinib or acalabrutinib resulted in a switch to an anergic pattern of B-cell receptor function with reduced signaling capacity, surface IgM expression and more efficient internalization.
Introduction It is now clear that signaling through the B-cell receptor (BCR) plays a key role in the pathogenesis of chronic lymphocytic leukemia (CLL) and other lymphomas. Several components of this pathway, including Syk,1 Erk,2 Akt,3 NFAT4 and NFκB5 can be constitutively activated and drugs that target BCR signaling, such as the Bruton’s tyrosine kinase inhibitors (BTKi), ibrutinib and acalabrutinib, are proving extremely effective in the clinic.6,7 BCR responsiveness varies markedly between patients with CLL and is linked to prognosis.8 Some cases show features of anergy,4,9 a pattern that is associated with lack of ability to transduce a downstream signal in response to BCR ligation and the presence of markers of good prognosis, including low levels of CD38 and mutated immunoglobulin heavychain variable (IGHV) genes. In contrast, cases with responsive or signaling competent BCRs usually express high levels of CD38, have unmutated IGHV genes and a more unfavorable clinical course;10 interestingly, these patients tend to respond more rapidly to BCR antagonists than those with anergic BCRs. Although BTKi therapy is very successful in controlling CLL, it is not curative and many haematologica | 2018; 103(3)
Correspondence: eve.coulter@kcl.ac.uk
Received: July 21, 2017. Accepted: December 12, 2017. Pre-published: December 14, 2017. doi:10.3324/haematol.2017.176164 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/497 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
497
E.M. Coulter et al.
patients are left with low level residual disease, which regrows on discontinuation of drug or when resistance mutations develop.11,12 This persistent disease also suggests that, within individual patients, the tumor may not behave in a homogeneous manner.13 Despite the central importance of BCR signaling in CLL and the efficacy of drugs that block this pathway, there is relatively little known about BCR dynamics in leukemic B cells. Surface levels of IgM and other BCR components are generally lower in CLL compared to normal B cells, and it has been suggested that this might be due to a failure to properly assemble the sIg α/β subunits CD79A and CD79B.14 Recent studies have shown that total IgM and CD79A levels are near normal in CLL but that CD79B expression, which is required for the transport of BCR to the cell surface,15 is reduced, thus trapping IgM within the cell.16 Exposure to interleukin 4 (IL4) increases CD79B expression and allows sIgM levels to increase and BCR signaling capacity to improve.16,17 CLL cell surface BCRs have an immature pattern of glycosylation that matures following ex vivo incubation18 or exposure to IL4,17 in keeping with accelerated BCR turnover induced by chronic activation. It has also been reported that, within the peripheral blood (PB) of individual patients with CLL, leukemic cells with the lowest sIgM expression show biochemical features of recent activation and proliferation, presumably because they have recently been released from lymphoid tissues where BCR stimulation and activation are thought to occur.19,20 Taken together, these previous data suggest that the reduced sIgM levels observed in CLL are due to a combination of increased turnover consequent to chronic activation coupled with defective transport to the cell surface resulting from a deficiency of CD79B. The ability of CLL BCRs to become internalized also has implications for how the tumor interacts with other cells, such as T cells. We, and others, have previously shown that, as in normal lymph nodes (LNs), activated CD4+ T cells colocalize with proliferating tumor cells and, in vitro, can supply signals that cause tumor proliferation.2124 This process normally involves endocytosis and the processing of antigen bound to BCR, however it is not known whether CLL B cells are capable of providing this function. Herein we investigated whether, like normal B cells, CLL cells can internalize their BCR and to what extent this is linked to downstream signaling in both untreated patients and those receiving therapy with the BTKis ibrutinib and acalabrutinib. Our results shed new light on BCR function in CLL and have implications for understanding the mode of action of this important new class of drugs.
Flow cytometry Cells were stained according to the manufacturer’s recommendations using fluorochrome-coupled antibodies (Online Supplementary Table S4). Viable CD19+CD5+ CLL and CD19+ normal B cells were acquired on a FACS Canto II flow cytometer (BD Biosciences) and analyzed using FlowJo software (TreeStar). Surface (s)IgM and IgD expression was assessed using Quantum™ fluorescein isothiocyanate molecules of equivalent soluble fluorochrome (FITC MESF) microsphere kits and the QuickCal v. 2.3 program, according to the manufacturer’s recommendations. Surface CD79B was quantified using R-phycoerythrin (R-PE) MESF microsphere kits. BCR signaling competence was determined using the ratiometric Ca2+ detector Indo1-AM (Life Technologies) to measure intracellular calcium (Ca2+) levels and Ca2+ influx following αIgM stimulation, as previously described.4,9 ERK1/2 phosphorylation activation was analyzed in both treatment-naïve and BTKi-treated (collected before and during BTKi therapy) CLL patient B cells. Cells were incubated with or without αIgM, or with phorbol 12-myristate 13-acetate (PMA; served as positive control) for ten minutes at 37°C prior to intracellular and surface staining.
B-cell receptor internalization in normal and CLL B cells BCR internalization was assessed in two ways. The uptake and retention of ligand/receptor complexes in acidified endosomes was measured using the pH sensitive fluorescent sensor, pHrodo™ Red avidin (Life Technologies) linked to agonistic αIgM or IgD (pHrodo-αIgM or D). Target cells were incubated with either pHrodo-αIgM or D for 30 minutes at 4°C and then 37°C for 1h. Results are expressed as the pHrodo mean fluorescent intensity (MFI) after subtraction of the background signal (MFI of unlabeled anti-IgM). BCR internalization in normal and CLL B cells was also directly assessed by measuring the rate of disappearance of sIgM following ligation by agonistic αIgM. Full methodology is provided in the Online Supplementary Information.
Immunofluorescence Staining CLL B cells were isolated using a human B-cell negative selection kit without CD43 depletion (StemCell Technologies) according to the manufacturer’s instructions, and purity was confirmed by flow cytometry. Cells were labeled with pHrodo-avidin or pHrodo-αIgM, deposited onto poly-l-lysine coated glass slides by cytospin and stained with CytoPainter Phalloidin-iFluor 488 reagent. Images were acquired on a Nikon Eclipse Ti-E inverted microscope equipped with the Nikon A1R Si confocal imaging system. Image analysis was with Nikon Elements v4.2 software (see Online Supplementary Information).
Statistical analysis Methods Patients and samples PB samples were obtained from 19 healthy volunteers, 40 untreated patients with confirmed CLL (Online Supplementary Table S1), and an additional 15 patients receiving BTKi therapy (Online Supplementary Table S2). LN fine needle aspirate (FNA) and paired matched PB samples were derived from seven untreated CLL patients (Online Supplementary Table S3). Ethical approval was obtained from the National Research Ethics Service (08/H0906/94); all patients provided written informed consent. Patients were classified as having unmutated IGHV genes if homology with germline was >98%25 and as CD38+ if expression levels were 7% or higher (Online Supplementary Tables S1-S3).26 498
Statistical analyses were performed using GraphPad Prism software version 5 (GraphPad Software, La Jolla, CA, USA). The Shapiro–Wilk test, t-test, Mann–Whitney and/or Wilcoxon’s test were used where indicated. P-values of <0.05 were considered significant.
Results Internalization of normal and CLL BCRs The uptake and retention of pHrodo-αIgM labeled BCRs in acidified endosomes was similar in normal and CLL B cells (MFI ± standard deviation: normal control: 147.5±16.8, n=19, P=0.825, CLL: 151.3±12.0, n=40; Figure haematologica | 2018; 103(3)
BCR internalization and signaling in CLL
1A) and correlated with sIgM expression (r2=0.405, P=0.0001, n=40; Figure 1B). Confocal immunofluorescence microscopy confirmed that the labeled BCRs accumulate in an intracellular compartment (Online Supplementary Figure S1A) and pre-incubation with excess unlabeled anti-IgM, sodium azide and cytochalasin D showed the process to be specific and dependent on energy and the cytoskeleton (Online Supplementary Figures S1BS1D). Repeated measurements confirmed the reproducibility of the assay within individual CLL cases (r2=0.874, P<0.0001, n=30; Online Supplementary Figure
A
C
S1E). It has previously been reported that sIgM expression is lower in CLL compared to normal B cells, and this was also the case in our patients (Figure 1C). Since BCR uptake by normal and CLL B cells is similar and dependent on sIgM expression, this suggested that the process might be more efficient in CLL compared to normal B cells. Correction of the pHrodo-αIgM uptake value for the number of surface IgM molecules (uptake index) confirmed this to be the case (Figure 1D). After correcting for the level of surface IgM expression, the pHrodo-αIgM signal, which reflects accumulation in acidified endosomes, was
B
D
E
Figure 1. B-cell receptor expression and internalization in CLL patients and healthy controls. (A) PBMCs from CLL patients (n=40) and healthy controls (n=19) were incubated with pHrodo-αIgM and the MFI of cells internalizing pHrodo-labeled avidin was measured within the CD19+5+ (CLL) and CD19+ (normal) B-cell population by flow cytometry. (B) B-cell receptor (BCR) internalization correlates with sIgM expression (Pearson’s correlation); correlation plot comparing the relationship between pHrodo-αIgM uptake and surface IgM (sIgM) expression in CD19+5+ cells derived from 40 CLL patients. (C) Although the level of pHrodo-αIgM uptake was comparable between CLL patients and healthy controls, marked differences in surface expression (sIgM) were observed (P=0.0001). (D) After correcting the level of uptake per molecule of sIgM, the uptake index was measured in both CLL patients and healthy controls. Within the CLL patient subsets, CLL B cells from CD38 negative, anergic and mutated patients were identified as having a greater uptake index than their counterparts, although mutational status was not significant (P=0.05, P=0.05 and P=0.436, respectively; unpaired t-test). (E) To assess the rate of BCR internalization more directly, the disappearance of sIgM following ligation of agonistic αIgM was measured. Accelerated BCR endocytosis was detected (2 min time point) in anergic B cells (n=6) compared to signaling competent (n=6) and normal (n=6) B cells (P=0.02 and P=0.01, respectively*; Mann-Whitney test). MFI: mean fluorescent intensity; MESF: molecules of equivalent soluble fluorochrome; CLL: chronic lymphocytic leukemia; SAV: streptavidin-APC.
haematologica | 2018; 103(3)
499
E.M. Coulter et al.
3.04 times more efficient in CLL compared to normal B cells (Figure 1D, P<0.0001). The uptake index varied considerably between patients, but was significantly higher in those whose BCRs lacked the ability to mobilize calcium in response to BCR ligation and cases with low CD38 expression (see Online Supplementary Figure S2 and Online Supplementary Table S1). Prolonged ex vivo incubation has previously been shown to reverse features of anergy, namely, re-expression of sIgM and restoration of BCR responsiveness.4,9 We therefore examined BCR expression and internalization after 24 hours ex vivo incubation, however, results were heterogeneous with no significant recovery in sIgM expression (Online Supplementary Figure S3A, P=0.83, n=7) or change in BCR internalization efficiency (Online Supplementary Figure S3B, P=0.88, n=7). It was not possible to assess internalization efficiency at later time points as the assay critically depends on metabolic integrity of the cells, which was not consistently maintained after 24 hours. Since the pHrodo-αIgM signal reflects both uptake into and retention within acidic endosomes, we also assessed the BCR internalization rate more directly by measuring loss from the cell surface following ligation with an agonistic antibody. This showed that all cases of CLL internalize their BCRs more rapidly than normal B cells (Figure 1E; P=0.04, 2 minute time point). More subtle differences were observed between anergic and signal competent cases of CLL, with more rapid initial loss from the surface in the former (Figure 1E; P=0.02 at 2 minute time point). We also measured sIgD expression and the levels of pHrodo-αIgD uptake using the same assays. All CLL B cells internalized pHrodo-αIgD, and pre-incubation with unlabeled αIgD reduced the level of uptake (Online Supplementary Figure S4A; n=18; P=0.001). No significant correlation between pHrodo-αIgD uptake and surface sIgD expression was observed (Online Supplementary Figure S4B) and there was no difference in the pHrodo-αIgD uptake index between CD38high and low, anergic/non anergic and IGHV-mutated and unmutated patients (Online Supplementary Figure S4C). This suggests that uptake and retention mechanisms differ between IgD and IgM; this was not addressed further herein.
Mechanism of dissociation of BCR signaling and internalization We next investigated the role of sCD79B, a molecule that is closely associated with the BCR and essential for signal transduction, in BCR internalization. As previously reported,27 CLL B cells express lower levels of sCD79B compared to normal B cells, with a particularly reduced expression in anergic compared to signaling competent cases (normal B-cell MFI=2561.2±400.9, anergic CLL MFI=318.2±63.4, signaling competent MFI=397.5±71.1). We also found that sIgM expression correlates with sCD79B both between (Figure 2A) and within a patient with CLL (Figures 2B,C). Despite these findings and the fact that sIgM and sCD79B are known to be associated during BCR signaling, no significant reduction in sCD79B expression occurred following ligation and downregulation of sIgM with agonistic antibody in both anergic and signaling competent cases of CLL (Figure 2D). In normal B cells, a slight but significant reduction in sCD79B expression was observed 30 and 60 minutes after BCR ligation with αIgM. Similarly, downregulation of CD79B with an agonistic anti-CD79B had no effect on the level of sIgM in 500
anergic and signaling competent CLL cells and normal B cells (Figure 2E,F). Thus, although both normal and CLL B cells are capable of internalizing IgM and CD79B following ligation with cognate agonistic antibody, the two molecules do not cointernalize, and thus cannot be closely associated during membrane trafficking. Having established that sCD79B is not required for sIgM internalization, we proceeded to assess its role in BCR signaling by measuring the effect of prior sCD79B downregulation on αIgM ERK phosphorylation in CLL cells. As expected, prior sCD79b downregulation reduced, but did not abolish αIgM-induced ERK phosphorylation (Figure 2G). In contrast, depletion of sCD79B was without effect on pHrodo-αIgM uptake (Figure 2H).
In vivo relationship between BCR signaling and internalization We next investigated the relationship between BCR internalization and signaling in vivo in CLL by studying subsets of PB cells, those from the LN where BCR activation and proliferation is thought to take place and longitudinal samples obtained from patients being treated with BTKis ibrutinib and acalabrutinib.
Recently proliferated and quiescent subsets It has previously been shown that recently proliferated LN emigrants express low levels of CXCR4 and high levels of CD5 (CXCR4dimCD5bright).28 Since proliferation within LNs is linked to BCR signaling, we compared BCR expression and internalization efficiency on CXCR4dimCD5bright and CXCR4brightCD5dim subsets. Since the BCR internalization assay involves incubation with agonistic αIgM, we first investigated whether 1-hour incubation with pHrodo-αIgM altered the expression of these markers. As expected, CXCR4 expression was reduced and CD5 increased following ligation of BCR, however, over the 1hour assay period, the magnitude of change was small (CXCR4 mean fold change = 0.93±0.02 and CD5 mean fold change 1.24±0.04, data not shown). Since the MFI of the top and bottom decade of CXCR4 and CD5 expression differed by 11.8±5.4-fold and 8.7±6.7-fold, respectively, assay-induced changes could not have changed the composition of these subsets. Since the CXCR4dimCD5bright subset are thought to have recently undergone BCR activation within the LN, we expected to find downregulation of sIgM in this fraction, however, this was not the case and levels were higher than in the CXCR4dimCD5bright subset (Figure 3A, P<0.0001). In keeping with post-activation anergy, in the majority of patients (14/21) there was more efficient uptake of pHrodo-αIgM in the CXCR4dimCD5bright subset, however, there was significant variability and, in the whole population, this did not reach statistical significance (Figure 3B; n=21; P=0.281).
BCR expression and internalization in the PB and LN of CLL patients Since our results suggested that BCR internalization is influenced by the capacity to initiate downstream signaling, which occur within LNs, we went on to directly compare sIgM levels and uptake of pHrodo-αIgM by LN CLL cells to those derived from simultaneously obtained PB from the same patients. As we have previously shown,29 LN CLL cells expressed higher levels of CD5 than those derived from the PB, in keeping with BCR activation at these sites (data not shown). As was the case for the recently proliferated CXCR4dimCD5bright subset, sIgM levels were haematologica | 2018; 103(3)
BCR internalization and signaling in CLL
significantly higher on cells derived from the LN compared to PB (Figure 3C; P=0.03, n=7; Online Supplementary Figure S5). Again, there was great variability in pHrodoαIgM uptake efficiency with higher levels in the LN than PB in five out of seven patients, but no significant difference overall (Figure 3D; P=0.218, n=7).
BTKi-treated CLL B-cells display features of anergy We next investigated the effect of BCR pathway blockade on BCR expression and function. PB samples were col-
A
C
E
B
D
F
G
haematologica | 2018; 103(3)
lected over time from BTKi-treated patients (ibrutinib and acalabrutinib) with a prolonged lymphocytosis for a minimum of 12 months (Online Supplementary Table S2). On comparison of CD19+CD5+ CLL B-cells before and after one month of treatment, most cases showed an initial increase in sIgM expression (Figure 3E; P=0.02, n=7) followed by a decrease in sIgM levels after at least 12 months of treatment (Figure 3F; P=0.008, n=7). After 12 months or more of BTKi therapy, CLL B cells exhibited more efficient BCR internalization (Figure 3G; P=0.023, n=7) and had
H
Figure 2. Dissociation of BCR signaling and internalization in CLL cells. (A) A correlation was detected between the expression levels of surface (s)IgM and sCD79B on CD19+5+ cells derived from 21 CLL patients (both anergic and signaling competent B cells, Pearson’s correlation). Values are expressed as molecules of equivalent soluble fluorochrome (MESF) and derived from mean fluorescence intensity (MFI) values. (B) To examine the relationship within an individual patient, 15 subsets of CD19+5+ cells were created, as defined by increasing sIgM expression, and mean sCD79B MFI values were recorded within the same gate. (C) A representative patient demonstrating a correlation between sIgM and sCD79B. (D) The percentage of sCD79B receptor expression was measured on CLL and normal B cells following agonistic αIgM ligation; a gradual reduction and internalization of sCD79B on normal CD19+ B cells was detected after 60 min incubation (n=6; P=0.03*; Wilcoxon matched pairs test). In addition, CD79B internalization (E) and sIgM receptor expression (F) was measured upon αCD79B ligation, and compared between anergic, signaling competent and normal B cells. CD79B internalization occurred more rapidly (2 min time point) in anergic B cells compared to signaling competent and normal B cells (P=0.01 and P=0.001, respectively*; Mann-Whitney test), however, the percentage of sIgM receptor expression remained unchanged in all CLL cases. Finally, signaling competent CLL B cells were pre-incubated with agonistic αCD79B for 10mins at 37°C prior to αIgM stimulation to determine the effect of CD79B internalization on αIgM-induced pERK activation (G: n=9; pERK levels were normalized to the positive control), as well as pHrodoαIgM uptake (H: n=9). Statistical analysis was performed via Wilcoxon matched pairs test. CLL: chronic lymphocytic leukemia; SAV: streptavidinAPC.
501
E.M. Coulter et al.
reduced ability to activate ERK following BCR stimulation (Figure 3H; P=0.024, n=12).
Discussion In the study herein, we investigated the capacity of normal and CLL B cells to internalize ligands that bind to the BCR. Using two complementary techniques, we showed that BCR internalization and transit to acidified endo-
somes occurs in both normal and CLL B cells. When corrected for the level of sIgM expression, we found that BCR internalization and accumulation in endosomes is three times more efficient in CLL than normal B cells, and is highest of all in cases with anergic BCRs. Using agonistic antibodies to sIgM and sCD79B we also showed that internalization of sIgM is not accompanied by internalization of CD79B and vice versa. A comparison of LN with PB CLL cells and PB CXCR4dimCD5bright cells, representing
A
B
C
D
E
F
G
H
Figure 3. In vivo relationship between BCR signaling and internalization in CLL cells. BCR expression and uptake index were investigated firstly in subsets of peripheral blood (PB) cells (CXCR4brightCD5dim and CXCR4dimCD5bright expressing CLL cells (A-B), and secondly in CD19+CD5+ cells from the matched lymph node (LN) and PB samples from seven unmutated CLL patients (C-D). Thirdly, sIgM expression was measured in CD19+CD5+ PB cells from CLL patients before, one month and at least 12 months after commencing BTKi treatment (ibrutinib, n=6 or acalabrutinib, n=1; presented as a percentage change in surface (s)IgM expression (E-F). Uptake index (G) and the capacity to induce pERK activation following BCR stimulation (H: pERK levels normalized to PMA/positive control) was also examined and compared in CLL B cells, pre- and during treatment. Statistical analysis was performed via Wilcoxon matched pairs test. MFI: mean fluorescent intensity; MESF: molecules of equivalent soluble fluorochrome.
502
haematologica | 2018; 103(3)
BCR internalization and signaling in CLL
recently proliferated LN emigrants,28 with the CXCR4dimCD5bright subset showed that both the LN and recent emigrants express higher levels of sIgM, however, no consistent difference in BCR internalization efficiency was observed. This was a surprising finding given that BCR activation takes place in the LN and should result in the downregulation of sIgM. Finally, long-term in vivo inhibition of BCR signaling using BTKis resulted in an increase in the number of CLL cells exhibiting features of anergy, including a reduction in sIgM expression and signaling competence as well as an increase in the efficiency of BCR internalization. These results have a number of implications. First, it is
clear that in CLL, BCR internalization is uncoupled from downstream signaling, since anergic CLL B cells internalize their BCRs more efficiently than non-anergic cases and normal B cells. These findings are in keeping with previous murine studies which showed that reduced surface BCR expression and uncoupling of signaling in normal anergic B cells is due to rapid internalization and retention in endosomes.30 Since an enhanced accumulation of BCR in endosomes was observed in every patient studied, the present results show that, at least in this respect, all cases of CLL show some features of anergy. Second, although CD79B is known to associate with sIgM and is essential for the export of IgM to the cell sur-
Figure 4. Proposed mechanism for the dissociation of BCR signaling and internalization in CLL. Two alternative configurations of the BCR are proposed. Type 1 in which CD79B and surface (s)IgM are already closely associated or become associated following ligand binding. This form of the receptor can transduce downstream signals, but is not internalized until the subunits dissociate. Type 2 BCRs do not contain CD79B either before or after ligand binding. These receptors therefore cannot signal but can become internalized. Possible BCR compositions in anergic and non-anergic CLL are illustrated. In anergic CLL, both sIgM and sCD79B levels are low and there is little potential for association between the two either before or after ligand binding. In this scenario, downstream BCR signaling is minimal but internalization is efficient. In non-anergic cases of CLL, both sIgM and sCD79B levels are higher and there is a greater likelihood of the two becoming associated. Signaling is thus relatively more efficient but the capacity for internalization is less. CLL: chronic lymphocytic leukemia; BCR: B-cell receptor.
haematologica | 2018; 103(3)
503
E.M. Coulter et al.
face and downstream BCR signaling, our data strongly suggest that in CLL B cells, they internalize independently, and thus cannot be associated during endocytosis. In normal B cells a small proportion of sCD79B and sIgM cointernalized, again suggesting minimal association of the two during endocytosis. A plausible explanation for these observations is that BCRs exist in two configurations in CLL (Figure 4); the first in which sIgM and sCD79B are not and do not become associated following ligand binding and a second in which sIgM and sCD79B are already colocalized or are induced to associate following activation. The first configuration could not transduce downstream signals but, because retention of sIgM at the surface requires association with CD79B, would internalize efficiently. Adversely, BCRs in which sIgM and CD79B are associated could initiate signaling, however, the association with CD79B would favor retention at the cell surface, at least until the subunits undergo phosphorylationinduced dissociation.31 Such a model has already been proposed for normal B cells,32 and is thought to favor signaling in response to low-affinity ligands, such as autoantigens of the type recognized by CLL BCRs.33 It is currently believed that BCR activation takes place in LNs and that this, combined with other signals from the microenvironment, leads to tumor proliferation. In other receptor systems, ligand binding is generally accompanied by downregulation of cell-surface receptors,34 and our observation that sIgM levels are actually higher in the LN than PB and in recently proliferated CXCR4dimCD5bright compared to CXCR4dimCD5bright PB CLL cells was therefore unexpected. As recently suggested by others, it is possible that the elevated sIgM levels observed within the LN and CXCR4dimCD5bright subsets are due to IL4-induced upregulation of CD79B within lymphoid tissues.16,17 Our findings also shed further light on the mechanism of action of BTKis. As noted by others,13,35 a persistent low-level lymphocytosis is frequently seen in such patients and complete remissions are rare. In the early period following the commencement of BTKi therapy, we observed an increase in sIgM expression in PB CLL cells. Since we have shown that cells in the LNs express higher levels of sIgM than those in the PB, this most likely reflects the previously documented redistribution of the tumor from the former compartment to the later.35 Over the longer term, however, the opposite is the case, with a significant reduction in sIgM that is accompanied by an increase in BCR internalization efficiency and a reduced capacity to phosphorylate ERK. These findings indicate that long-term BTKi therapy causes the emergence of a population of neoplastic B cells with anergic phenotypic
References 1. Gobessi S, Laurenti L, Longo PG, et al. Inhibition of constitutive and BCR-induced Syk activation downregulates Mcl-1 and induces apoptosis in chronic lymphocytic leukemia B cells. Leukemia. 2009; 23(4):686697. 2. Muzio M, Apollonio B, Scielzo C, et al. Constitutive activation of distinct BCR-signaling pathways in a subset of CLL patients: a molecular signature of anergy. Blood. 2008;112(1):188-195. 3. Zhuang J, Hawkins SF, Glenn MA, et al. Akt
504
and functional properties. This may occur either through reprograming of the tumor into a more anergic state or because there is subclonal heterogeneity within the tumor and selection of cells with anergic features by BTKi therapy. The latter possibility is supported by in vitro studies of leukemic B-cell migration29 and in vivo observations using heavy water or glucose labeling36 that suggest the existence of subclonal heterogeneity at a functional level in CLL. Finally, since BCR internalization into endosomes is the first event in antigen processing, our data support the theory that, under some circumstances, CLL B cells might act as antigen presenting cells. We have recently shown that LN derived CLL B cells express higher levels of costimulatory molecules, form immune synapses and stimulate an allogeneic mixed lymphocyte reaction more efficiently than those from the PB.29 In addition, elution of peptides from CLL major histocompatibility complex (MHC) class II reveals presentation of a range of autologous peptides,37 with evidence for expansion of cognate T-cell clones in CLL but not normal PB. Furthermore, a number of groups have documented the presence of T cells in CLL patients that are capable of responding to a range of other molecules, including Rh antigen,38 tumor idiotype23 immunoglobulin framework,39 or CDR3 motifs40 as well as broader responses against tumor lysates41 or intact leukemic cells.42 It is therefore plausible that the abnormal phenotype, repertoire and function of CLL T cells might be a consequence of excessive and aberrant antigen presentation occurring within lymphoid tissues. In summary, we have shown that, as in normal B-cell anergy, CLL B cells internalize ligands that bind to the BCR more efficiently than normal. This process is uncoupled from downstream signaling and does not involve association with CD79B. We demonstrate that BTKi therapy induces or selects for cells with anergic properties that persist in the long term. Understanding how this occurs will be important in order to optimize the efficacy of this important new class of drugs. Funding This work was supported by research funding from Bloodwise (grant number: 15012) and the British Society of Haematology (BSH; grant number: 34721/start up). The authors also acknowledge financial support from the UK Department of Health via the National Institute for Health Research (NIHR) comprehensive Biomedical Research Centre award to Guy's & St Thomas' NHS Foundation Trust in partnership with King's College London and Kingâ&#x20AC;&#x2122;s College Hospital NHS Foundation Trust.
is activated in chronic lymphocytic leukemia cells and delivers a pro-survival signal: the therapeutic potential of Akt inhibition. Haematologica. 2010;95(1):110-118. 4. Apollonio B, Scielzo C, Bertilaccio MT, et al. Targeting B-cell anergy in chronic lymphocytic leukemia. Blood. 2013;121(19):38793888, S1-8. 5. Hewamana S, Alghazal S, Lin TT, et al. The NF-kappaB subunit Rel A is associated with in vitro survival and clinical disease progression in chronic lymphocytic leukemia and represents a promising therapeutic target. Blood. 2008;111(9):4681-4689.
6. Byrd JC, Harrington B, O'Brien S, et al. Acalabrutinib (ACP-196) in Relapsed Chronic Lymphocytic Leukemia. N Engl J Med. 2016;374(4):323-332. 7. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-42. 8. Cesano A, Perbellini O, Evensen E, et al. Association between B-cell receptor responsiveness and disease progression in B-cell chronic lymphocytic leukemia: results from single cell network profiling studies. Haematologica. 2013;98(4):626-634.
haematologica | 2018; 103(3)
BCR internalization and signaling in CLL
9. Mockridge CI, Potter KN, Wheatley I, Neville LA, Packham G, Stevenson FK. Reversible anergy of sIgM-mediated signaling in the two subsets of CLL defined by VH-gene mutational status. Blood. 2007;109(10):4424-4431. 10. Stevenson FK, Krysov S, Davies AJ, Steele AJ, Packham G. B-cell receptor signaling in chronic lymphocytic leukemia. Blood. 2011;118(16):4313-4320. 11. Jain P, Keating M, Wierda W, et al. Outcomes of patients with chronic lymphocytic leukemia after discontinuing ibrutinib. Blood. 2015;125(13):2062-2067. 12. Guieze R, Robbe P, Clifford R, et al. Presence of multiple recurrent mutations confers poor trial outcome of relapsed/refractory CLL. Blood. 2015; 126(18):2110-2117. 13. Woyach JA, Smucker K, Smith LL, Lozanski A, Zhong Y, Ruppert AS, et al. Prolonged lymphocytosis during ibrutinib therapy is associated with distinct molecular characteristics and does not indicate a suboptimal response to therapy. Blood. 2014 Mar 20;123(12):1810-1817. 14. Vuillier F, Dumas G, Magnac C, et al. Lower levels of surface B-cell-receptor expression in chronic lymphocytic leukemia are associated with glycosylation and folding defects of the mu and CD79a chains. Blood. 2005;105(7):2933-2940. 15. Williams GT, Venkitaraman AR, Gilmore DJ, Neuberger MS. The sequence of the mu transmembrane segment determines the tissue specificity of the transport of immunoglobulin M to the cell surface. J Exp Med. 1990;171(3):947-952. 16. Guo B, Zhang L, Chiorazzi N, Rothstein TL. IL-4 rescues surface IgM expression in chronic lymphocytic leukemia. Blood. 2016;128(4):553-562. 17. Aguilar-Hernandez MM, Blunt MD, Dobson R, et al. IL-4 enhances expression and function of surface IgM in CLL cells. Blood. 2016;127(24):3015-3025. 18. Krysov S, Potter KN, Mockridge CI, et al. Surface IgM of CLL cells displays unusual glycans indicative of engagement of antigen in vivo. Blood. 2010;115(21):4198-4205. 19. Coelho V, Krysov S, Steele A, et al. Identification in CLL of circulating intraclonal subgroups with varying B-cell receptor expression and function. Blood. 2013;122(15):2664-2672. 20. Herishanu Y, Perez-Galan P, Liu D, et al. The lymph node microenvironment promotes Bcell receptor signaling, NF-kappaB activa-
haematologica | 2018; 103(3)
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
tion, and tumor proliferation in chronic lymphocytic leukemia. Blood. 2011;117(2):563574. Ghia P, Strola G, Granziero L, et al. Chronic lymphocytic leukemia B cells are endowed with the capacity to attract CD4+, CD40L+ T cells by producing CCL22. Eur J Immunol. 2002;32(5):1403-1413. Patten PE, Buggins AG, Richards J, et al. CD38 expression in chronic lymphocytic leukemia is regulated by the tumor microenvironment. Blood. 2008;1 11(10):5173-5181. Os A, Burgler S, Ribes AP, et al. Chronic lymphocytic leukemia cells are activated and proliferate in response to specific T helper cells. Cell Rep. 2013;4(3):566-577. Pascutti MF, Jak M, Tromp JM, et al. IL-21 and CD40L signals from autologous T cells can induce antigen-independent proliferation of CLL cells. Blood. 2013; 122(17):30103019. Ghia P, Stamatopoulos K, Belessi C, et al. ERIC recommendations on IGHV gene mutational status analysis in chronic lymphocytic leukemia. Leukemia. 2007;21(1):13. Krober A, Seiler T, Benner A, et al. V(H) mutation status, CD38 expression level, genomic aberrations, and survival in chronic lymphocytic leukemia. Blood. 2002; 100(4):1410-1416. Cabezudo E, Carrara P, Morilla R, Matutes E. Quantitative analysis of CD79b, CD5 and CD19 in mature B-cell lymphoproliferative disorders. Haematologica. 1999; 84(5):413418. Calissano C, Damle RN, Hayes G, et al. In vivo intraclonal and interclonal kinetic heterogeneity in B-cell chronic lymphocytic leukemia. Blood. 2009;114(23):4832-4842. Pasikowska M, Walsby E, Apollonio B, et al. Phenotype and immune function of lymph node and peripheral blood CLL cells are linked to transendothelial migration. Blood. 2016;128(4):563-573. Blery M, Tze L, Miosge LA, Jun JE, Goodnow CC. Essential role of membrane cholesterol in accelerated BCR internalization and uncoupling from NF-kappa B in B cell clonal anergy. J Exp Med. 2006; 203(7):1773-1783. Vilen BJ, Nakamura T, Cambier JC. Antigenstimulated dissociation of BCR mIg from Igalpha/Ig-beta: implications for receptor desensitization. Immunity. 1999;10(2):239248. Hou P, Araujo E, Zhao T, et al. B cell antigen
33.
