Haematologica, Volume 103, Issue 9 by Haematologica

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 EHA-SAH Hematology Tutorial on lymphoid Malignancies and Plasma Cell Dyscrasias September 14-15, 2018 Buenos Aires, Argentina Highlights of Past EHA - Cairo 2018 Chairs: P Sonneveld, J Gribben, M Qari, M Mattar, A ElBeshlawy, A Kamel, September 27 - 29, 2018 Cairo, Egypt 14th Educational Course of the Lymphoma Working Party European Society for Blood and Marrow Transplantations (EBMT) Lymphoma Working Party Chairs: S Montoto, A Sureda, L Bento September 27-28, 2018 Palma, Spain EHA-SWG Scientific Meeting on Aging and Hematology Chair: D Bron October 12-14, 2018 Warsaw, Poland

EHA-Baltic Hematology Tutorial on Lymphoid malignancies, including WaldenstrĂśm's Macroglobulinemia EHA in close collaboration with the Estonian Society of Haematology, the Lithuanian Society of Hematology and the Latvian Hematology Society Chairs: J Gribben, E Laane, S Lejniece, V Peceliunas October 18-19, 2018 Tallinn, Estonia 13th Congress of the Albanian Association of Hematology Albanian Association of Hematology Chairs: A Ivanaj, M Dimopoulos, G Gaidano, X Pivot November 5-6, 2018 Tirana, Albania

Calendar of Events updated on July 25, 2018â&#x20AC;?

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

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

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

Scientific Latin

haematologicus (adjective) = related to blood

Scientific Latin

haematologica (adjective, plural and neuter, used as a noun) = hematological subjects

Modern English

The oldest hematology journal, publishing the newest research results. 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

Table of Contents Volume 103, Issue 9: September 2018 Cover Figure

Bone marrow smear showing hyperplasia of the granuloblastic lineage with predominance of promyelocytes with heavy granulation in a patient with drug-induced neutropenia. Courtesy of Prof. Rosangela Invernizzi.

Editorials 1415

Targeting dihydroorotate dehydrogenase in acute myeloid leukemia Zhihong Zeng and Marina Konopleva

1417

The complexity of stem cell transplants: can we improve our understanding? Andrea Bacigalupo and Francesca Bonifazi

1419

Arterial thrombosis and cancer: the neglected side of the coin of Trousseau syndrome Valerio De Stefano

Review Article 1422

Cardiovascular adverse events in modern myeloma therapy – Incidence and risks. A review from the European Myeloma Network (EMN) and Italian Society of Arterial Hypertension (SIIA) Sara Bringhen et al.

1433

Hereditary hemorrhagic telangiectasia: diagnosis and management from the hematologist’s perspective Athena Kritharis et al.

Articles Hematopoiesis

1444

Transient inhibition of NF-κB signaling enhances ex vivo propagation of human hematopoietic stem cells Mehrnaz Safaee Talkhoncheh et al.

Bone Marrow Failure

1451

Hematopoietic stem cell loss and hematopoietic failure in severe aplastic anemia is driven by macrophages and aberrant podoplanin expression Amanda McCabe et al.

Myelodysplastic Syndromes

1462

Transforming growth factor β1-mediated functional inhibition of mesenchymal stromal cells in myelodysplastic syndromes and acute myeloid leukemia Stefanie Geyh et al.

Acute Myeloid Leukemia

1472

Pharmacological inhibition of dihydroorotate dehydrogenase induces apoptosis and differentiation in acute myeloid leukemia cells Dang Wu et al.

1484

Clofarabine, high-dose cytarabine and liposomal daunorubicin in pediatric relapsed/refractory acute myeloid leukemia: a phase IB study Natasha K.A. van Eijkelenburg et al.

Acute Lymphoblastic Leukemia

1493

Investigating chemoresistance to improve sensitivity of childhood T-cell acute lymphoblastic leukemia to parthenolide Benjamin C. Ede et al.

Haematologica 2018; vol. 103 no. 9 - September 2018 http://www.haematologica.org/

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

1502

Sustained efficacy and detailed clinical follow-up of first-line ibrutinib treatment in older patients with chronic lymphocytic leukemia: extended phase 3 results from RESONATE-2 Paul M. Barr et al.

1511

Real-world outcomes and management strategies for venetoclax-treated chronic lymphocytic leukemia patients in the United States Anthony R. Mato et al.

Plasma Cell Disorders

1518

A phase I/II dose-escalation study investigating all-oral ixazomib-melphalan-prednisone induction followed by single-agent ixazomib maintenance in transplant-ineligible newly diagnosed multiple myeloma Jesús F. San-Miguel et al.

Stem Cell Transplantation

1527

Competing-risk outcomes after hematopoietic stem cell transplantation from the perspective of time-dependent effects Daniel Fuerst et al.

Quality of Life

1535

Patient-reported outcomes and health status associated with chronic graft-versus-host disease Stephanie J. Lee et al.

Blood Transfusion

1542

Transfusion of packed red blood cells at the end of shelf life is associated with increased risk of mortality – a pooled patient data analysis of 16 observational trials Monica S.Y. Ng et al.

Coagulation & ita Disorders

1549

Frequency, risk factors, and impact on mortality of arterial thromboembolism in patients with cancer Ella Grilz et al.

Platelet Biology & its Disorders

1557

Impaired mitochondrial activity explains platelet dysfunction in thrombocytopenic cancer patients undergoing chemotherapy Constance C. F. M. J. Baaten et al.

1568

NLRP3 regulates platelet integrin αIIbβ3 outside-in signaling, hemostasis and arterial thrombosis Jianlin Qiao et al.

Errata Corrige 1577

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

Letters to the Editor Letters are available online only at www.haematologica.org/content/103/9.toc

e384

Fetal hemoglobin induction in sickle erythroid progenitors using a synthetic zinc finger DNA-binding domain Biaoru Li et al. http://www.haematologica.org/content/103/9/e384

Haematologica 2018; vol. 103 no. 9 - September 2018 http://www.haematologica.org/

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation e388

Microenvironmental M1 tumor-associated macrophage polarization influences cancer-related anemia in advanced ovarian cancer: key role of interleukin-6 Clelia Madeddu et al. http://www.haematologica.org/content/103/9/e388

e392

Value of cytogenetic abnormalities in post-polycythemia vera and post-essential thrombocythemia myelofibrosis: a study of the MYSEC project Barbara Mora, et al. http://www.haematologica.org/content/103/9/e392

e395

GFI1 is required for RUNX1/ETO positive acute myeloid leukemia Anna E. Marneth et al. http://www.haematologica.org/content/103/9/e395

e400

Clonal genetic evolution at relapse of favorable-risk acute myeloid leukemia with NPM1 mutation is associated with phenotypic changes and worse outcomes Carmen Martínez-Losada et al. http://www.haematologica.org/content/103/9/e400

e404

Efficacy of the combination of venetoclax and hypomethylating agents in relapsed/refractory acute myeloid leukemia Ibrahim Aldoss et al. http://www.haematologica.org/content/103/9/e404

e408

MYD88 mutated and wild-type Waldenström’s Macroglobulinemia: characterization of chromosome 6q gene losses and their mutual exclusivity with mutations in CXCR4 Maria Luisa Guerrera et al. http://www.haematologica.org/content/103/9/e408

e412

Dramatically improved survival in multiple myeloma patients in the recent decade: results from a Swedish population-based study Sigrun Thorsteinsdottir et al. http://www.haematologica.org/content/103/9/e412

e416

A phase I study of romidepsin and ifosfamide, carboplatin, etoposide for the treatment of patients with relapsed or refractory peripheral T-cell lymphoma Paolo Strati et al. http://www.haematologica.org/content/103/9/e416

e419

Short-hairpin RNA against aberrant HBB IVSI-110(G>A) mRNA restores β-globin levels in a novel cell model and acts as mono- and combination therapy for β-thalassemia in primary hematopoietic stem cells Petros Patsali et al. http://www.haematologica.org/content/103/9/e419

Case Reports Case Reports are available online only at www.haematologica.org/content/103/9.toc

e424

NKG2D-based chimeric antigen receptor therapy induced remission in a relapsed/refractory acute myeloid leukemia patient David A. Sallman et al. http://www.haematologica.org/content/103/9/e424

e427

Clinical efficacy of ruxolitinib and chemotherapy in a child with Philadelphia chromosome-like acute lymphoblastic leukemia with GOLGA5-JAK2 fusion and induction failure Yang Y. Ding et al. http://www.haematologica.org/content/103/9/e427

e432

t(11;14)-positive mantle cell lymphomas lacking cyclin D1 (CCND1) immunostaining because of a CCND1 mutation or exclusive expression of the CCND1b isoform Ingram Iaccarino et al. http://www.haematologica.org/content/103/9/e432

Haematologica 2018; vol. 103 no. 9 - September 2018 http://www.haematologica.org/

EDITORIALS Targeting dihydroorotate dehydrogenase in acute myeloid leukemia Zhihong Zeng1 and Marina Konopleva1,2 1

Department of Leukemia and 2Department of Stem Cell Transplantation and Cellular Therapy, The University of Texas MD Anderson Cancer Center, Houston, TX, USA E-mail: mkonople@mdanderson.org doi:10.3324/haematol.2018.197806

ith their discovery of a promising inhibitor of dihydroorotate dehydrogenase (DHODH), Wu and Wong et al. have made progress toward effective differentiation therapy for acute myeloid leukemia (AML).1 Their paper entitled “Pharmacological inhibition of dihydroorotate dehydrogenase induces apoptosis and differentiation of acute myeloid leukemia cells,” published in this issue of Haematologica, reports that this inhibitor had significant anti-leukemia efficacy in vitro and in vivo.1 DHODH is an enzyme that catalyzes the oxidation of dihydroorotate into orotate in intracellular de novo pyrimidine synthesis. Pyrimidine bases are essential for cellular metabolism and cell growth, and are considered important precursors in nucleotide, glycoprotein, and phospholipid biosynthesis and nucleotide recycling.2 Dysregulation of and functional dependency on pyrimidine biosynthesis have been found in diverse solid tumors.3 As a key enzyme regulating this process, DHODH has been identified as a synthetically lethal target in tumors that carry specific genetic mutations, including PTEN-deficient and triple-negative breast cancer,4,5 BRAFV600E–mutant melanoma,6 RAS/LKB1 double-mutant lung cancer,7 and KRAS-mutated pancreatic cancer.8 AML is a highly heterogeneous hematologic malignancy, characterized by uncontrolled growth of immature hematopoietic stem/progenitor cells and impaired differentiation.9,10 The long-term survival rate of AML patients who receive standard therapies remains poor.11 Differentiation therapy was proposed decades ago with the goal of promoting the normal process of hematopoietic maturation from self-renewing progenitors to terminally differentiated effector cells,12 but its success has been limited except in specific genetic subtypes of AML. For example, all-trans retinoic acid induces differentiation and remissions in patients with acute promyelocytic leukemia.13 Inhibitors of isocitrate dehydrogenase (IDH) have been approved more recently for treatment of AML carrying IDH1/2 mutations, and the activity of these inhibitors is associated with clinical and morphological signs of myeloid differentiation.14 Recent pre-clinical findings by Sykes et al. showed that DHODH is a metabolic regulator in the pyrimidine synthesis pathway and a new metabolic target in differentiation therapy for AML.15 Inhibition of DHODH effectively enabled cell differentiation and had anti-leukemia efficacy in vitro and in vivo.15 This discovery opens up a new perspective in differentiation therapy for AML carrying complex and heterogeneous combinations of gene mutations. It leads to pharmacological intervention via development of novel, potent, and optimized DHODH inhibitors for clinical applications and to studies aiming to determine the mechanisms underlying DHODH inhibition–mediated AML differentiation. In this issue of Haematologica, Wu and Wang et al. confirm the essential role of DHODH in AML survival and differenhaematologica | 2018; 103(9)

tiation.1 Using CRISPR-Cas9–mediated gene knockout, the authors observed induction of apoptosis and cellular differentiation in DHODH-knockout AML cells. Through analysis of RNA sequencing data for 173 primary AML samples, they found that DHODH and MYC, a key regulator of myeloid cell proliferation and differentiation, are coordinately expressed in AML. Using structure-based virtual screening of 337 natural products, the research team identified isobavachalcone, a chalcone derived from the traditional Chinese medicine Psoralea corylifolia, as a potent and selective inhibitor of DHODH. Isobavachalcone, a competitive inhibitor of CoQ0 and a noncompetitive inhibitor of the DHODH substrate dihydroorotate, binds the docking site of the “ubiquinone channel” of the DHODH complex, interacting structurally with and stabilizing DHODH. Treatment with isobavachalcone led to dose-dependent induction of apoptosis and myeloid differentiation in the cell lines tested. The on-target effect mediated by isobavachalcone’s inhibition of DHODH is supported by the findings that uridine, a downstream metabolite of DHODH in the pyrimidine synthesis pathway, was depleted in isobavachalcone-treated cells and that growth inhibition was rescued by uridine supplementation.1 Mechanistically, the apoptosis and cell differentiation induced by isobavachalcone’s inhibition of DHODH was found to occur, at least in part, through cMYC suppression, by directly reducing its gene transcription and degrading c-MYC protein via proteasome dependent degradation and downregulation of the enzyme O-linked Nacetylglucosamine transferase that normally stabilizes MYC via transfer of GlcNAc from uridine diphosphate-GlcNAc to MYC.16,17 Notably, O-GlcNAc post-translational modification similarly stabilizes additional proteins such as AKT and the TET family of proteins.18 In a subcutaneous AML cell line xenograft model, isobavachalcone treatment suppressed tumor growth, and this effect was accompanied by inhibition of DHODH activity and induction of apoptosis of tumor cells.1 Oral administration of isobavachalcone was well tolerated in mice. A combination of isobavachalcone and adriamycin (doxorubicin), an agent widely used in AML therapy, synergistically induced cell death and restored sensitivity of doxorubicin-resistant AML cells in vitro; this combination also led to cell differentiation and prolonged mouse survival in vivo in a systemic AML cell line xenograft leukemia model, analogous to the report of the efficacy of doxorubicin combined with the approved DHODH inhibitor leflunomide in triple-negative breast cancer.5 These data support the idea that isobavachalcone, an orally active DHODH inhibitor, might have clinical benefit in AML differentiation therapy. The work by Wu and Wang et al. advances our knowledge of DHODH regulation in AML and offers a novel inhibitor with promise for differentiation therapy. The detailed mechanisms underlying this activity, such as how the functional 1415

Editorials

Figure 1. Isobavachalcone in the regulation of dihydroorotate dehydrogenase in pyrimidine biosynthesis and myeloid blast differentiation. Left, the pyrimidine biosynthesis pathway and its biological functions. Isobavachalcone (IBC) prevents DHODH from catalyzing dihydroorotate into orotate, blocking pyrimidine biosynthesis, resulting in myeloid lineage differentiation and induction of apoptosis. The role of DHODH and the negative impact of c-MYC on myeloid blast differentiation is indicated by the dotted line/arrow. DHODH: dihydroorotate dehydrogenase; UMP: uridine 5â&#x20AC;˛-monophosphate; UTP: uridine triphosphate; CTP: cytidine triphosphate; dTTP: deoxythymidine triphosphate; UDP-GlcNAc: uridine diphosphate N-acetylglucosamine; CDP-DAG: cytidine diphosphate diacylglycerol; CDP-choline: cytidine diphosphate choline; O-GlcNAc: O-linked N-acetylglucosamine; OGT: O-GlcNAc transferase. CTP to dTTP requires multiple steps.

interplay between DHODH and c-MYC in AML triggers differentiation and whether this is dependent on cellular context, is worthy of future exploration. Importantly, the capacity of isobavachalcone to facilitate blast differentiation in patient-derived xenografts and AML patients with different cytogenetic abnormalities needs to be addressed. In studies similar to that of Wu and Wong et al., two groups recently reported new findings related to anti-leukemia mechanisms of DHODH inhibition in different types of hematologic malignancies. DHODH blockade by the chiral tetrahydroindazole ([R]-HZ00), a newly developed DHODH inhibitor, increased p53 activation and enhanced the anti-leukemic effect of the MDM2 inhibitor, Nutlin3a, in p53-WT chronic myeloid leukemia.19 PTC299, identified as a VEGFA inhibitor, targeted DHODH, resulting in cell growth inhibition and differentiation in leukemias, including AML, linking DHODH regulation and stress-induced VEGFA and angiogenesis.20 Notably, these studies demonstrated that DHODH inhibition alone is not sufficient to eliminate leukemia, requiring combination with standard chemotherapeutic agents or target-specific inhibitors. Given the genetic complexity of AML, future use of DHODH inhibitors in combination with selected drugs 1416

with non-overlapping mechanisms of action may tackle specific aspects of the complex pathogenesis of AML and ultimately improve patientsâ&#x20AC;&#x2122; outcomes.