34. 35.
36.
37.
38.
39.
40.
41.
42.
receptor signaling and internalization are mutually exclusive events. PLoS Biol. 2006;4(7):e200. Lanemo Myhrinder A, Hellqvist E, Sidorova E, et al. A new perspective: molecular motifs on oxidized LDL, apoptotic cells, and bacteria are targets for chronic lymphocytic leukemia antibodies. Blood. 2008;111(7): 3838-3848. Goh LK, Sorkin A. Endocytosis of receptor tyrosine kinases. Cold Spring Harb Perspect Biol. 2013;5(5):a017459. Herman SE, Niemann CU, Farooqui M, et al. Ibrutinib-induced lymphocytosis in patients with chronic lymphocytic leukemia: correlative analyses from a phase II study. Leukemia. 2014;28(11):2188-2196. Herndon TM, Chen SS, Saba NS, et al. Direct in vivo evidence for increased proliferation of CLL cells in lymph nodes compared to bone marrow and peripheral blood. Leukemia. 2017;31(6):1340-1347. Kowalewski DJ, Schuster H, Backert L, et al. HLA ligandome analysis identifies the underlying specificities of spontaneous antileukemia immune responses in chronic lymphocytic leukemia (CLL). Proc Natl Acad Sci U S A. 2015;112(2):E166-175. Hall AM, Vickers MA, McLeod E, Barker RN. Rh autoantigen presentation to helper T cells in chronic lymphocytic leukemia by malignant B cells. Blood. 2005;105(5):20072015. Trojan A, Schultze JL, Witzens M, et al. Immunoglobulin framework-derived peptides function as cytotoxic T-cell epitopes commonly expressed in B-cell malignancies. Nat Med. 2000;6(6):667-672. Rezvany MR, Jeddi-Tehrani M, Rabbani H, et al. Autologous T lymphocytes may specifically recognize leukaemic B cells in patients with chronic lymphocytic leukaemia. Br J Haematol. 2000;111(2):608617. Goddard RV, Prentice AG, Copplestone JA, Kaminski ER. Generation in vitro of B-cell chronic lymphocytic leukaemia-proliferative and specific HLA class-II-restricted cytotoxic T-cell responses using autologous dendritic cells pulsed with tumour cell lysate. Clin Exp Immunol. 2001;126(1):16-28. Krackhardt AM, Harig S, Witzens M, Broderick R, Barrett P, Gribben JG. T-cell responses against chronic lymphocytic leukemia cells: implications for immunotherapy. Blood. 2002;100(1):167173.
505
ARTICLE
Plasma Cell Disorders
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):506-513
Outcome and survival of myeloma patients diagnosed 2008-2015. Real-world data on 4904 patients from the Swedish Myeloma Registry
Cecilie Hveding Blimark,1 Ingemar Turesson,2 Anna Genell,3 Lucia Ahlberg,4 Bo Björkstrand,5 Kristina Carlson,6 Karin Forsberg,7 Gunnar Juliusson,8 Olle Linder,9 Ulf-Henrik Mellqvist,1,10 Hareth Nahi11 and Sigurdur Y. Kristinsson12,13
Department of Hematology, Sahlgrenska University Hospital and Institution of Internal Medicine, Sahlgrenska Academy at University of Gothenburg, Sweden; 1Department of Hematology, Oncology and Radiation Physics, Skåne University Hospital, Lund-Malmö, Sweden; 3Regional Cancer Center West, Western Sweden Health Care Region, Gothenburg, Sweden; 4Division of Hematology, Linkoping University Hospital, Linkoping, Sweden; 5Internal Medicine /Hematology, Karolinska Institutet, Stockholm, Sweden; 6 Department of Hematology, Uppsala University Hospital, Sweden; 7Department of Hematology, Umeå University Hospital, Sweden; 8Hematology/Transplantation, Stem Cell Center, Lund University, Sweden; 9Department of Hematology, Örebro University Hospital, Sweden; 10Department of Hematology, Borås Hospital, Sweden; 11Division of Hematology, Karolinska University Hospital and Karolinska Institutet, Stockholm, Sweden; 12Department of Medicine and Division of Hematology, University of Iceland, Reykjavik, Iceland and 13Karolinska University Hospital and Karolinska Institutet, Stockholm, Sweden; for the Swedish Myeloma Registry 1
ABSTRACT
E
Correspondence: cecilie.blimark@vgregion.se
Received: August 4, 2017. Accepted: December 7, 2017. Pre-published: December 7, 2017. doi:10.3324/haematol.2017.178103 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/506 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
506
pidemiology and outcome of myeloma are mainly reported from large university centers and collaborative groups, and do not represent 'real-world' patients. The Swedish Myeloma Registry is a prospective population-based registry documenting characteristics, treatment and outcome in newly diagnosed myeloma, including asymptomatic and localized forms, with the purpose of improving disease management and outcome. This report presents information on patients diagnosed between 2008 and 2015, including data on first-line treatment in patients diagnosed up to 2014, with a follow up until December 2016. We present age-adjusted incidence, patients' characteristics at baseline, treatment, response, and survival. Baseline data were available with a 97% coverage in 4904 patients (median age 71 years, males 70 years, females 73 years; 72% were 65 years or older), and at 1-year follow up in 3558 patients with symptomatic disease (92% of patients initially reported). The age-adjusted incidence was 6.8 myeloma cases per 100,000 inhabitants per year. Among initially symptomatic patients (n=3988), 77% had osteolytic lesions or compression fractures, 49% had anemia, 18% impaired kidney function, and 13% hypercalcemia. High-dose therapy with autologous stem cell transplantation was given to 77% of patients aged up to 66 years, and to 22% of patients aged 66-70 years. In the study period, 68% received bortezomib, thalidomide, and/or lenalidomide as part of the first-line treatment, rising from 31% in 2008 to 81% in 2014. In active myeloma, the median relative survival of patients aged 65 years or under was 7.7 years, and 3.4 years in patients aged 66 years and over. Patients diagnosed with myeloma in more recent years were associated with significantly higher rates of complete or very good partial remission (P<0.05), and with a significantly higher survival, with a Hazard Ratio (HR) of 0.84 (95%CI: 0.77-0.92; P<0.05). There was a small, but significant survival benefit in patients treated at university hospitals (HR 0.93; 95%CI: 0.87-0.99; P<0.05). We report here on a near complete 'real-world' population of myeloma patients during an 8-year period; a period in which newer drugs were implemented into standard practice. The overall incidence and median age were both higher than in most previous studies, indicating a more complete coverage of older patients. Myeloma survival in Sweden is comparable to other large registry studies, and responses and survival improved during the study period. haematologica | 2018; 103(3)
Experiences from the Swedish Myeloma Registry
Introduction Over recent decades, new treatment options have emerged in myeloma, with great expectations of improved survival. The introduction of high-dose melphalan with autologous stem cell support (HDM-ASCT) and newer drugs, such as the immunomodulatory agents (thalidomide, lenalidomide, and pomalidomide), proteasome inhibitors (bortezomib and carfilzomib), monoclonal antibodies, and other classes, has led to a rapid implementation of these drugs under international guidelines.1-7 To date, most studies on myeloma are based on selected patients from large referral centers and collaborative groups, with defined inclusion and exclusion criteria. But these often omit elderly patients, and thus do not reflect the true 'real-world' population.8 Also, there is limited information available on the use of new therapies and their efficacy and tolerability in standard practice, supporting the need for representative population-based prospective studies on characteristics, diagnostics, treatment and outcome in myeloma patients. Survival data from cancer registries are available, but often lack information on baseline characteristics and treatment. EUROCARE, covering nearly 50% of patients diagnosed with plasma cell neoplasms in Europe in the period 2000-2007, reports an age-standardized 5-year relative survival (RS) of 39.2%, an increase from 29.8% in 1997. Outcome was significantly better in the younger patients (68.6% vs. 21.8% 5-year relative survival), and in women (40.4% vs. 38.1%).9 These results have later been confirmed by other cancer registry data.10-12 A 2010 Swedish study of retrospective data regarding baseline characteristics and treatment of consecutive patients in MalmĂś found a similar trend in improved survival, which correlated with the introduction of new treatment modalities.13,14 The Swedish Myeloma Registry was established in 2008, and the first Swedish guidelines on diagnostics and treatment of myeloma were published in 2010. This is the first report on our population-based data on characteristics, treatment and survival in Swedish myeloma patients diagnosed from January 2008 through December 2015.
Methods The Swedish Cancer Registry The Swedish Cancer Registry is a nation-wide compulsory dual-report system developed in 1958, which is supported by the personal identification code system used for all Swedish citizens which was established in 1947. First, all pathology specimens indicating malignancy are reported by the pathologist to the Regional Tumor Registry. Second, data on date and type of cancer diagnosis of all patients with a newly diagnosed cancer are reported by clinicians, with missing data actively requested to secure a high level of completeness. In a validation study, the completeness (95%) and diagnostic accuracy (98%) of the Swedish Cancer Registry was found to be very high for multiple myeloma patients.15
Table 1. Characteristics of active myeloma (MM) and smoldering myeloma (SMM) patients in the Swedish Myeloma Registry.
Characteristics Total, n (%) Diagnosis, n (%) Multiple myeloma Smoldering multiple myeloma Age in years at dx, median All Male Female Immunoglobuline class n (%) IgG IgA Bence-Jones MM Non-secretory MM IgD IgM Not known More than one Ig IgE
Patients 4904 (100%) 3988 (81.3%) 916 (18.6%) 71 71 73 2882 (58.8) 1033 (22.3) 688 (14.0) 143 (2.9) 19 (0.4) 14 (0.3) 23 (0.5) 41 (0.8) 1 (0.0)
n: number; dx: diagnosis.
through the compulsory Swedish Cancer Registry. Survival data are obtained from the Swedish Population Registry. Patients diagnosed by autopsy are included in the Swedish Cancer Registry, but not in the Swedish Myeloma Registry. The registry is publicly financed, and the patients are reported by treating physicians and nurses. Courses are held for those responsible for reporting patient data to assure coherent reporting in all regions and hospitals. These courses cover inclusion criteria, parameters, and the manual of the Swedish Myeloma Registry. Criteria for the diagnosis of active myeloma (MM), smoldering myeloma (SMM), plasmocytoma, and plasma cell leukemia are defined according to the International Myeloma Working Group (2003).16 Other gammopathies, such as monoclonal gammopathy of undetermined significance (MGUS) and AL-amyloidosis are not included in the registry. Age-specific incidence, age distribution at diagnosis, median time from diagnosis to registry report, and distribution of the diagnoses in the registry are reported. Adherence to treatment guidelines concerning diagnostics and ISS-staging (International Staging System) is checked by studying the use of different diagnostic tools such as bone marrow sample, cytogenetics including fluorescence in situ hybridization (FISH), β 2-microglobulin (β2m) and s-albumin. Baseline characteristics at diagnosis are collected, including M-protein isotype, percentage of plasma cells in the bone marrow, serum free-light chain (FLC), and laboratory parameters capturing CRAB criteria (CRAB; Calcium, Renal insufficiency, Anemia or Bone lesions). One year after diagnosis of symptomatic MM, data on first-line therapy, occurrence and date of first relapse or complications are requested. The study was performed in agreement with the ethics committee of Stockholm and the Swedish Society of Hematology.
The Swedish Myeloma Registry The Swedish Cancer Registry comprises web-reported clinical and laboratory data on all patients diagnosed with active myeloma, smoldering myeloma, plasma cell leukemia, and solitary bone and extramedullary plasmocytomas in Sweden since 2008, at time of diagnosis and after a 1-year follow up. Coverage is analyzed haematologica | 2018; 103(3)
Treatment of MM in Sweden In Sweden, patients with myeloma are typically diagnosed and followed clinically by physicians at hospital-based hematology centers, and no patients are seen at private hospitals. In the study period, the treatment of MM was guided by the British/Nordic 507
C.H. Blimark et al. A
B
Figure 1. Age distribution in the Swedish Myeloma Registry in men and women in (A) active myeloma (MM) and (B) smoldering myeloma (SMM). n: number.
treatment program for multiple myeloma (2005),17 and the Swedish 2010 National Guidelines (up-dated in 2013). Briefly, high-dose melphalan and autologous transplantation (HDMASCT) was recommended as up-front treatment for all MM patients aged 65 years or under, and in patients aged 66-70 years if they had good performance status. In 2005, vincristine, adriamycin, and dexamethasone (VAD) or similar combinations were recommended as induction treatment before HDM-ASCT, and later, in the 2010 guidelines, bortezomib and thalidomide became part of standard induction, following an introduction period subsequent to approval in 2004. Patients at smaller hospitals are, as a rule, only referred to university hospitals for the ASCT procedure and afterwards return to their hospital of origin. For patients aged 66 years and older, melphalan and prednisone (MP) or cyclophosphamide and dexamethasone (CyDex) was standard up-front treatment until 2004 when melphalan, prednisone and thalidomide (MPT) was incorporated as a treatment option. In 2010, MPT was the standard for patients not eligible for ASCT, and MP and bortezomib (MPV) were treatment options. In the 2013 version, both MPT and MPV were standard up-front treatments in those patients not eligible for ASCT.
Statistical analysis Incidence was extracted from the Swedish National Board of Health statistical database on cancer 1970-2015, which includes all patients with the diagnosis ICD 203*.18 All other analyses were performed on patients reported to the Myeloma Registry with a 97% coverage compared to the Swedish Cancer Registry.19 For diagnoses of MM and SMM, we summarized descriptive statistics 508
Table 2. Prevalence of myeloma-related organ and tissue impairment (ROTI) and International Staging System (ISS) stage at diagnosis in patients with active myeloma at diagnosis in the Swedish Myeloma Registry.
Patients n=3988 ROTI (%) Anemia* Renal impairment** Hypercalcemia*** Skeletal disease ISS stage (%) Stage I Stage II Stage III
49% 18% 13% 77% 23 44 33
n: number; in patients with report on: *anemia defined as hemoglobin < 10g/dL and reduction of 2g/dL from the normal value; **renal failure defined as creatinine >173 mol/L; ***hypercalcemia defined as s-calcium (uncorrected) > 2.75 mmol/L or ionized calcium>1.45 mmol/L.
at diagnosis. We tabulated categorical variables such as sex, Igclass and use of new drugs. Summary statistics, for example, median and range, were calculated for continuous variables such as age and β2M. The Ď&#x2021;2 test was used as significance test of difference in proportions. Statistical analysis of treatment was only carried out on MM patients with a reported 1-year follow up, including patients who had developed symptomatic disease after SMM haematologica | 2018; 103(3)
Experiences from the Swedish Myeloma Registry
A
B
Figure 2. Survival in active myeloma (MM) in the Swedish Myeloma Registry: observed (A) and relative (B) survival, by 10-year age cohorts. n: number.
or plasmacytoma. We estimated observed survival using the Kaplan-Meier method. When estimating relative survival (RS), relative to the general Swedish population, we used the Ederer II method for expected survival. For observed survival (OS), we estimated Hazard Ratios (HR) using Cox proportional hazards regression modeling. Also for RS, we estimated HR using proportional hazards regression, but in transformed time.20 Survival time was calculated from date of diagnosis to death or censoring. Patients were censored at the end of follow up in the study or loss to follow up. Age-standardized RS was calculated in each age group separately and then weighted together using weights from a standard population, in this case, the International Cancer Survival Standard (ICSS) 1. We used a proportional hazard model of RS by year of diagnosis in all patients to estimate changes in survival over time. The survival analysis by year of diagnosis included both SMM and MM, and the date of diagnosis refers to the date of the primary diagnosis, whether this was SMM or MM. To evaluate the impact of the treating hospital, we estimated a proportional hazard model of RS by hospital type, in the categories “university hospital” or “not”, and hospitals reporting treatment on more or less than 10 patients per year. The survival analysis by treatment haematologica | 2018; 103(3)
response and by hospital type was carried out on symptomatic MM patients only (including patients who had developed symptomatic disease after SMM or plasmacytoma) with reported 1year follow up, to enable comparison with statistics on treatment. When adjusting for ISS stage in regression analysis, we treated patients with missing values in the stage variable as a category within the ISS stage variable in order to not exclude the cohort of patients with missing data on ISS stage. P<0.05 was considered statistically significant. All data preparation and analysis were carried out using R statistical software.21
Results A total of 5222 patients with plasma cell diseases diagnosed in the period 2008-2015 had been reported to the Swedish Myeloma Registry as of December 31st 2016, with 97% coverage when compared with the Swedish Cancer Registry. Clinical data at diagnosis were available for 4904 MM and SMM patients diagnosed in the period 2008-2015 509
C.H. Blimark et al. Table 3. Proportion of patients who received novel drugs (thalidomide, bortezomib or lenalidomide) as first-line line treatment among active myeloma patients with reported follow up, by year of diagnosis and by age group (-65, 66–80, >80) years in the Swedish Myeloma Registry.
Patients with novel drugs first line
All ages n=2400 (%)
≤65 years n=913 (%)
66-80 years n=1212 (%)
>80 years n=275 (%)
67.5 31.1 56.1 69.1 75.2 77.0 81.0 81.1
81.3 24.4 76.8 91.6 93.8 98.1 95.6 92.2
72.6 42.0 55.7 74.1 76.0 83.3 88.1 86.2
35.6 17.7 24.8 32.6 34.5 37.3 49.2 54.3
2008-2014 2008 2009 2010 2011 2012 2013 2014*
*2014 has less follow up on patients reported (at data cut off 78.7% of initially reported).
Table 4. Proportion with very good partial remission (VGPR) or better among active myeloma patients with reported follow up after first-line treatment in patients diagnosed 2008-2014 in the Swedish Myeloma Registry, by year of diagnosis and by age group (-65, 66-80, >80 years).
Patients VGPR or better 2008-2014 2008 2009 2010 2011 2012 2013 2014*
All ages n (%)
≤65 years n (%)
66-80 years n (%)
>80 years n (%)
1415 (45.8) 152 (36.1) 173 (40.3) 209 (47.5) 223 (45.4) 223 (46.4) 230 (51.6) 205 (53.5)
725 (68.3) 87 (55.1) 97 (62.2) 104 (67.5) 123 (72.4) 114 (73.5) 117 (78.0) 83 (70.3)
575 (38.9) 56 (28.0) 58 (23.4) 86 (43.9) 91 (35.1) 91 (39.6) 94 (46.8) 99 (50.5)
115 (20.8) 9 (14.3) 18 (23.4) 19 (21.1) 9 (14.5) 18 (18.8) 19 (20.0) 23 (33.3)
*2014 has less follow up on patients reported (at data cut off 78.7% of initially reported).
(Table 1), and at 1-year follow up for 3558 of all MM cases diagnosed 2008-2014, being 92% of all MM initially reported 2008-2014. Data were reported from 74 different centers in Sweden, approximately 40% from university hospitals, and 60% from regional and smaller hospitals, all public care institutions. The median time of follow up of all SMM and MM patients was 4.9 years. The total crude and age-adjusted (to the population in Sweden in the year 2000) incidence was 7.0 and 6.8 cases per 100,000 inhabitants, respectively (8.0 and 8.2 for men, and 6.0 and 5.3 for women per 100,000 inhabitants, respectively). The corresponding incidences for European and World standard populations are 4.8 and 3.2, respectively. Due to the difference in age distribution in the population, the total number of women was higher in the cohort aged over 85 years (Figure 1). However, the agespecific incidence was higher amongst men in all ages, and the difference increased with advancing age (Online Supplementary Figure S1). The median age of patients reported to the registry with a diagnosis of MM or SMM was 71 years (70 years for men and 73 years for women; 71 years for all MM and 72 years for all SMM). Twentyfour percent of patients were 80 years or older at the time of diagnosis. Notably, the percentage of patients aged under 65 years was 28.3%; 61.4% of these were men and 38.6 women.
510
Baseline characteristics Serum protein electrophoresis was performed in 99.5% of all patients and a skeletal survey was performed in 97%. A bone marrow sample was taken in 97% of patients at diagnosis, with a median of 27% plasma cells in MM patients and 15% in SMM. Among patients with MM at diagnosis (n=3988), 77% had reported osteolytic lesions and/or compression fractures at diagnosis, and this did not increase over the study period. Anemia was seen in 49%, renal insufficiency (s-creatinine >173 μmol/L) in 18%, and creatinine levels more than 110 μmol/L were reported in 33% of MM patients. Hypercalcemia was reported in 13% of MM patients at the time of diagnosis (Table 2). The number of patients aged 80 years and under who had FISH performed at diagnosis increased over the study period, from 30% in the period 2008-2010, to 43% in 2011-2015. Staging according to the ISS was reported in 71% of patients with MM in the study period. In MM patients with reported ISS-stage, 23% were ISS stage I, 44% stage II, and 33% stage III (Table 2).
Treatment Of all patients with reported follow up, 77% of patients aged 65 years or under at diagnosis and 5% of patients aged over 66 years received HDM-ASCT as first-line treatment. In patients aged 66-70 years, HDM/ASCT was performed in 22%. Allogeneic transplantation as part of firstline treatment was performed in only 1% of patients in haematologica | 2018; 103(3)
Experiences from the Swedish Myeloma Registry
A
B
Figure 3. Relative survival in active myeloma (MM) by treatment response in the Swedish Myeloma Registry in age cohorts: (A) under 65 years of age and (B) 66 years of age and over. CR: complete remission; VGPR: very good partial remission; PR: partial remission; n: number.
the study period. A total of 5.2% of reported MM patients did not receive any anti-myeloma treatment the first year after diagnosis and, notably, this involved 11% of patients over 80 years of age. Bisphosphonates were given in 79% of patients aged 65 years or under, and in 67% in patients over 65 years of age. There was an increase in the use of one or more of the novel drugs (thalidomide, lenalidomide, and bortezomib) over the study period (Table 3).
Table S1 and Figure S2. Early death (<1 year after diagnosis) was observed in 19% of patients. The 3-year RS was 62% (95%CI: 59.7-64.6) in women, and 67% (95%CI: 65.069.3) in men. After age standardization, the 3-year RS in women was 67% (95%CI: 65.1-69.6) and 70% in men (95%CI: 67.8-71.8). Survival per SMM and MM diagnosis is shown in Online Supplementary Tables S2 and S3.
Survival in MM Response The proportion of patients achieving very good partial remission (VGPR) or better after first-line treatment increased from 36% in patients diagnosed in 2008 to 54% in 2014 (P<0.05). The increase was seen in all age groups, but was more pronounced in patients aged over 80 years, where the proportion of patients reaching VGPR or better rose from 14% to 33% (Table 4).
Survival in all myeloma patients The 1-, 3-, and 5-year OS in all patients (SMM+MM) was 81%, 59%, 42%, and the corresponding RS was 84%, 65%, and 49%, respectively. Survival in 10-year cohorts in all myeloma patients is shown in Online Supplementary haematologica | 2018; 103(3)
In patients with MM and reported follow up (n=3558), the median OS varied considerably depending on age at diagnosis, ranging from 7.8 years in patients aged 60 years and under, to 1.5 years for patients aged 80-89 years (Online Supplementary Table S4). After a median follow up of 5.5 years, the median OS in the youngest cohort (<50 years) had not yet been reached (Figure 2). The median RS of patients aged 65 years or under was 7.7 years, and 3.4 years in those aged 66 years and over. The 5-year OS and RS in MM patients was 38.3% and 44.9%, respectively. The median RS according to ISS stage was 3.2 years and 5.6 years for stages III and II, and 8.2 years for stage I. Patients with no reported stage had a similar median RS as stage III patients of three years. 511
C.H. Blimark et al.
Survival according to response Overall, better response to first-line treatment was significantly associated with superior survival (P<0.05) (Online Supplementary Table S5). In younger patients, there was no significant difference in 5-year RS in patients in PR, VGPR and CR (Figure 3)
Survival according to year of diagnosis Patients diagnosed in the period 2011-2015 had a trend to better 1-, 3- and 5-year RS compared to patients diagnosed 2008-2010. In patients aged over 65 years, this trend was more evident than in younger patients (Online Supplementary Tables S5 and S6, and Figure S3). In a proportional hazard model of RS by year of diagnosis in all patients, later calendar year of diagnosis was significantly associated with improved RS, with an HR of 0.93 (95%CI: 0.92-0.95; P<0.05).
Survival according to treating hospital The 1, -3 â&#x20AC;&#x201C; and 5-year survival was significantly higher in university hospitals (Online Supplementary Table S7). In a proportional hazards model for the RS, the HR was 0.93 (95%CI: 0.87-0.99; P<0.005). Even when adjusting for age, sex, and ISS-stage, the HR was of borderline statistical significance (HR=0.91; 95%CI: 0.83-1.0; P=0.04). Similar results were obtained when analyzing centers that treated 10 or more MM patients per year (data not shown).
Discussion In this study from the Swedish Myeloma Registry, we report incidence, baseline characteristics and survival of an unselected population comprising more than 97% of all myeloma patients diagnosed in Sweden in the period 2008-2015. We found an age-adjusted incidence of 6.8 per 100,000 inhabitants; this translates into 4.8 and 3.2 per 100,000 inhabitants in European and World standards, respectively. This is higher than figures previously reported by most population-based studies,22,23 but is in agreement with data from a previous large Swedish study.24 The high age-adjusted incidence might be explained by better case ascertainment in the elderly. Overall, the proportion of elderly (65 years and older) myeloma patients at diagnosis was 72%, and this exceeds the number of reported elderly patients in most known registries today, but is supported by population-based data from the Danish Myeloma Registry25 and a recent report on a large cohort of European patients.26 We observed a median age of 71 years at diagnosis, which is higher than other myeloma studies,8 and a steep increase in age-specific incidence extending to the oldest age cohorts. This indicates that our population, given the very high coverage provided by the Swedish Myeloma Registry, reflects the 'real-world' situation in myeloma today. Our study shows encouraging survival rates in the MM population. In our population-based study, the 5-year OS was 38%, similar to the data from the EUROCARE study.27 In a 2014 report from the Mayo clinic based on 1084 MM patients (median age 66 years), the median OS from diagnosis was 5.2 years and the 6-year OS estimate was 45%.8 We show that with the increased use of novel agents there was an improvement in response rates. We also show that, over the study period, the proportion of elderly patients receiving novel drugs increased. The difference in survival between the different age cohorts was 512
less pronounced in RS compared to OS, which demonstrates the importance of including RS in survival analyses in MM. In the European Registry data from 2008 (EUROCARE),28 a 2% survival advantage was seen in women. However, in our more recent study covering the period from 2008 to 2015, after age standardization, there was no difference in survival between men and women. As expected, and as shown before,29,30 achievement of response was predictive of prolonged survival. There was a significant difference in survival in patients aged over 65 years. Given this, we investigated the impact of response grade on survival in different age cohorts in patients with MM at diagnosis. The analysis revealed that responding patients in all age groups had a better outcome than nonresponding patients, and that patients achieving CR had the longest survival. However, interestingly, in patients aged 65 years or under there was no significant difference in survival according to the degree of response (CR, VGPR or PR). This is contrary to results from many randomized studies,31-33 and may indicate that achievement of a high quality response to first-line treatment may not have the same importance for survival in a young, unselected myeloma population where the majority of patients will eventually receive multiple lines of treatment. We found a survival benefit in patients reported from university hospitals and those hospitals treating a large number of MM patients. This is not surprising given the speed of progress in diagnostics and the new treatments of recent years, and has, in fact, also been reported in other studies.34,35 We could not detect a significant difference in referral patterns, but in spite of this, our results should be interpreted with caution, as residual confounding factors may have influenced outcome. However, this does underline the importance of high volume centers with expert knowledge in MM treatment and the need for further studies to monitor access to care for myeloma patients. The strength of this study is the large, population-based cohort and excellent coverage provided by the Swedish Cancer Registry. Another strength is the public Swedish health care system. In Sweden, all patients with a diagnosis of cancer are treated in public hospitals, enabling publicly financed and equal treatment for all MM patients; this reduced the risk of information- and selection-bias in this study. The Swedish Myeloma Registry has provided valuable information on how new treatments have been introduced and have been established as standard of care in clinical practice, leading to improved response rates in all age groups. Importantly, we have been able to show that there is good adherence to guidelines in all regions of Sweden, both with regards to diagnostics and to management, and the registry has helped define areas where improvement is needed. The proportion of patients with prognostic classification according to ISS and for whom FISH was performed as part of diagnostic workup has increased; however, FISH has still not been established as standard clinical practice in all hospitals. One limitation is that treatment data on 8% of patients were incomplete, and some baseline characteristics, such as ISS-stage, were also missing. In addition, we do not have detailed data on cytogenetics and comorbidities. Finally, we did not have sufficient follow-up data to perform analyses on progression-free survival after first-line treatment, which is a further limitation of this study. Many large and important studies on characteristics and haematologica | 2018; 103(3)
Experiences from the Swedish Myeloma Registry
survival in MM patients are compromised by the reporting bias of referral centers, either because they are university hospital registries with a low median age at MM diagnosis, or because they report on selected patients in clinical trials who do not necessarily, therefore, reflect the 'real-world' scenario in myeloma. Great efforts are being made to ensure the data available in the Swedish Myeloma Registry are complete, and to present popula-
References 1. Bird J, Behrens J, Westin J, et al. UK Myeloma Forum (UKMF) and Nordic Myeloma Study Group (NMSG): guidelines for the investigation of newly detected Mproteins and the management of monoclonal gammopathy of undetermined significance (MGUS). Br J Haematol. 2009;147(1):22-42. 2. Rajkumar SV. Multiple myeloma: 2013 update on diagnosis, risk-stratification, and management. Am J Hematol. 2013; 88(3):226-235. 3. Kristinsson SY, Anderson WF, Landgren O. Improved long-term survival in multiple myeloma up to the age of 80 years. Leukemia. 2014;28(6):1346-1348. 4. Gay F, Larocca A, Wijermans P, et al. Complete response correlates with longterm progression-free and overall survival in elderly myeloma treated with novel agents: analysis of 1175 patients, Blood. 2011;117(11):3025-3031. 5. San Miguel JF, Schlag R, Khuageva NK, et al. Bortezomib plus melphalan and prednisone for initial treatment of multiple myeloma. N Engl J Med. 2008;359(9):906-917. 6. Dimopoulos MA, Chen C, Spencer A, et al. Long-term follow-up on overall survival from the MM-009 and MM-010 phase III trials of lenalidomide plus dexamethasone in patients with relapsed or refractory multiple myeloma. Leukemia. 2009; 23(11):21472152. 7. Kristinsson SY, Landgren O, Dickman PW, Derolf AR, Bjorkholm M. Patterns of survival in multiple myeloma: a populationbased study of patients diagnosed in Sweden from 1973 to 2003. J Clin Oncol. 2007;25(15):1993-1999. 8. Kumar SK, Dispenzieri A, Lacy MQ, et al. Continued improvement in survival in multiple myeloma: changes in early mortality and outcomes in older patients. Leukemia. 2014;28(5):1122-1128. 9. De Angelis R, Minicozzi P, Sant M, et al. Survival variations by country and age for lymphoid and myeloid malignancies in Europe 2000-2007: Results of EUROCARE-5 population-based study. Eur J Cancer. 2015;51(15):2254-2268. 10. Kristinsson S, et al. Improved long-term survival in multiple myeloma up to the age of
haematologica | 2018; 103(3)
tion-based data on management and outcome in Sweden. However, we can now present a near complete 'realworld' population of myeloma patients, and show that the overall incidence and median age is higher than in most previous studies, indicating a more complete coverage of older patients. Myeloma survival in Sweden was similar to other large registry studies, and responses and survival improved over the study period.