References 1. Wu D, Wang W, Chen W, et al. Pharmacological inhibition of dihydroorotate dehydrogenase induces apoptosis and differentiation of acute myeloid leukemia cells. Haematologica. 2018;103(9):1472-1483 2. Levine RL, Hoogenraad NJ, Kretchmer N. A review: biological and clinical aspects of pyrimidine metabolism. Pediatr Res. 1974;8 (7):724-734. 3. DeBerardinis RJ, Lum JJ, Hatzivassiliou G, Thompson CB. The biology of cancer: metabolic reprogramming fuels cell growth and proliferation. Cell Metab. 2008;7:11-20. 4. Mathur D, Stratikopoulos E, Ozturk S, et al. PTEN regulates glutamine flux to pyrimidine synthesis and sensitivity to dihydroorotate dehydrogenase inhibition. Cancer Discov. 2017;7(4):380-390. 5. Brown KK, Spinelli JB, Asara JM, Toker A. Adaptive reprogramming of de novo pyrimidine synthesis is a metabolic vulnerability in triple-negative breast cancer. Cancer Discov. 2017;7(4):391-399. 6. Thomas NE, Berwick M, Cordeiro-Stone M. Could BRAF mutations in melanocytic lesions arise from DNA damage induced by ultraviolet radiation? J Invest Dermatol. 2006;126(8):1693-1696. 7. Kim J, McMillan E, Kim HS, et al. XPO1-dependent nuclear export is a druggable vulnerability in KRAS-mutant lung cancer. Nature 2016;538(7623):114-117. 8. Koundinya M, Sudhalter J, Courjaud A, et al. Dependence on the pyrimidine biosynthetic enzyme DHODH is a synthetic lethal vulnerability in mutant KRAS-Driven cancers. Cell Chem Bio. 2018;25(6):705717.

haematologica | 2018; 103(9)

Editorials 9. 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. 10. Saultz JN, Garzon R. Acute myeloid leukemia: a concise review. J Clin Med 2016;5(3). 11. Dombret H, Gardin C. An update of current treatments for adult acute myeloid leukemia. Blood 2016;127(1):53-61. 12. Nowak D, Stewart D, Koeffler HP. Differentiation therapy of leukemia: 3 decades of development. Blood 2009;113(16):3655-3665. 13. Sanz MA, Grimwade D, Tallman MS, et al. Management of acute promyelocytic leukemia: recommendations from an expert panel on behalf of the European LeukemiaNet. Blood 2009;113(3):1875-1891. 14. DiNardo CD, Stein EM, de Botton S, et al. Durable remissions with ivosidenib in IDH1-mutated relapsed or refractory AML. N Engl J Med. 2018;378(25):2386-2398.

15. Sykes DB, Kfoury YS, Mercier FE, et al. Inhibition of dihydroorotate dehydrogenase overcomes differentiation blockade in acute myeloid leukemia. Cell 2016;167(1):171-186. 16. Dang CV. MYC on the path to cancer. Cell. 2012;149(1):22-35. 17. Itkonen HM, Minner S, Guldvik IJ, et al. O-GlcNAc transferase integrates metabolic pathways to regulate the stability of c-MYC in human prostate cancer cells. Cancer Res. 2013;73(16):5277-5287. 18. Jozwiak P, Forma E, Brys M, Krzeslak A. O-GlcNAcylation and metabolic reprograming in cancer. Front Endocrinol (Lausanne) 2014;5:145. 19. Ladds M, van Leeuwen IMM, Drummond CJ, et al. A DHODH inhibitor increases p53 synthesis and enhances tumor cell killing by p53 degradation blockage. Nat Commun. 2018;9(1):1107. 20. Cao L, Branstrom A, Baird J, et al. PTC299 is a novel DHODH inhibitor that modulates VEGFA mRNA translation and inhibits proliferation of a broad range of leukemia cells. Blood. 2017;130(Suppl 1):1371.

The complexity of stem cell transplants: can we improve our understanding? Andrea Bacigalupo1 and Francesca Bonifazi2 1 Fondazione Policlinico Universitario Gemelli IRCCS, Roma and 2Istituto di Ematologia “Seràgnoli”, Azienda Ospedaliera Universitaria Sant’Orsola, Bologna, Italy

E-mail: apbacigalupo@yahoo.com doi:10.3324/haematol.2018.198010

ox regression analysis can be considered a robust, easy and universal way to calculate the role of variables on outcome endpoints, such as survival, disease-free survival, and so on. The Cox model is a semiparametric approach based on the strong assumption that the effects of different variables on survival (or on the particular endpoint) are constant over time and are additive in a particular scale. The setting of allogeneic stem cell transplantation is, however, complicated by two additional levels that limit the application of Cox analysis and call for new, more complex, statistical methods: the first is that some variables in allogeneic stem cell transplantation are not timefixed covariates (such as age, gender, or type of donor) but develop after a certain interval of time from transplantation, and need to be accounted for as time-dependent. In other words, with a starting population of patients, some will develop an event (e.g., cytomegalovirus infection) and some will not: a comparison of patients with and without cytomegalovirus infection will need to consider the infection as a time-dependent variable. A further level of complexity is provided by competing events: a competing event is one that precludes the event of interest from occurring, or significantly changes its probability. Death before cytomegalovirus infection, is a clear example of a competing event for cytomegalovirus infection. Relapse and non-relapse mortality is another clear example of competing events. So, there are time-fixed covariates, time-dependent events, and competing events. In a study published in this issue of Haematologica, Fuerst and colleagues have added a fourth level of complexity: they hypothesized that the effect of different covariates may be different at different intervals from transplantation, and this is exactly what they found.1 One example is the stem cell source: bone marrow and haematologica | 2018; 103(9)

peripheral blood as sources of stem cells have been compared in numerous prospective and retrospective studies, including meta-analyses, to define which is better, and results have often been conflicting. Again the complexity of transplantation does not make comparisons easy: in the first randomized study2 of patients with low-risk disease, receiving a myeloablative regimen and HLA identical sibling grafts, the hazard risk (HR) of death was 1.20 for recipients of peripheral blood compared to bone marrow (P=0.2). In a more recent prospective study3 with unrelated donor grafts, using both myeloablative and reduced intensity conditioning regimens for patients with low, intermediate and high-risk disease, the risk of death was 1.20 for bone marrow versus peripheral blood (P=0.2). Fuerst and colleagues offer a new way of looking into this particular issue: they found that peripheral blood has a significant protective effect on non-relapse mortality early after transplantation, and a significant detrimental effect later on.1 The time point for a change of effect on non-relapse mortality was set at 8 months: this means that patients receiving peripheral blood grafts had a lower non-relapse mortality within 8 months (HR: 0.75) and a higher non-relapse mortality beyond 8 months after transplantation (HR:1.38), which were both highly significant effects (Figure 1). There was no protective effect of peripheral blood on relapse, which is the competing event (Figure 1). The authors also looked at a second model of competing events (transplant-related mortality and non-transplant related, or death due to other causes, including relapses), disproving common beliefs; they found no protective effect of peripheral blood as compared to bone marrow grafts on deaths due to other causes, which raises the question of whether peripheral blood should remain the preferred stem cell source in allogeneic stem cell transplants. Indeed an increased risk of chronic graft-versus-host disease seems not to be compensated by 1417

Editorials

Figure 1. Time-dependent effect of peripheral blood grafts compared to bone marrow grafts. The box plots represent the hazard ratio (HR) for nonrelapse mortality (NRM) <8 months from transplant (0.75; range, 0.680.84) (P<0.001), for NRM >8 months from transplant (1.38; range, 1.141.66) (P<0.001) and for relapse any time after transplantation (1.04; range, 0.94-1.15) (P=0.4). This analysis illustrates a protective effect of peripheral blood (PB) on NRM early after transplant; a detrimental effect of PB on NRM later on, and no effect of PB on relapse, when compared to bone marrow (BM) as a source of stem cells.

reduced deaths from other causes, and non-relapse mortality is significantly increased in the long-term. Another debated issue is the comparison between reduced intensity and myeloablative conditioning regimens, and their effect on relapse and survival.4-5 The authors found that reduced intensity conditioning regimens protect patients from early non-relapse mortality (as expected), but this effect is lost after 4 months, and its competing event, relapse, unfortunately, increases constantly over time. Thus, when using a reduced intensity conditioning regimen, the clinician must be aware that the beneficial effect is short-lived and that in the longterm there is no protection against non-relapse mortality, with significantly greater risk of relapse. In the era of personalized medicine the statistical approach suggested by Fuerst et al. provides a tool to disentangle the effects of different transplant components. This in turn gives new answers, sometimes unexpected, to important questions, such as the lack of reduced relapse risk using peripheral blood cells, or the significantly increased risk of relapse with reduced intensity condi-

1418

tioning regimens. A better understanding of these components lays the basis for a reconsideration of transplant protocols and the design of tailored clinical trials.

References 1. Fuerst D, Frank S, Mueller C, et al. Competing-risk outcomes after hematopoietic stem cell transplantation from the perspective of timedependent effects. Haematologica. 2018;103(9):1527-1534. 2. Friedrichs B, Tichelli A, Bacigalupo A, et al. Long-term outcome and late effects in patients transplanted with mobilised blood or bone marrow: a randomised trial. Lancet Oncol. 2010;11(4):331-338. 3. Anasetti C, Logan BR, Lee SJ, et al. Peripheral-blood stem cells versus bone marrow from unrelated donors. N Engl J Med. 2012;367(16):14871496. 4. BornhĂ¤user M, Kienast J, Trenschel R, et al. Reduced-intensity conditioning versus standard conditioning before allogeneic haemopoietic cell transplantation in patients with acute myeloid leukaemia in first complete remission: a prospective, open-label randomised phase 3 trial. Lancet Oncol. 2012;13(10):1035â&#x20AC;&#x201C;1044. 5. Scott BL, Pasquini MC, Logan BR, et al. Myeloablative versus reducedintensity hematopoietic cell transplantation for acute myeloid leukemia and myelodysplastic syndromes. J Clin Oncol. 2017;35 (11):1154-1161.

haematologica | 2018; 103(9)

Editorials

Arterial thrombosis and cancer: the neglected side of the coin of Trousseau syndrome Valerio De Stefano Fondazione Policlinico Universitario A. Gemelli IRCCS, Istituto di Ematologia, Università Cattolica, Roma, Italia E-mail: valerio.destefano@unicatt.it doi:10.3324/haematol.2018.197814

he clinical observation of the association between cancer and thrombosis was first reported in 1823 by Jean Baptiste Bouillaud when he was still a medical student.1 However, Armand Trousseau in 1865 was the first to study systematically clinical and autopsy findings and formulate the theory of the close and frequent association between cancer and increased risk of thrombosis. He re-examined the previous observations: “I have just reperused the memoir of Dr. Bouillaud, and have pleasure in stating that the work of my colleague is as complete as if it had been written yesterday.”2 In his 95th lecture on clinical medicine, delivered at the Hôtel Dieu in Paris, he noted: “I have long been struck with the frequency with which cancerous patients are affected with painful oedema in the superior or inferior extremities, whether one or other was the seat of cancer. [...] I have since that period had an opportunity of observing other cases of painful oedema, in which, at the autopsy, I found visceral cancer, but in which during life, there was

no appreciable cancerous tumor; and in which there existed a cachexia referable neither to the tubercular diathesis, the puerperal state, nor chlorosis. I have thus been led to the conclusion, that when there is a cachectic state not attributable to the tuberculous diathesis or to the puerperal state, there is most probably a cancerous tumor in some organ. [...] So great, in my opinion, is the semiotic value of phlegmasia in the cancerous cachexia, that I regard this phlegmasia as a sign of the cancerous diathesis as certain as sanguinolent effusion into the serous cavities.”2 The procoagulant mechanisms underlying cancer-associated thrombophilia are complex and multifactorial. However, the expression of tumor cell-associated clot promoting properties leads to the activation of the clotting cascade, with the generation of thrombin and fibrin, and the stimulation of platelets, leukocytes and endothelial cells, enhancing their cellular procoagulant features.3 A possible particular role of neutrophil extracellular traps

Table 1. Studies investigating the incidence of arterial thrombosis conducted in patient cohorts composed of different cancer groups.

Reference

Cancer patients, n

Time frame

Design

Khorana et al.10

66,106

1995-2002

Multicenter retrospective cohort

Di Nisio et al.11

1934

2003-2009

Zoller et al.12,13

820,491

1987-2008

Navi et al.14

279,719

2002-2011

Brenner et al.15

5717

2009-2014

Grilz et al.8

1880

2003-2013

Inclusion criteria

Source of information

ATE, n (%)

Diagnosis of cancer, Discharge 1135 hospitalization, database (1.72%) chemotherapy, neutropenia Multicenter Diagnosis of cancer, Medical 5 retrospective ambulatory follow up, records (0.27%) cohort chemotherapy, no history of ATE Nationwide Diagnosis of National N/A retrospective cohort cancer computerized registry (MigMeg2 database) based on unique person number Retrospective Age >65 years, SEER-Medicare Cumulative matched cohort primary breast, dataset incidence lung, prostate, colorectal, at 1 yr: 6.5% bladder, pancreatic, at 2 yrs:9.1% gastric cancer or non-Hodgkin lymphoma Multicenter Active cancer, RIETE 63 prospective cohort previous VTE registry (1.10%) Monocenter Age >18 years, newly CATS dataset 48 prospective diagnosed or relapsed (2.55%) cohort cancer, no indication Cumulative to long-term Incidence anticoagulation at 1 yr: 1.7% at 2 yrs:2.6%

Risk vs. controls

AMI, n (%)

N/A

577 (0.87%)

N/A

4 (0.20%)

38 (1.96%)

At 6 months 1.7 (AMI) 2.2 (IS)

34,666 (4.2%)

31,524 (3.8%)

N/A

At 6 months 2.2 (ATE) 2.9 (AMI) 1.9 (IS)

Cumulative Incidence at 1 yr: 2.6% at 2 yrs:3.7%

N/A

15 (0.26%) 20 (1.06%)

N/A

IS, n (%)

VTE, n (%)

422 4255 (0.64%) (6.44%)

Cumulative N/A Incidence at 1 yr: 4.3% at 2 yrs:5.8%

42 499 (0.73%) (8.72%) 16 157 (0.85%) (8.35%)

n: number; yr/yrs: year(s); N/A: not available; AMI: acute myocardial infarction; ATE: arterial thromboembolic event; CATS: Vienna Cancer and Thrombosis Study; IS: ischemic stroke; RIETE: Registro Informatizado de Enfermedad TromboEmbolica; SEER: Surveillance Epidemiology and End Results-Medicare; VTE: venous thromboembolism.

haematologica | 2018; 103(9)

1419

Editorials

Figure 1. Interaction of multiple factors involved in the pathogenesis of arterial thromboembolism in cancer patients. IMIDs: immunomodulatory drugs; TKI: tyrosine kinase inhibitors; VEGF: vascular endothelial growth factor.

(NETs) in promoting cancer-associated arterial thrombosis has been reported in patients with ischemic stroke and cancer.4 According to Trousseauâ&#x20AC;&#x2122;s first observations, traditionally the association between thrombosis and malignancy has been focused on the occurrence of venous thromboembolism (VTE). Overall, cancer has been reported to account for 17-29% of all VTE cases;5 on the other hand, in the prospective Vienna Cancer and Thrombosis Study (CATS), 7.7% of the cancer patients developed VTE within one year after diagnosis or progression of the disease.6 The epidemiology of arterial thrombosis in cancer patients has received much less attention. Indeed, thrombosis in arteries has long been recognized in cancer patients.7 In this issue of the Journal, Grilz et al. report the up-dated results of the CATS study, aimed to investigate the incidence of arterial thromboembolism (ATE), the related risk factors, and the impact on mortality in a cohort of 1880 patients with active newly diagnosed or relapsed cancer.8 The global burden of arterial thrombosis in the general population for the two common artery acute diseases (i.e. acute myocardial infarction and ischemic stroke) has been estimated to be 139.3 and 114.3 per 100,000 people, accounting for an overall incidence rate of 0.25% of individuals.9 In the CATS cohort, the frequency of ATE was 2.6% during a median prospective observation time of two years. In agreement with previous reports, the rate of VTE was 8.4%.8 It is difficult to compare this with earlier reports due to the considerable heterogeneity in the design, inclusion criteria, and source of information (Table 1).10-15 In the paper of Grilz et al., the multivariable analysis demonstrated among the types of cancer a significant association only with kidney cancer [Hazard Ratio 1420

(HR) 1.6];8 this association had already been reported in some cohorts12,13 but not in others.10 Lung cancer was found to be significantly associated with ATE in most reports;10,12-15 this was confirmed in the CATS cohort but only under univariable analysis.8 Cancer stage has been reported to be associated with the risk of ATE,12-15 but this increase in risk has not been found in the CATS cohort,8 in agreement with a report on hospitalized patients under chemotherapy.10 Grilz et al. addressed the role of cardiovascular risk factors in the CATS cohort as determinants of ATE in cancer; age, male sex, and smoking were independent risk factors for ATE. Moreover, history of cardiovascular disease, hypertension, and medical conditions needing treatment with lipid-lowering agents or antiplatelet agents were all associated with a 2.2-3.7-fold increase in risk of ATE.8 This was aligned with the findings obtained in the RIETE cohort of cancer patients with VTE, in which history of arterial thrombosis, hypertension, diabetes, and use of statins were over-represented in the patients with subsequent ATE compared with those without.15 The role of previous ATE as a predictor of a novel ATE is indirectly confirmed by the low rate of ATE in the study of Di Nisio et al.,11 in which patients with a history of cardiovascular events were excluded. The median age of the patients recruited in the CATS cohort was 61 years, explaining the relatively lower rate of ATE with respect to the study based on the SEER-Medicare dataset, which collected data from patients over 65 years of age (Table 1).14 Notably, the presence of two or more cardiovascular risk factors (hypertension, diabetes, known arterial cardiovascular disease, and dyslipidemia) resulted in an as high as 5.6-12.5% higher risk of ATE at two years from diagnosis in comparison with patients with none or a sinhaematologica | 2018; 103(9)