80 years. Leukemia. 2014;28(6):1346-1348. 11. Pulte D, Gondos A, Brenner H. Improvement in Survival of Older Adults with Multiple Myeloma: Results of an Updated Period Analysis of SEER Data. Oncologist. 2011;16(11):1600-1603. 12. Brenner H, Gondos A, Pulte D. Recent major improvement in long-term survival of younger patients with multiple myeloma. Blood. 2008;111(5):2521-2526. 13. Turesson I, Velez R, Kristinsson SY, Landgren O. Patterns of multiple myeloma during the past 5 decades: stable incidence rates for all age groups in the population but rapidly changing age distribution in the clinic. Mayo Clin Proc. 2010;85(3):225-230. 14. Turesson I, Velez R, Kristinsson SY, Landgren O. Patterns of improved survival in patients with multiple myeloma in the twenty-first century: a population-based study. J Clin Oncol. 2010;28(5):830-834. 15. Turesson I, Linet MS, Bjorkholm M, et al. Ascertainment and diagnostic accuracy for hematopoietic lymphoproliferative malignancies in Sweden 1964-2003. Int J Cancer. 2007;121(10):2260-2266. 16. Greipp PR, San Miguel J, Durie BG, et al. International staging system for multiple myeloma. J Clin Oncol. 2005;23(15):34123420. 17. Smith A, Wisloff F, Samson D, et al. Guidelines on the diagnosis and management of multiple myeloma 2005. Br J Haematol. 2006;132(4):410-451. 18. Socialstyrelsen. Available from: http: //www.socialstyrelsen.se/statistics/statisticaldatabase/cancer. 19. The completeness of the Swedish Cancer Register â&#x20AC;&#x201C; a sample survey for year 1998. Acta Oncol. 2009;48(1):27-33. 20. Pohar M, Stare J. Relative survival analysis in R. Comput Programs Biomed. 2006;81(3):272-278. 21. R Core Team R Foundation for Statistical Computing V, Austria. R: A language and environment for statistical computing, 2016. 22. Alexander DD, Mink PJ, Adami HO, et al. Multiple myeloma: a review of the epidemiologic literature. Int J Cancer. 2007;120(Suppl 12):40-61. 23. 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.
24. Velez R, Turesson I, Landgren O, Kristinsson SY, Cuzick J. Incidence of multiple myeloma in Great Britain, Sweden, and Malmo, Sweden: the impact of differences in case ascertainment on observed incidence trends. BMJ. 2016;6(1):e009584. 25. Gimsing P, Holmstrom MO, Klausen TW, et al. The Danish National Multiple Myeloma Registry. Clin Epidemiol. 2016;8:583-587. 26. Yong K, Delforge M, Driessen C, et al. Multiple myeloma: patient outcomes in real-world practice. Br J Haematol. 2016; 175(2):252-264. 27. 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. 28. Micheli A, Ciampichini R, Oberaigner W, et al. The advantage of women in cancer survival: An analysis of EUROCARE-4 data. Eur J Cancer. 2009;45(6):1017-1027. 29. Lahuerta JJ, Paiva B, Vidriales MB, et al. Depth of Response in Multiple Myeloma: A Pooled Analysis of Three PETHEMA/GEM Clinical Trials. J Clin Oncol. 2017;35(25):2900-2910. 30. van de Velde H, Londhe A, Ataman O, et al. Association between complete response and outcomes in transplant-eligible myeloma patients in the era of novel agents. Eur J Haematol. 2017;98(3):269-279. 31. Landgren O, Iskander K. Modern multiple myeloma therapy: deep, sustained treatment response and good clinical outcomes. J Intern Med. 2017;281(4):365-382. 32. Landgren O, Iskander K. Modern multiple myeloma therapy: deep, sustained treatment response and good clinical outcomes. J Intern Med. 2017;281(4):365-382. 33. Lonial S, Anderson KC. Association of response endpoints with survival outcomes in multiple myeloma. Leukemia. 2014;28(2):258-268. 34. Go RS, Bartley AC, Crowson CS, et al. Association Between Treatment Facility Volume and Mortality of Patients With Multiple Myeloma. J Clin Oncol. 2017;35(6):598-604. 35. Ailawadhi S, Advani P, Yang D, et al. Impact of access to NCI- and NCCN-designated cancer centers on outcomes for multiple myeloma patients: A SEER registry analysis. Cancer. 2016;122(4):618-625.
513
ARTICLE
Plasma Cell Disorders
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):514-521
Melphalan 140 mg/m2 or 200 mg/m2 for autologous transplantation in myeloma: results from the Collaboration to Collect Autologous Transplant Outcomes in Lymphoma and Myeloma (CALM) study. A report by the EBMT Chronic Malignancies Working Party Holger W. Auner,1 Simona Iacobelli2, Giulia Sbianchi,2 Cora Knol-Bout,3 Didier Blaise,4 Nigel H. Russell,5 Jane F. Apperley,1 David Pohlreich,6 Paul V. Browne,7 Guido Kobbe,8 Cecilia Isaksson,9 Stig Lenhoff,10 Christof Scheid,11 Cyrille Touzeau,12 Esa Jantunen,13 Achilles Anagnostopoulos,14 Ibrahim Yakoub-Agha,15 Alina Tanase,16 Nicolaas Schaap,17 Wieslaw Wiktor-Jedrzejczak,18 Marta Krejci,19 Stefan O. Schönland,20 Curly Morris,21 Laurent Garderet22 and Nicolaus Kröger23
1 Department of Medicine, Imperial College London, UK; 2Department of Biology, Tor Vergata University of Rome, Italy; 3EBMT Data Office, Leiden, the Netherlands; 4Institut Paoli Calmettes, Marseille, France; 5Nottingham University, UK; 6Charles University Hospital, Prague, Czech Republic; 7St. James's Hospital, Trinity College Dublin, Ireland; 8Heinrich Heine Universität, Düsseldorf, Germany; 9Umeå University Hospital, Sweden; 10Skåne University Hospital, Lund, Sweden; 11University of Cologne, Germany; 12CHU Nantes, France; 13 Kuopio University Hospital, Finland; 14George Papanicolaou General Hospital, Thessaloniki, Greece; 15CHU de Lille, LIRIC, INSERM U995, France; 16Fundeni Clinical Institute, Bucharest, Romania; 17Radboud University Medical Centre, Nijmegen, the Netherlands; 18Medical University, Warsaw, Poland; 19University Hospital Brno, Czech Republic; 20University of Heidelberg, Germany; 21Queens University of Belfast, Northern Ireland; 22Hôpital Saint-Antoine, Paris, France and 23University Hospital Hamburg-Eppendorf, Hamburg, Germany
ABSTRACT
Correspondence: holger.auner04@imperial.ac.uk
Received: September 29, 2017. Accepted: December 1, 2017. Pre-published: December 7, 2017.
doi:10.3324/haematol.2017.181339 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/514 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
514
M
elphalan at a dose of 200 mg/m2 is standard conditioning prior to autologous hematopoietic stem cell transplantation for multiple myeloma, but a dose of 140 mg/m2 is often used in clinical practice in patients perceived to be at risk of excess toxicity. To determine whether melphalan 200 mg/m2 and melphalan 140 mg/m2 are equally effective and tolerable in clinically relevant patient subgroups we analyzed 1964 first single autologous transplantation episodes using a series of Cox proportional-hazards models. Overall survival, progression-free survival, cumulative incidence of relapse, non-relapse mortality, hematopoietic recovery and second primary malignancy rates were not significantly different between the melphalan 140 mg/m2 (n=245) and melphalan 200 mg/m2 (n=1719) groups. Multivariable subgroup analysis showed that disease status at transplantation interacted with overall survival, progression-free survival, and cumulative incidence of relapse, with a significant advantage associated with melphalan 200 mg/m2 in patients transplanted in less than partial response (adjusted hazard ratios for melphalan 200 mg/m2 versus melphalan 140 mg/m2: 0.5, 0.54, and 0.56). In contrast, transplantation in very good partial or complete response significantly favored melphalan 140 mg/m2 for overall survival (adjusted hazard ratio: 2.02). Age, renal function, prior proteasome inhibitor treatment, gender, or Karnofsky score did not interact with overall/progression-free survival or relapse rate in the melphalan dose groups. There were no significant survival or relapse rate differences between melphalan 200 mg/m2 and melphalan 140 mg/m2 patients with high-risk or standard-risk chromosomal abnormalities. In conclusion, remission status at the time of transplantation may favor the use of melphalan 200 mg/m2 or melphalan 140 mg/m2 for key transplant outcomes (NCT01362972).
haematologica | 2018; 103(3)
Melphalan dose pre-transplant in MM
Introduction High-dose chemotherapy followed by autologous hematopoietic stem cell transplantation (ASCT) has been the standard consolidation treatment for patients up to the age of 65 years with newly diagnosed multiple myeloma for over two decades. Initially, high-dose chemotherapy plus ASCT proved superior to conventional chemotherapy.1,2 More recently, the benefit of upfront high-dose chemotherapy plus ASCT has been confirmed in treatment approaches incorporating thalidomide analogs and proteasome inhibitors.3-6 High-dose chemotherapy plus ASCT is also commonly used in older patients over the age of 65 years.7,8 However, the superiority of ASCT over non-intensive therapies in older patients remains to be established.9,10 In the trials that demonstrated superiority of autologous transplantation over non-intensive approaches, patients received high-dose chemotherapy with melphalan at a dose of 200 mg/m2 (Mel200).1-6 Mel200 was less toxic than other high-dose combination regimens11,12 and associated with longer progression-free survival and overall survival in patients younger than 60 years, when compared to melphalan 100 mg/m2 in a tandem transplant approach.13 Mel200 has therefore been recommended and widely used as standard conditioning therapy for ASCT.14-16 Some studies have linked Mel200 to excess toxicity in older patients and those with renal impairment.17-19 Consequently, a dose of 140 mg/m2 (Mel140) is widely used in clinical practice in older patients and in patients with renal impairment.20-25 However, Mel140 was associated with inferior response or survival rates compared to Mel200 in two very recent studies.24,26 It therefore remains to be determined whether Mel140 and Mel200 are equally effective and tolerable across subgroups of patients. To address this question we analyzed outcomes of almost 2000 first single autologous transplants for multiple myeloma after conditioning with either Mel140 or Mel200 which were reported to the European Society for Blood and Marrow Transplantation (EBMT). The results of the study indicate that the selection of Mel200 versus Mel140 may have a significant effect on key transplant outcomes, including overall survival.
Methods Study criteria and data management The Collaboration to Collect Autologous Transplant Outcomes in Lymphoma and Myeloma (CALM) study (NCT01362972) is an observational clinical outcome analysis of a defined cohort of patients with lymphoma or multiple myeloma who underwent ASCT between 2008 and 2012, with data reported retrospectively to the EBMT. Patients were eligible for the CALM study if they were ≥18 years old and received their first autologous peripheral blood stem cell transplant using cells mobilized with one of the following mobilization regimens: plerixafor plus granulocyte colony-stimulating factor (G-CSF), plerixafor plus G-CSF plus chemotherapy, G-CSF plus chemotherapy, or G-CSF alone. For this non-planned subgroup analysis, patients were selected from the CALM study population in the EBMT registry if they had a diagnosis of multiple myeloma and received a first single ASCT. Tandem transplants (defined as an ASCT followed by a second transplant within 6 months of the first and no relapse/progression between the two transplants), and patients who received melphahaematologica | 2018; 103(3)
lan doses other than 200 or 140 mg/m2, were not included. A total of 2253 patients from the CALM study EBMT registry fulfilled these general criteria. We excluded 289 of these patients from further analysis because of missing or inconclusive data regarding subsequent transplants (n=213), relapse date (n=67), or renal function (n=9), resulting in a final study population of 1964 patients. The database for this study was closed on December 14, 2016. The study was performed in accordance with the principles of the Declaration of Helsinki and approved by the Chronic Malignancies Working Party of the EBMT, a non-profit scientific society representing more than 600 transplant centers mainly located in Europe. Data reported to the EBMT are entered, managed, and maintained in a central database with internet access housed in Leiden University Medical Center, the Netherlands. Each EBMT center is represented in this database, and all patients whose transplant data are reported by participating centers provide informed consent for transplant-related data to be used for research purposes in an anonymous way.
Statistical analysis Patients’ characteristics between the two groups (Mel140 and Mel200) were compared using the χ2 test for categorical variables and the Mann-Whitney test for continuous variables. P-values for variables with more than two levels refer to an overall test for the presence of any difference. Overall survival was defined as the time from the date of ASCT to death from any cause. Patients still alive were censored at their last follow up. Progression-free survival was defined as the time between transplantation and progression of disease or death, censoring patients who did not develop an event. The probabilities of overall survival and progressionfree survival were obtained using the Kaplan–Meier estimator and comparisons were made with the log-rank test. The probabilities of relapse (cumulative incidence of relapse) and death without prior relapse (non-relapse mortality) were calculated by the proper non-parametric estimator for outcomes with competing risk and comparisons made with the Gray test. These methods were also used to compute the cumulative incidence of second primary malignancy considering death without such a prior malignancy as a competing event. Cox proportional hazards models were used to estimate adjusted hazard ratios (HR) for Mel140 compared to Mel200 in terms of overall survival, progression-free survival and the cumulative incidence of relapse. Factors included in the multivariable analysis were age at transplant (<65 versus ≥65 years), renal function (normal glomerular filtration rate >50 mL/min versus impaired glomerular filtration rate ≤50 mL/min), prior proteasome inhibitor treatment (yes versus no), status of disease at transplant (complete response/very good partial response versus partial response versus less than partial response), Karnofsky performance score (<90 versus ≥90) and gender. Age was dichotomized with a cut-off of 65 years for comparability with other studies considering that Martingale residuals analysis did not suggest other cut-off points (data not shown). There was no evidence that exclusion of missing values from multivariable analysis induced any bias in the estimation of regression coefficients (data not shown). In order to explore any possible modification of the effect of the melphalan dose in different subgroups, we then fitted a secondary series of Cox models. Each model included melphalan dose, the selected adjustment variables, and the interaction between melphalan dose and one of the factors. This procedure returned estimated adjusted hazard ratios for Mel140 compared to Mel200 in each subgroup defined by the selected factors, and the results are shown in forest plots. Due to the partial availability of International Staging System (ISS) and cytogenetic data, the interactions of ISS stage and chro515
H.W. Auner et al.
mosomal abnormalities with melphalan dose were analyzed separately. Chromosomal abnormalities were classified as high-risk [t(4;14), t(14;16), and del(17)] or standard risk (all other cytogenetic findings). All P-values shown are from two-sided tests, and the reported confidence intervals (CI) refer to 95% boundaries. A P-value <0.05 was regarded as statistically significant. A value up to 0.2 was used to determine the significance of interaction terms.
Results Patient- and treatment-related characteristics Patient-related and treatment characteristics are shown in Table 1. Patients in the Mel140 group (n=245, 12.5%) were older than patients in the Mel200 group (n=1719, 87.5%) at the time of ASCT [median 64 years (range, 2773) versus 59 years (range, 25-76); P<0.001]. Compared to the Mel200 patients, those in the Mel140 group more often had light chain myeloma, were more often in ISS III, and were less often transplanted within 12 months of diagnosis. The two groups differed significantly in terms of body mass index, with a higher proportion of normal weight patients in the Mel140 group. Mel140 patients had received proteasome inhibitor-containing induction therapy more often, and a greater proportion had a Karnofsky score of <90. Finally, more Mel140 patients had impaired renal function, defined as a glomerular filtration rate of ≤50 mL/min, and a greater proportion of Mel140 patients underwent ASCT in partial remission or worse.
Efficacy Overall survival was not significantly different between the two melphalan dose groups (Mel140, median not reached; 95% CI, 70.6 months to indeterminate; Mel200, 78 months; 95% CI, 74.0 months to indeterminate) (Figure 1A). The overall adjusted hazard ratio (HR) for death from all causes was 1.10 (95% CI: 0.79-1.54; P=0.56) for the Mel140 group (Figure 1B). Multivariable analysis of different subgroups showed that age, renal function, prior proteasome inhibitor treatment, gender, or Karnofsky score did not interact with overall survival in the melphalan dose groups (Figure 1B). However, disease status at transplant significantly modified the risk of death (P=0.006). In patients transplanted in less than partial response, Mel200 was associated with a significant overall survival advantage (adjusted HR 0.5 for Mel200 versus Mel140). In contrast, transplantation in very good partial response/complete response significantly favored Mel140 (adjusted HR 2.02). Transplantation in partial response did not modify the effect of melphalan dose on overall survival (adjusted HR 0.98). The median progression-free survival was 29 months (95% CI: 24.6-33.7) in the Mel140 group and 26.3 months (95% CI: 24.6-28.1) in the Mel200 group (Figure 2A). The adjusted HR for disease progression or death was 1.0 (95% CI: 0.79-1.25; P=0.98) for the Mel140 group (Figure 2B). The multivariable models with interaction terms indicated that the HR of Mel200 versus Mel140 was not significantly modified by age, renal function, prior proteasome inhibitor treatment, gender, or Karnofsky score. However, in line with the overall survival analysis, there was a statistically significant change (P=0.043) according to disease status at transplantation. Among the patients transplanted in partial response or less, Mel200 was associated with a 516
significant progression-free survival advantage (adjusted HR 0.54 for Mel200 versus Mel140), while Mel140 was linked to a numerically better outcome in those trans-
Table 1. Patient- and transplant-related characteristics.
Patients’ characteristics
Mel140 n=245 (12.5%)
P value
997 (58.0%) 722 (42.0%)
144 (58.8%) 101 (41.2%)
0.818
59 (25–76)
64 (27–73)
0.001*
15 (1.1%) 484 (35.2%) 587 (42.7%) 216 (15.7%) 72 (5.2%)
4 (2.0%) 95 (48.5%) 66 (33.7%) 22 (11.2%) 9 (4.6%)
0.004*
1117 (71.9%) 436 (28.1%)
139 (61.8%) 86 (38.2%)
0.002*
1316 (96.3%) 51 (3.7%)
123 (63.1%) 72 (36.9%)
0.001*
159 (65.7%) 79 (32.6%) 4 (1.7%)
0.001*
31 (23.5%) 48 (36.4%) 53 (40.2%)
0.001*
78 (85.7%) 13 (14.3%)
0.979
7 (3.4%) 35 (17.1%)
0.001*
Mel200 n=1719 (87.5%)
Gender Male Female Age at ASCT (median and range) years Body mass index (BMI) Underweight (BMI <18.5) Normal weight (18.5 ≤ BMI < 25) Overweight (25 ≤ BMI < 30) Obese (30 ≤ BMI < 35) Severely obese (BMI ≥ 35) Karnofsky score ≥ 90 < 90 Estimated GFR at ASCT >50 mL/min ≤50 mL/min
Myeloma characteristics Myeloma type Common type 1328 (78.9%) Light chain 328 (19.5%) Non-secretory 27 (1.6%) International Staging System stage I 453 (43.1%) II 373 (35.5%) III 225 (21.4%) High-risk chromosomal abnormalities# No 466 (85.8%) Yes 77 (14.2%)
Treatment-related parameters Pre-transplant treatment Alkylating agent(s) 155 (10.6%) Alkylating agent(s) 236 (16,2%) + proteasome inhibitor Alkalyting agent(s) + IMiD(s) 322 (22.1%) Alkylating agent(s) + proteasome 88 (6.0%) inhibitor + IMiD(s) Proteasome inhibitor 233 (16.0%) IMiD(s) 92 (6.3%) Proteasome inhibitor + IMiD(s) 280 (19.2%) Other 50 (3.4%) Time from diagnosis to ASCT < 12 months 1366 (79.5%) > 12 months 353 (20.5%) Disease status at ASCT CR/VGPR 765 (45.2%) PR 797 (47.1%) <PR 130 (7.7%) + CD34 cells infused (per kg) < 3 x 106 405 (33.4%) 3 – 5 x 106 508 (41.9%) > 5 x 106 299 (24.7%)
37 (18.0%) 14 (6.8%) 44 (21.5%) 4 (2.0%) 54 (26.3%) 10 (4.9%) 179 (73.1%) 66 (26.9%)
0.022*
105 (43.9%) 103 (43.1%) 31 (13.0%)
0.020*
79 (40.9%) 71 (36.8%) 43 (22.3%)
0.123
*Denotes statistically significant P values (<0.05); # includes t(4;14), t(14;16), or del(17p). ASCT: autologous stem cell transplantation; GFR: glomerular filtration rate; IMiD: immunomodulatory drug; CR: complete response; VGPR: very good partial response; PR: partial response.
haematologica | 2018; 103(3)
Melphalan dose pre-transplant in MM
planted in very good partial response/complete response (adjusted HR 1.19). The cumulative incidence of relapse at 3 years was not significantly different between the Mel140 (55.1%; 95 % CI: 48.6-61.6) and Mel200 (59.9%; 95% CI: 57.5-62.3) groups (Figure 3A). The adjusted HR for relapse was 0.99 (95% CI: 0.78-1.25; P=0.935) for Mel140 (Figure 3B). Subgroup analysis again showed a significant interaction of melphalan dose with disease status at the time of ASCT (P=0.07), in that transplantation in partial response or less significantly favored Mel200 (adjusted HR 0.56 for Mel200 versus Mel140). The adjusted HR for transplantation in partial response was 0.98 while that for transplantation in very good partial response/complete response was 1.2. Patients with high-risk chromosomal abnormalities had poorer overall and progression-free survival, and a higher cumulative incidence of relapse, compared with those with other chromosomal aberrations, but we observed no statistically significant differences between Mel140 and Mel200 in high-risk or standard-risk patients (Figure 4). Similarly, while ISS stage was associated with overall and
progression-free survival, and cumulative incidence of relapse (Online Supplementary Figure S1), there was no interaction between melphalan dose and ISS stage (Online Supplementary Figure S2).
Toxicity Non-relapse mortality was not significantly different between the Mel140 and Mel200 groups [1-year nonrelapse mortality 1.3% (95% CI: 0.0-2.7) and 0.9% (95% CI: 0.4 â&#x20AC;&#x201C; 1.3), respectively; P=0.20]. The early non-relapse mortality rate at 3 months after ASCT was not significantly different either (0.8 and 0.5%, respectively; P=0.198). The main cause of death within 12 months of the transplant was relapse/progression, being the cause in 77.8% of patients in the Mel140 group and 80.0% of patients in the Mel200 group. Median times to neutrophil and platelet recovery were not significantly different between the Mel140 and Mel200 groups, being 12 (Mel140 95% CI: 1213; Mel200 95% CI: 12-12) days in both groups for neutrophil recovery (P=0.283) and 16 (95% CI: 15-17) and 15 (95% CI: 15-16) days for platelet recovery (P=0.468). Second primary malignancy rates 5 years after ASCT were
A
B Figure 1. Overall survival after autologous stem cell transplantation for patients who received conditioning with melphalan 140 mg/m2 (Mel140) or 200 mg/m2 (Mel200). (A) KaplanMeier curves. (B) Multivariate analyses. The overall hazard ratio and corresponding P value (bottom line) refers to the comparison between Mel140 and Mel200 performed in a Cox model without interactions; the other hazard ratios and P values refer to interaction terms between melphalan dose and the indicated factors. All comparisons are adjusted for: age at transplant, renal function, prior proteasome inhibitor treatment, gender, status of disease, and Karnofsky score. CR/VGPR: complete response/very good partial response; PR: partial response.
haematologica | 2018; 103(3)
517
H.W. Auner et al.
very similar between the Mel140 (4.8%; 95% CI: 1.1-8.5) and Mel200 groups (4.8%; 95% CI: 3.6-6.0) (P=0.61).
Discussion While Mel200 is generally recommended as standard conditioning prior to ASCT for multiple myeloma,27 Mel140 is often used in clinical practice in those perceived to be at risk of excess toxicity from Mel200. However, the effect of melphalan dose on transplant outcomes remains undetermined. Here we present data from a large number of patients undergoing ASCT as part of real-world treatment practice. These data suggest that remission status at the time of transplantation may favor Mel200 or Mel140 for key transplant outcomes including overall survival. One of the key findings of the study is that transplantation in less than partial response favored Mel200 over Mel140 in terms of overall survival, progression-free survival, and relapse risk. This may be explained by a greater dose-dependency of melphalan-induced anti-myeloma effects in cells with limited chemosensitivity. However, we observed no benefit of Mel200 over Mel140 for patients with high-risk chromosomal aberrations or higher ISS
stage. Thus, while the better outcomes with Mel200 may at least partly be explained by the ability of the higher dose to overcome clinical resistance to induction therapies, Mel200 does not overcome the effects of poor-risk cytogenetics or advanced ISS stage. While the number of patients with known high-risk aberrations in our study was limited, the findings may be considered in line with preliminary data from an ongoing study which suggest a possible benefit of tandem ASCT for high-risk patients.28 It remains to be determined whether molecular risk profiles other than those based on cytogenetic findings, or clinical features such as extramedullary disease, favor Mel200. These data were not available for analysis in this study. In contrast to transplantation in poor clinical responders to induction therapy, transplantation in very good partial response/complete response appeared to favor Mel140. Considering that Mel200 was not linked to delayed hematopoietic recovery, increased early or late nonrelapse mortality, or second primary malignancy rate, an explanation for these findings is not apparent. It is conceivable that Mel200 resulted in moderately increased toxicities that were not clinically apparent or were not captured in our study, such as delayed physical recovery, or organ-specific toxicities such as cardiac arrhythmias.26
A
B Figure 2. Progression-free survival after autologous stem cell transplantation for patients who received conditioning with melphalan 140 mg/m2 (Mel140) or 200 mg/m2 (Mel200). (A) KaplanMeier curves. (B) Multivariate analyses. The overall hazard ratio and corresponding P value (bottom line) refers to the comparison between Mel140 and Mel200 performed in a Cox model without interactions; the other hazard ratios and P values refer to interaction terms between melphalan dose and the indicated factors. All comparisons are adjusted for: age at transplant, renal function, prior proteasome inhibitor treatment, gender, status of disease, and Karnofsky score. CR/VGPR: complete response/very good partial response; PR: partial response.
518
haematologica | 2018; 103(3)
Melphalan dose pre-transplant in MM
Such effects may have affected physiciansâ&#x20AC;&#x2122; and patientsâ&#x20AC;&#x2122; attitude towards the nature, intensity, or duration of posttransplant treatment. However, they are not likely to fully explain the favorable outcomes linked to Mel140 in patients transplanted in very good partial response/complete response. Differences in melphalan pharmacokinetics in certain subgroups of patients may also have accounted for some of the effects we observed, given that diverse factors such as creatinine clearance, fat free mass and hematocrit influence melphalan exposure.29 Melphalan exposure can vary considerably in myeloma patients treated with high-dose melphalan and ASCT, and higher exposure has been linked to greater toxicity and better disease responses.30 In a recent study, high melphalan exposure was associated with significantly improved overall survival in myeloma patients undergoing ASCT.31 However, despite the clear survival benefit, melphalan exposure was not associated with time to progression or progressionfree survival, suggesting a possible link between melphalan exposure and long-term outcomes that is not directly attributed to immediate anti-myeloma effects.
While some studies have suggested that older age and renal impairment can be linked to excess toxicity with Mel200, others have not reported such an association.19,20,2224,26,32-34 The results of this study support the notion that older age and impaired renal function do not favor the use of a lower melphalan dose with regards to non-relapse mortality, hematopoietic recovery, or second primary malignancy rate. Moreover, we found no interaction of Karnofsky performance score with melphalan dose. However, the paucity in our study of data on comorbidities and frailty scores, and on the nature and grading of specific adverse events such as mucositis, means that we cannot exclude that Mel200 may be linked to an increase in some toxic effects in certain patients, or that Mel140 may have avoided such effects. Nonetheless, the data provide further support for the notion that ASCT is safe and effective in fit, older patients, and they are in line with those of a recent study demonstrating the value of Mel140 tandem ASCT as an independent component of therapy in older patients.35 The application of objective criteria to determine patientsâ&#x20AC;&#x2122; fitness in the context of co-morbidi-
A
B Figure 3. Cumulative incidence of relapse after autologous stem cell transplantation for patients who received conditioning with melphalan 140 mg/m2 (Mel140) or 200 mg/m2 (Mel200). (A) Kaplan-Meier curves. (B) Multivariate analyses. The overall hazard ratio and corresponding P-value (bottom line) refers to the comparison between Mel140 and Mel200 performed in a Cox model without interactions; the other hazard ratios and P values refer to interaction terms between melphalan dose and the indicated factors. All comparisons are adjusted for: age at transplant, renal function, prior proteasome inhibitor treatment, gender, status of disease, and Karnofsky score. CR/VGPR: complete response/very good partial response; PR: partial response.
haematologica | 2018; 103(3)
519
H.W. Auner et al. A
B
C
Figure 4. Survival and relapse risk by cytogenetic risk. (A) Overall survival, (B) progression-free survival and (C) relapse risk estimates and confidence intervals at 2 years after ASCT are shown for patients with high-risk or standard-risk chromosomal abnormalities.
ties should aid optimal selection of both younger and older patients.36-40 The CALM study is based on the retrospective analysis of registry data that were collected in a defined cohort of patients. Thus, the choice of Mel140 or Mel200 was made by transplant physicians and influenced or determined by local practice, thereby introducing a potential for biased treatment decisions. The paucity of data on post-transplant treatments, including maintenance, is another limitation of the analysis. However, the large number of patients from multiple centers across Europe is likely to have formed a representative â&#x20AC;&#x2DC;real-worldâ&#x20AC;&#x2122; sample of myeloma patients undergoing up-front ASCT. This notion is supported by the distribution of baseline clinical and cytogenetic features and the outcomes of patients with high-risk compared to standard-risk disease. Moreover, we applied robust statistical methods for the estimation of hazard ratios. Our data indicate that the vast majority of patients undergoing upfront ASCT in a real-world setting receive Mel200 conditioning, and that patients with poor clinical responses to induction therapies derive more benefit from
References 1. Attal M, Harousseau JL, Stoppa AM, et al. A prospective, randomized trial of autologous and transplantation marrow bone chemotherapy in multiple myeloma.
520
Mel200 than from Mel140. However, the results of this study also indicate that transplantation in very good partial response/complete response may favor Mel140 over Mel200. While the reasons for this unexpected finding remain to be determined, the data raise the challenging question of whether more patients should receive Mel140. This is relevant given that modern induction regimens achieve high very good partial response/complete response rates. The data presented here suggest that a randomized trial to define the optimal melphalan dose is warranted. Such a trial could also investigate the use of alternative conditioning approaches that incorporate novel agents41-44 and the potential role of melphalan dosing in tandem transplant approaches, which was not feasible in this analysis. In the meantime, although Mel200 should remain the standard of care for ASCT conditioning, Mel140-based transplants could be considered as a valid alternative to offer patients an effective combination of ASCT plus novel therapies.45 In conclusion, our findings indicate that remission status at the time of a first ASCT may need to be considered when deciding the melphalan dose.
Intergroupe Francais du Myelome. N Engl J Med. 1996;335(2):91-97. 2. Child JA, Morgan GJ, Davies FE, et al. Highdose chemotherapy with hematopoietic stem-cell rescue for multiple myeloma. N Engl J Med. 2003;348(19):1875-1883. 3. Palumbo A, Cavallo F, Gay F, et al.
Autologous transplantation and maintenance therapy in multiple myeloma. N Engl J Med. 2014;371(10):895-905. 4. Gay F, Oliva S, Petrucci MT, et al. Chemotherapy plus lenalidomide versus autologous transplantation, followed by lenalidomide plus prednisone versus
haematologica | 2018; 103(3)
Melphalan dose pre-transplant in MM
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16. 17.