Editorials

gle risk factor (cumulative incidence of ATE 1.4-2.7% at 2 years) (Online Supplementary Appendix).8 The risk of cancer-related ATE has been reported to vary during the disease, being highest during the first 612 months after diagnosis and then declining.12-14 In the CATS cohort, the peak of ATE soon after diagnosis was modest and was then subsequently constant over the entire follow-up period, in contrast to VTE, which peaked during the first six months after diagnosis.8 This distribution could be due to different compositions of the patient cohorts as regards type of cancer, as well as the presence of cardiovascular risk factors. The time distribution of cancer-related ATE has been reported to vary according to the nature of the disease. Patients with colorectal cancer, bladder cancer, and non-Hodgkin lymphoma showed an increased risk of myocardial infarction but not of ischemic stroke beyond one year.14 On the other hand, attenuation of the incidence rate of ATE over time or persistence up to ten years from the diagnosis has been variously described according to different cancer diagnoses.12,13 However, the constant rate of ATE over time could be explained by the role of cardiovascular risk factors as a persistent trigger independently of the course of the cancer. It should be pointed out that, in the CATS cohort, cardiovascular risk factors that are significantly associated with ATE, such as hypertension or known arterial cardiovascular disease, are present in 37.5% and 8.5% of the cohort, respectively. Moreover, 303 patients (16.1% of the entire cohort) were at higher risk (i.e. presence of 2 or more cardiovascular risk factors) (Online Supplementary Appendix).8 In cancer patients, the development of ATE has been reported to be associated with a 4-5-fold increased hazard for death.10,14 This was substantially confirmed in the CATS cohort, where the occurrence of ATE was associated with a 3.2-fold increased risk of death.8 The major limitation of the study of Grilz et al.,8 as of the other studies, is the lack of information about cancer treatment (chemotherapy or radiotherapy). Many therapeutic agents are associated with a significant risk of ATE or VTE, as recently reviewed (Figure 1).16,17 Therefore the drugs used for the active treatment of cancer are essential co-variates in the estimate of the hazard risk of the cancer-related thrombosis. Moreover, antithrombotic prophylaxis administered during active therapy should be taken into account. In the RIETE cohort of cancer patients who have had a VTE event, 86% of the subsequent ATE events occurred during heparin treatment and only 6.3% during antiplatelet treatment.15 In conclusion, ATE is fully recognized as a less frequent but key part of Trousseau syndrome, complicating the course of cancer and having a significant impact on the prognosis. As has been well established for VTE, ATE can be the first signal of an otherwise occult malignancy and can precede the overt diagnosis,4,18 such that it should arouse suspicion of a potential cancer, especially in those patients without cardiovascular risk factors. The awareness that cancer itself could be an additional risk factor for ATE, and that cancer therapy can also cause

haematologica | 2018; 103(9)

short-term and long-term cardiovascular complications, makes it mandatory to identify high-risk patients, to modify the pre-existent cardiovascular risk factors, and to adopt effective antithrombotic prophylaxis.19,20 Early diagnosis and timely intervention in response to these issues can significantly improve the management of cancer patients and lead to a better outcome, especially for those who are elderly or who have cardiovascular risk factors, and those undergoing therapies with cardiovascular toxicity.

References 1. Bouillaud JB. De l'obliteration des veines et de son influence sur la formation des hydropisies partielles: consideration sur la hydropisies passive et general. Arch Gen Med. 1823;1(2):188-204. 2. Trousseau A. Lecture XCV. Phlegmasia alba dolens. In: Lectures on Clinical Medicine, delivered at the Hôtel-Dieu, Paris, translated from the Third Edition of 1868 by Sir J. R. Cormack. London: The New Sydenham Society; 1872. vol. 5, p. 281-332. 3. Ay C, Pabinger I, Cohen AT. Cancer-associated venous thromboembolism: burden, mechanisms, and management. Thromb Haemost. 2017;117(2):219-230. 4. Thålin C, Demers M, Blomgren B, et al. NETosis promotes cancer-associated arterial microthrombosis presenting as ischemic stroke with troponin elevation. Thromb Res. 2016;139:56-64. 5. Timp JF, Braekkan SK, Versteeg HH, Cannegieter SC. Epidemiology of cancer-associated venous thrombosis. Blood. 2013;122(10):1712-1723. 6. Ay C, Dunkler D, Marosi C, et al. Prediction of venous thromboembolism in cancer patients. Blood. 2010;116(24):5377-5382. 7. Sack GH Jr, Levin J, Bell WR. Trousseau's syndrome and other manifestations of chronic disseminated coagulopathy in patients with neoplasms: clinical, pathophysiologic, and therapeutic features. Medicine (Baltimore). 1977;56(1):1-37. 8. Grilz E, Königsbrügge O, Posch F, et al. Frequency, risk factors, and impact on mortality of arterial thromboembolism in patients with cancer. Haematologica. 2018 May 24. [Epub ahead of print] 9. Wendelboe AM, Raskob GE. Global Burden of Thrombosis: epidemiologic aspects. Circ Res. 2016;118(9):1340-1347. 10. Khorana AA, Francis CW, Culakova E, Fisher RI, Kuderer NM, Lyman GH. Thromboembolism in hospitalized neutropenic cancer patients. J Clin Oncol. 2006;24(3):484-490. 11. Di Nisio M, Ferrante N, Feragalli B, et al. Arterial thrombosis in ambulatory cancer patients treated with chemotherapy. Thromb Res. 2011;127(4):382-383. 12. Zöller B, Ji J, Sundquist J, Sundquist K. Risk of coronary heart disease in patients with cancer: a nationwide follow-up study from Sweden. Eur J Cancer. 2012;48(1):121-128. 13. Zöller B, Ji J, Sundquist J, Sundquist K. Risk of haemorrhagic and ischaemic stroke in patients with cancer: a nationwide follow-up study from Sweden. Eur J Cancer. 2012;48(12):1875-1883. 14. Navi BB, Reiner AS, Kamel H, et al. Risk of Arterial Thromboembolism in Patients With Cancer. J Am Coll Cardiol. 2017;70(8):926-938. 15. Brenner B, Bikdeli B, Tzoran I, et al; RIETE Investigators. Arterial ischemic events are a major complication in cancer patients with venous thromboembolism. Am J Med. 2018 May 25. [Epub ahead of print] 16. Aronson D, Brenner B. Arterial thrombosis and cancer. Thromb Res. 2018;164 Suppl 1:S23-S28. 17. Tuzovic M, Herrmann J, Iliescu C, Marmagkiolis K, Ziaeian B, Yang EH. Arterial thrombosis in patients with cancer. Curr Treat Options Cardiovasc Med. 2018;20(5):40. 18. Sundbøll J, Veres K, Horváth-Puhó E, Adelborg K, Sørensen HT. Risk and prognosis of cancer after lower limb arterial thrombosis. Circulation. 2018 Mar 14. [Epub ahead of print] 19. Chang HM, Moudgil R, Scarabelli T, Okwuosa TM, Yeh ETH. Cardiovascular complications of cancer therapy: best practices in diagnosis, prevention, and management: part 1. J Am Coll Cardiol. 2017;70(20):2536-2551. 20. Chang HM, Okwuosa TM, Scarabelli T, Moudgil R, Yeh ETH. Cardiovascular complications of cancer therapy: best practices in diagnosis, prevention, and management: part 2. J Am Coll Cardiol. 2017;70(20):2552-2565.

1421

REVIEW ARTICLE Ferrata Storti Foundation

Haematologica 2018 Volume 103(9):1422-1432

Cardiovascular adverse events in modern myeloma therapy – Incidence and risks. A review from the European Myeloma Network (EMN) and Italian Society of Arterial Hypertension (SIIA)

Sara Bringhen,1 Alberto Milan,2 Claudio Ferri,3 Ralph Wäsch,4 Francesca Gay,1 Alessandra Larocca,1 Marco Salvini,1 Evangelos Terpos,5 Hartmut Goldschmidt,6 Michele Cavo,7 Maria Teresa Petrucci,8 Heinz Ludwig,9 Holger W. Auner,10 Jo Caers,11 Martin Gramatzki,12 Mario Boccadoro,1 Hermann Einsele,13 Pieter Sonneveld14 and Monika Engelhardt4 on behalf of the European Hematology Association, the European Myeloma Network and the Italian Society of Arterial Hypertension

Myeloma Unit, Division of Hematology, University of Torino, Azienda OspedalieroUniversitaria Città della Salute e della Scienza di Torino, Italy; 2Department of Medical Sciences, Internal Medicine and Hypertension Division, University of Torino, Azienda Ospedaliera Universitaria Città della Salute e della Scienza di Torino, Italy; Rete Oncologica Piemontese, Italy; 3University of L’Aquila, MeSVA Department, San Salvatore Hospital, Division of Internal Medicine & Nephrology, Coppito, Italy; 4Department of Medicine I, Hematology, Oncology & Stem Cell Transplantation, Medical Center, Faculty of Medicine, University of Freiburg, Germany; 5Department of Clinical Therapeutics, University of Athens School of Medicine, Greece; 6Medizinische Klinik, Abteilung Innere Medizin V, Universitätsklinikum Heidelberg und National Centrum für Tumorerkrankungen (NCT), Heidelberg, Germany; 7“Seràgnoli” Institute of Hematology and Medical Oncology, University of Bologna, Italy; 8Hematology, Department of Cellular Biotechnologies and Hematology, Sapienza University of Rome, Italy; 91. Medical Department and Oncology, Wilhelminenspital Wien, Austria, 10Department of Medicine, Imperial College London, UK; 11 Department of Hematology, Domaine University, Liege, Belgium; 12Division of Stem Cell Transplantation and Immunotherapy, University of Kiel, Germany; 13Department of Internal Medicine II, University Hospital of Würzburg, Germany and 14Department of Hematology, Rotterdam, the Netherlands 1

Correspondence: sarabringhen@yahoo.com

Received: February 15, 2018. Accepted: June 5, 2018. Pre-published: July 26, 2018.

doi:10.3324/haematol.2018.191288 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1422 ©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.

1422

ABSTRACT

ardiovascular disease in patients with multiple myeloma may derive from factors unrelated to the disease (age, diabetes, dyslipidemia, obesity, prior cardiovascular diseases), related to the disease (cardiac AL-amyloidosis, hyperviscosity, high-output failure, arteriovenous shunting, anemia, renal dysfunction) and/or related to antimyeloma treatment (anthracyclines, corticosteroids, alkylating agents, immunomodulatory drugs, proteasome inhibitors). Good knowledge of cardiovascular events, effective dose reductions, prevention and management of early and late cardiovascular side effects of chemotherapeutic agents are essential in current clinical practice. Myeloma experts are obliged to carefully balance the efficacy and toxicity of drugs for each individual patient. This review summarizes current data and novel insights into cardiovascular adverse events of today's anti-myeloma treatment, focusing on carfilzomib, as a starting point for developing consensus recommendations on preventing and managing cardiovascular side effects in patients with multiple myeloma.

Introduction Multiple myeloma (MM) is a plasma cell dyscrasia accounting for 1% of neoplastic diseases. It typically affects the elderly population, with the median age at diagnosis being 70 years.1,2 Cardiovascular disease is one of the most frequent comorbidities in MM patients,3 being the main cause of death in western countries.4 Since haematologica | 2018; 103(9)

Cardiovascular safety in MM

the global population is aging, the prevalence of both MM and cardiovascular disease is expected to increase in the near future.5,6 Cardiovascular disease in MM may derive from factors unrelated to the disease (age, diabetes, dyslipidemia, obesity, history of cardiovascular diseases), or those related to the myeloma (cardiac AL-amyloidosis, hyperviscosity, high-output failure, arteriovenous shunting, anemia, renal dysfunction) and/or be related to the treatment of the disease [anthracyclines, corticosteroids, alkylating agents, immunomodulatory drugs, proteasome inhibitors, highdose conditioning and autologous/allogeneic stem cell transplantation (SCT)].7 Of interest, the incidence of cardiovascular disease has been described to be higher in patients with MM (60.1%) than in non-MM patients (54.5%)7 and a recent review and meta-analysis of patients enrolled in clinical trials with carfilzomib showed that the incidence of all grade and grades â&#x2030;Ľ3 cardiovascular adverse events (CVAEs) was 18.1% and 8.2%, respectively.8 Although carfilzomib is one of the most cardiotoxic and, at the same time, effective novel agents used in MM treatment, it is not the only anti-myeloma drug characterized by a high risk of CVAEs. Moreover, it is difficult to estimate the actual incidence of chemotherapy-induced cardiovascular diseases, because current data about druginduced cardiotoxicity were generated in clinical trials, from which patients with severe cardiovascular comorbidities were excluded. However, in real-life clinical practice, MM patients may indeed suffer from cardiovascular disease, have cardiovascular risk factors, and/or may have already received several cardiotoxic drugs. Physicians managing patients with MM are, therefore, obliged to carefully balance expected drug efficacy with toxicity. This review is published as a consensus paper by the European Hematology Association (EHA), the European Myeloma Network (EMN) and the Italian Society of Arterial Hypertension (SIIA), and it aims to describe CVAEs related to anti-myeloma treatment, especially novel agents and the proteasome inhibitor carfilzomib. Furthermore, the paper intends to be a starting point for future debates on how to manage CVAEs appropriately. At present, studies to validate recommendations on prevention and management of CVAEs during MM treatment are lacking, and overcoming this lacuna is particularly relevant. Besides focusing on toxicity issues, this review also aims to provide a more complete overview of MM treatments by clarifying the efficacy and outcome advantages associated with such treatments. Of note, this paper is not only a review of the current literature, but it is also the result of a thorough discussion involving a panel of EMN clinical experts in MM and cardiology experts.

Cardiovascular toxicity of multiple myeloma treatment Antineoplastic therapy is frequently complicated by the onset of CVAEs, which in turn may increase treatmentrelated morbidity and mortality.9 Cardiovascular disease or events are among the most frequent adverse events during chemotherapy, leading to the concept of cardiotoxicity, which may be due to the direct effects of antineoplastic treatment on the structure and function of the heart or may accelerate the onset of cardiovascular disease in highhaematologica | 2018; 103(9)

risk patients with predisposing cardiovascular risks.10 Cardiovascular complications of antineoplastic therapy may be acute, subacute or chronic and can be classified depending on the type of damage,11 the organ structure or function targeted by the damage,10 or the drug responsible for the damage.12 Depending on the damage, two types of cardiotoxicity have been described: type I and type II. In type I cardiotoxicity, mostly associated with traditional anticancer therapies, there is irreversible destruction of myocytes, leading to congestive heart failure. Type II cardiotoxicity may involve a reversible loss of cardiac contractility and is mainly caused by new targeted therapies, such as vascular endothelial growth factor inhibitors and tyrosine kinase inhibitors.11 With regards to the organ structure or function that is the target of the damage, the classification of cardiovascular chemotherapy-induced complications recognizes myocardial dysfunction and heart failure, coronary artery disease and myocardial ischemia, arrhythmias, hypertension, thromboembolic disease, peripheral vascular disease and stroke, pulmonary hypertension, valvular diseases and pericardial complications.12 In MM, drug-related cardiotoxicity may be caused by chemotherapeutic agents or by new targeted therapies, such as immunomodulatory drugs (thalidomide, lenalidomide, pomalidomide) and proteasome inhibitors (bortezomib, ixazomib, carfilzomib).

Cardiovascular toxicity induced by chemotherapy Myocardial dysfunction and heart failure are common side effects of chemotherapy and may appear early during antineoplastic treatment or years thereafter. The risk of heart failure is increased in elderly patients with cardiovascular risks. Anthracyclines and alkylating agents are associated with direct cytotoxic cardiac injury. Cardiac damage induced by anthracyclines involves myocyte cell membrane injury by oxygen free radicals and lipid peroxidation, leading to oxidative stress10 and reduced cardiac contractility. In <1% of patients the effects may be manifested acutely (as supraventricular arrhythmia, transient left ventricular dysfunction or electrocardiographic changes). More commonly, the manifestations occur early within the first year of treatment [decline in left ventricular ejection fraction (LVEF)] or late (chronic heart failure). Cardiotoxic effects are dosedependent and increased in patients with pre-existing cardiovascular diseases and in elderly patients. Cardiotoxicity is amplified by combination treatment, such as with alkylating or antimicrotubule agents, immuno- and targeted drugs. Cyclophosphamide induces heart failure in 7-28% of cases,9 although mainly in relation to high doses before autologous SCT. The mechanism of damage may be direct endothelial injury, extravasation of toxic metabolites with interstitial hemorrhage, edema and structural damage to cardiomyocytes. Anthracyclines and alkylating agents may induce arrhythmias through direct cardiotoxicity, cardiac ischemia or metabolic changes caused by the chemotherapy.13 Moreover, all cancers lead to a prothrombotic state, release of pro-angiogenic cytokines and drug-induced hepatotoxicity, which may contribute further to thromboembolism. Thromboembolic risks are particularly high in patients with other predisposing factors, in those with disseminated tumors and in those with hematologic neoplasms, in particular MM.14 Arterial thromboembolism is 1423