18.
lenalidomide maintenance, in patients with multiple myeloma: a randomised, multicentre, phase 3 trial. Lancet Oncol. 2015;16(16): 1617-1629. Attal M, Lauwers-Cances V, Hulin C, et al. Lenalidomide, bortezomib, and dexamethasone with transplantation for myeloma. N Engl J Med. 2017;376(14):1311-1320. Gay F, Oliva S, Petrucci MT, et al. Autologous transplant vs oral chemotherapy and lenalidomide in newly diagnosed young myeloma patients: a pooled analysis. Leukemia. 2017;31(8):1727-1734. Auner HW, Szydlo R, Hoek J, et al. Trends in autologous hematopoietic cell transplantation for multiple myeloma in Europe: increased use and improved outcomes in elderly patients in recent years. Bone Marrow Transplant. 2015;50(2):209-215. Costa LJ, Zhang MJ, Zhong X, et al. Trends in utilization and outcomes of autologous transplantation as early therapy for multiple myeloma. Biol Blood Marrow Transplant. 2013;19(11):1615-1624. Facon T, Mary JY, Hulin C, et al. Melphalan and prednisone plus thalidomide versus melphalan and prednisone alone or reducedintensity autologous stem cell transplantation in elderly patients with multiple myeloma (IFM 99-06): a randomised trial. Lancet. 2007;370(9594):1209-1218. Palumbo A, Bringhen S, Petrucci MT, et al. Intermediate-dose melphalan improves survival of myeloma patients aged 50 to 70: results of a randomized controlled trial. Blood. 2004;104(10):3052-3057. Lahuerta JJ, Mateos MV, Martinez-Lopez J, et al. Busulfan 12 mg/kg plus melphalan 140 mg/m2 versus melphalan 200 mg/m2 as conditioning regimens for autologous transplantation in newly diagnosed multiple myeloma patients included in the PETHEMA/GEM2000 study. Haematologica. 2010;95(11):1913-1920. Moreau P, Facon T, Attal M, et al. Comparison of 200 mg/m(2) melphalan and 8 Gy total body irradiation plus 140 mg/m(2) melphalan as conditioning regimens for peripheral blood stem cell transplantation in patients with newly diagnosed multiple myeloma: final analysis of the Intergroupe Francophone du Myelome 9502 randomized trial. Blood. 2002;99(3):731-735. Palumbo A, Bringhen S, Bruno B, et al. Melphalan 200 mg/m(2) versus melphalan 100 mg/m(2) in newly diagnosed myeloma patients: a prospective, multicenter phase 3 study. Blood. 2010;115(10):1873-1879. Jagannath S, Vesole DH, Glenn L, Crowley J, Barlogie B. Low-risk intensive therapy for multiple myeloma with combined autologous bone marrow and blood stem cell support. Blood. 1992;80(7):1666-1672. Shah N, Callander N, Ganguly S, et al. Hematopoietic stem cell transplantation for multiple myeloma: guidelines from the American Society for Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2015;21(7):1155-1166. Giralt S. 200 mg/m(2) melphalan--the gold standard for multiple myeloma. Nat Rev Clin Oncol. 2010;7(9):490-491. Badros A, Barlogie B, Siegel E, et al. Autologous stem cell transplantation in elderly multiple myeloma patients over the age of 70 years. Br J Haematol. 2001;114(3):600607. Badros A, Barlogie B, Siegel E, et al. Results
haematologica | 2018; 103(3)
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
of autologous stem cell transplant in multiple myeloma patients with renal failure. Br J Haematol. 2001;114(4):822-829. Sweiss K, Patel S, Culos K, Oh A, Rondelli D, Patel P. Melphalan 200 mg/m2 in patients with renal impairment is associated with increased short-term toxicity but improved response and longer treatment-free survival. Bone Marrow Transplant. 2016;51(10):13371341. Katragadda L, McCullough LM, Dai Y, et al. Effect of melphalan 140 mg/m(2) vs 200 mg/m(2) on toxicities and outcomes in multiple myeloma patients undergoing single autologous stem cell transplantation-a single center experience. Clin Transplant. 2016;30(8):894-900. Dhakal B, Nelson A, Guru Murthy GS, et al. Autologous hematopoietic cell transplantation in patients with multiple myeloma: effect of age. Clin Lymphoma Myeloma Leuk. 2017;17(3):165-172. Gertz MA, Lacy MQ, Dispenzieri A, et al. Impact of age and serum creatinine value on outcome after autologous blood stem cell transplantation for patients with multiple myeloma. Bone Marrow Transplant. 2007;39(10):605-611. Kumar SK, Dingli D, Lacy MQ, et al. Autologous stem cell transplantation in patients of 70 years and older with multiple myeloma: results from a matched pair analysis. Am J Hematol. 2008;83(8):614-617. Garderet L, Beohou E, Caillot D, et al. Upfront autologous stem cell transplantation for newly diagnosed elderly multiple myeloma patients: a prospective multicenter study. Haematologica. 2016;101(11):13901397. Dimopoulos MA, Terpos E, Chanan-Khan A, et al. Renal impairment in patients with multiple myeloma: a consensus statement on behalf of the International Myeloma Working Group. J Clin Oncol. 2010;28(33):4976-4984. Muchtar E, Dingli D, Kumar S, et al. Autologous stem cell transplant for multiple myeloma patients 70 years or older. Bone Marrow Transplant. 2016;51(11):1449-1455. Moreau P, San Miguel J, Sonneveld P, et al. Multiple myeloma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2017;28(suppl_4):iv52iv61. Cavo M, Petrucci MT, Di Raimondo F, et al. Upfront single versus double autologous stem cell transplantation for newly diagnosed multiple myeloma: an intergroup, multicenter, phase III study of the European Myeloma Network (EMN02/HO95 MM Trial). Blood. 2016;128(22):991. Nath CE, Shaw PJ, Trotman J, et al. Population pharmacokinetics of melphalan in patients with multiple myeloma undergoing high dose therapy. Br J Clin Pharmacol. 2010;69(5):484-497. Shaw PJ, Nath CE, Lazarus HM. Not too little, not too much-just right! (better ways to give high dose melphalan). Bone Marrow Transplant. 2014;49(12):1457-1465. Nath CE, Trotman J, Tiley C, et al. High melphalan exposure is associated with improved overall survival in myeloma patients receiving high dose melphalan and autologous transplantation. Br J Clin Pharmacol. 2016;82(1):149-159. Merz M, Neben K, Raab MS, et al. Autologous stem cell transplantation for eld-
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
erly patients with newly diagnosed multiple myeloma in the era of novel agents. Ann Oncol. 2014;25(1):189-195. Bashir Q, Shah N, Parmar S, et al. Feasibility of autologous hematopoietic stem cell transplant in patients aged >/=70 years with multiple myeloma. Leuk Lymphoma. 2012;53(1):118-122. Auner HW, Garderet L, Kroger N. Autologous haematopoietic cell transplantation in elderly patients with multiple myeloma. Br J Haematol. 2015;171(4):453-462. Straka C, Liebisch P, Salwender H, et al. Autotransplant with and without induction chemotherapy in older multiple myeloma patients: long-term outcome of a randomized trial. Haematologica. 2016;101(11): 1398-1406. Engelhardt M, Ihorst G, Caers J, Gunther A, Wasch R. Autotransplants in older multiple myeloma patients: hype or hope in the era of novel agents? Haematologica. 2016;101(11):1276-1278. Zweegman S, Engelhardt M, Larocca A, EHA SWG on "Aging and Hematology". Elderly patients with multiple myeloma: towards a frailty approach? Curr Opin Oncol. 2017;29(5):315-321. Engelhardt M, Domm AS, Dold SM, et al. A concise revised Myeloma Comorbidity Index as a valid prognostic instrument in a large cohort of 801 multiple myeloma patients. Haematologica. 2017;102(5):910921. Bruno B, Auner HW, Gahrton G, et al. Stem cell transplantation in multiple myeloma and other plasma cell disorders (report from an EBMT preceptorship meeting). Leuk Lymphoma. 2016;57(6):1256-1268. Palumbo A, Bringhen S, Mateos MV, et al. Geriatric assessment predicts survival and toxicities in elderly myeloma patients: an International Myeloma Working Group report. Blood. 2015;125(13):2068-2074. Roussel M, Moreau P, Huynh A, et al. Bortezomib and high-dose melphalan as conditioning regimen before autologous stem cell transplantation in patients with de novo multiple myeloma: a phase 2 study of the Intergroupe Francophone du Myelome (IFM). Blood. 2010;115(1):32-37. Blanes M, Lahuerta JJ, Gonzalez JD, et al. Intravenous busulfan and melphalan as a conditioning regimen for autologous stem cell transplantation in patients with newly diagnosed multiple myeloma: a matched comparison to a melphalan-only approach. Biol Blood Marrow Transplant. 2013;19(1):69-74. Rodriguez TE, Hari P, Stiff PJ, Smith SE, Sterrenberg D, Vesole DH. Busulfan, melphalan, and bortezomib versus high-dose melphalan as a conditioning regimen for autologous hematopoietic stem cell transplantation in multiple myeloma. Biol Blood Marrow Transplant. 2016;22(8): 1391-1396. Nieto Y, Valdez BC, Pingali SR, et al. Highdose gemcitabine, busulfan, and melphalan for autologous stem-cell transplant in patients with relapsed or refractory myeloma: a phase 2 trial and matched-pair comparison with melphalan. Lancet Haematol. 2017;4(6):e283-e292. Morgan GJ, Rasche L. Haematological cancer: Where are we now with the treatment of multiple myeloma? Nat Rev Clin Oncol. 2017;14(8):461-462.
521
ARTICLE
Stem Cell Transplantation
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):522-530
Bortezomib-based immunosuppression after reduced-intensity conditioning hematopoietic stem cell transplantation: randomized phase II results John Koreth,1 Haesook T. Kim,2 Paulina B. Lange,1 Samuel J. Poryanda,1 Carol G. Reynolds,1 Sharmila Chamling Rai,1 Philippe Armand,1 Corey S. Cutler,1 Vincent T. Ho,1 Brett Glotzbecker,1 Rushdia Yusuf,1 Sarah Nikiforow,1 Yi-Bin Chen,3 Bimalangshu Dey,3 Malgorzata McMasters,4 Jerome Ritz,1 Bruce R. Blazar,5 Robert J. Soiffer,1 Joseph H. Antin1 and Edwin P. Alyea III1
Division of Hematologic Malignancies, Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA; 2Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute and Harvard School of Public Health, Boston, MA; 3 Massachusetts General Hospital Cancer Center, Boston, MA; 4Beth Israel Deaconess Hospital Cancer Center, Boston, MA and 5University of Minnesota Masonic Cancer Center and Department of Pediatrics, Division of Blood and Marrow Transplantation, Minneapolis, MN, USA 1
ABSTRACT
A
Correspondence: john_koreth@dfci.harvard.edu
Received: July 18, 2017. Accepted: January 10, 2018. Pre-published: January 11, 2018. doi:10.3324/haematol.2017.176859 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/522 Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
522
prior phase I/II trial of bortezomib/tacrolimus/methotrexate prophylaxis after human leukocyte antigen (HLA)-mismatched reduced intensity conditioning allogeneic hematopoietic stem cell transplantation documented low acute graft-versus-host disease incidence, with promising overall and progression-free survival. We performed an open-label three-arm 1:1:1 phase II randomized controlled trial comparing grade II-IV acute graft-versus-host disease between conventional tacrolimus/methotrexate (A) versus bortezomib/ tacrolimus/methotrexate (B), and versus bortezomib/sirolimus/ tacrolimus (C), in reduced intensity conditioning allogeneic transplantation recipients lacking HLA-matched related donors. The primary endpoint was grade II-IV acute graft-versus-host disease incidence rate by day +180. One hundred and thirty-eight patients (A 46, B 45, C 47) with a median age of 64 years (range: 24-75), varying malignant diagnoses and disease risk (low 14, intermediate 96, high/very high 28) received 7-8/8 HLA-mismatched (40) or matched unrelated donor (98) grafts. Median follow up in survivors was 30 months (range: 14-46). Despite early immune reconstitution differences, day +180 grade II-IV acute graft-versus-host disease rates were similar (A 32.6%, B 31.1%, C 21%; P=0.53 for A vs. B, P=0.16 for A vs. C). The 2-year non-relapse mortality incidence was similar (A 14%, B 16%, C 6.4%; P=0.62), as were relapse (A 32%, B 32%, C 38%; P=0.74), chronic graft-versus-host disease (A 59%, B 60% C 55%; P=0.66), progression-free survival (A 54%, B 52%, C 55%; P=0.95), and overall survival (A 61%, B 62%, C 62%; P=0.98). Overall, the bortezomib-based regimens evaluated did not improve outcomes compared with tacrolimus/methotrexate therapy. clinicaltrials.gov Identifier: 01754389
Introduction Allogeneic hematopoietic stem cell transplantation (HSCT) is curative in advanced or aggressive hematologic malignancies despite associated toxicities. While a sibling matched at human leukocyte antigen (HLA)-A, -B, -C, and -DRB1 is optimal, only a minority of patients who may benefit from HSCT have such a donor available.1 Utilizing a 7/8 HLA-matched graft increases the likelihood of obtaining an adult donor for all racial and ethnic groups,1 but at the expense of worse outcomes. In reduced-intensity conditioning (RIC) HSCT, retrospective registry studies document increased rates of grade II-IV acute graft-versus-host disease haematologica | 2018; 103(3)
Bortezomib-based RIC HSCT: Phase II RCT
(aGvHD) and non-relapse mortality (NRM), with worse progression-free survival (PFS) and overall survival (OS) for 7/8 vs. 8/8 HLA-matched donors.2 The proteasome-inhibitor bortezomib (bort) can selectively deplete proliferating alloreactive T lymphocytes, reduce Th1 cytokines and Interleukin-6 levels, and block antigen-presenting cell (APC) activation.3-6 It can also spare regulatory T cells (Tregs) that are relevant in GvHD control.7 Administered early after graft infusion, short-course bort can control GvHD in major histocompatibility complex (MHC)-mismatched mouse HSCT and maintain graft-versus-tumor responses,6,8,9 while avoiding the proinflammatory colonic toxicity of delayed or prolonged bort.5,8,10 In a phase I/II study of RIC we documented that bortbased GvHD prophylaxis (bort 1.3 mg/m2 IV on d +1, +4, +7, plus conventional tacrolimus [tax] and methotrexate [mtx]) was safe and potentially efficacious in T-replete HLA-mismatched donor (MMD) transplantation, with survival comparable to HLA-matched cohorts.11,12 These data were used to support the inclusion of the bort-based regimen as one novel arm of an ongoing multicenter phase II randomized trial, comparing it to two other novel regimens: post-transplant cyclophosphamide (PTCy), and maraviroc, each compared to a non-randomized ‘conventional care’ cohort (The Blood and Marrow Transplant Clinical Trials Network [BMTCTN] 1203). We additionally conducted a phase II randomized control trial (RCT) for patients lacking 8/8 HLA-matched sibling donors (DFCI 12-404), directly randomizing conventional tac/mtx vs. two novel bort-based GvHD prophylaxis regimens, whose mature results we report herein. The regimens comprised bort plus tac/mtx (arm B), directly based on our phase I/II data; and bort plus sirolimus (sir)/tac (arm C), to explore the tolerizing effects of enhanced Treg sparing with combined proteasome- and mechanistic target of rapamycin (mTOR)-inhibition during early immune reconstitution.7,13 We chose bort/sir/tac instead of bort-based calcineurin inhibitors (CNI)-free prophylaxis (e.g., bort/sir/mycophenolate mofetil [MMF]) as the sir/MMF doublet has limited clinical efficacy,14 and in our experience, adding bort to sir/MMF did not provide adequate efficacy in GvHD prevention, while the sir/tac doublet has documented efficacy comparable to conventional tac/mtx.15
Methods This prospective clinical trial was approved by the institutional review board of the Dana-Farber Cancer Institute/Harvard Cancer Center. Written informed consent was obtained prior to patient enrollment.
Study Design The study was a one-stage randomized phase II trial with the primary objective of comparing the incidence of aGvHD in two bort-based GvHD prophylaxis regimens to conventional tac/mtx. The study was designed for the primary comparisons of the grade II-IV aGvHD rates in Arm A vs. B and Arm A vs. C in parallel, and was powered to test the hypothesis of superiority in Arm B and C (15%), compared to that in Arm A (40%). The accrual goal was 138 patients, randomizing all eligible patients in 1:1:1 ratio to the three regimens. Randomization was stratified by degree of HLAmatch (8/8 vs. 7/8). haematologica | 2018; 103(3)
Patients Adult hematologic malignancy patients lacking an available HLA-matched sibling with an available 8/8 (HLA-A, -B, -C, DRB1) matched unrelated donor (MUD) or 1-antigen/allele mismatched related donor (MMRD) or mismatched unrelated donor (MMUD) were enrolled on the trial between January 2013 and November 2015. Patients with HIV infection, active hepatitis B or C disease, abnormal renal (serum creatinine > upper limit of normal [ULN] with clearance <40 mL/min/1.73m2 body surface area [BSA]) or liver function (serum total bilirubin >ULN, serum alanine/aspartate aminotransferase >2x ULN), Eastern Cooperative Oncology Group (ECOG) performance status >2, hyperlipidemia (serum cholesterol >300mg/dL; trigylcerides >400mg/dL) despite therapy, and peripheral neuropathy ≥grade 2 within 21 days prior to transplantation were excluded.
Transplantation Conditioning comprised busulfan (0.8 mg/kg twice daily IV) and fludarabine (30 mg/m2 once daily IV) from days -5 to -2. Treplete donor peripheral blood stem cells (PBSC) dosed at ≥ 2x106 CD34+ cells/kg were infused on day 0. GvHD prophylaxis regimens were: tac/mtx (arm A), bort/tac/mtx (arm B), and bort/sir/tac (arm C). Dosing was: bort (1.3 mg/m2 IV on day +1, +4, +7), mtx (10 mg/m2 IV on day +1, 5 mg/m2 on day +3, +6, +11), sir (target trough level 5-12 ng/ml) and/or tac (target trough level 5-10 ng/ml) from day -3. The tapering of immunosuppression started at around d+100, with the aim of being off immune suppression (IS) by day +180, as applicable per treatment arm. Participants received levetiracetam for seizure prophylaxis from day -5 to -1 and filgrastim 5 μg/kg daily from day +1 until an absolute neutrophil count (ANC) >1500 cells/μl was attained, and at least 12 months of Pneumocystis jiroveci and herpes simplex virus (HSV)/ varicella zoster virus (VZV) prophylaxis. Anti-fungal prophylaxis was not routine.
Immunophenotyping CD4+ T cells were defined as CD3+CD4+; CD8+ T cells were defined as CD3+CD8+; CD8+ naïve cells were defined as CD8+CD45RO-CD62L+; CD4 Tregs were defined as CD3+CD4+CD25med-highCD127low; CD4 conventional T cells (Tcons) were defined as CD3+CD4+CD25low-negCD127med-high; natural killer (NK) cells as CD56+CD3–; and B cells as CD19+. Aliquots of anticoagulated whole blood (ethylenediaminetetraacetic acid [EDTA]) were incubated with fluorophore-conjugated monoclonal antibodies: anti-CD3 V450 (clone UCHT1, BD Biosciences), anti-CD4 APC-H7 (clone RPA-T4, BD Biosciences), anti-CD8 Pacific-Orange (clone RPA-T8, Biolegend), anti-CD25 PE-Cy7 (clone M-A251, BD Biosciences), anti-CD127 PE-Cy5 (clone eBioRDR5, eBioscience) for T-cell subsets; anti-CD56 PE (clone B159, BD Biosciences), anti-CD3 V450 (clone UCHT1, BD Biosciences) for NK/NKT cells, and anti-CD19 APC (clone HIB19, BD Biosciences) for B cells. RBC lysis with BD Pharm Lyse was performed either prior to or following incubation with conjugated antibodies. Flow cytometry analysis utilized FACSCanto II (BD Bioscience) or the Fortessa (BD Bioscience) and FACSDiva software (BD Bioscience). There was a change in the use of flow cytometry machines over the course of the study. Both flow cytometers were validated and results were comparable.
Statistical considerations The primary endpoint was a grade II-IV aGvHD rate by day 180 after stem cell infusion. Secondary endpoints included cumulative incidence of aGvHD, NRM, relapse, chronic (c)GvHD, and PFS, OS and GvHD-free/relapse-free survival (GRFS). The study was designed for the primary comparisons of the grade II-IV aGvHD 523
J. Koreth et al.
rates in arm A vs. B and arm A vs. C in parallel. We projected the incidence of grade II-IV aGvHD as 40% in arm A and 15% in arm B and in arm C. With the sample size of 46 per arm, there is an 80% power to detect a 25% difference in grade II-IV aGvHD rate between two arms. This power calculation is based on Fisherâ&#x20AC;&#x2122;s exact test at one-sided type I error rate of 0.05. The primary analysis was performed per the modified intent-to-treat principle (mITT), i.e., all patients who were randomized and received any
amount of the study treatment were included in the primary analysis. Cumulative incidence of aGvHD and cGvHD, relapse and NRM were estimated in the competing risk framework. Relapse with or without death was considered a competing risk for NRM, and similarly NRM for relapse. Death or relapse without developing GvHD was a competing event for GvHD. GvHD incidences occurring after relapse, in the setting of relapse interventions (e.g., immunosuppression [IS] taper, donor lymphocyte infu-
Table 1. Baseline characteristics of the study cohorts: by treatment arms A, B, C, and overall.
Arm A (N=46) (Tac/Mtx) N % Age, median (range) Patient Sex Male Female Donor Sex Male Female Patient-Donor Sex Match MM MF FM FF ECOG Performance Status 0 1 2 Primary Disease AML CLL/SLL/PLL CML Hodgkin Disease ALL MDS MPD Mixed MDS/MPD NHL Other Acute Leukemia HLA Type 8/8 match 7/8 match Patient-Donor CMV Serostatus Positive Graft Source PB Disease Risk Index Low Intermediate High/Very high
Arm B (N=45) (Bort/Tac/Mtx) N %
65 (29, 74) 28 18
60.9 39.1
Arm C (N=47) (Bort/Sir/Tac) N %
65 (30, 75) 32 13
71.1 28.9
62 (24, 73) 24 23
51.1 48.9
All (N=138) N
%
64 (24, 75) 84 54
P 0.13 0.15
60.9 39.1 0.75
32 14
69.6 30.4
33 12
73.3 26.7
36 11
76.6 23.4
101 37
73.2 26.8
20 8 12 6
43.5 17.4 26.1 13.0
26 6 7 6
57.8 13.3 15.6 13.3
21 3 15 8
44.7 6.4 31.9 17.0
67 17 34 20
48.6 12.3 24.6 14.5
7 21 18
15.2 45.7 39.1
10 20 15
22.2 44.4 33.3
10 30 7
21.3 63.8 14.9
27 71 40
19.6 51.4 29.0
19 5 2 3 12 2 2 1
41.3 10.9 4.4 6.5 26.1 4.4 4.4 2.2
16 3 1 2 12 1 1 9 -
35.6 6.7 2.2 4.4 26.7 2.2 2.2 20.0 -
18 3 2 3 3 9 9 -
38.3 6.4 4.3 6.4 6.4 19.2 19.2 -
53 11 3 7 6 33 3 1 20 1
38.4 8.0 2.2 5.1 4.4 23.9 2.2 0.7 14.5 0.7
34 12
73.9 26.1
32 13
71.1 28.9
32 15
68.1 31.9
98 40
71.0 29.0
36
78.3
35
77.8
25
53.2
96
69.6
46
100
45
100
47
100
138
100
5 33 8
10.9 71.7 17.4
4 30 11
8.9 66.7 24.4
5 33 9
10.6 70.2 19.2
14 96 28
10.1 69.6 20.3
0.39
0.11
0.48
0.83
0.01 NA 0.94
Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus; MM: male patient/male donor; MF: male patient/female donor; FM: female patient/male donor; FF: female patient/female donor; AML: acute myeloid leukemia; CLL/SLL/PLL: chronic lymphocytic leukemia/small lymphocytic lymphoma/prolymphocytic leukemia; CML: chronic myeloid leukemia; MDS: myelodysplastic syndromes; MPD: myeloproliferative syndromes; NHL: non-Hodgkin lymphoma; PB: peripheral blood; ALL: acute lymphoblastic leukemia; HLA: human leukocyte antigen; CMV: cytomegalovirus.
524
haematologica | 2018; 103(3)
Bortezomib-based RIC HSCT: Phase II RCT
sion [DLI]), were counted in the estimation of cumulative incidence of GvHD in order to obtain comprehensive estimates, and both inclusive and exclusive estimates are presented for aGvHD endpoints. GRFS, PFS and OS were estimated using the KaplanMeier method. GRFS was defined as the time from stem cell infusion to incidence of grade III/IV aGvHD, cGvHD requiring systemic immunosuppression agents, relapse or death, whichever occured first. PFS and OS have been defined elsewhere. Cumulative incidences in the presence of competing events were compared using the Gray test, and Kaplan-Meier estimates were compared using the log-rank test.16 For time-to-event endpoints, P-values reflect comparing entire distributions along with point estimates for ease of presentation. For the primary endpoint, incidence rates of grade II-IV aGvHD by day +180 were compared at one-sided significance level of 0.05 using Fisherâ&#x20AC;&#x2122;s exact test; P-values for secondary endpoints are two-sided at the significance level of 0.05, without adjusting for multiple comparisons. Multivariable analysis adjusting for variables, as listed in Table 1, were performed for OS and PFS using a Cox model, and a Fine and Gray model was used for grade II-IV aGvHD.17 Immunophenotype data
were compared using the Wilcoxon rank-sum test at each time point without adjusting for multiple comparisons. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC, USA) and R version 3.1.3 (the CRAN project).
Results Patient and transplant variables: one hundred and fortytwo subjects enrolled and 138 randomized subjects were evaluable per the protocol-specified mITT criteria (four cancelled study participation before receiving transplantation conditioning: two due to acute infection (one with subsequent disease relapse), one due to provider preference to continue protocol-excluded medications, and another due to disease relapse at transplantation admission) (Figure 1). The treatment arms (A 46, B 45, C 47) were balanced for pre-transplant variables (Table 1), except lower cytomegalovirus (CMV) seropositivity in arm C (A, 78.3% vs. B, 77.8% vs. C, 53.2%, P=0.01).
Figure 1. CONSORT Diagram. *1 patient with relapse and infection. Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus.
Table 2. Summary of the primary endpoint
(A) aGvHD after relapse with IS taper included A (Tac/MTX) Day 180 aGvHD rate* grade II-IV (LL, UL) grade III-IV (LL, UL)
32.6%(21.3, 45.7) 2.2% (0.1, 9.9)
Arm B (Bort/Tac/MTX) C (Bort/Sir/Tac)
P A vs. B
A vs. C
31.1%(19.9, 44.3) 8.9% (3.1, 19.2)
21% (12, 33.4) 14.9% (7.2, 26.2)
0.53 (1.0) 0.97 (0.2)
0.16 (0.25) 0.997 (0.06)
28.9% (18, 42) 8.9% (3.1, 19.2)
14.9% (7.2, 26.2) 10.6% (4.3, 21.1)
0.44 (0.82) 0.97 (0.2)
0.04 (0.054) 0.99 (0.2)
(B) aGvHD after relapse with IS taper excluded Day 180 aGvHD rate* grade II-IV (LL, UL) grade III-IV (LL, UL)
32.6%(21.3, 45.7) 2.2% (0.1, 9.9)
*proportion of aGvHD at Day +180. **P-values without parenthesis indicate one-sided testing of whether the aGvHD rate in arm B or C is lower than the aGvHD rate in arm A. P-values in parenthesis indicate two-sided testing. UL: upper one-sided confidence limit at 5% level; LL: lower one-sided confidence limit at 5% level; aGvHD: acute graft-versushost disease; Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus.
haematologica | 2018; 103(3)
525
J. Koreth et al.
Subjects had a median age of 64 years (range: 24-75), varying diagnoses (53 acute myeloid leukemia [AML], 33 myelodysplastic syndrome [MDS], 20 non-Hodgkin lymphoma [NHL], 11 chronic lymphocytic leukemia [CLL],
etc.) and disease-risk indices (low 14, intermediate 96, high/very high 28). They received T-replete 8/8 MUD (n=98) or 7/8 MMD (n=40) PBSC grafts. Median follow up in survivors was 30 months (range: 14-46).
Table 3. Summary of the secondary endpoints
Arm B (Bort/Tac/MTX)
C (Bort/Sir/Tac)
A vs. B
P A vs. C
A vs. B vs. C
33%(20, 47) 2.2% (0.2, 10)
31%(18, 45) 8.9% (2.8, 19)
21% (11, 34) 15% (6.5, 27)
0.71 0.64
0.33 0.07
0.36 0.1
33%(20, 47) 2.2% (0.2, 10)
29%(16, 43) 8.9% (2.8, 19)
15% (6.5, 27) 11% (3.8, 21)
0.54 0.64
0.11 0.2
0.15 0.6
59% (43, 72) 39% (24, 53) 14% (5, 26) 32% (18, 46) 54% (38, 68) 61% (45, 74)
60% (43, 73) 49% (33, 64) 16% (7, 29) 32% (19, 47) 52% (35, 65) 62% (45, 74)
55% (39, 68) 48% (33, 62) 6.4% (1.6, 16) 38% (24, 52) 55% (40, 68) 62% (46, 74)
0.9 0.37 0.92 0.64 0.76 0.96
0.43 0.57 0.42 0.75 0.87 0.9
0.66 0.64 0.62 0.74 0.95 0.98
A (Tac/MTX) Day 180 aGvHD* (a) grade II-IV (95% CI) grade III-IV (95% CI) Day 180 aGvHD* (b) grade II-IV (95% CI) grade III-IV (95% CI) 2y cGvHD* All cGvHD (95% CI) Ext cGvHD (95% CI) 2y NRM* (95% CI) 2y Relapse* (95% CI) 2y PFS (95% CI) 2y OS (95% CI)
*cumulative incidence. P-values for cumulative incidence, PFS and OS are from Grayâ&#x20AC;&#x2122;s test and log-rank test measuring the difference of entire distributions between two groups. All P-values are two-sided. (a): aGvHD after relapse with IS taper included; (b): aGvHD after relapse with IS taper excluded; aGvHD: acute graft-versus-host disease; cGvHD: chronic graft-versus-host disease; Ext: extensive; IS: immunosuppression; NRM: non-relapse mortality; PFS: progression-free survival; OS: overall survival; CI: confidence interval; Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus.
A
Grade II-IV Acute GvHD
B
P=0.36
C
Grade III-IV Acute GvHD
P=0.1
D
P=0.62
P=0.74
Figure 2. aGvHD: non-relapse mortality and relapse outcomes. Cumulative incidence of (A) grade II-IV aGvHD*, (B) grade III-IV aGvHD*, (C) non-relapse mortality (NRM), and (D) relapse per treatment arm. Black indicates arm A (tac/mtx), red indicates arm B (bort/tac/mtx), and blue indicates arm C (bort/sir/tac). Grayâ&#x20AC;&#x2122;s test for comparing the entire distributions was used. *: Acute graft-versus-host disease (GvHD) after relapse with IS taper included. Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus.
526
haematologica | 2018; 103(3)
Bortezomib-based RIC HSCT: Phase II RCT
Engraftment, chimerism, safety: the median time to neutrophil (>500/μl) and platelet engraftment (>20,000/μl) among patients who experienced count nadir (i.e., neutrophil count <500/vl and/or platelet count <20,000/μl) was 11 days (range: 2-45) and 19 days (range: 10-57), respectively, with no significant difference between treatment arms (P=0.9 and P=0.55 for neutrophils and platelets, respectively). For the entire cohort, 31% of patients did not experience count nadir (24% in arm A, 27% in arm B, 43% in arm C, P=0.11) and no subject failed neutrophil engraftment. Median total nucleated cell donor chimerism by day 30 was 96% (range: 42-100) and by day 100 was 97% (range: 0-100, with no significant difference between treatment arms (P=0.84 and P=0.83, respectively). The bort-based regimens were well tolerated. No bort doses were delayed or reduced due to toxicity. No serious adverse event (SAE) attributable to bort (e.g., neuropathy) was documented. Common Terminology Criteria for Adverse Events (CTCAE) grade ≥3 organ dysfunction (hepatic, renal, pulmonary) did not differ significantly between treatment arms, including the incidence of acute kidney injury (AKI, 8.7% overall), thrombotic microangiopathy/hemolytic-uremic syndrome (TMA/HUS, 6.5% overall) or hepatic veno-occlusive disease (VOD, 3.3% overall) (P=0.14, P=0.16 and P=0.41, respectively). Acute GvHD: The primary endpoint of grade II-IV aGvHD incidence rate by day +180 did not differ significantly between treatment arms, at 32.6% (A) vs. 31.1% (B, one-sided P=0.53 for arm A vs. B) vs. 21% (C, one-sided P=0.16 for arm A vs. C) (Table 2). This result is consistent
A
All chronic GvHD
with the cumulative incidence of aGvHD (P=0.36) (Table 3, Figure 2A). For seven patients, grade II-IV aGvHD occurred after hematologic malignancy relapse (six after early IS taper, and 1 after donor lymphocyte infusions [DLI]). Four of these (B, 1; C, 3) had aGvHD onset prior to day +180, and are included in the primary endpoint. This result was consistent with the result from multivariable analysis: subdistribution hazard ratio (sHR) was 0.87 (95% confidence interval [CI] 0.42-1.78, P=0.7) for arm B vs. arm A, and 0.68 (95% CI 0.33-1.4, P=0.29) for arm C vs. arm A. If patients with aGvHD onset after relapse are excluded from the analysis, the grade II-IV aGvHD incidence rate by day +180 was 32.6% (A) vs. 28.9% (B, onesided P=0.44 for arm A vs. B) vs. 14.9% (C, one-sided P=0.04 for arm A vs. C) (Table 2). On multivariable analysis, sHR was 0.8 (95% CI 0.39-1.66, P=0.55) for arm B vs. A, and 0.53 (95% CI 0.23-1.18, P=0.12) for arm C vs. A. aGvHD incidence continued to rise after day +180 across arms, and the 1-year cumulative incidence of grade II-IV aGvHD was 40% in arm A, 34% in arm B, and 26% in arm C (P=0.71 for arm A vs. B, P=0.33 for arm A vs. C) (Figure 2A). For grade III-IV severe aGvHD, day +180 cumulative incidence was 2.2% (A) vs. 8.9% (B, P=0.64 for arm A vs. B) vs. 15% (C, P=0.07 for arm A vs. C) (Table 3, Figure 2B). In an exploratory analysis in line with our protocol stratification, we repeated the analysis by HLA match status (Online Supplementary Table S1A). For the 8/8 HLAmatched recipients, when all grade II-IV aGvHD were counted, the six-month cumulative incidence was 33% (A) vs. 16% (B, P=0.1 for arm A vs. B) vs. 19% (C, P=0.21
B
P=0.66
P=0.95
C
D
GvHD-free, relapse-free survival
P=0.53
P=0.98
haematologica | 2018; 103(3)
Figure 3. cGvHD: survival and GRFS outcomes. Cumulative incidence of (A) all cGvHD, and Kaplan-Meier survival plots of (B) progression-free survival (PFS), (C) overall survival (OS), and (D) grade III-IV aGvHD/cGvHD requiring systemic IS agents/relapse-free survival (GRFS) per treatment arm. Black indicates arm A (tac/mtx), red indicates arm B (bort/tac/mtx), and blue indicates arm C (bort/sir/tac). Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus; GvHD: graftversus-host disease.