S. Bringhen et al.

a rare event and may occur in patients receiving anthracyclines. Venous thromboembolism is frequent, occurring in 15-20% of patients with MM.10 Anthracyclines and cyclophosphamide may also cause acute pericarditis,11 and high doses of alkylating agents have been associated with pulmonary veno-occlusive disease.10

Cardiovascular toxicity induced by new targeted therapies, such as immunomodulatory drugs Immunomodulatory drugs may induce arrhythmias, such as bradycardia and atrioventricular block. The bradycardia and heart block may be caused by the antineoplastic drugs themselves, age-related fibrosis or AL-amyloidosis. Thalidomide is associated with sinus bradycardia in 5% of patients. Sinus bradycardia can lead to syncope and may require placement of a pacemaker.15 Immunomodulatory drugs are characterized by an increased risk of thromboembolic events. The mechanisms underlying this phenomenon are direct damage to endothelial cells, increased platelet aggregation and higher von Willebrand factor serum levels. In the absence of thromboprophylaxis, the risk of venous thromboembolism is 1.3 and 4.1 per 100 patient-cycles in newly diagnosed patients receiving thalidomide alone and in combination with dexamethasone, respectively. The corresponding risk at relapse is 0.4 and 0.8 per 100 patient-cycles and it increased to 6.7 in patients treated with thalidomide, dexamethasone and doxorubicin.16 In the absence of thromboprophylaxis, the risk of venous thromboembolism is 0.8 and 0.7 per 100 patient-cycles in newly diagnosed and relapsed patients, respectively, receiving lenalidomide and dexamethasone. Venous thromboembolism may be fatal if complicated by pulmonary embolism. Adequate prophylaxis, which is based on risk evaluation, is therefore mandatory. Risk factors can be related to characteristics of the patient (age, obesity, history of venous thromboembolism, centralvenous catheter, comorbidities – such as diabetes, infections, cardiac diseases –, surgical procedures – including vertebroplasty and kyophoplasty –, and inherited thrombophilia); related to the myeloma (diagnosis per se, hyperviscosity) and related to therapy (high-dose dexamethasone, doxorubicin, or multi-agent chemotherapy). Prophylaxis in low-risk patients (≤1 patient/myelomarelated risk factor) consists of aspirin 100 mg, unless contraindicated. If more than one risk factor is present, low molecular weight heparin or full-dose warfarin should be used and continued for at least 4 months; subsequently, switching to aspirin may be an option.17,18 New oral anticoagulants are increasingly being considered, although these drugs have not been systematically studied in myeloma yet. Specific recommendations cannot, therefore, be made as yet.19 Recently, a pilot study with the novel oral anticoagulation agent apixaban was conducted to evaluate the feasibility of venous thrombo-prophylaxis in 104 patients with MM, during treatment with immunomodulatory drugs. Results were encouraging but further evaluation is needed.

Cardiovascular toxicity of proteasome inhibitors The metabolism of cellular proteins has a pivotal role in regulating cell function and homeostasis. Intracellular protein degradation is mainly regulated by the ubiquitin-proteasome pathway, which acts on proteins involved in cell cycle functions, apoptosis, transcription and DNA repair.20 Proteins are tagged by ubiquitin and recognized by the 1424

26S proteasome complex, which degrades proteins into small peptides.21 If proteasome activity is altered, cells stop growing and undergo apoptosis at an increased rate because of an increasingly aberrant proteome.22 In several cancers, neoplastic cells are more sensitive than untransformed cells to proteasome inhibition. Based on this evidence, proteasome inhibitors therefore represent a highly relevant therapeutic strategy in MM treatment. The pathophysiology of the cardiotoxicity induced by proteasome inhibitors is not entirely clear. One mechanism could be the inhibition of sarcomeric protein turnover, resulting in myocyte death.23 Other hypotheses are increased apoptosis of endothelial progenitor cells, endothelial nitric oxide synthase dysfunction, functional/structural abnormalities of cardiomyocytes secondary to protein accumulation due to impaired protein degradation, inhibition of physiological NFκB activity, potentiation of the effects of other cardiotoxic agents, suppression of the adaptive/cytoprotective activity of the ubiquitin-proteasome pathway in an underlying cardiomyopathy, myocardial scarring and fibrosis, endoplasmic reticulum stress induction in cardiomyoblasts, dysfunction of cardiac mitochondria, contractile left ventricular dysfunction and/or increased smooth muscle cell apoptosis causing atherosclerotic plaque instability.24-29

Bortezomib Bortezomib is a dipeptide boronic acid that inhibits the proteolytic activity of the proteasome chymotrypsin-like site, via the formation of a reversible interaction at the 26S proteasome. With regards to bortezomib-related cardiotoxicity, the information leaflet for patients contains a warning concerning the possibility of acute development or exacerbation of chronic heart failure and a new onset of decreased LVEF. There have also been isolated cases of QT-interval prolongation, but causality has not been established.30 The incidence of all grades of cardiac dysfunction in patients taking bortezomib shows marked variability from 2% to 17.9%, depending on the clinical trial.9,29 In the meta-analysis conducted by Xiao et al., the incidence of all grades and high-grade bortezomib-related cardiotoxicity was 3.8% and 2.3%, respectively.29 Of interest, bortezomib does not seem to significantly increase the risk of all grades or high-grade cardiotoxicity, if compared with control medications. Nevertheless, the overall incidence of bortezomib-related cardiotoxicity of all grades was higher in MM patients (4.3%) than in non-MM patients.29 The main limitation of this meta-analysis was, however, that many cardiovascular comorbidities, as well as concomitant therapies, were unknown or unrecorded. Moreover, cardiotoxicity was misreported in many studies and the trials' treatment designs were very different, giving rises to heterogeneity of data regarding cardiovascular effects of bortezomib and a high probability of confounding effects secondary to other prior treatments or clinical conditions. In the ENDEAVOR study, patients were randomized to receive carfilzomib and dexamethasone (Kd) or bortezomib and dexamethasone (Vd).31 The patients’ median age was fairly low (65 years), with only about 15% of patients aged ≥75 years. The incidence of CVAEs with Vd treatment is summarized in Table 1. During treatment or within 30 days of receiving the last study-dose, 5% of patients in the Vd group died and cardiovascular events were the second cause of death. Interestingly, dyspnea haematologica | 2018; 103(9)

Cardiovascular safety in MM

Table 1. Incidence (in %) of cardiovascular events in patients with relapsed/refractory multiple myeloma treated with carfilzomib in phase 2 and 3 studies

Hypertension All grades Grade ≥3 Phase 3 studies ASPIRE37 # KRd group (n=392) Rd group (n=389) ENDEAVOR31 § Kd group (n=463) Vd group (n=456) FOCUS40 Carfilzomib group (n=157) CS±cyclophosphamide group (n=158) Phase 2 studies38 * Carfilzomib (n=526)

Cardiac failure All grades Grade ≥3

Ischemic heart disease All grades Grade ≥3

Dyspnea All grades Grade ≥3

14.3 6.9

4.3 1.8

6.4 4.1

3.8 1.8

5.9 4.6

3.3 2.1

19.4 14.9

2.8 1.8

25 9

9 3

<9 <4

<6 <3

<3 <4

<2 <3

28 13

5 2

15 6

3 0

5 1

2 1

15 9

1 0

7.2

5.7

3.4

1.3

K: carfilzomib; R: lenalidomide; d: dexamethasone; V: bortezomib; CS: corticosteroids. #In the ASPIRE study, the category of cardiac failure included (in descending order of frequency) cardiac failure, congestive cardiac failure, pulmonary edema, hepatic congestion, cardiopulmonary failure, acute pulmonary edema, acute cardiac failure, and right ventricular failure. The category of ischemic heart disease included (in descending order of frequency) angina pectoris, myocardial infarction, acute myocardial infarction, an increased serum creatine kinase level, coronary artery disease, myocardial ischemia, coronary artery occlusion, an increased troponin level, an increased level of troponin T, an acute coronary syndrome, abnormal results on a cardiac stress test, cardiomyopathy stress, unstable angina, coronary-artery stenosis, an abnormal ST-T segment on electrocardiography, and an abnormal T wave on electrocardiography. §In the ENDEAVOR study, cardiac failure included (in descending order of frequency): cardiac failure, ejection fraction decreased, pulmonary edema, acute cardiac failure, congestive cardiac failure, acute pulmonary edema, acute left ventricular failure, chronic cardiac failure, cardiopulmonary failure, hepatojugular reflex, right ventricular failure, and left ventricular failure. Ischemic heart disease included (in descending order of frequency): angina pectoris, acute coronary syndrome, myocardial infarction, increased troponin T, coronary artery disease, increased troponin I, acute myocardial infarction, myocardial ischemia, and cardiomyopathy stress. *Phase 2 studies38 = pooled analysis of safety data from four phase 2 studies (PX171-003-A0, PX-171-003-A1, PX-171-004, and PX-171-005) in 526 patients with relapsed and refractory multiple myeloma treated with carfilzomib. Carfilzomib was administered intravenously on days 1, 2, 8, 9, 15, and 16 in 28-day cycles. The planned dose regimen was 20/27 mg/m2 (starting dose of 20 mg/m2 in cycle 1 escalating to 27 mg/m2 in cycle 2) for all studies except PX171-005 (15/20/27 mg/m2 in cycles 1-3 n=50).

was reported as a CVAE, but its etiology (cardiac, pulmonary versus other causes) remained undetermined, so it could be classified as either a cardiovascular or pulmonary adverse event.