527
J. Koreth et al.
for arm A vs. C) (Online Supplementary Figure S1). When arms B and C were combined, the six-month cumulative incidence was 17% (P=0.08 for arm A vs. arm B or C). When we repeated the analysis after excluding aGvHD that occurred after relapse and IS taper, the six-month cumulative incidence was 33% (A) vs. 13% (B, P=0.046 for arm A vs. B) vs. 16% (C, P=0.13 for arm A vs. C). When arms B and C were combined, the cumulative incidence was 15% (P=0.03 for arm A vs. arm B or C). However, for 7/8 MMD recipients such a benefit was not appreciable (albeit with limited sample size) (Online Supplementary Table S1A, Online Supplementary Figure S1). NRM, relapse, chronic GvHD, and survival: NRM did not differ significantly between treatment arms (Table 3), with a 2-year cumulative incidence of 14% (A, 95% CI, 5-26) vs. 16% (B, 95% CI, 7-29) vs. 6.4% (C, 95% CI, 1.6-16; P=0.62) (Figure 2C). Relapse did not differ significantly between treatment arms, with a 2-year cumulative incidence of 32% (A, 95% CI, 18-46) vs. 32% (B, 95% CI, 1947) vs. 38% (C, 95% CI, 24-52; P=0.74) (Figure 2D). The 2year cumulative incidence of cGvHD did not differ signif-
A Absolute CD3+ T-cell count per μL
icantly between treatment arms, at 59% (A) vs. 60% (B) vs. 55% (C; P=0.66) (Figure 3A). For five patients, cGvHD occurred after documented hematologic malignancy relapse (A, 1; B, 2; C, 2) and these were included in the estimation of cumulative incidence of cGvHD. Two had early IS taper, none received DLI. The 2-year PFS did not differ significantly between treatment arms, at 54% (A) vs. 52% (B) vs. 55% (C; P=0.95) (Figure 3B). The 2-year OS did not differ significantly between treatment arms, at 61% (A) vs. 62% (B) vs. 62% (C; P=0.98) (Figure 3C). The composite 2-year GRFS endpoint did not differ significantly between treatment arms, at 11% (A) vs. 12% (B) vs. 8% (C; P=0.53) (Figure 3D). Immune reconstitution: The median total CD3+ T-cell count/µl at one month after transplantation was 401 in arm A (Q1-Q3, 248-681) vs. 414 in arm B (Q1-Q3, 195898) vs. 190 in arm C (Q1-Q3, 137-340; P<0.0001), and remained lower in arm C through months two and three (A, 602 and 728 vs. B, 639 and 571 vs. C, 221 and 269, respectively; P<0.0001). At six months after transplantation, CD3+ T-cell counts were similar in all three arms at 763 (A, Q1-Q3, 463-1120) vs. 535 (B, Q1-Q3, 333-907) vs. 508 (C, Q1-Q3, 322-728; P=0.10) (Figure 4A). This reflected both a lower median CD8+ and CD4+ T-cell count/μl at one to three months after transplantation in arm C (P<0.0001, data not shown). Similarly, the median CD19+ Bcell and the median CD56+CD3– NK cell count/μl was lower for arm C during the one to three months following transplantation (P<0.05, data not shown). While the total CD4+ Tcon cell count/μl at one, two, and three months after transplantation was lower for arm C (P=0.005, P=0.0006, P=0.024 at one, two and three months after transplantation, respectively), the total CD4+ Treg cell count/µl was unimpaired in arm C at those time points, resulting in an improved ratio of Treg:Tcon reconstitution at one and three months after transplantation (A: 0.049 and 0.05 vs. B: 0.051 and 0.032 vs. C: 0.088 and 0.067; P=0.006 and P=0.012, respectively) (Figure 4B).
Discussion B Treg: Tcon Ratio
Figure 4. Immune reconstitution outcomes. Reconstitution of (A) median of absolute CD3+ T-cell count/μL, and (B) median values of CD4+ Treg:Tcon cell ratio per treatment arm. Blue indicates arm A (tac/mtx), red indicates arm B (bort/tac/mtx), and green indicates arm C (bort/sir/tac). Treg: regulatory T cells; Tcon: conventional T cells; Tac: tacrolimus; Mtx: methotrexate; Bort: bortezomib; Sir: sirolimus; W1: week 1; W2: week 2; M: month.
528
Most adult hematologic malignancy patients who may benefit from allogeneic HSCT lack an available sibling donor and are considered for a MUD or 1-locus MMD, with umbilical cord blood (UCB) and haploidentical (haplo) donors being additional alternative options. Regarding 8/8 MUD, as a result of improvements in DNAbased typing and supportive care, survival outcomes are considered similar to those of MRD HSCT,18,19 although studies indicate that MUD HSCT is associated with increased grade II-IV aGvHD (52% vs. 34%), grade III-IV aGvHD (21% vs.16%), and NRM (RR 2.76; P<0.01).20 The use of 1-locus MMD adds risk. A large registry analysis of 2,588 patients with acute leukemias, MDS or chronic myeloid leukemia (CML) undergoing RIC HSCT compared 7/8 with 8/8 HLA-matched donors, and documented higher rates of grade II-IV aGvHD and NRM, and lower 3-year OS (30% vs. 38%, respectively; P=0.01) with a mismatched graft.2 Novel regimens to improve GvHD outcomes for patients lacking a preferred sibling donor would represent a major advance, and is the focus of our efforts. Proteasome inhibition with bort has immunomodulatory properties relevant to allogeneic HSCT, as previously highlighted. Based on encouraging phase I/II results in haematologica | 2018; 103(3)
Bortezomib-based RIC HSCT: Phase II RCT
HLA-mismatched T-replete RIC HSCT, we undertook a prospective randomized evaluation of aGvHD prophylaxis with conventional tac/mtx (arm A) vs. two novel regimens of short-course bort for transplantation recipients lacking 8/8 HLA-matched related donors: bort plus tac/mtx (arm B), and bort plus sir/tac (arm C). In our study, Bort, limited to three doses peritransplantation (day +1, +4 and +7), did not add toxicity. No bort doses were missed or modified. No subjects developed toxicities of prolonged/delayed bort administration (e.g., neuropathy, colonic necrosis). Comparing treatment arms, no increase in hepatic VOD, AKI or TMA/HUS was noted for bort-based regimens vs. tac/mtx. Engraftment was rapid and sustained, with robust donor chimerism by day 30 that was sustained and comparable across treatment arms. The addition of bort alone did not appear to impair immunologic recovery, with similar counts of T cells (CD4, CD8), B cells and NK cells in the conventional tac/mtx vs. bort/tac/mtx arms. Count recovery for arm C (bort/sir/tac) was lower, as we previously documented for other sir-based cohorts.10 As hypothesized, there was relative sparing of Treg reconstitution with arm C (bort/sir/tac), with no impairment of Treg count recovery and elevated Treg:Tcon ratios. The clinical impact of these systematic immunologic differences, however, remains unclear. Importantly, the 2-year NRM incidence was low (ranging from 6.4% to 16%) and did not differ significantly between treatment arms. Relapse was in the expected range after RIC HSCT (ranging from 24% to 36%), and did not differ significantly between treatment arms. In this study, survival, while not differing significantly between treatment arms, with 2-year OS ranging from 61% to 62% (P=0.98), appeared better than anticipated for 7/8 and 8/8 HLA-matched recipients, plateauing at 50% and 66%, respectively, in contrast to 1- and 3-year registry benchmark survivals recently reported for 7/8 and 8/8 donor recipients, of 48%→30% and 55%→38%, respectively.2 We highlight a lack of significant difference between treatment arms with regard to the primary endpoint of grade II-IV aGvHD rate by day +180. In the context of delayed T-cell reconstitution noted for arm C, this suggests that the combination of a CNI (tac) plus mTOR inhibitor (sir) provides IS without long term tolerizing effects. While it is possible that longer duration of tac/sir immunosuppression in arm C beyond day +180 may have better prevented aGvHD, ultimately, aGvHD deferral rather than long-term amelioration remains a possibility. In subgroup analysis for the 8/8 HLA-matched recipients, bort-based regimens had a borderline trend towards grade II-IV aGvHD benefit vs. tac/mtx, with a cumulative incidence of 17% vs. 33%, respectively, (P=0.08). However, grade III-IV severe aGvHD rates were not improved, and for 7/8 MMD grafts a similar trend for aGvHD benefit of bort-based regimens was not appreciable, though sample size was limited.
References 1. Gragert L, Eapen M, Williams E, et al. HLA match likelihoods for hematopoietic stemcell grafts in the U.S. registry. N Engl J Med. 2014;371(4):339-348. 2. Verneris MR, Lee SJ, Ahn KW, et al. HLA
haematologica | 2018; 103(3)
Our trial has strengths due to its appropriate size in the phase II context, and its direct prospective randomization to conventional tac/mtx vs. two novel bort-based GvHD prophylaxis regimens. Concerns regarding open-label treatment assignment are ameliorated by the minimal dropout rate and the mITT analysis to further avoid bias, while the ‘hard’ endpoints of GvHD, NRM, relapse and survival additionally obviate assessment bias concerns. Our primary finding is that, as tested, bort-based regimens did not appear to provide additional benefit for grade II-IV aGvHD prevention in T-replete PBSC RIC HSCT, failing to meet the protocol-specified 25% reduction in aGvHD incidence for success. Our phase II RCT requirements for success were stringent, which limits our ability to detect lower, albeit potentially clinically meaningful benefit with bort. However, it is also notable that conventional tac/mtx outcomes were better than anticipated, a finding similar to that recently reported in other contemporary randomized HSCT trials.21 These data highlight the inadequacy of non-randomized comparators for evaluating early phase single arm interventional studies, and document, as an updated standard, the improved MMD and MUD RIC HSCT outcomes achievable with conventional tac/mtx in this study (2-year OS of 58% and 62%, respectively). In the future, prospective trials of alternative donors (e.g., UCB, haplo) and novel GvHD prophylaxis regimens (e.g., maraviroc, PTCy) may need to benchmark these outcomes. In contrast, the bar for the novel GRFS endpoint appears far lower, ranging from 8% to 12% in our study, with no significant difference between treatment arms (P=0.53). In summary, mature data from this open-label 1:1:1 three-arm phase II RCT indicates that the bort-based regimens evaluated appear to provide lower than anticipated grade II-IV aGvHD benefit compared to conventional tac/mtx in T-replete PBSC RIC HSCT. While we note the potential benefit of bort for 8/8 HLA-matched transplants, direct phase III prospective randomization is required in order to confirm such a benefit. Overall, however, the lack of benefit for other transplantation outcomes (NRM, relapse, chronic GvHD, and survival) suggests limited utility for bort prophylaxis in the doses and combinations assessed. Alternative proteasome inhibitor combination regimens (e.g., with PTCy) should also be considered. Acknowledgments We thank clinical research nurses Susan Stephenson RN and Mildred Pasek RN. JK is a Scholar in Clinical Research of the Leukemia and Lymphoma Society. Funding This study was supported in part by Millennium Pharmaceuticals Inc., the Jock and Bunny Adams Research and Education Endowment, and the National Institutes of Health CA183560, CA183559, and P01CA142106.
mismatch is associated with worse outcomes after unrelated donor reduced-intensity conditioning hematopoietic cell transplantation: an analysis from the Center for International Blood and Marrow Transplant Research. Biol Blood Marrow Transplant. 2015;21(10):1783-1789. 3. Nencioni A, Schwarzenberg K, Brauer KM,
et al. Proteasome inhibitor bortezomib modulates TLR4-induced dendritic cell activation. Blood. 2006;108(2):551-558. 4. Blanco B, Perez-Simon JA, Sanchez-Abarca LI, et al. Bortezomib induces selective depletion of alloreactive T lymphocytes and decreases the production of Th1 cytokines. Blood. 2006;107(9):3575-3583.
529
J. Koreth et al. 5. Pai CC, Hsiao HH, Sun K, et al. Therapeutic benefit of bortezomib on acute graft-versus-host disease is tissue specific and is associated with interleukin-6 levels. Biol Blood Marrow Transplant. 2014;20(12):1899-1904. 6. Sun K, Welniak LA, Panoskaltsis-Mortari A, et al. Inhibition of acute graft-versus-host disease with retention of graft-versustumor effects by the proteasome inhibitor bortezomib. Proc Natl Acad Sci USA. 2004; 101(21):8120-8125. 7. Kim JS, Lee JI, Shin JY, et al. Bortezomib can suppress activation of rapamycin-resistant memory T cells without affecting regulatory T-cell viability in non-human primates. Transplantation. 2009;88(12):13491359. 8. Sun K, Wilkins DE, Anver MR, et al. Differential effects of proteasome inhibition by bortezomib on murine acute graftversus-host disease (GVHD): delayed administration of bortezomib results in increased GVHD-dependent gastrointestinal toxicity. Blood. 2005;106(9):3293-3299. 9. Vodanovic-Jankovic S, Hari P, Jacobs P, Komorowski R, Drobyski WR. NF-kappaB as a target for the prevention of graft-versus-host disease: comparative efficacy of bortezomib and PS-1145. Blood. 2006;107(2):827-834. 10. Liang Y, Ma S, Zhang Y, et al. IL-1beta and TLR4 signaling are involved in the aggravated murine acute graft-versus-host disease caused by delayed bortezomib admin-
530
11.
12.
13.
14.
15.
16.
istration. J Immunol. 2014;192(3):12771285. Koreth J, Stevenson KE, Kim HT, et al. Bortezomib, tacrolimus, and methotrexate for prophylaxis of graft-versus-host disease after reduced-intensity conditioning allogeneic stem cell transplantation from HLAmismatched unrelated donors. Blood. 2009; 114(18):3956-3959. Koreth J, Stevenson KE, Kim HT, et al. Bortezomib-based graft-versus-host disease prophylaxis in HLA-mismatched unrelated donor transplantation. J Clin Oncol. 2012;30(26):3202-3208. Zeiser R, Leveson-Gower DB, Zambricki EA, et al. Differential impact of mammalian target of rapamycin inhibition on CD4+CD25+Foxp3+ regulatory T cells compared with conventional CD4+ T cells. Blood. 2008;111(1):453-462. Johnston L, Florek M, Armstrong R, et al. Sirolimus and mycophenolate mofetil as GVHD prophylaxis in myeloablative, matched-related donor hematopoietic cell transplantation. Bone Marrow Transplant. 2012;47(4):581-588. Cutler C, Logan B, Nakamura R, et al. Tacrolimus/sirolimus vs tacrolimus/methotrexate as GVHD prophylaxis after matched, related donor allogeneic HCT. Blood. 2014;124(8):1372-1377. Gray R. A class of k-sample tests for comparing the cumulative incidence of a competing risk. Annals of Statistics. 1988;16(3):1141-1154.
17. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496-509. 18. Kiehl MG, Kraut L, Schwerdtfeger R, et al. Outcome of allogeneic hematopoietic stem-cell transplantation in adult patients with acute lymphoblastic leukemia: no difference in related compared with unrelated transplant in first complete remission. J Clin Oncol. 2004;22(14):2816-2825. 19. Yakoub-Agha I, Mesnil F, Kuentz M, et al. Allogeneic marrow stem-cell transplantation from human leukocyte antigen-identical siblings versus human leukocyte antigen-allelic-matched unrelated donors (10/10) in patients with standard-risk hematologic malignancy: a prospective study from the French Society of Bone Marrow Transplantation and Cell Therapy. J Clin Oncol. 2006;24(36):5695-5702. 20. Ringden O, Pavletic SZ, Anasetti C, et al. The graft-versus-leukemia effect using matched unrelated donors is not superior to HLA-identical siblings for hematopoietic stem cell transplantation. Blood. 2009;113(13):3110-3118. 21. Soiffer RJ, Kim HT, McGuirk J, et al. A prospective randomized double blind phase 3 clinical trial of anti-T lymphocyte globulin (ATLG) to assess impact on chronic graft-versus-host disease (cGVHD) free survival in patients undergoing HLA matched unrelated myeloablative hematopoietic cell transplantation (HCT). Blood. 2016;128(22):505.
haematologica | 2018; 103(3)
ARTICLE
Cell Therapy & Immunotherapy
In vivo IL-12/IL-23p40 neutralization blocks Th1/Th17 response after allogeneic hematopoietic cell transplantation
Joseph Pidala,1,2 Francisca Beato,1 Jongphil Kim,3 Brian Betts,1,2 Heather Jim,1,4 Elizabeth Sagatys,5 John E. Levine,6 James L.M. Ferrara,6 Umut Ozbek,6 Ernesto Ayala,1,2 Marco Davila,1,2 Hugo F. Fernandez,1,2 Teresa Field,1,2 Mohamed A. Kharfan-Dabaja,1,2 Divis Khaira,1,2 Farhad Khimani,1,2 Frederick L. Locke,1,2 Asmita Mishra,1,2 Michael Nieder,1,2 Taiga Nishihori,1,2 Lia Perez,1,2 Marcie Riches,1,2 and Claudio Anasetti1,2
Blood and Marrow Transplantation, Moffitt Cancer Center, Tampa, FL; Oncologic Sciences, College of the Medicine at University of South Florida, Tampa, FL; 3 Biostatistics, Moffitt Cancer Center, Tampa, FL; 4Health Outcomes and Behavior, Moffitt Cancer Center, Tampa, FL; 5Hematopathology and Laboratory Medicine, Moffitt Cancer Center, Tampa, FL and 6Tisch Cancer Institute, the Icahn School of Medicine at Mt. Sinai, New York, NY, USA 1
2
Ferrata Storti Foundation
Haematologica 2018 Volume 103(3):531-539
ABSTRACT
T
-helper 1 and T-helper 17 lymphocytes mediate acute graft-versushost disease (GvHD). Interleukin 12 is critical for T-helper 1 differentiation and interleukin 23 for T-helper 17 maintenance. Interleukin 12 and 23 are heterodimeric cytokines that share the p40 subunit (IL-12/IL-23p40). In a randomized, blinded, placebo-controlled trial, we examined the biological impact and clinical outcomes following IL-12/IL-23p40 neutralization using ustekinumab. Thirty patients received peripheral blood mobilized hematopoietic cell transplantation (HCT) from HLA-matched sibling or unrelated donors, received sirolimus plus tacrolimus as GvHD prophylaxis, and were randomized to ustekinumab versus placebo with 1:1 allocation after stratification by donor type. The primary end point of the trial was the mean percentage (%) T-regulatory (Treg) cells on day 30 post HCT. Ustekinumab was delivered by subcutaneous injection on day -1 and day +20 after transplantation. On day 30 post transplant, no significant difference in % Treg was observed. Ustekinumab suppressed serum IL-12/IL-23p40 levels. Host-reactive donor alloresponse at days 30 and 90 after transplantation was polarized with significant reduction in IL-17 and IFN-Îł production and increase in IL-4. No toxicity attributed to ustekinumab was observed. Overall survival and National Institute of Health moderate/severe chronic GvHD-free, relapse-free survival were significantly improved among ustekinumab-treated patients. No significant improvements were observed in acute or chronic GvHD, relapse, or non-relapse mortality. These data provide first evidence that IL-12/IL23p40 neutralization can polarize donor anti-host alloresponse in vivo and provide initial clinical efficacy evidence to be tested in subsequent trials. (Trial registered at clinicaltrials.gov identifier: 01713400.) Introduction Differentiation of CD4+ T cells into distinct lineages [Th1, Th2, Th17, T-regulatory (Tregs)] is co-ordinated by specific cytokine programs.1-3 IL-12 plays a key role in Th1 differentiation, and IL-23 stabilizes the Th17 phenotype. These cytokines share a common p40 subunit. Previous data have implicated Th1 and Th17 in acute graft-versus-host disease (GvHD) pathogenesis,4-6 and demonstrated that regulatory Tregs control alloreactivity.7 In experimental systems, disruption of Tbet and ROR t transcription factors,8 or neutralization of IL-12/IL-23p40, reduces Th1 and Th17 differentiation, increases Th2 and Tregs, and reduces GvHD mortality. Clinical translation of this work is made possible through neutralization of IL12/IL-23p40 using the monoclonal antibody ustekinumab (Stelara, Janssen Biotech Inc.), which is approved for therapy of plaque psoriasis, psoriatic arthritis, and haematologica | 2018; 103(3)
Correspondence: joseph.pidala@moffitt.org
Received: April 20, 2017. Accepted: December 6 2017. Pre-published: December 14, 2017. doi:10.3324/haematol.2017.171199 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/531 Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
531
J. Pidala et al.
Crohn disease. In humans, ustekinumab has clinical activity in several immune-mediated disorders, including steroid-refractory acute GvHD.10 A sirolimus (SIR)-based pharmacological immune suppression program provides an ideal platform for examining the biological effect of IL12/IL-23p40 neutralization: SIR inhibits differentiation of Th1 and Th17 cells, and promotes generation of Tregs from naĂŻve T cells. Tregs proliferate and survive in the presence of SIR, as they do not activate the phosphatidyl inositol 3-kinase (PI3-K)/AKT pathway.11,12 The primary objective of the trial was to examine the biological impact of IL-12/IL-23p40 neutralization. We anticipated disruption of Th1 and Th17 with sparing of
Th2 and Tregs. Secondary objectives were to determine safety of this investigational GvHD prevention approach and examine initial evidence of clinical efficacy.
Methods See Online Supplementary Appendix for full details.
Inclusion criteria Patients were aged 18-70 years with adequate vital organ function and Karnofsky performance status. Exclusion criteria were: uncontrolled infection, active HIV, hepatitis B or C infection, or
Table 1. Patient, disease, and transplantation variables.
Variable Patient age (median, range) Donor age (median, range) Donor:recipient sex F:M Others Donor:recipient CMV N:N N:P P:N P:P Disease ALL AML CLL HD MDS MPN NHL Remission status CR PIF PR REL SD Disease risk index29 Low Intermediate High Graft type PBSC BM Conditioning regimen FluBu (6.4 mg/kg) FluBu (AUC 3500) FluBu (AUC 5300) FluMel Donor relation MRD MUD DQB1-mismatch No Yes
Ustekinumab
Placebo
P
53 (22-69) 33 (22-72)
59 (38-69) 29 (22-66)
P=0.16 P=0.64
4 11
4 11
P=0.88
3 7 2 3
5 2 2 6
P=0.22
3 6 1 2 2 0 1
1 6 2 1 1 2 2
P=0.67
9 1 3 0 2
9 0 0 2 4
P=0.18
2 12 1
0 12 3
P=0.22
15 0
15 0
3 1 8 3
1 3 9 2
P=0.57
7 8
6 9
P=1
13 2
15 0
P=0.12
F: female; M: male; CMV: cytomegalovirus; N: negative; P: positive; ALL: acute lymphoblastic leukemia; AML: acute myelogenous leukemia; CLL: chronic lymphocytic leukemia; HD: Hodgkin lymphoma; MDS: myelodysplastic syndrome; MPN: myeloproliferative neoplasm; NHL: non-Hodgkin lymphoma; CR: complete remission; PIF: primary induction failure; PR: partial response; REL: relapsed/refractory disease; SD: stable disease. PBSC: peripheral blood mobilized stem cells; BM: bone marrow; Flu: fludarabine; Bu: busulfan; Mel: melphalan; AUC: average daily exposure of busulfan in ÂľMol*minute (min); MRD: matched related donor; MUD: matched unrelated donor.
532
haematologica | 2018; 103(3)
In vivo IL-12/IL-23p40 neutralization
hematopoietic stem cell transplantation (HCT) comorbidity (HCT-CI) â&#x2030;Ľ 3.13 Eligible related or unrelated donors matched with the patient at HLA-A, B, C, and DRB1 by high resolution typing provided filgrastim-mobilized peripheral blood stem cells (PBSC) for transplantation according to standard practices. All patients received tacrolimus and sirolimus for GvHD prophylaxis. The trial was conducted under the approval of the University of South Florida Institutional Review Board.
Trial design and treatment plan Both HCT recipients and their donors provided consent for participation in the study. Eligible patients were randomized (1:1 allocation, stratified by donor relation) to ustekinumab versus placebo (clinicaltrials.gov identifier: 01713400). Study investigators, clinical staff, and patients were blinded throughout to study arm assignment. Ustekinumab was delivered as a subcutaneous injection on day -1, and again on day +20 post HCT at a dose of 45 mg (for those with weight â&#x2030;¤ 100 kg) or 90 mg (weight > 100 kg). Placebo
was a subcutaneous injection of sterile saline of identical volume, character, and packaging as the investigational agent administered on the same schedule.
Clinical outcomes Neutrophil engraftment was defined by an absolute neutrophil count over 500 per microliter of blood sustained for at least three days, and platelet engraftment was defined by platelet count over 20 per microliter of blood sustained for at least seven days without transfusion support. Mucositis was graded per Common Toxicity Criteria (CTC) v.4.0. Diagnosis and severity grading of thrombotic microangiopathy (TMA) adhered to Bone Marrow Transplantation Clinical Trials Network consensus.14 Hepatic veno-occlusive disease (VOD) was diagnosed according to standard clinical criteria.15 Acute GvHD was scored weekly from HCT to day 100 according to consensus guidelines.16 Chronic GvHD scoring used National Institutes of Health (NIH) Consensus Criteria for diagnosis and staging.17
A 4000 P = 0.003
Ustekinumab Placebo
P = 0.001
3000
ng/mL
P = 0.001
P = 0.004
2000
P = 0.004
P = 0.005
1000 P = 0.021
0 Base
7
21
35
49
63
77
91
Collection day
B 30 Ustekinumab Placebo
25
ng/mL
20
15
10 5 0 Base
7
21
35
49
Collection day haematologica | 2018; 103(3)
63
77
91
Figure 1. Pharmacokinetic and pharmacodynamics measurements. (A) Concentration of anti-IL-12/IL-23p40 antibody over time post hematopoietic stem cell transplantation (HCT). (B) Concentration of circulating IL-12/IL-23p40 over time post HCT.
533
J. Pidala et al.
Statistical analysis The primary end point of the study was the peripheral blood mean Treg/total CD4+ ratio at day 30 following HCT. The expected Treg/total CD4+ was 19% at 30 days among SIR/TAC/placebo patients based on observed data in SIR/TAC-treated patients in a previous trial.18 With 15 patients per study arm, standard deviation of 11, and type I error of 0.05, the study had 80% power to detect an increase in Treg/total CD4+ to 31% among ustekinumab-treated patients using a two-sided Mann-Whitney test. Baseline characteristics were summarized using descriptive statistics including mean, median, standard deviation and range for continuous measures and counts, and frequencies for categorical measures. Comparisons between study arms were made by the Cochran-Mantel-Haenszel test for categorical measures and by the two-way analysis of variance (ANOVA) or the Friedman twoway ANOVA for continuous measures including biological correlative data and Quality of Life over the study period, adjusting for the stratification variable donor type: matched sibling versus
matched unrelated donor. The difference in cumulative incidence of grade II-IV acute GvHD was estimated using the stratified Gray test.19 Survival data were analyzed using the Kaplan-Meier method, and stratified comparisons used the log-rank test.
Results Patientsâ&#x20AC;&#x2122; characteristics Randomization resulted in a balanced distribution of patients', disease, and HCT variables (Table 1). Included patients were adults with an anticipated representation of hematologic malignancies. No significant differences were observed between groups for the studied variables.
Pharmacokinetic and pharmacodynamic studies The mean serum concentration-time plots for ustekinumab are shown in Figure 1A. Sustained levels of ustek-
P â&#x2030;¤ 0.0001
A
P = 0.28
P = 0.11
P = 0.0009
P = 0.37
P = 0.4
B P = 0.18
C
P = 0.08
P = 0.38
P = 0.03
P = 0.029
P = 0.015 P = 0.09
P = 0.02
Figure 2. Donor responder cell cytokine production after stimulation with host or 3rd-party stimulators. *IFN-Îł ELISPOT and supernatant cytokines (IL-4, IL-17) shown as in (A), (B) and (C). P-values for comparison by Mann-Whitney test.
534
haematologica | 2018; 103(3)
In vivo IL-12/IL-23p40 neutralization inumab above 1 μg/mL were observed throughout the treatment period, and were associated with neutralization of circulating IL-12/IL-23p40 (Figure 1B). From day 0 to peak level post HCT, placebo-treated patients had a 14.1fold increase in IL-12/IL-23p40, while ustekinumab-treated patients had only a 2.7-fold increase.
Donor alloreactive T-cell polarization When stimulated with third-party alloantigen, donor T cells collected on day 30 from blood of ustekinumab-treated patients produced less IFN-γ (P=0.001) and IL-17 (P=0.03), and similar amounts of IL-4 (P=0.38) compared to donor T cells from blood of placebo-treated patients (Figure 2). Other Th2 cytokines (IL-5, IL-13) were not tested. The sensitivity to detect host-specific alloresponses was low because the transplants were from HLA compatible donors. However, lower production of IL-17 by hostspecific donor T cells was also observed in ustekinumabtreated patients compared to placebo at day 30 (P=0.02). Findings on day 30 and day 90 samples were similar (Figure 2). IFN-γ production in response to PMA/ionomycin was higher among the ustekinumab group at day 30, yet similar at day 90. Among other tested supernatant cytokines, there were no significant differences in TNF-α, TGF-β, IL-10, IL-2, IL-6, IL-12p40, or IL-23. We did not observe significant differences in Th1/Th2/Th17 phenotype by flow cytometry at days 30 and 90 post HCT, and did not detect significant differences in serum cytokines at days 7 and 28 post HCT, except a significant reduction in IL-12/IL-23p40 in the ustekinumab-treated group at day 28 (Online Supplementary Figures S2-S4). No significant differences in gene expression were observed among peripheral blood mononuclear cells assayed with the NanoString nCounter GX Human Immunology V2 Kit (data not shown).
Regulatory T-cell frequency and suppressive function There was no significant difference in Treg/total CD4 at day 30 (ustekinumab: median 13, range 2-22; placebo:
median 11, range 3-20), both without (P=0.4) and with (P=0.4) adjustment for donor type (Figure 3). There was also no significant difference in absolute number of Treg or Treg suppressive function between groups (Figure 3).
Engraftment and early toxicity There were no significant differences in neutrophil (P=0.2) or platelet (P=0.5) engraftment between the ustekinumab- and placebo-treated patients, and all cases achieved these milestones promptly. No differences in thrombotic microangiopathy (TMA) (n=1 for each arm), TMA grade, or median time to TMA onset (48 vs. 62 days) were observed. No difference was observed in hepatic VOD (n=3 vs. n=2), VOD severity (moderate in all cases), or median time to VOD onset (28 vs. 31 days). In all cases, the syndrome resolved without intervention beyond diuresis, and no deaths occurred from VOD.
Acute and chronic GvHD Acute GvHD organ staging and overall grade are presented in Table 2. The day 100 cumulative incidence of grade II-IV acute GvHD was 33% for the ustekinumab arm versus 40% for the placebo arm (Figure 4A). Median time to acute GvHD onset was significantly longer among ustekinumab-treated patients compared to placebo (56 days vs. 28 days; P=0.02). Examining validated serum biomarkers (REG3α, ST2) of lethal GvHD and non-relapse mortality (NRM) at day 7 post HCT (analyzed by ELISA using methods previously described),21,22 we found ustekinumab-treated subjects had significantly lower REG3α compared to the placebo group [median: 55 (range 8-234) vs. 135 (21-387) ng/mL; P=0.014]; ST2 was not significantly different [median: 11,817 (range 3493-40,898) vs. 19,923 (range 3791-97,662) pg/mL; P=0.68] (Figure 5). There was no significant difference in cumulative incidence of any grade chronic GvHD or NIH moderate/severe chronic GvHD (Figure 4B) between groups. In addition, no significant differences were observed between groups for individual organ involvement and severity or overall NIH 0-3
Table 2. Acute graft-versus-host disease (GvHD) organ staging and overall grade.
Overall acute GvHD grade 0 I II III Skin stage 0 1 2 3 GI stage 0 1 2 Liver stage 0
Ustekinumab
Study arm Placebo
9 1 5 0
6 3 5 1
12 1 0 2
8 4 1 2
P=0.35
12 3 0
10 4 1
P=0.57
15
15
P=0.51
GI: gastrointestinal tract.
haematologica | 2018; 103(3)
535
J. Pidala et al.
score at both chronic GvHD onset and maximal severity (Online Supplementary Table S2).
Infectious complications Standard institutional infectious prophylaxis and monitoring was performed in all cases. There was no significant difference between ustekinumab and placebo groups for time from HCT to first systemic infection (median 27 vs. 26 days; P=0.7), total number of infections (n=14 vs. n=10; P=0.9), infection sites or organisms identified, or infection density (defined as infections per survival time, median 0.21 vs. 0.22; P=0.8). There was no significant difference in number with cytomegalovirus (CMV) reactivation more than 1000 copies (n=2 vs. n=1) and median time to CMV reactivation (57 vs. 24 days; P=0.3), Epstein-Barr virus (EBV) reactivation more than 1000 copies (n=0 vs. n=2) and median time to EBV reactivation (45 vs. 80 days; P=1), or HHV-6 reactivation more than 1000 copies (n=1 vs. n=2) and median time to HHV-6 reactivation (83 vs. 73 days; P=1.00).
Patient-reported outcomes There was no significant difference in change in Quality of Life between study arms (P=0.15) (Online Supplementary Figure S1). The Functional Assessment of Cancer TherapyBone Marrow Transplant (FACT-BMT) Trial Outcome Index (TOI) scores demonstrated transient worsening in both groups (P<0.001). Between-group comparisons at
A
each time point revealed no significant group differences in Quality of Life pre-HCT (P=0.32) or 30 days post HCT (P=0.98), although the placebo arm demonstrated better Quality of Life at 90 days post HCT (P=0.04). Mean Standard of Error (SE) FACT-BMT TOI scores at 90 days post HCT were 58.47 (3.19) in the ustekinumab arm and 68.51 (3.43) in the placebo arm.
Discussion Th1 and Th17 lymphocytes have been implicated in acute GvHD pathogenesis, and prior pre-clinical work has demonstrated that targeted disruption of key lineage-defining transcription factors (Tbet, RORy) or IL-12/IL-23p40 cytokine neutralization polarizes T-cell differentiation and mitigates GvHD lethality. In an original translational effort, targeted IL-12/IL-23p40 neutralization was tested in a blinded randomized trial to discern the biological impact of this intervention and to examine the clinical safety and initial efficacy in acute GvHD prevention. With the examined dose and schedule of ustekinumab delivered, IL-12/IL23p40 was neutralized through the early post-HCT period in vivo. Consistent with murine experimental data, donor alloresponse was polarized. Specifically, IFN-y production was significantly reduced among ustekinumab-treated patients in response to third-party stimulus at day 30, while response to HLA-matched host was low, as expected. IL-4
B
C D
Figure 3. Regulatory T cell (Treg) numbers and suppressive function at 30 days post HCT. (A) number of peripheral blood Treg/total CD4+ T cells at day 30 post HCT; (B) absolute number of Treg/ÎźL at day 30 post-HCT. (C) Representative Treg suppression assay, demonstrating reduced proliferation of T-responder cells in presence of escalating Treg:T-effector ratio; D) absolute Treg numbers to achieve IC-25.
536
haematologica | 2018; 103(3)
In vivo IL-12/IL-23p40 neutralization
production was significantly increased among ustekinumab-treated patients in response to third-party stimulus at day 90. IL-17 production was decreased among ustekinumab-treated patients in response to host at day 30, and to third-party at both days 30 and 90. These findings were likely detected due to the sensitivity of this functional assay, in contrast to that of circulating serum cytokine or peripheral blood phenotype assays. In contrast, this intervention failed to increase peripheral blood Treg. This could be due to several factors. In contrast to murine models, patients in both arms of this trial received sirolimus and tacrolimus throughout. The effect of this pharmacological immune suppression platform may have exerted effects on Treg reconstitution post HCT that overwhelmed any alteration from IL-12/IL-23p40 neutralization. In addition, cytokines relevant to Treg homeostasis were not targeted by the study intervention, and IL-6 (central to Th17/Treg balance) was not targeted by this approach. While anti-IL-6 receptor antibody (tocilizumab) therapy (together with cyclosporine and methotrexate) produced a low incidence of grade II-IV acute GvHD in a previous trial, peripheral blood Tregs were similarly not altered.23 In total, these data support the concept that targeted neutralization of IL-12/IL-23p40 or inhibition of IL-6 receptor signaling in the context of conventional pharmacological immune suppression fails to augment peripheral blood Tregs, and suggests that clinical benefit of these approaches is not dependent upon
A
increased Tregs. These data support an acceptable safety profile for combined IL-12/IL-23p40 neutralization among HCT recipients. Compared to the randomized placebo control in this trial, we observed no evidence of increased toxicity, impaired donor engraftment, infectious morbidity, or death. In contrast to particular infectious risks observed among patients with genetic deficiency in IL-12/IL-23p40 or IL-12Rβ,1,24 no cases of Mycobacterium tuberculosis or Salmonella were seen, and all Candida albicans infections (oral n=4, esophageal n=1) on trial were seen among the placebo arm. In addition, there was no evidence that IL12/23p40 depletion increased disease relapse. This finding is important, given the relevance of IL-12 to natural killer (NK) and CD8+ cytotoxic T lymphocytes and tumor control,25,26 and anti-tumor effects of IL-23 (although tumorpromoting effects of IL-23 have been demonstrated).24,27 A recent analysis of 3117 patients (with 8998 person-years of follow up) from 4 major randomized phase II and III ustekinumab trials supports the overall safety profile of this therapy with no evidence for opportunistic infections, or increased rates of malignancies or mortality above those of the general US population.28 The trial was not powered for clinical end points, and provides only initial estimates to be formally tested in a subsequent randomized trial. Specific analysis of acute GvHD target-organ differences is restricted by limited (skin) or no (liver) involvement of sites other than gas-
B
C
P=0.41
P=0.57
D
E
P=0.87
F
P=0.085 P=0.017
P=0.027
Figure 4. Comparison of clinical outcomes for ustekinumab versus placebo-treated subjects. (A) Cumulative incidence of grade II-IV acute graft-versus-host disease (GvHD). (B) Cumulative incidence of Naitonal Institutes of Health (NIH) moderate/severe chronic GvHD. (C) Cumulative incidence of malignancy relapse. (D) Cumulative incidence of non-relapse mortality. (E) Overall survival. (F) NIH moderate/severe chronic GvHD-, relapse-free survival (Conditional Random Fields Score, CRFS). *P-value for comparisons of cumulative incidence by Gray method; P-value for survival plots by log-rank test.
haematologica | 2018; 103(3)
537
J. Pidala et al.
P=0.68
P=0.014
Figure 5. Serum concentration of ST2 and REG3Îą at day 7 post-HCT.
trointestinal tract (GI). REG3Îą was significantly reduced in the ustekinumab-treated subjects at day 7 post HCT, suggesting that neutralization of IL-12/IL-23p40 may ameliorate early damage to the GI tract that ultimately cascades into GvHD lethality. This finding also requires confirmation in a subsequent trial. While there are conflicting data regarding the relative contribution of Th2 (vs. other Th subsets) to chronic GvHD development, our study demonstrates no evidence of worsened chronic GvHD after IL-12/IL-23p40 neutralization. We acknowledge as a limitation that IL-12/IL-23p40 neutralization together with sirolimus/tacrolimus did not offer patients complete protection from acute GvHD. Additional dosing of ustekinumab beyond the studied
References 1. Coghill JM, Sarantopoulos S, Moran TP, Murphy WJ, Blazar BR, Serody JS. Effector CD4+ T cells, the cytokines they generate, and GVHD: something old and something new. Blood. 2011;117(12):3268-3276. 2. Muranski P, Restifo NP. Essentials of Th17 cell commitment and plasticity. Blood. 2013;121(13):2402-2414. 3. O'Shea JJ, Paul WE. Mechanisms underlying lineage commitment and plasticity of helper CD4+ T cells. Science. 2010; 327(5969):1098-1102. 4. Carlson MJ, West ML, Coghill JM, Panoskaltsis-Mortari A, Blazar BR, Serody JS. In vitro-differentiated TH17 cells mediate lethal acute graft-versus-host disease with severe cutaneous and pulmonary pathologic manifestations. Blood. 2009; 113(6):1365-1374. 5. Iclozan C, Yu Y, Liu C, Liang Y, Yi T, Anasetti C, Yu XZ. T helper17 cells are sufficient but not necessary to induce acute graft-versus-host disease. Biol Blood Marrow Transplant. 2010;16(2):170-178.
538
approach may offer benefit, as anti-IL-12/IL-23p40 antibody levels declined and IL-12/IL-23p40 cytokine levels increased in the range of 50-60 days post HCT onward. Future trials could incorporate prolonged maintenance dosing modeled after approved maintenance therapy in psoriasis and Crohn disease. Funding We acknowledge the following funding support: Gateway for Cancer Research (to JP) G12-900 (inclusive of trial costs and study drug costs), American Cancer Society MRSG-11-149-01LIB, Moffitt Cancer Center Support Grant P30 CA076292 (flow cytometry, analytical pharmacology, biostatistics, and analytic microscopy cores).
6. Kappel LW, Goldberg GL, King CG, et al. IL-17 contributes to CD4-mediated graftversus-host disease. Blood. 2009; 113(4):945-952. 7. Hoffmann P, Ermann J, Edinger M, Fathman CG, Strober S. Donor-type CD4(+)CD25(+) regulatory T cells suppress lethal acute graft-versus-host disease after allogeneic bone marrow transplantation. J Exp Med. 2002;196(3):389-399. 8. Yu Y, Wang D, Liu C, et al. Prevention of GVHD while sparing GVL effect by targeting Th1 and Th17 transcription factor T-bet and RORgammat in mice. Blood. 2011; 118(18):5011-5020. 9. Wu Y, Bastian D, Schutt S, et al. Essential Role of Interleukin-12/23p40 in the Development of Graft-versus-Host Disease in Mice. Biol Blood Marrow Transplant. 2015;21(7):1195-1204. 10. Pidala J, Perez L, Beato F, Anasetti C. Ustekinumab demonstrates activity in glucocorticoid-refractory acute GVHD. Bone Marrow Transplant. 2012;47(5):747-748. 11. Bensinger SJ, Walsh PT, Zhang J, et al. Distinct IL-2 receptor signaling pattern in CD4+CD25+ regulatory T cells. J
Immunol. 2004;172(9):5287-5296. 12. Zeiser R, Leveson-Gower DB, Zambricki EA, et al. Differential impact of mammalian target of rapamycin inhibition on CD4+CD25+Foxp3+ regulatory T cells compared with conventional CD4+ T cells. Blood. 2008;111(1):453-462. 13. Sorror ML, Maris MB, Storb R, et al. Hematopoietic cell transplantation (HCT)specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. 14. Ho VT, Cutler C, Carter S, et al. Blood and marrow transplant clinical trials network toxicity committee consensus summary: thrombotic microangiopathy after hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2005;11(8): 571-575. 15. McDonald GB, Hinds MS, Fisher LD, et al. Veno-occlusive disease of the liver and multiorgan failure after bone marrow transplantation: a cohort study of 355 patients. Ann Intern Med. 1993;118(4):255-267. 16. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant.
haematologica | 2018; 103(3)
In vivo IL-12/IL-23p40 neutralization
1995;15(6):825-828. 17. Filipovich AH, Weisdorf D, Pavletic S, et al. National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: I. Diagnosis and staging working group report. Biol Blood Marrow Transplant. 2005;11(12):945-956. 18. Pidala J, Kim J, Jim H, et al. A randomized phase II study to evaluate tacrolimus in combination with sirolimus or methotrexate after allogeneic hematopoietic cell transplantation. Haematologica. 2012; 97(12):1882-1889. 19. Gray R. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Annals Of Statistics 1988; 16(3):1141-1154. 20. Hartwell MJ, Ozbek U, Holler E, et al. An early-biomarker algorithm predicts lethal graft-versus-host disease and survival. JCI
haematologica | 2018; 103(3)
Insight. 2017;2(3):e89798. 21. Vander Lugt MT, Braun TM, Hanash S, et al. ST2 as a marker for risk of therapy-resistant graft-versus-host disease and death. N Engl J Med. 2013;369(6):529-539. 22. Ferrara JL, Harris AC, Greenson JK, et al. Regenerating islet-derived 3-alpha is a biomarker of gastrointestinal graft-versus-host disease. Blood. 2011;118(25):6702-6708. 23. Kennedy GA, Varelias A, Vuckovic S, et al. Addition of interleukin-6 inhibition with tocilizumab to standard graft-versus-host disease prophylaxis after allogeneic stemcell transplantation: a phase 1/2 trial. Lancet Oncol. 2014;15(13):1451-1459. 24. Teng MW, Bowman EP, McElwee JJ, et al. IL-12 and IL-23 cytokines: from discovery to targeted therapies for immune-mediated inflammatory diseases. Nat Med. 2015; 21(7):719-729. 25. Trinchieri G. Interleukin-12 and the regula-
26.
27. 28.
29.
tion of innate resistance and adaptive immunity. Nat Rev Immunol. 2003; 3(2):133-146. Colombo MP, Trinchieri G. Interleukin-12 in anti-tumor immunity and immunotherapy. Cytokine Growth Factor Rev. 2002; 13(2):155-168. Ngiow SF, Teng MW, Smyth MJ. A balance of interleukin-12 and -23 in cancer. Trends Immunol. 2013;34(11):548-555. Papp KA, Griffiths CE, Gordon K, et al. Long-term safety of ustekinumab in patients with moderate-to-severe psoriasis: final results from 5 years of follow-up. Br J Dermatol. 2013;168(4):844-854. Armand P, Kim HT, Logan BR, et al. Validation and refinement of the Disease Risk Index for allogeneic stem cell transplantation. Blood. 2014;123(23):3664-3671.
539
ARTICLE
Platelet Biology & Its Disorders
Ferrata Storti Foundation
Variable impairment of platelet functions in patients with severe, genetically linked immune deficiencies Magdolna Nagy,1 Tom G. Mastenbroek,1* Nadine J.A. Mattheij,1* Susanne de Witt,1 Kenneth J. Clemetson,2 Janbernd Kirschner,3 Ansgar S. Schulz,4 Thomas Vraetz,5 Carsten Speckmann,6 Attila Braun,7 Judith M.E.M. Cosemans,1 Barbara Zieger6* and Johan W.M. Heemskerk1*
Haematologica 2018 Volume 103(3):540-549
Department of Biochemistry, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, Maastricht, the Netherlands; 2Department of Haematology, Inselspital, University of Bern, Switzerland; 3Department of Neuropediatrics and Muscle Disorders, Medical Center, University of Freiburg, Germany; 4Department of Pediatrics and Adolescent Medicine, University Medical Centre Ulm, Germany; 5Department of Pediatrics and Adolescent Medicine, Medical Center-University of Freiburg, Faculty of Medicine, Germany; 6Center for Chronic Immunodeficiency and Department of Pediatrics and Adolescent Medicine, Medical Centre, University of Freiburg, Germany and 7Institute of Experimental Biomedicine, University Hospital and Rudolf Virchow Centre, University of WĂźrzburg, Germany 1
*TGM, NJAM, BZ and JWMH contributed equally to this work.
ABSTRACT
I
Correspondence: jwm.heemskerk@maastrichtuniversity.nl
Received: July 26, 2017. Accepted: December 7, 2017. Pre-published: December 14, 2017. doi:10.3324/haematol.2017.176974 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/540 Š2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
540
n patients with dysfunctions of the Ca2+ channel ORAI1, stromal interaction molecule 1 (STIM1) or integrin-regulating kindlin-3 (FERMT3), severe immunodeficiency is frequently linked to abnormal platelet activity. In this paper, we studied platelet responsiveness by multi-parameter assessment of whole blood thrombus formation under high-shear flow conditions in 9 patients, including relatives, with confirmed rare genetic mutations of ORAI1, STIM1 or FERMT3. In platelets isolated from 5 out of 6 patients with ORAI1 or STIM1 mutations, storeoperated Ca2+ entry (SOCE) was either completely or partially defective compared to control platelets. Parameters of platelet adhesion and aggregation on collagen microspots were impaired for 4 out of 6 patients, in part related to a low platelet count. For 4 patients, platelet adhesion/aggregation and procoagulant activity on von Willebrand Factor (VWF)/rhodocytin and VWF/fibrinogen microspots were impaired independently of platelet count, and were partly correlated with SOCE deficiency. Measurement of thrombus formation at low shear rate confirmed a greater impairment of platelet functionality in the ORAI1 patients than in the STIM1 patient. For 3 patients/relatives with a FERMT3 mutation, all parameters of thrombus formation were strongly reduced regardless of the microspot. Bone marrow transplantation, required by 2 patients, resulted in overall improvement of platelet function. We concluded that multiparameter assessment of whole blood thrombus formation in a surface-dependent way can detect: i) additive effects of low platelet count and impaired platelet functionality; ii) aberrant ORAI1-mediated Ca2+ entry; iii) differences in platelet activation between patients carrying the same ORAI1 mutation; iv) severe platelet function impairment linked to a FERMT3 mutation and bleeding history.
Introduction Severe, genetically linked immunodeficiency can be accompanied by platelet function defects, especially in cases of rare mutations in the ORAI1, STIM1 and FERMT3 genes on platelet properties, in spite of solid evidence for a role of the mouse orthologs in arterial thrombosis. In platelets and other blood cells, stromal interaction molecule 1 (STIM1) acts as a major Ca2+ sensor located in the endoplasmic reticulum. When the reticular Ca2+ haematologica | 2018; 103(3)
Platelet function in genetic immune deficiencies
concentration is lower, it assembles with the ORAI1 Ca2+ (ICRAC) channels in the plasma membrane to mediate store-operated Ca2+ entry (SOCE).1-4 Therefore, the conventional test for SOCE (i.e. for STIM1 and ORAI1 activity) is to provoke depletion of the STIM1-linked Ca2+ store with the endoplasmic Ca2+ -ATPase inhibitor thapsigargin, and then measure Ca2+ entry through ORAI1 channels upon addition of extracellular CaCl2.5,6 Both ORAI1 and STIM1 have non-redundant roles in patroling and defense functions of white blood cells. In platelets, ORAI1 as well as STIM1 is considered to enhance the Ca2+ signal generation, induced in particular by protein tyrosine kinase-linked receptors, such as glycoprotein VI (GPVI). In mouse, both ORAI1 and STIM1 are implicated in hemostasis and arterial thrombus formation.7-10 Murine knockout studies have indicated that the ORAI1-STIM1 Ca2+ signaling contributes to multiple platelet activation processes, such as adhesiveness via integrin activation, granule release, aggregation of platelets, and procoagulant activity.8,11,12 However, in humans, the consequences of defective ORAI1 or STIM1 activity in platelets have not been thoroughly investigated. In humans, dysfunctional mutations in the ORAI1 or STIM1 genes are very rare.13 The patients described with such mutations usually suffer from severe immunodeficiency, congenital myopathy, ectodermal dysplasia or other Ca2+ -linked abnormalities.14-16 The immune deficiency can be attributed to the loss-of-function of leukocyte and lymphocyte subsets. Given the severity of the symptoms, often already evident at a young age, these patients are commonly treated by allogeneic hematopoietic stem cell transplantation. The few patients described do not have an overt bleeding history, although they may periodically show autoimmune thrombocytopenia.16 Leukocyte adhesion deficiency type III (LAD-III) forms another severe immune disease, in this case accompanied by epistaxis or petechiae. It is associated with dysfunctional mutations in the FERMT3 gene of the integrin-regulation protein, kindlin-3.13 LAD-III patients present with normal platelet count, but impaired platelet adhesion, which may explain the bleeding symptoms.17,18 These patients may require hematopoietic stem cell transplantation during childhood.19 Recently, we developed a microspot-based assay for multiparameter assessment of whole blood thrombus formation under flow conditions.20 This assay proved to be valuable in characterizing platelet count and function abnormalities in patients with a variety of genetic bleeding disorders. Here, we used this integrative method to investigate overall platelet functions in 9 patients with severe immunodeficiencies, including parents, with a confirmed dysfunctional mutation in ORAI1, STIM1 or FERMT3. For 2 patients, we also investigated the effects of bone marrow transplantation. The results suggest a variable phenotypic penetrance on platelet properties in patients and relatives carrying these mutations.
committees. Healthy controls had normal blood cell counts, had not received anti-platelet medication for at least two weeks, and did not have a known history of bleeding or immunodeficiency. Blood samples from patients with immunodeficiency and their parents (all genotyped) were collected at the Department of Pediatrics and Adolescent Medicine, University Medical Center in Freiburg, and at the Children’s Hospital, University of Ulm, Germany (Table 1). Simultaneously, blood samples were taken from healthy donors serving as daily travel controls. Relevant patients' characteristics, including identified mutations and clinical history, are summarized in Table 1. Patient P1 showed homozygosity in the R91W mutation in the ORAI1, known to be linked to severe combined immunodeficiency and defective T-cell Ca2+ signaling.1 Heterozygosity in this mutation was confirmed for both parents (P2 and P3). Patient P4 carried a heterozygous G98S mutation in ORAI1, which was confirmed in the mother (P5). This mutation is reported to be linked to tubular aggregated myopathy-2.21 Patient P6 carried a heterozygous R429C mutation of STIM1, which is also linked to T-cell immunity.22 Patient P7 suffered from severe immunodeficiency, linked to a homozygous R573X mutation in kindlin-3 encoded by the FERMT3 gene. The 2 parents (P8 and P9) were heterozygous for this mutation. Two of the patients with severe immunodeficiency, P1 and P7, were eligible for bone marrow transplantation. Blood samples in these cases were obtained before and at two or three months after transplantation, respectively. Only limited volumes of blood were available for investigation due to the young age of the patients.
Platelet Ca2+ responses Rises in cytosolic [Ca2+]i were measured in platelets after loading with Fura-2 acetoxymethyl ester (2.5 μM) by calibrated ratio fluorometry.23 Data are presented as nM increases in cytosolic [Ca2+].24
Multiparameter thrombus formation on microspots under flow Whole blood thrombus formation was assessed under flow conditions, basically as described before.20 In brief, blood samples (0.5 mL) were re-calcified in the presence of thrombin inhibitors, and perfused over a coverslip coated with three microspots (spot 1: type I collagen; spot 2: von Willebrand Factor (VWF)/rhodocytin; spot 3: VWF/fibrinogen) in a transparent parallel-plate perfusion chamber. Perfusion was at high wall-shear rate of 1600 s-1 for 3.5 minutes (min), or at a low shear rate of 150 s-1 for 6.0 min. The thrombi formed on the microspots were immediately post stained with FITC-labeled anti-fibrinogen mAb (1:100), FITC-anti-CD62P mAb (25 µg/mL) and AF647annexin A5 (0.25 μg/mL). Representative brightfield and fluorescence images were captured from each microspot in real-time without fixation. Duplicate runs were performed whenever possible. Analysis of brightfield and fluorescence images, giving 7 parameters per microspot, was performed with pre-defined scripts in Fiji software.25 (Un)supervised heatmaps using scaled parameters (range 0-10) of thrombus formation were constructed using R software.
Statistical analysis Methods Patients and controls Blood was drawn from patients and healthy controls after full informed consent in compliance with the Declaration of Helsinki. The studies were approved by the local medical ethics haematologica | 2018; 103(3)
Significance of differences was determined with the paired sample t-test (intervention effects) and by principal component analysis (PCA), using the statistical package for social sciences (SPSS, v.11.0). A detailed description of the methods used is available in the Online Supplementary Appendix. 541
M. Nagy et al.
Table 1. Subjects’ characteristics.
Subject
Mutation
Platelet count (x 109/L)
MPV (fL)
Hematocrit (L/L)
Clinical history
None assumed None assumed ORAI1 (R91W) W/W#
166-390 156-291 207
9.1-12.1 9.0-11.6 n.d.
0.36-0.49 0.36-040 0.41
BM transplanted
74+
8.9+
0.23+
P2 (mother P1) P3 (father P1) P4 (child)
ORAI1 (R91W) R/W## ORAI1 (R91W) R/W## ORAI1 (G98S) G/S##
115+ 156+ 114+
8.8+ 9.2 7.7+
0.36 0.32+ 0.34+
P5 (mother P4)
ORAI1 (G98S) G/S##
70+
9.2
0.37
P6 (adult) P7 (child)
STIM1 (R429C) R/C## FERMT3 (R573X) X/X#
150 182
n.d. 7.5+
n.d. 0.31+
P7 (BM) P8 (father P7) P9 (mother P7)
BM transplanted FERMT3 (R573X) R/X## FERMT3 (R573X) R/X##
168 126+ 142+
6.0+ 8.3 6.8+
0.33+ 0.42 0.39
None None Immunodeficiency, myopathy, anhydrosis Thrombocytopenia, stabilized myopathy Mild thrombocytopenia No known history of bleeding Thrombocytopenia (recurrent), myopathy with tubular aggregates, anhydrosis Thrombocytopenia, myopathy with tubular aggregates, anhydrosis No known history of bleeding Immunodeficiency, severe vaginal bleeding, leukocytosis No known clinical symptoms No known history of bleeding No known history of bleeding
HC1-12 (home C) C1-5 (travel C) P1 (child) P1 (BM)
HC: home control; C: travel control; n.d.: not determined; MPV: mean platelet volume; P: patient; BM: bone marrow. +Outside of normal range; #homozygosity in the indicated mutation; ##heterozygosity in the indicated mutation.
Results Variable aberrances in SOCE in platelets from patients with ORAI1 or STIM1 mutations Blood samples were obtained from 3 patients with a mutation in the Ca2+ flux-regulating proteins ORAI1 or STIM1, as well as from parents carrying the same mutation. For comparison, blood samples were also taken from two cohorts of healthy subjects, i.e. a group of normal home controls (HC1-12) and a group of normal travel controls (C1-6). Using Fura-2-loaded platelets, we first evaluated the alterations in Ca2+ signaling in response to the GPVI receptor agonist, convulxin, the PAR receptor agonist, thrombin, and the sarco/endoplasmic reticulum Ca2+ATPase (SERCA) inhibitor, thapsigargin. In each case, the platelets were first stimulated with agonist in Ca2+-free medium, after which extracellular CaCl2 was added to measure secondary Ca2+ entry. In comparison to control platelets, platelets from patient P1, carrying a homozygous ORAI191W/W mutation, responded normally to each agonist, but showed a substantially reduced Ca2+ entry following stimulation with convulxin but not thrombin (Figure 1A and B). Markedly, in this patient's platelets, Ca2+ entry after stimulation with thapsigargin (as a default condition for SOCE) was completely abolished. This pointed to a complete absence of the STIM1-ORAI1 pathway, similar to that established for platelets from STIM1- or ORAI1-deficient- mice.7,8 Flow cytometric evaluation indicated that convulxin-stimulated platelets from P1 showed a reduced phosphatidylserine 542
(PS) exposure [7±1% vs. 28±2% for control platelets, mean±SEM, n=3; P<0.05], but were not altered in integrin αIIbβ3 activation or P-selectin expression (P>0.10). Platelets from the patient's parents (both with confirmed heterozygosity) were diminished in SOCE, but to a different extent (Figure 2). Platelets from the mother (P2) showed a nearly annulled Ca2+ entry, whereas platelets from the father (P3) were less severely reduced when compared to platelets from two cohorts of healthy controls. It was possible to obtain a second blood sample from patient P1 at two months after bone marrow transplantation. The platelets showed a 50% recovery in SOCE signal after thapsigargin stimulation (SOCEP1: 22 nM vs. SOCEP1-BM: 575 nM). Platelets from one patient (P4), carrying a heterozygous mutation ORAI189G/S in the same protein region, responded differently. These platelets displayed a high SOCE signal (Ca2+ entry after thapsigargin). In contrast, platelets from parent P5 (also carrying the mutation) were greatly reduced in SOCE (Figure 2). This difference was confirmed for a second set of blood samples (data not shown). A different picture was obtained with patient P6 with a heterozygous mutation in STIM1429R/C.22 In these platelets, increases in Ca2+ evoked by convulxin/CaCl2 or thrombin/CaCl2 were in the normal range, whereas SOCE after thapsigargin was completely abolished (Figure 3A and B). Phenotypic analysis of platelets was also performed of a LAD-III patient (P7) who carried a homozygous FERMT3573X/X mutation, and presented with immunodeficiency and a history of bleeding.26 Flow cytometry indicated an almost complete lack of integrin αIIbβ3 activation in haematologica | 2018; 103(3)
Platelet function in genetic immune deficiencies
A
B
Figure 1. Defective Ca2+ entry in platelets from a patient with a homozygous R91W mutation in ORAI1. Fura-2-loaded platelets from patient P1 (ORAI1W/W) and travel control C1 were used, suspended in Hepes buffer with 0.1 mM EGTA. Rises in cytosolic Ca2+ were measured in time upon stimulation with convulxin (Cvx, 50 ng/mL), thrombin (Thr, 4 nM) or thapsigargin (Thaps, 1.0 μM). After a defined time, CaCl2 (2 mM) was added to induce Ca2+ entry. Representative traces (A) and quantification (B) of Ca2+ increases in platelets from control subject and patient P1. Note that the maximal Ca2+ rise in response to CaCl2 following thapsigargin was used as a measure of store-operated Ca2+ entry (SOCE) capacity.
response to receptor agonists, as observed in another patient carrying this mutation.27 In platelets from the 2 heterozygous parents (P8, P9), αIIbβ3 integrin activation was in the normal range (data not shown).
Lower range blood cell counts in patients with ORAI1, STIM1 or FERMT3 mutations Table 1 provides an overview of the clinical histories of patients P1-9 including hematologic parameters determined in freshly isolated blood samples. In all patients, with the exception of P1 and P7, platelet counts were between 70 and 150x109/L, which are slightly below the normal ranges presented in both control cohorts. In the majority of patients, also hematocrit levels were in the lower range of normal, i.e. between 0.31 and 0.42 L/L. After bone marrow transplantation (2-3 months), platelet count in P1 and P7 had restored to 74 and 168x109/L, respectively.
Different patterns between patients of aberrant whole-blood thrombus formation at high shear rate To obtain detailed insight into the hemostatic potential of the patients' platelets, whole blood samples were used for multiparameter assessment of thrombus formation under flow at high wall-shear rate of 1600 s-1. Samples from corresponding control subjects (C1-6) and home control subjects (HC1-12) were again used for comparison. This high-throughput method, previously established,20 allowed simultaneous examination of platelet adhesion, aggregation and activation at three microspots: spot 1 with coated type I collagen (involving platelet receptors: GPIb-V-IX, GPVI and integrin α2β1); spot 2 with coated VWF/rhodocytin (receptors GPIb-V-IX and CLEC-2); and spot 3 with coated VWF/fibrinogen (receptors GPIb-V-IX and αIIbβ3). From each microspot, microscopic brightfield images were recorded to assess platelet adhesion and aggregation (parameters V1-4); in parallel, fluorescence images were recorded in three colors for detection of PS haematologica | 2018; 103(3)
exposure, P-selectin expression and αIIbβ3 activation (parameters V5-7). Representative examples of brightfield and fluorescence images from spot 1 for blood samples of control C2 and immunodeficient patients with ORAI1 or STIM1 mutation (P1, P4, and P6) are shown in Figure 4. Regarding the patient blood samples, the images show patterns of platelet adhesion, aggregation and activation that vary from from normal to reduced, as was also observed for spot 2 (VWF/rhodocytin) and spot 3 (VWF/fibrinogen) (see below). A heatmap was constructed with data from all analyzed images (3 spots × 7 parameters) for the cohort of 12 normal home controls (HC1-12), the 6 travel controls (C1-6), and the 9 individual patients/relatives with mutations in ORAI1 (P1-5), STIM1 (P6) or FERMT3 (P7-9), in which all values were normalized to a scale of 0-10 per parameter (Figure 5A). In the derived subtraction heatmap of Figure 5B, the patient data were expressed relative to those from the home controls. This subtraction confirmed an overall high similarity between the values of the two control groups (HC1-12 and C1-6), with the exception of a parameter reflecting platelet deposition on spot 2. Datasets for individual subjects C1-6 were within the normal ranges (data not shown). This confirmed the usefulness of the multi-parameter test,20 and underscored the quality of the analyzed blood samples. For each patient/relative, we arbitrarily set a relevant effect threshold, i.e. when outside the range of mean±2 SD of the control group (HC1-12). This filter produced a 'relevant' subtraction heatmap, indicating distinct patterns of altered parameters of thrombus formation for individual patients P1-9 (Figure 5C). Figure 5C underlines that, overall, multiple parameters on spot 1 were reduced for patients P2-5, whereas especially parameters on spots 2-3 were reduced for patients P1 and P5. Interestingly, patient P4 carrying the assumed gain-of-function mutation ORAI1G98S (but not relative P5 with low SOCE) showed a typical increase in platelet PS 543
M. Nagy et al. Table 2. Overall comparison of defective store-operated Ca2+ entry (SOCE) and impairment in thrombus formation of individual patients.
# ## ## ## ## # ## ##
For indicated patients, measurements of store-operated Ca2+ entry (SOCE) (see Figure 2) and differential heatmap data relative to control cohort HC1-12 (see Figure 5C) were tabled to obtain an overview of the changes in platelet function. For thrombi (1600 s-1) formed on the three microspots (Sp1-3), all parameters (V1-7) were scored as within normal range (0), decreased (-1) or increased (+1). Per patient, summed scores are indicated (column 'Sum'). In addition, fractions of altered scores are indicated for three phases of thrombus formation. Color intensity corresponds with the relative changes in parameters. Phase 1 refers to platelet adhesion and aggregation (Sp1-3,V1-4), Phase 2 to platelet integrin activation and granule secretion (Sp1-3,V6-7), and Phase 3 to platelet procoagulant activity (Sp1-3,V5). n.d.: not determined; #homozygosity in the indicated mutation; ##heterozygosity in the indicated mutation.
exposure. For the patient with homozygous FERMT3 mutation (P7), a more severe reduction on all three spots was seen in comparison to the 2 heterozygous relatives (P8 and P9).
Effects of low platelet count Considering that whole blood thrombus formation can be influenced by not only the inherited platelet disorder, but also a low platelet count,28 the present heatmaps may reflect both platelet-related properties. To examine this in more detail, the subtraction heatmap data were extended with values of platelet count and SOCE. Unsupervised clustering of the extended dataset indicated that particularly spot 1, parameters (V1-4) clustered with platelet count, while spot 2 parameters (V3-5) and PS exposure (Sp1V5, Sp2V5) clustered with altered SOCE (Online Supplementary Figure S1). As an alternative approach, we used the primary (nonnormalized) data of thrombus formation (Sp1-3, V1-7), platelet count and SOCE for PCA. This revealed a similar pattern, in that the majority of spot 1 parameters together with platelet count determined component 1, whereas most spot 2 parameters determined component 2 (Online Supplementary Figure S2A). Pearson regression coefficient confirmed a correlation of spot 1 parameters with platelet count, and also a correlation of platelet PS exposure (Sp1V5, Sp3V5) with SOCE (Online Supplementary Figure S2B). This information was then used to interpret the alterations in thrombus formation for individual patients (see clustered heatmap in Online Supplementary Figure S1). Concerning the patient (P1) with homozygous R91W mutation in ORAI1, with normal platelet count (207x109/L) and near abolished SOCE, thrombus formation was near normal on spot 1 (collagen), but markedly reduced on spot 2 (VWF/rhodocytin) and spot 3 (VWF/fibrinogen). The impaired Ca2+ signal was linked to a low PS exposure on all spots (Online Supplementary Table S1). On the other hand, for the patients P2 and P3 with heterozygous R91W mutation in ORAI1 (relatively low platelet counts of 115-156x109/L, and 20-60% of normal SOCE, 544
Figure 2. Variably defective Ca2+ entry in platelets from patients with ORAI1W/W, ORAI1G/S or STIM1R/C mutations. Fura-2-loaded platelets in Hepes buffer with 0.1 mM EGTA were stimulated with thapsigargin (1.0 ÎźM), and after a defined time with CaCl2 (2 mM). Platelets were analyzed from 12 healthy home controls (HC1-12), 6 healthy travel controls (C1-6), and the indicated patients/relatives (P1-6). Platelets from patient P1 were also analyzed after bone marrow (BM) transplantation. Shown are maximal increases in Ca2+. Dotted lines indicate range of store-operated Ca2+ entry (SOCE) levels (CaCl2-induced Ca2+ rise after thapsigargin) for platelets from home controls (mean Âą SD, for HC1-12).
respectively), particularly parameters for spot 1 were below normal, with a reduced PS exposure. Concerning patients P4 and P5, heterozygous for the G98S mutation in ORAI1 (low platelet counts of 70 and 114x109/L; 100% and 30% of normal SOCE, respectively), thrombus formation on spot 1 was reduced, while parameters of thrombus formation on spots 2-3 (including PS exposure) were only reduced for patient P5. In P4 (but not P5), a gain-of-function of Ca2+ channel activity was apparent from an increased PS exposure on spots 2-3. For patient P6, heterozygous for the R429C mutation in haematologica | 2018; 103(3)
Platelet function in genetic immune deficiencies
A
B
Figure 3. Defective Ca2+ entry in platelets from a patient with a heterozygous R429C mutation in STIM1. Fura-2-loaded platelets from patient P6 (STIM1R/C) and travel control C6, suspended in Hepes buffer with 0.1 mM EGTA, were stimulated with convulxin (Cvx, 50 ng/mL), thrombin (Thr, 4 nM) or thapsigargin (Thaps, 1.0 ÎźM), after which CaCl2 (2 mM) was given to induce Ca2 entry. Representative traces (A) and quantification (B) of Ca2+ rises in platelets from control subject and patient P6.
STIM1, thrombus formation on spot 1 (collagen) was normal in spite of the abolished SOCE. Together, this reinforced the idea that a low platelet count rather than altered SOCE determines the thrombus formation on collagen, but not on the other surfaces. Control experiments with reconstituted blood confirmed that, for control donors, lowering of the platelet count to 100x109/L had only a limited effect on particular parameters of thrombus formation limited to spot 1, whereas lowering to 50x109/L affected thrombus formation on all spots 1-3 (Online Supplementary Figure S3). Taken together with the predictive effect of platelet count in the PCA, this suggests that a relatively low count affects thrombus formation under flow conditions more severely when combined with a reduced platelet functionality.
Patterns of aberrant whole-blood thrombus formation at low shear rate Using remaining samples from patients/relatives P1-6, we went on to perform whole blood flow measurements at a low shear rate of 150 s-1. Subtraction heatmap analysis of the normalized parameter values, in comparison to control data from HC1-12 cohort, indicated for all patients with ORAI1 mutations (P1-5) a consistent pattern of reduced thrombus formation parameters on spot 1 (Online Supplementary Figure S4A-C). In particular for patients P2 and P3, markers of platelet activation appeared to be reduced (PS exposure, P-selectin expression and ιIIbβ3 activation). Once again, for patient P6 with STIM1 mutation, most parameters were unchanged. Principal component analysis of the low shear data again pointed to a linkage of spot 1 parameters with platelet count and a linkage of Sp1V5 (PS exposure) with SOCE activity (Online Supplementary Figure S5A and B).
Partial recovery of whole blood thrombus formation after bone marrow transplantation Blood samples from 2 patients were also obtained after bone marrow transplantation (P1 after 2 months, P7 after 3 months), and used for platelet activation and thrombus haematologica | 2018; 103(3)
formation studies. Transplantation of patient P1 led to a partially recovered SOCE from 20% to 60% of normal (Figure 2). After transplantation, parameters of thrombus formation were particularly reduced on spot 1 (paralleling a reduction in platelet count from 207 to 74x109/L), but were unchanged for spot 2, and enhanced for spot 3 (Figure 6). Transplantation of patient P7 resulted in an overall improvement of most parameters on all spots (platelet count decreased from 182 to 168x109/L).
Discussion In the present study, we used a multiparameter test of whole blood thrombus formation under flow conditions, as a proxy measurement of hemostatic activity20 to characterize quantitative and qualitative platelet abnormalities in rare patients and their relatives with severe immunodeficiencies, linked to signaling protein defects and mutations in the ORAI1, STIM1 and FERMT3 genes. So far, functional effects of the ORAI1 and STIM1 mutations have only been described for human immune cells or cell lines. Hence, the present data are the first to report on comparative alterations in platelet SOCE and platelet functions in as many as 9 genotyped patients/relatives.
Aberrations in SOCE accompanied by mutations in ORAI1 or STIM1 Earlier work with bone marrow chimeric mice with megakaryocytic deficiency in Orai1 or Stim1 demonstrated a prominent role of these Ca2+-entry regulating proteins in platelet calcium homeostasis and activation, including PS exposure.11,29 Our human data are compatible with these findings in that SOCE after thapsigargin or GPVI stimulation appeared to be impaired in platelets from patient P1 with a homozygous ORAI1W/W mutation. A similar impairment of SOCE has been reported in platelets from Orai1R93W mice, i.e. a loss-of-function mutation orthologous to the human ORAI1 R91W variant.29 Platelets from the latter mice showed a defect in Ca2+ flux545
M. Nagy et al.
es and other responses, when triggered with low concentrations of (thrombin or collagen) receptor agonists. These platelets displayed normal aggregation under flow conditions, but a decreased procoagulant activity (PS exposure).29 Relatives P2 and P3, heterozygous for the R91W mutation, showed approximately 20-50% residual SOCE activity, which is compatible with co-expression of the non-mutated allele in the platelets. Interestingly, Ca2+ entry after thrombin stimulations remained unaffected in patients/relatives P1-3, which is explained by involvement of the SOCE-independent Ca2+ entry mechanism through TRPC6 channels.30 On the other hand, the R98S mutation in ORAI1 carried by patient P4, with assumed gain-of-function,31 was not accompanied by altered SOCE activity after thapsigargin or GPVI stimulation, whereas SOCE was substantially reduced in the relative P5 (confirmed in two independent blood samples). The difference in SOCE activity between P4 and P5 might be explained by a different 'penetration' of the mutated allele in ORAI1 expression in megakary-
ocytes and platelets. However, an alternative explanation is the presence of other modifying genetic or acquired factors between P4 and P5, with possible effects on platelet SOCE activity, on which we can only speculate. Concerning patient P6 with a heterozygous R429C mutation in STIM1, Ca2+ entry in platelets was impaired after thapsigargin, but not after GPVI or PAR1/4 stimulation. In mammalian cell lines, the R429 mutation modulates the C-terminal oligomerization and puncta formation of STIM1 with ORAI1.32 Our results suggest that, in platelets, this mutation strongly inhibits the interaction of STIM1 with ORAI1 after full Ca2+ store depletion, such as that provoked by thapsigargin.
Altered platelet functions in thrombus formation accompanied by mutations in ORAI1, STIM1 or FERMT3 Previous murine studies have pointed out that deficiency in platelet ORAI1 or STIM1 led to a moderately reduced collagen-dependent thrombus formation with minimal effect on bleeding,8,11 whereas murine deficiency
Figure 4. Altered thrombus formation of blood from patients with ORAI1W/W, ORAI1G/S or STIM1R/C mutations. Whole blood from indicated control subjects and patients was perfused over three microspots [(Sp1, collagen type I; Sp2, von Willebrand Factor (VWF)/rhodocytin; Sp3, VWF/fibrinogen; downstream → upstream)] at wall-shear rate of 1600 s-1. After 3.5 minutes of perfusion, brightfield images were taken from thrombi on all microspots, and platelets were stained for phosphatidylserine (PS) exposure (AF568-annexin A5), P-selectin expression (AF647 anti-CD62P mAb), and integrin αIIbβ3 activation (FITC anti-fibrinogen mAb). Shown are representative brightfield and fluorescence images at spot 1, obtained with blood from control C2 and patients P1, P4 and P6 (bars, 50 µm).
546
haematologica | 2018; 103(3)
Platelet function in genetic immune deficiencies
in FERMT3 (kindlin-3) resulted in impaired thrombus formation and a clear bleeding phenotype.33 Furthermore, both ORAI1-deficient and ORAI1R93W mouse platelets were found to be partly defective in procoagulant activity (PS exposure), along with the annulled SOCE activity.8,11,29 In the present study, we observed even more variable platelet phenotypes in the 9 patients/relatives with a
mutation in ORAI1, STIM1 or FERMT3. Heatmap analysis indicated extensive but distinct patterns of reduced thrombus formation between patients, compared to blood from cohorts of healthy control subjects. Blood analysis further indicated that most patients/relatives carrying such a mutation had platelet counts below or near the lower range of normal. Low platelet count was found to corre-
A
B
C
Figure 5. Integrated analysis of thrombus formation for patients with ORAI1, STIM1 or FERMT3 mutations. Thrombus formation on three microspots was measured with blood from home controls (HC1-12), travel controls (C1-6) and indicated patients/relatives with a genetic mutation in ORAI1 (P1-5), STIM1 (P6) or FERMT3 (P79), at wall-shear rate of 1600 s-1, as for Figure 4. Coding of microspots: Sp1, collagen type I; Sp2, von Willebrand Factor (VWF)/rhodocytin; Sp3, VWF/fibrinogen. Coding of outcome parameters: V1, thrombus morphological score (scale 0-5); V2, platelet surface area coverage (% SAC); V3, thrombus contraction score (scale 03); V4, thrombus multilayer score (scale 0-3); V5, phosphatidylserine (PS) exposure (% SAC); V6, P-selectin expression (% SAC); V7, αIIbβ3 activation (% SAC). Data were scaled per parameter from 0-10. (A) Heatmap of scaled values for control groups HC1-12 and C1-6 (means); and of scaled values for individual patients (*) and relatives. (B) Subtraction heatmap of scaled values, compared to those from HC1-12. (C) Subtraction heatmap after filtering for differences considered to be relevant, i.e. outside the range of mean ± 2 SD (HC1-12).
haematologica | 2018; 103(3)
547
M. Nagy et al.
Figure 6. Altered thrombus formation after bone marrow transplantation. Thrombus formation (1600 s-1) on three microspots measured with blood from indicated home controls (HC1-12), travel controls (C1-6) and 2 patients P1 (ORAI1W/W) and P7 (FERMT3R/X), both before (*) and after bone marrow transplantation (BMT). For description of microspots and parameters, see Figure 4. Subtraction heatmap of scaled values after filtering for differences considered to be relevant, i.e. outside the range of mean ± 2 SD (HC1-12).
late with low thrombus formation parameters, especially on spot 1. By comparison with flow studies using blood from healthy controls at lower counts, it appears that mild thrombocytopenia can lead to a more severe reduction in thrombus formation if combined with lower platelet functionality. Mild thrombocytopenia for patients with a lossof-function mutation in ORAI1 or STIM1 has been reported before.16 So far, published papers on patients with a FERMT3 mutation have not reported on thrombocytopenia.26,27 Because of the relatively large number of patients in our study, we were able to separate the thrombus formation parameters linked to qualitative or quantitative platelet defects. Both unsupervised clustering of the heatmap data and PCA of the raw data indicated that mostly thrombus parameters on spot 1 (collagen; platelet receptors GPIb-VIX, GPVI, α2β1) showed a dependency on platelet count, whereas those of spot 2 (VWF/rhodocytin; receptors GPIb-V-IX, CLEC-2) and spot 3 (VWF/fibrinogen; receptors GPIb-V-IX, αIIbβ3) were not dependent on platelet count. Markedly, this was true for both the high shear and low shear flow tests. An explanation for this finding is that, with collagen as a relatively strong agonist for GPVI, on spot 1 platelet delivery (thus, count) rather than platelet activation is a limiting factor for thrombus-forming parameters. In contrast, the immobilized ligands of spots 2 and 3, being less platelet-stimulating, may rely more on full platelet activation including normal SOCE and αIIbβ3 integrin activation. Correlation analysis also indicated that PS exposure was a key parameter linked to SOCE, in agreement with earlier mouse data.10,11
Improved platelet functions after bone marrow transplantation Bone marrow transplantation of patient P1 (ORAI1W/W), two months before, resulted in improved SOCE activity
References 1. Feske S, Gwack Y, Prakriya M, et al. A mutation in Orai1 causes immune deficiency by abrogating CRAC channel function. Nature. 2006;441(7090):179-185.
548
and normalized thrombus formation on spot three but not on spot 1 (linked to a low platelet count). Transplantation of P7 (FERMT3X/X), three months before, led to an overall improvement of thrombus formation parameters (at normal platelet count). The partial restoration of platelet count at two months post transplantation is compatible with reports that a full normalization can take several months.26,34 Based on the relevant differential analysis of thrombus formation parameters (Figure 5C), Table 2 provides a summative overview for each patient of alterations in: platelet adhesion/aggregation, integrin activation/secretion and procoagulant activity (phases 1-3). This approach is based on the rationale that this whole blood flow assay senses additive effects of low platelet count and impaired platelet functionality. Table 2 shows an overall defect in adhesion/aggregation for all patients/relatives with ORAI1 (R91W) mutation, as well as a defect in procoagulant activity. It also underlines the fact that the assumed gain-of-function ORAI1 (G98S) mutation in patient P4 is accompanied by a typical increase in procoagulant activity (but not in relative P5 with low SOCE). Furthermore, in patients carrying the FERMT3 (R573X) mutation, all platelet responses appeared to be more severely reduced in cases of homozygosity than those of heterozygosity. This may be of clinical relevance, since only the homozygous carrier P7 had a history of bleeding (Table 1). Funding Financial support from the Netherlands Centre for Translational Molecular Medicine (CTMM, MICRO-BAT), the Interreg V Euregio Meuse-Rhine program (Poly-Valve), Dutch Heart Foundation (2015T79 to TGM and JMEMC) and the Netherlands Organization for Scientific Research (NWO Vidi 91716421 to JMEMC).
2. Penna A, Demuro A, Yeromin AV, et al. The CRAC channel consists of a tetramer formed by STIM-induced dimerization of Orai dimers. Nature. 2008;456(7218):116120. 3. Luik RM, Wang B, Prakriya M, Wu M, Lewis RS. Oligomerization of STIM1 cou-
ples ER calcium depletion to CRAC channel activation. Nature. 2008;454(7203):538542. 4. Soboloff J, Rothberg BS, Madesh M, Gill DL. STIM proteins: dynamic calcium signal transducers. Nat Rev Mol Cell Biol. 2012;13(9):549-565.
haematologica | 2018; 103(3)
Platelet function in genetic immune deficiencies
5. Varga-Szabo D, Braun A, Nieswandt B. STIM1 and Orai1 in platelet function. Cell Calcium. 2011;50(3):270-278. 6. Ambily A, Kaiser WJ, Pierro C, et al. The role of plasma membrane STIM1 and Ca2+entry in platelet aggregation. STIM1 binds to novel proteins in human platelets. Cell Signal. 2014;26(3):502-511. 7. Varga-Szabo D, Braun A, Kleinschnitz C, et al. The calcium sensor STIM1 is an essential mediator of arterial thrombosis and ischemic brain infarction. J Exp Med. 2008; 205(7):1583-1591. 8. Braun A, Varga-Szabo D, Kleinschnitz C, et al. Orai1 (CRACM1) is the platelet SOC channel and essential for pathological thrombus formation. Blood. 2009; 113(9):2056-2063. 9. Gilio K, Harper MT, Cosemans JM, et al. Functional divergence of platelet protein kinase C (PKC) isoforms in thrombus formation on collagen. J Biol Chem. 2010; 285(30):23410-23419. 10. Van Kruchten R, Braun A, Feijge MA, et al. Antithrombotic potential of blockers of store-operated calcium channels in platelets. Arterioscler Thromb Vasc Biol. 2012;32(7):1717-1723. 11. Gilio K, van Kruchten R, Braun A, et al. Roles of platelet STIM1 and Orai1 in glycoprotein VI- and thrombin-dependent procoagulant activity and thrombus formation. J Biol Chem. 2010;285(31):23629-23638. 12. Heemskerk JW, Mattheij NJ, Cosemans JM. Platelet-based coagulation: different populations, different functions. J Thromb Haemost. 2013;11(1):2-16. 13. Malacards: human disease database. 2016. Available from: www.malacards/org 14. Shaw PJ, Feske S. Regulation of lymphocyte function by ORAI and STIM proteins in infection and autoimmunity. J Physiol. 2012;590(17):4157-4167. 15. Nakamura L, Sandrock-Lang K, Speckmann C, et al. Platelet secretion defect in a patient with stromal interaction molecule 1 deficiency. Blood. 2013;122(22):3696-3698. 16. Lacruz RS, Feske S. Diseases caused by
haematologica | 2018; 103(3)
17.
18.
19.
20.
21.
22.
23.
24.
25.
mutations in ORAI1 and STIM1. Ann N Y Acad Sci. 2015;1356:45-79. Kuijpers TW, van de Vijver E, Weterman MAJ, et al. LAD-1/variant syndrome is caused by mutations in FERMT3. Blood. 2009;113(19):4740-4746. Van de Vijver E, De Cuyper IM, Gerrits AJ, et al. Defects in Glanzmann thrombasthenia and LAD-III (LAD-1/v) syndrome: the role of integrin 1 and 3 in platelet adhesion to collagen. Blood. 2012;119(2):583586. Stepensky PY, Wolach B, Gavrieli R, et al. Leukocyte adhesion deficiency type III: clinical features and treatment with stem cell transplantation. J Pediatr Hematol Oncol. 2015;37(4):264-268. De Witt SM, Lamers MME, Swieringa F, et al. Identification of platelet function defects by multi-parameter assessment of thrombus formation. Nat Commun. 2014;5:4257. Endo Y, Noguchi S, Hara Y, et al. Dominant mutations in ORAI1 cause tubular aggregate myopathy with hypocalcemia via constitutive activation of store-operated Ca2+ channels. Hum Mol Genet. 2015;24(3):637648. Fuchs S, Rensing-Ehl A, Speckmann C, et al. Antiviral and regulatory T cell immunity in a patient with stromal interaction molecule 1 deficiency. J Immunol. 2012; 188(3):1523-1533. Feijge MA, van Pampus EC, LacabaratzPorret C, et al. Inter-individual varability in Ca2+ signalling in platelets from healthy volunteers, relation with expression of endomembrane Ca2+-ATPases. Br J Haematol. 1998;102(3):850-859. Heemskerk JW, Vis P, Feijge MA, et al. Roles of phospholipase C and Ca2+ATPase in calcium responses of single, fibrinogen-bound platelets. J Biol Chem. 1993;268(1):356-363. Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012; 9(7):676-682.
26. Crazzolara R, Maurer K, Schulze H, et al. A new mutation in the KINDLIN-3 gene ablates integrin-dependent leukocyte, platelet, and osteoclast function in a patient with leukocyte adhesion deficiency-III. Pediatr Blood Cancer. 2015;62(9):16771679. 27. Jurk K, Schulz AS, Kehrel BE, et al. Novel integrin-dependent platelet malfunction in siblings with leukocyte adhesion deficiency-III (LAD-III) caused by a point mutation in FERMT3. Thromb Haemost. 2010; 103(5):1053-1064. 28. Cauwenberghs S, Feijge MA, Theunissen E, et al. Novel methodology for the assessment of platelet transfusion therapy by measuring increased thrombus formation and thrombin generation. Br J Haematol. 2007;136(3):480-490. 29. Bergmeier W, Oh-Hora M, McCarl CA, et al. R93W mutation in Orai1 causes impaired calcium influx in platelets. Blood. 2009;113(3):675-678. 30. Hassock SR, Zhu MX, Trost C, Flockerzi V, Authi KS. Expression and role of TRPC proteins in human platelets: evidence that TRPC6 forms the store-independent calcium entry channel. Blood. 2002;100(8):28012811. 31. Bohm J, Bulla M, Urquhart JE, et al. ORAI1 mutations with distinct channel gating defects in tubular aggregate myopathy. Hum Mutat. 2017;38(4):426-438. 32. Maus M, Jairaman A, Stathopulos PB, et al. Missense mutation in immunodeficient patients shows the multifunctional roles of coiled-coil domain 3 (CC3) in STIM1 activation. Proc Natl Acad Sci USA. 2015; 112(19):6206-6211. 33. Moser M, Nieswandt B, Ussar S, Pozgajova M, Fassler R. Kindlin-3 is essential for integrin activation and platelet aggregation. Nat Med. 2008;14(3):325-330. 34. Takami A, Shibayama M, Orito M, et al. Immature platelet fraction for prediction of platelet engraftment after allogeneic stem cell transplantation. Bone Marrow Transplant. 2007;39(8):501-507.
549
ARTICLE
Coagulation and its Disorders
Ferrata Storti Foundation
Desmopressin in moderate hemophilia A patients: a treatment worth considering
Janneke I. Loomans,1 Marieke J.H.A. Kruip,2 Manuel Carcao,3 Shannon Jackson,4 Alice S. van Velzen,1 Marjolein Peters,1 Elena Santagostino,5 Helen Platokouki,6 Erik Beckers,7 Jan Voorberg,8 Johanna G. van der Bom9,10 and Karin Fijnvandraat1,8 for the RISE consortium
Department of Pediatric Hematology, Immunology and Infectious diseases, Emma Children’s Hospital, Amsterdam, the Netherlands; 2Department of Hematology, Erasmus University Medical Center, Rotterdam, the Netherlands; 3Division of Haematology/Oncology, Department of Paediatrics and Child Health Evaluative Sciences, Research Institute, The Hospital for Sick Children, Toronto, ON, Canada; 4 Division of Hematology, Department of Medicine, St. Paul’s Hospital and University of British Columbia, Vancouver, BC, Canada; 5A. Bianchi Bonomi Hemophilia and Thrombosis Center, IRCCS Ca’ Granda Foundation, Maggiore Hospital Policlinico, Milan, Italy; 6Aghia Sofia Children’s Hospital, Athens, Greece; 7Maastricht University Medical Centre, the Netherlands; 8Department of Plasma Proteins, Sanquin Research, Amsterdam, the Netherlands; 9Leiden University Hospital, the Netherlands and 10Sanquin Research, Leiden, the Netherlands 1
Haematologica 2018 Volume 103(3):550-557
ABSTRACT
D
Correspondence: c.j.fijnvandraat@amc.uva.nl
Received: September 1, 2017. Accepted: December 27, 2017. Pre-published: January 5, 2018. doi:10.3324/haematol.2017.180059 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/3/550 ©2018 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.
550
esmopressin increases endogenous factor VIII levels in hemophilia A. Large inter-individual variation in the response to desmopressin is observed. Patients with a lower baseline factor VIII activity tend to show a reduced response, therefore, desmopressin is less frequently used in moderate hemophilia A patients (baseline factor VIII activity 1-5 international units/deciliter), even though factor VIII levels may rise substantially in some of them. We aim to describe the response to desmopressin in moderate hemophilia A patients and to identify predictors. We selected data on 169 patients with moderate hemophilia from the multicenter Response to DDAVP In non-severe hemophilia A patients: in Search for dEterminants (RISE) cohort study. Adequate response to desmopressin was defined as a peak factor VIII level ≥ 30, and excellent response as ≥ 50 international units/deciliter after desmopressin administration. We used univariate and multiple linear regression techniques to analyze predictors of the peak factor VIII level. Response was considered adequate in 68 patients (40%), of whom 25 showed excellent response (15%). Intravenous administration, age, pre-desmopressin factor VIII activity and von Willebrand factor antigen, peak von Willebrand factor activity and desmopressin-induced rise in von Willebrand factor antigen were significant predictors of peak factor VIII level and explained 65% of the inter-individual variation. In 40% of moderate hemophilia A patients, desmopressin response was adequate, thus it is important not to withhold this group of patients from desmopressin responsiveness. Among the six predictors that we identified for desmopressin-induced factor VIII rise, factor VIII activity and desmopressin-induced rise in von Willebrand factor antigen had the strongest effect. Introduction Hemophilia A (HA) is a hereditary clotting disease caused by mutations in the F8 gene, leading to a deficiency of clotting factor VIII (FVIII) that occurs in one out of 5,000 men. Patients are classified based on residual levels of FVIII activity (FVIII:C). Severe patients have no detectable FVIII:C, non-severe patients have some activity (moderate FVIII:C 1-5 and mild 6-40 IU/dL). Severe and moderate HA patients are generally treated with FVIII concentrates, whereas most mild HA patients may be successfully treated with 1-Deamino-8-DArgininVasoPressin (desmopressin; DDAVP) for minor injuries or procedures. Using DDAVP, and thereby avoiding FVIII concentrates, has two important advantages: depending on the country, DDAVP is much cheaper than FVIII concentrate, and DDAVP does not carry the risk of inhibitor development associated with the use of exogenous (allogeneic) sources of FVIII present in concentrates.1-3 haematologica | 2018; 103(3)
DDAVP response in moderate hemophilia A
DDAVP is a synthetic vasopressin analogue and can be administered intravenously, subcutaneously or intranasally. The drug increases endogenous FVIII plasma concentrations by an average of three- to five-fold by inducing the release of von Willebrand factor (VWF), the carrier protein of FVIII, and the direct release of FVIII from Weibel-Palade bodies (WPBs) in endothelial cells.4,5 FVIII is primarily synthesized in liver sinusoidal endothelial cells.6 Extrahepatic FVIII is believed to be made by cells in the spleen, by lymphatic tissue, and especially by endothelial cells.7-11 The effect of DDAVP is dependent on the vasopressin type 2 receptor which is highly expressed in lung endothelial cells, but not in other populations of vascular endothelial cells.12 It is currently unknown from which sites FVIII and VWF are released upon DDAVP stimulation. Interestingly, liver transplantation in HA patients eliminates DDAVP response for FVIII but not for VWF, suggesting that extrahepatic FVIII synthesis may be necessary for DDAVP response.13 Large inter-individual variation in the response to DDAVP is observed. The variability of biological response within the same individual is smaller than between individuals.14 Although the relative increase in VWF and FVIII levels may be similar between non-severe HA patients, as moderates start at a much lower baseline FVIII:C, they may not reach a sufficient peak FVIII level to allow for treatment of minor procedures or trauma. Nevertheless, peak FVIII:C levels reaching 30 IU/dL may be clinically relevant for minor procedures or bleeding events. Several single-center studies described DDAVP in moderate HA.15-22 A total of 21% of the moderate patients who were tested showed an increase of FVIII:C to at least 30 IU/dL and identified the following predictors of response: age, route of administration, blood group, disease severity, and F8 mutations. However, these studies were hampered by small sample sizes and provided heterogeneous outcomes due to differences in patient characteristics and route of administration. Moreover, VWF was not studied as a potential determinant and the outcome variable which was principally studied was peak FVIII:C. In addition to the peak FVIII:C, the incremental response (proportional rise) may reveal important information on the biological mechanisms underlying DDAVP response, with possibly different predictors. We aim to describe the response to DDAVP in moderate HA patients and to identify predictors in a large, international cohort of moderate HA patients. Our results show that DDAVP provides a valuable treatment option in a large proportion of patients with moderate HA.
Methods Study population We selected data on all 169 patients with moderate HA from the multicenter Response to DDAVP In non-severe hemophilia A patients: in Search for dEterminants (RISE) cohort study, consisting of 1,474 non-severe HA patients from 24 hemophilia treatment centers (Figure 1). The aim of the RISE project was to assess the predictive value of clinical and genetic factors on the DDAVP response in non-severe HA patients. This international retrospective cohort study includes all consecutive non-severe HA patients with DDAVP administration between 1980 and 2012. Participating centers (listed in the Online Supplementary Appendix) were located in Canada, Australia and ten European countries. The institutional review boards of all centers approved the study. Since haematologica | 2018; 103(3)
this project involves retrospective data collection, all review boards indicated that informed consent was not required. This study was conducted in accordance with the Declaration of Helsinki.
Data collection We collected demographic and clinical data from available medical records using a standardized electronic case report form. The following data on baseline characteristics were collected: date of birth, ethnicity, ABO blood group, family history of DDAVP response, F8 mutation (Human Genome Variation Society [HGVS] numbering was used), reason for DDAVP administration, lifetime lowest FVIII:C (one-stage clotting assay), pre-DDAVP FVIII:C, preDDAVP VWF antigen (VWF:Ag) and activity (VWF:Act), date of DDAVP response, DDAVP dose, Body Mass Index (BMI), inhibitor status, route of administration, FVIII:C/VWF:Ag/VWF:Act after DDAVP, and potential side effects. In cases where the patient was treated to prevent or stop bleeding, we also collected information on the reason for treatment, location and severity of bleeding and therapeutic response.
Patient selection We selected patients from the RISE study with moderate disease severity. Patients were defined as moderate if one of the available FVIII:C measurements was 5 IU/dL or lower (lifetime lowest FVIII:C). In case of multiple DDAVP administrations, we selected the most recent DDAVP administration. It is important to mention that 13 patients from Seary et al. and 17 patients from the study conducted by Stoof et al. with FVIII:C ≤5 IU/dL are also included in our population.15,16 We explored selection bias of our study population by comparing the RISE population to 357 moderate patients from the Intervention as a Goal in Hypertension Treatment (INternational Study on etiology of inhibitors in patients with a moderate or mild form of hemophilia A, influences of Immuno Genetic & Hemophilia Treatment factors ([INSIGHT]) study population that did not receive DDAVP.23 We compared: FVIII:C, inhibitors, cumulative exposure days to FVIII, date of birth, and age.
Definition of response The main study outcome is the peak FVIII:C after DDAVP (in IU/dL). We classified peak response as none (<20), partial (20-29), complete (30-49) or excellent (≥50). With these classifications we were able to compare our findings to previously reported response rates. For further univariate analyses of the determinants of response, we compared patients with inadequate response to patients with at least a complete response. Incremental response was calculated by dividing peak FVIII:C by pre-DDAVP FVIII:C. Data was collected on the therapeutic response, which is defined in Online Supplementary Table S1.
Statistical analyses Summary statistics include frequencies and percentages for categorical variables and medians and interquartile ranges for continuous variables. An unpaired t-test and χ2 test were used to compare means between patients with inadequate response and at least a complete response. Furthermore, we used multiple linear regression to model relationships between potential explanatory variables and peak FVIII:C and incremental response in the patients tested for DDAVP responsiveness. Potential explanatory variables included in the model were: blood group, route of administration, dose, lifetime lowest FVIII:C, age, pre-DDAVP FVIII:C/VWF:Ag/VWF:Act, peak VWF:Ag/VWF:Act, and DVWF:Ag. DVWF:Ag was defined as peak VWF:Ag minus preDDAVP VWF:Ag. We restricted multivariate analyses to DDAVP test results as only nine patients were exclusively treated with 551
J.I. Loomans et al. Table 1. Baseline characteristics and 1-Deamino-8-D-ArgininVasoPressin (desmopressin; DDAVP) response in moderate patients.
All patients n=169
Calendar year of birth, median (IQR) Age at DDAVP administration, median years (IQR) Lifetime lowest FVIII:C, median IU/dL (IQR) Pre-DDAVP FVIII:C, median IU/dL (IQR) Pre-DDAVP VWF:Act, median IU/dL (IQR) Pre-DDAVP VWF:Ag, median IU/dL (IQR) Bloodgroup, n (%) O Non-O Unknown Weight, median kg (IQR) BMI, median (IQR) Total amount of DDAVP, median µg (IQR) DDAVP dose per kg, median µg (IQR) Route of administration Intraveneous Intranasal Subcutaneous Peak FVIII:C, median IU/dL (IQR) Incremental response, median (IQR) Time after DDAVP administration for peak levels, median minutes (IQR) Peak VWF:Act median IU/dL (IQR) Peak VWF:Ag median IU/dL (IQR) Test, n (%) Treatment, n (%) Mutation known, n (%)
1977 (1961-1991) 23 (12-41)
Inadequate response <20 IU/dL n=70 (41%)
Partial response 20-29 IU/dL n=31 (18%)
1975 (1960-1992) 1971 (1961-1990) 21 (8-43) 25 (17-35)
Complete response 30-49 IU/dL n=43 (25%)
Excellent response ≥50 IU/dL n=25 (15%)
1979 (1959-1991) 23 (9-49)
1974 (1964-1990) 28 (16-41)
4
(3-5)
4
(2-5)
4
(3-5)
4
(3-5)
4
(3-5)
5 87 93
(4-9) (63-110) (75-118)
4 95 103
(3-5) (82-113) (81-127)
5 84 93
(4-6) (62-112) (78-116)
7 73 96
(5-12) (44-132) (69-120)
11 73 76
(8-17) (60-98) (59-100)
33 (20) 33 (20) 103 (61) 75 (54-86) 24 (19-27) 20 (16-24) 0.3 (0.29-0.31)
9 14 47 72 22 20 0.3
(13) (20) (67) (48-83) (19-27) (14-24) (0.30-0.32)
7 9 15 72 26 20 0.3
(23) (29) (48) (60-88) (19-29) (18-24) (0.30-0.31)
104 9 56 24 4 60
(62) (5) (33) (14-39) (2.7-6.3) (60-75)
33 3 34 13 3 60
(47) (4) (49) (8-17) (2.1-4) (60-82)
19 3 9 24 4.8 60
203 209 159 9 106
(186-260) (188-257) (94) (5) (63)
200 212 65 4 40
(192-234) (188-267) (93) (6) (57)
210 206 28 3 17
14 (33) 5 (12) 24 (56) 76 (45-85) 25 (17-26) 20 (14-24) 0.3 (0.25-0.31)
3 (12) 5 (20) 17 (68) 77 (62-96) 25 (18-27) 20 (19-29) 0.3 (0.29-0.30)
(61) (10) (29) (21-27) (3.9-6.5) (52-78)
32 1 10 36 5.1 60
(74) (2) (23) (33-42) (3.4-8) (45-75)
20 2 3 63 5,5 60
(80) (8) (12) (57-87) (4.4-8.7) (30-60)
(137-260) (184-215) (90) (10) (55)
205 197 41 2 31
(118-326) (153-216) (95) (5) (72)
212 248 25 0 18
(178-284) (184-371) (100) (72)
*Interquartile Range (IQR). **Number. FVIII: factor VIII; VWF:Act: von Willebrand factor activity; VWF:Ag: von Willebrand factor antigen; BMI: body mass index.
DDAVP (Figure 1). We added pre-DDAVP FVIII:C times age as an interaction term to the model, as younger patients might have lower pre-DDAVP FVIII:C and VWF:Ag levels, and pre-DDAVP levels are known to affect the DDAVP response. A P-value <0.05 was considered statistically significant. Logistic transformation was used for variables that were not-normally distributed. Missing data were analyzed by Little’s missing completely at random (MCAR) test. If Little’s MCAR test P-value was not significant, missing data was imputed by multiple imputation.
Results Baseline characteristics Eleven percent of the source population (n=169) had moderate disease severity (Figure 1). Table 1 shows the baseline characteristics and response rates of the 169 moderates with DDAVP. Administration 552
was mostly intravenous (62%). Patients with subcutaneous versus intravenous administration were comparable, except for weight (mean 76 kg intravenous and 52 kg subcutaneous, P<0.001). In total, 99 patients (59%) were responsive to DDAVP (peak FVIII:C >20 IU/dL). The responses were excellent, complete and partial in 25 (15%), 43 (25%), and 31 (18%) patients, respectively. Table 2A displays the response rates per lowest lifetime FVIII and Table 2B shows the discrepencies between lowest lifetime FVIIII and preDDAVP FVIII. Further information on treatment outcomes is displayed in Online Supplemental Table S1. To address selection bias, we compared moderates that received and did not receive DDAVP. The only characteristic that differed between the groups was lifetime lowest FVIII:C. This was higher in patients who received DDAVP (median 4 (interquartile range [IQR] 3-5) vs. 3 (IQR 3-4) IU/dL, P=0.002). Missing data are displayed in Online Supplementary Table S2. haematologica | 2018; 103(3)
DDAVP response in moderate hemophilia A
Table 2A. Response rates per *Lowest lifetime factor VIII (FVIII:C).
FVIII:C (IU/dL)* 1 2 3 4 5 Total
Number of patients (%)
Inadequate Response
Partial Response
Complete Response
Excellent Response
% At least comlete
7 (4) 16 (9) 43 (25) 47 (28) 56 (33) 169
6 12 16 15 21 70
0 2 7 10 12 31
0 2 14 12 15 43
1 0 6 10 8 25
14 13 47 47 41
Table 2B. Pre-1-Deamino-8-D-ArgininVasoPressin (desmopressin; DDAVP) factor VIII (FVIII:C) minus lowest lifetime FVIII:C, per response group. Pre-DDAVP FVIII:C-lowest lifetime FVIII:C, median IU/dL (IQR)
All patients
Inadequate
Partial
Complete
Excellent
0 (0-4)
0 (0-0)
0 (0-1)
4 (0-8)
7 (3.5-13.5)
IQR: interquartile range.
Side effects Where data was available, side effects were present in 3/119 cases. Three patients reported skin flushing and one patient had an abnormally low blood pressure (defined as <2SD for age) with an increased heart rate (>100/min) following DDAVP administration.
incremental response decreases by 0.026 IU/dL, 1.469 IU/dL, and 0.187 year*IU/dL, respectively.
Discussion
Complete and excellent responders had significantly higher pre-DDAVP FVIII:C (P<0.001) and a higher proportion of intravenous administration. Pre-DDAVP VWF showed a trend towards lower levels in excellent responders (Ag: P=0.06, Act: P=0.07). No differences were observed between the response groups for other characteristics.
Herein, we present DDAVP response rates and predictors in a large, international cohort of moderate HA patients. In total, 68 patients (40%) achieved a peak FVIII:C of at least 30 IU/dL, among these 25 responses were excellent (FVIII:C â&#x2030;Ľ 50IU/dL). We identified six predictors of peak FVIII:C, which, taken together, explain 65% of the variation in peak FVIII:C. The pre-DDAVP FVIII:C and DDAVP-induced rise in VWF:Ag were the most important. The incremental response could be explained for 29% by different predictors, other than for peak FVIII:C.
Mutations
Response rates
Genotype was known in 107 patients (63%). We identified 58 different mutations of which nine were present in at least three patients (Table 3, Figure 2). The Arg2169His mutation was most prevalent (n=21). Responses are scattered among the different mutation groups.
Eight single-center studies previously reported DDAVP response rates in moderate HA patients and the characteristics of included patients (Table 5).15-22 The number of moderate patients in these studies varied from one to 17. Taken together, a total of 12 out of 56 patients from the eight studies showed a response of at least 30 IU/dL after DDAVP administration (21%). The difference between the response rates that we report herein (40%) might be due to differences in selection, population characteristics and routes of administration.
Univariate analyses of peak FVIII:C: no versus at least complete response
Multivariate analysis of peak FVIII:C and increment The following predictors explain 65% (Adjusted R2=0.65) of the variation in peak FVIII:C: intravenous administration, pre-DDAVP FVIII:C and VWF:Ag, DVWF:Ag, peak VWF:Act and age (Table 4A). Both pre-FVIII:C and DVWF:Ag have strong effects; the peak FVIII:C increases by 2.5 IU/dL for every unit increase in pre-DDAVP FVIII:C, and by 0.165 IU/dL for every unit increase in DVWF:Ag (DDAVP induced rise of VWF:Ag). Peak FVIII:C increased with intravenous compared to subcutaneous and intranasal administration (β=3.7). Remarkably, for every unit increase in baseline VWF:Ag, peak FVIII:C decreases by 0.117 IU/dL. The incremental response of FVIII:C can be explained for 29% (Adjusted R2=0.29) by DVWF:Ag, pre-DDAVP VWF:Ag, lowest lifetime FVIII:C, and the interaction term age*pre-DDAVP FVIII:C (Table 4B). For every unit increase in D VWF:Ag, the incremental response increases by 0.047 IU/dL, whereas for every unit increase of pre-DDAVP VWF:Ag, lifetime lowest FVIII:C, the interaction term, the haematologica | 2018; 103(3)
Predictors of peak FVIII:C We identified six predictors explaining 65% of the variation in peak FVIII:C. Intravenous administration predicts higher peak FVIII:C compared to subcutaneous or intranasal administration in our cohort, as well as in other studies.17,24-31 However, it is unknown, as of yet, whether this difference is clinically relevant, and data on moderate patients are scarce. Subcutaneous administration is believed to be biologically equivalent to the intravenous route, but this is based on only one paper.32 The rate of subcutaneous absorption could affect either the FVIII:C peak or its timing. If there is a clinically relevant difference, then this effect might be more critical in moderate patients due to their lower baseline levels. 553
J.I. Loomans et al.
Our observation that higher pre-DDAVP FVIII:C predicts higher peak FVIII:C in moderate HA patients is supported by other studies.15,18 DVWF:Ag was the second strongest predictor of peak FVIII:C. In consistency with this finding, we identified lower pre-DDAVP VWF:Ag as a predictor of peak FVIII:C. We can only speculate on the biological mechanism explaining this observation. Potentially, the stronger increase in VWF provides more binding sites for FVIII released upon DDAVP administration. Some patients may have higher pre-DDAVP VWF:Ag due to stress. Stress is known to increase endogenous FVIII:C and VWF. Hence, patients may have already released some of their stored VWF due to stress, with with less potential for VWF to rise further with DDAVP. Finally, it is conceivable that some patients have a phenotype of “greedy” endothelium, whereby VWF stores (or storage compartments) are only depleted upon extra (DDAVP) stimulation. There is still more to learn about the exact sources and secretion of both FVIII and VWF upon DDAVP stimulation, and how they interact. It is important to stress that DVWF:Ag is derived from a post hoc parameter (DDAVP-induced rise in VWF:Ag), and can therefore not be used by the clinician to predict response adequacy in advance. We do not have an explanation for the lower peak VWF:Act which predicts a higher FVIII:C peak. This finding is inconsistent with the DVWF:Ag finding.
Our finding that younger age predicted higher peak FVIII:C in our cohort of moderate HA patients is inconsistent with the literature reporting moderate patients. Stoof et al. and Knöfler et al. did not find an effect of age on DDAVP response, which they both attribute to the fact that the age of their population was higher than in the studies of Seary et al. and Revel-Vilk et al. (Table 5).15-18 Both Seary et al. and Revel-Vilk et al. found that responders had a higher mean age compared to non-responders, however, they did not adjust for pre-DDAVP FVIII:C. Furthermore, their study was performed exclusively in children whereas our study focuses predominantly on adults (Table 5).15,18
Determinants of incremental response We identified four predictors of the incremental response; DVWF:Ag, pre-DDAVP VWF:Ag and lifetime lowest FVIII:C levels increase the incremental response. Furthermore, a smaller product of the interaction term age*pre-DDAVP FVIII:C predicted a higher incremental response. The interaction term indicates that the effect of pre-DDAVP FVIII:C on the incremental change of FVIII following DDAVP varies for age, or that the effect of age on the incremental change of FVIII following DDAVP is altered by pre-DDAVP FVIII:C. This is the first study which used incremental response as an outcome variable reflecting the biological mechanisms underlying DDAVP response. Unlike for peak FVIII:C, for incremental response the pre-DDAVP is not a predictor and
Figure 1. Patient selection of 169 moderate hemophilia A (HA) patients with 1-Deamino-8-DArgininVasoPressin (desmopressin; DDAVP) administration. The 169 patients are from 23 different hemophilia treatment centers. *Reason for treatment was unknown for one patient.
Table 3. Mutations in at least three patients.
Pro149Arg Tyr450Asn Arg550Cys Arg612Cys Arg1960Gln Gly1979Val Arg2169His Trp2248Cys Gln2265Arg Total
Number of patients (%)
No Response
Partial Response
Complete Response
Excellent Response
6 (4) 3 (2) 3 (2) 3 (2) 3 (2) 3 (2) 21 (12) 3 (2) 3 (2) 48 (28)
1 3 0 0 0 0 2 1 1 8
2 0 1 0 1 0 3 0 1 8
3 0 1 1 1 3 11 2 0 22
0 0 1 2 1 0 5 0 1 10
*Human Genome Variation Society nomenclature was used. Mutations in bold were additionally identified as high-risk mutations for inhibitor development.
554
haematologica | 2018; 103(3)
DDAVP response in moderate hemophilia A
lifetime lowest FVIII:C is inversely associated. This is most likely due to the reciprocal effect of higher pre-DDAVP FVIII:C on the relative increase, since it is the denominator of our outcome variable. There may be an effect of stressinduced increase of FVIII, depicted in Table 2B by a difference between lowest lifetime and pre-DDAVP FVIII:C. However, we did not adjust for age and mutation, as the effect of age on FVIII:C is dependent on mutation.33 The variation in the incremental response was explained
for only 29% by the four identified predictors. This is less than the explained variation in peak response (65%). We believe that further analyses of different mutation groups might help to reveal the observed discrepancy in explained variance.
Mutation analysis Two studies presented the DDAVP response for different mutations in non-severe patients.15,16 All moderate patients
Figure 2. Mutations and associated response to 1-Deamino-8-DArgininVasoPressin (desmopressin; DDAVP). All mutations were present in three patients, except for the Pro149Arg and Arg2169His mutation (n=6 and n=21, respectively). The size of the diagrams reflects the number of patients.
Table 4A. Multiple linear regression model of predictors of peak factor VIII (FVIII:C) explaining 65% of peak FVIII:C.
Unstandardized β Coefficients Intravenous administration Pre-DDAVP FVIII:C (IU/dL) DVWF:Ag (IU/dL) Pre- DDAVP VWF:Ag (IU/dL) Peak VWF:Act (IU/dL) Age (log, in yrs)*
3.7 2.51 0.17 -0.12 -0.03 -2.74
95%Confidence Interval 1.52 2.33 0.15 -0.14 -0.04 -5.02
5.88 2.7 0.18 -0.09 -0.01 -0.46
*All predictors had a P-value <0.001, except for that of age (P=0.019). Peak FVIII:C increases 2.5 IU/dL for every unit increase in pre-1-Deamino-8-D-ArgininVasoPressin (desmopressin; DDAVP) FVIII:C and 0.165 IU/dL for every unit increase in D von Willebrand factor antigen (VWF:Ag). In comparison to subcutaneous or intranasal administration, the response after intravenous administration appears to be 3.7-fold higher, although this is not statistically significant due to low patient numbers. For every unit increase in baseline VWF:Ag, peak von Willebrand factor activity (VWF:Act) and age, peak FVIII:C decreases by 0.12 IU/dL, 0.03 IU/dL and 2.74 log years, respectively.
haematologica | 2018; 103(3)
Table 4B. Multiple linear regression model of incremental response (peak factor VIII [FVIII:C]/pre-1-Deamino-8-D-ArgininVasoPressin [desmopressin; DDAVP] FVIII:C) explaining 29% of incremental response.
DVWF:Ag (IU/dL) Pre-DDAVP VWF:Ag (IU/dL) Lifetime lowest FVIII:C (IU/dL) Interaction term (age*pre-DDAVP FVIII:C)
Unstandardized β Coefficients 0.05 -0.03 -1.47 -0.19
95% Confidence Interval 0.04 -0.04 -1.87 -0.25
0.05 -0.15 -1.07 -0.13
*All predictors had a P-value <0.001. For every unit increase in D von Willebrand factor antigen (VWF:Ag), the incremental response increases by 0.047 IU/dL, whereas for every unit increase of pre-DDAVP VWF:Ag, lifetime lowest FVIII:C, and the interaction term, incremental response decreases by 0.026 IU/dL, 1.469 IU/dL, and 0.187 year*IU/dL, respectively.
555
J.I. Loomans et al.
Table 5. Previous studies on 1-Deamino-8-D-ArgininVasoPressin (desmopressin; DDAVP) response in moderate hemophilia A (HA) patients.
Author, year
N
Study design
Intervention
Stoof et al. 2014
17
Retrospective
Test and treatment
Knöfler et al. 2012
2
Seary et al. 2012
13
Revel-Vilk et al. 11 2002 Ghirardini et al. 1988 De La Fuente et al. 1985 Mariana et al. 1984 Warrier et al. 1983 Total Ɨ
2
Age Investigated DDAVP dose Time FVIII:C Pre-DDAVP After Partial (median determinants (μg/kg) and measurement FVIII:C DDAVP response years, IQR) of DDAVP administration (min) (IU/dL) FVIII:C responseƗ route (IU/dL)
Blood group O, IV and IN mutations, age, route of administration RetroTest 32.5 Age, 0.4 IV and spective (16.7–48.2) administration 0.4 SC route RetroTest < 18 Age, 0.3 IV spective and treatment disease severity, and mutations 0.3 SC RetroTest 2.8 Age, disease 0.3 IV spective (0.13–9.75)** severity, blood group O Prospective Treatment uk uk 0.3 SC
8 Prospective
Test and treatment 2 Prospective Treatment 1 Prospective
Test and treatment
28 (8–67)*
30 (2–66) uk
uk uk
0.3 IV
0.3 IV and 0.4 IV 23 Administration 2.0 IN (10–45)*** route
56
0; 60; 180; 600
0; 30; 60; 120; 240
1-5*
uk
2.6 14.75 (1.7–3.5) (14– 15.5)
At least complete response
uk
8 (47%)
0
0
0; 60; 240
1-5 *
uk
0
1 (8%)
0; 60
1-5 *
uk
0
0
11.5 (9–14) 24 (7–47) 15.7 (11.3–20) 13
0
0
2 (25%) 1 (50%) 0
3 (38%) 0
3 (5%)
12 (21%)
0; 60
3.5 (3–4) 0; 15; 30; 60; 3.5 180; 300; 1440 (2–5) 0; 60 4.8 (4.5–5) 0; 15; 90; 180; 270 2
0
Underlined factors were identified determinants. *Age is not specific for moderate patients, but for complete study population including mild HA patients. **Age is not specific for moderate HA patients, but for non-responders. ***Age is not specific for moderate HA patients, but for all patients with IN administration of DDAVP. IV: intravenous; IN: intranasal; SC: subcutaneous; uk: unknown; IQR: interquartile range; FVIII:C: factor VIII.
from these studies are also included in our cohort. Unexpectedly, we observed a discordance in DDAVP response among patients with the same mutation that was not due to differences in route of administration (Figure 2). Four mutations present in at least three patients were also identified in the INSIGHT cohort as high-risk mutations for inhibitor development (Arg550Cys, Arg612Cys, Arg2150His and Trp2248Cys HGVS numbering).23 As Figure 2 and Table 3 show, half of the patients with these mutations in our cohort showed at least a complete response to DDAVP. This is of clinical importance as these patients may be successfully treated with DDAVP in order to reduce hazardous exposure to FVIII concentrates.
Limitations and strengths This large, international cohort study provides data on the response to DDAVP in moderate HA patients. Our study is unique owing to the large number of patients included, thus increasing the statistical power. As lifetime lowest and pre-DDAVP FVIII:C was not measured in a standardized manner, this may have led to an overestimation of patients who were once classified as moderate, but who were not defined as moderate at the time of the DDAVP test. This can be seen in Table 1, where patients with complete and excellent responses have a higher pre-DDAVP FVIII:C irrespective of their lowest lifetime FVIII:C. 556
With respect to our outcome variables, it is important to state that the biological response is only a proxy of the clinical response to DDAVP. Additional data are warranted to establish when and to which extent desmopressin can be clinically used to treat patients with moderate HA. As only 32% of the moderate patients from our source population were tested or treated with DDAVP, this might lead to an overestimation of patients with a good response. The proportion of moderate patients receiving DDAVP per hemophilia treatment center ranged from 19 to 50. Moderate patients who received DDAVP had higher lowest lifetime FVIII:C compared to those who did not. For these reasons, caution is needed when extrapolating these results to all moderate HA patients. Although our study is unique in terms of its size, we still lacked the power to further analyze the effect of mutations on DDAVP response in moderate HA patients. As the F8 genotype is known to influence baseline FVIII:C, and thereby DDAVP response, the effect of mutations via baseline FVIII:C may contribute to clarifying 35% of the unexplained variation in peak FVIII:C levels.33 For further functional analyses of genotype, it would have been informative to have, in addition, the FVIII antigen (FVIII:Ag) levels. Castaman et al. demonstrated that the presence of a dysfunctional FVIII molecule (Antigen>Activity) per se does not prevent a response to DDAVP.34 haematologica | 2018; 103(3)
DDAVP response in moderate hemophilia A
Conclusion
Clinical significance and future studies As illustrated in Table 5, the use of DDAVP for moderate HA patients in clinical practice has been limited to a few cases over the last 40 years, despite the fact that this drug has been available since 1977. The study herein evinces that 40% of the patients with moderate HA in our cohort have a clinical relevant response to DDAVP for mild bleeds/injuries. Moreover, half of the patients with high inhibitor risk mutations respond to DDAVP; for this reason, it is important to always assess such patients for DDAVP responsiveness. Doing so might lead to less exposure to FVIII concentrates, which reduces risk for inhibitor development, and realizes a reduction in costs. In order to confirm our findings, future studies should focus on the prospective inclusion of moderate HA patients. Lastly, more data on mutations are needed in order to assess the effect of F8 missense mutations on DDAVP response in these patients.
References 1. Fijnvandraat K, Cnossen MH, Leebeek FWG, Peters M. Diagnosis and management of haemophilia. BMJ. 2012;344(2707): 1-5. 2. Schramm W, Royal S, Kroner B, et al. Clinical outcomes and resource utilization associated with haemophilia care in Europe. Haemophilia. 2002;8(1):33-43. 3. Mannucci P. Desmopressin (DDAVP) in the treatment of bleeding disorders: the first 20 years. Blood. 1997;90(7):2515-2521. 4. Mannucci P, Pareti F, Ruggeri Z, Capitanio A. DDAVP: a new pharmacological approach to the management of haemophilia and von Willebrand’s disease. Lancet. 1977;1(8017):869-872. 5. Rosenberg JB, Greengard JS, Montgomery RR. Desmopressin (DDAVP) in the treatment of bleeding disorders: the first 20 years. Arterioscler Thromb Vasc Biol. 2000; 20(12):2689-2695. 6. Shanani T, Covens K, Lavend’homme R, et al. Human liver sinusoidal endothelial cells but not hepatocytes contain factor VIII. J Thromb Haemost. 2013;12(1):36–42. 7. Groth CG, Hathaway WE, Gustafsson A, et al. Correction of coagulation in the hemophilic dog by transplantation of lymphatic tissue. Surgery. 1974;75(5):725-733. 8. Norman JC, Covelli VH, Sise HS. Transplantation of the spleen: experimental cure of hemophilia. Surgery. 1968;64(1):1-14. 9. Jacquemin M, Neyrinck A, Hermanns MI, et al. FVIII production by human lung microvascular endothelial cells. Blood. 2006;108(2):515-517. 10. Everett LA, Cleuren AC, Khoriaty RN, Ginsburg D. Murine coagulation factor VIII is synthesized in endothelial cells. Blood. 2014;123(24):3697-3705. 11. Fahs SA, Hille MT, Shi Q, Weiler H, Montgomery RR. A conditional knockout mouse model reveals endothelial cells as the principal and possibly exclusive source of plasma factor VIII. Blood. 2014;123(24):3706-3713. 12. Kaufmann JE, Oksche A, Wollheim CB, Günther G, Rosenthal W, Vischer UM. Vasopressin-induced von Willebrand factor secretion from endothelial cells involves V2 receptors and cAMP. J Clin Invest. 2000;106(1):107-116. 13. Lamont PA, Ragni MV. Lack of desmopressin (DDAVP) response in men with hemophilia
haematologica | 2018; 103(3)
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
In 40% of the administrations in moderate HA patients in our cohort, the use of DDAVP is adequate for treatment in case of minor bleeding or trauma; as such, it is important to consistently assess moderate patients for DDAVP responsiveness. Furthermore, we identified six predictors of peak FVIII:C, which, taken together, explain 65% of the variation in peak FVIII:C. Pre-DDAVP FVIII:C levels and the patients’ ability to raise VWF levels are the essential predictors for response. Acknowledgments Romy van Spengen, medical student in the Academic Medical Center Amsterdam, helped to construct Table 5. Funding This work was supported by the Access to Insight Grant from Novo Nordisk.
A following liver transplantation. J Thromb Haemost. 2005;3(10):2259-2263. Rodeghiero F, Castaman G, Di Bona E, Ruggeri M. Consistency of responses to repeated DDAVP infusions in patients with von Willebrand’s disease and hemophilia A. Blood. 1989;74(6):1997-2000. Seary ME, Feldman D, Carcao MD. DDAVP responsiveness in children with mild or moderate haemophilia A correlates with age, endogenous FVIII:C level and with haemophilic genotype. Haemophilia. 2012;18(1):50-55. Stoof SCM, Sanders YV, Cnossen MH, de Maat MPM, Leebeek FWG, Kruip MJHA. Desmopressin response in hemophilia A patients with FVIII: C < 0.10 IU mL-1. J Thromb Haemost. 2014;12(1):110-112. Knöfler R, Koscielny J, Tauer JT, et al. Desmopressin testing in haemophilia A patients and carriers: Results of a multi centre survey. Hamostaseologie. 2012;32(4):271275. Revel-Vilk S, Blanchette VS, Sparling C, Stain AM, Carcao MD. DDAVP challenge tests in boys with mild/moderate haemophilia A. Br J Haematol. 2002;117(4):947-951. Ghirardini A, Chistolini A, Tirindelli MC, Di Paolantonio T, Iacopino G, Marianio P. Clinical evaluation of subcutaneously administered DDAVP. Thromb Res. 1988;49(3):363-372. de la Fuente B, Kasper C, Rickles F, Hoyer L. Response of patients with mild and moderate hemophilia A and von Willebrand’s disease to treatment with desmopressin. Ann Intern Med. 1985; 103(1):6-14. Mariana G, Ciavarella N, Mazzucconi M, et al. Evaluation of the effectiveness of DDAVP in surgery and in bleeding episodes in haemophilia and von Willebrand’s disease. A study on 43 patients. Clin Lab Haematol. 1984;6(3):229-238. Warrier AI, Lusher JM. DDAVP. A useful alternative to blood components in moderate hemophilia A and von Willebrand disease. J Pediatr. 1983;102(2):228-233. Eckhardt CL, Van Velzen AS, Peters M, et al. for the INSIGHT study group. Factor VIII gene (F8) mutation and risk of inhibitor development in nonsevere hemophilia A. Blood. 2013;122(11):1954–1962. Jacquemin M, Lavend’homme R, Benhida A, et al. A novel cause of mild/moderate hemophilia A: mutations scattered in the factor VIII C1 domain reduce factor VIII binding to
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
von Willebrand factor. Blood. 2000;96(3): 958–965. Lethagen S, Harris S, Nilsson M. Intranasal desmopressin (DDAVP) by spray in mild hemophilia A and von Willebrand's disease type I. Blut. 1990;60(3):187-191. Christolini A, Dragoni F, Ferrari A, et al. Intranasal DDAVP: biological and clinical evaluation in mild factor VIII deficiency. Pathophysiol Haemost Thromb. 1991; 21(5):273-277. Rose EH, Aledort LM. Nasal spray desmopressin (DDAVP) for mild hemophilia A and von Willebrand disease. Ann Intern Med. 1991;114(7):563-568. Lethagen S, Egervall K, Berntorp E, Bengtsson B. The administration of desmopressin by nasal spray: a dose-determination study in patients with mild haemophilia A or von Willebrand’s disease. Haemophilia. 1995;1(2):97-102. Leissinger C, Becton D, Cornell C, Gill JC. High-dose DDAVP intranasal spray (Stimate) for the prevention and treatment of bleeding in patients with mild haemophilia A, mild or moderate type 1 von Willebrand disease and symptomatic carriers of haemophilia A . Haemophilia. 2001;7(3):258-266. Gill JC, Ottum M, Schwartz B. Evaluation of high concentration intranasal and intravenous desmopressin in pediatric patients with mild hemophilia A or mild-to-moderate type 1 von Willebrand disease. J Pediatr. 2002;140(5):595-599. LethagenS, Ragnarson Tennvall G. Self-treatment with desmopressin intranasal spray in patients with bleeding disorders: Effect on bleeding symptoms and socioeconomic factors. Ann Hematol. 1993;66(5):257-260. Mannucci P, Vicente V, Alberca I, et al. Intravenous and subcutaneous administration of desmopressin (DDAVP) to hemophiliacs: pharmacokinetics and factor VIII responses. Thromb Haemost. 1987;58(4): 1037-1039. Loomans JI, van Velzen AS, Eckhardt CL, et al. Variation in baseline factor VIII concentration in a retrospective cohort of mild/moderate hemophilia A patients carrying identical F8 mutations. J Thromb Haemost. 2017;15(2):246-254. Castaman G, Mancuso ME, Giacomelli SH, et al. Molecular and phenotypic determinants of the response to desmopressin in adult patients with mild hemophilia A. J Thromb Haemost. 2009;7(11):1824-1831.
557