Ixazomib ixazomib is an oral analog of bortezomib that reversibly inhibits the chymotrypsin-like site of the 20S proteasome, and is responsible for caspase-dependent induction of apoptosis and inhibition of cell cycle in MM cells. Moreover, ixazomib inhibits the NFκB pathways in MM supporting cells, thus influencing cytokines important for MM cell growth and survival.20 Kumar et al. reported an incidence of grade ≥3 hypertension of 5% in treatment-naïve patients given ixazomib in combination with lenalidomide and dexamethasone (Rd).32 The TOURMALINE MM1 study group analyzed the safety and efficacy profile of ixazomib in a phase 3 trial in patients with relapsed, refractory or relapsed and refractory MM.33 Patients received ixazomib-Rd or placebo-Rd. The rates of serious adverse events were similar in both groups (47% and 49%), as were death rates during the study (4% and 6%, respectively); grade 3 or higher adverse events occurred in 74% and 69% of cases, respectively. There were no differences in the incidence of heart failure (4% in each group), arrhythmias (16% and 15%), hypertension (6% and 5%) and myocardial ischemia (1% and 2%, respectively). At present, there is only one case report in which ixazomib was described as possibly responsible for cardiotoxicity with a mechanism similar to that observed with bortezomib (type I chemotherapy-induced cardiotoxicity).34

Carfilzomib Relapsed or refractory multiple myeloma This second-generation proteasome inhibitor is an irreversible cell-permeable tetrapeptide epoxyketone, an anahaematologica | 2018; 103(9)

log of epoxomicin.35 The irreversible binding leads to the activity of carfilzomib being more sustained than that of bortezomib, as new proteasome subunit synthesis and assembly are required for restoration of proteasome activity.21 Carfilzomib has been demonstrated to be more specific than bortezomib, and the irreversible carfilzomib binding may provide better inhibition of proteasome activity, which may overcome bortezomib-resistance.36 In a head to head study of carfilzomib versus standard treatments, bortezomib or lenalidomide, carfilzomib demonstrated superior efficacy and was associated with improved overall survival. Data regarding cardiac events in MM patients receiving carfilzomib are available from various phase 1, 2 and 3 studies, as well as retrospective and observational analyses, in both relapsed/refractory MM and newly diagnosed MM. Data regarding the dosing and treatment schedules of relevant studies are reported in Table 2. The most frequent adverse event was hypertension. In randomized clinical trials, the relative risks of allgrade and grade ≥3 CVAEs were 1.8 and 2.2, respectively, for patients receiving carfilzomib compared with control patients. Carfilzomib doses of 45 mg/m2 or higher were associated with high-grade CVAEs.8 The incidence of hypertension in patients with relapsed/refractory MM being treated with carfilzomib was 14.3-25% for all grades and 3-9% for grade ≥3, the cumulative incidence of cardiac failure of any grade was 59% and that of grade ≥3 was 2-6%. Corresponding cumulative values for ischemic heart disease were 3-5.9% and 1.3-3.3%. The incidence of dyspnea of any grade was 1519.4% while that of grade ≥3 was 1-5% (Table 1). The most frequent cardiovascular events were dyspnea and hypertension.31,37-43 Carfilzomib has been approved in Europe for the treatment of relapsed or refractory MM, in combination with Rd or dexamethasone, based on the randomized trials 1425

S. Bringhen et al. Table 2. Data from clinical trials of carfilzomib in relapsed/refractory multiple myeloma and newly diagnosed multiple myeloma.

Study

Phase

N. of patients

Relapsed and refractory MM ASPIRE37 3

ENDEAVOR

396 396 464 465

FOCUS40

157

158

PX-171-003-007-A038 PX-171-003-007-A138 PX-171-00438 PX-171-00538 CHAMPION-141 Phase 1

2 2 2 2 1-2

46 266 164 50 116 27

Phase 2

Newly diagnosed MM CLARION47,48 3

478 477

MYELOMA XI60

526

1021 1021

CHAMPION-249

3 (36 mg/m2) 3 (45mg/m2) 16 (56mg/m2)

NCT0220424150 NCT0185711550 NCT0134678750

1-2

148

Regimen

CFZ 20 mg/m2 cycle 1 d 1 + 2, 27 mg/m2 on d 8, 9, 15, 16 cycle 2+ 27 mg/m2 on d 1, 2, 8, 9, 15, 16 R 25 mg d 1–21, D 40 mg on d 1,8, 15, 22 R 25 mg days 1–21, D 40 mg on d 1,8, 15, 22 CFZ 20 mg/m2 cycle 1 d 1 + 2, 56 mg/m2 on d 8, 9, 15, 16 cycle 2+ 56 mg/m2 on d 1, 2, 8, 9, 15, 16 D 20 mg on d 1, 2, 8, 9, 15, 16, 22, 23 V 1.3 mg/m2 on d 1, 4, 8, 11, D 20 mg on d 1, 2, 4, 5, 8, 9, 11, 12 CFZ 20 mg/m2 cycle 1 d 1 + 2, 27 mg/m2 on d 8, 9, 15, 16 cycle 2-9 27 mg/m2 on d 1, 2, 8, 9, 15, 16 cycle 10+ 27 mg/m2 on d 1, 2, 15, 16; treatment on d 8 and 9 was optional per the investigator’s discretion P 30 mg or D 6 mg or other equivalent corticosteroid (not to exceed 84 mg D or equivalent per 28-d cycle). Patients could receive optional V 50 mg per the investigator’s discretion. CFZ 20 mg/m2 on d 1, 2, 8, 9, 15, 16; 28-d cycles CFZ 20/27 mg/m2 on d 1, 2, 8, 9, 15, 16; 28-d cycles CFZ 20 or 27 mg/m2 on d 1, 2, 8, 9, 15, 16; 28-d cycles CFZ 15, 20 or 27 mg/m2 on d 1, 2, 8, 9, 15, 16; 28-d cycles

Median prior lines (range) 2 (1–3) 2 (1–3)

Notes ≈50% of patients received a prior ASCT with Mel200 ≈60% previously received V, 20% lenalidomide

2 (1–2) 2 (1–2) 5 (3–15)

Prior antimyeloma drugs, median (range) 8 (5–17)

5 (3–17)

Prior antimyeloma drugs, median (range) 8.5 (4–14)

4 4 4 4

CFZ on d 1, 8, and 15 of 28-d cycles. 20 mg/m2 on 1 cycle 1, d 1 (C1D1); subsequent doses (beginning with C1D8) escalated at 45, 56, 70, or 88 mg/m2 to determine MTD CFZ at MTD according to the same schedule as phase 1. 1 D 40 mg d 1, 8, 15, 22 of a 28-d cycle for the first 8 cycles and omitted on d 22 during cycle 9 and beyond.

Median CFZ treatment duration 7.8 months

Median n of cycles CFZ 20 mg/m2 d 1-2 and 36 mg/m2 thereafter 9 Each regimen was given as 42-day treatment M 9 mg/m2 and P 60 mg/m2 d 1-4 of each treatment cycle for 9 cycles BTZ 1.3 mg/m2 on d 1, 4, 8, 11, 22, 25, 29, 32 M 9 mg/m2 and P 60 mg/m2 d 1-4 of each treatment cycle 9 CFZ 20 mg/m2 d 1-2 and 36 mg/m2 thereafter Patients were planned to receive a minimum of Cy 500 mg/m2 on d 1, 8 4 cycles of induction. Therapy was continued to R 25mg/d, d 1-21, D weekly (40/20mg cycles 1-4/5+) maximum response or intolerance. After in 28-d cycles induction eligible patients could undergo autologous transplant. Cy 500 mg/m2 on d 1, 8 R 25 mg/d, d 1-21, D weekly (40/20 mg cycles 1-4/5+) in 28-d cycles Cy 500 mg/m2 on d 1, 8, 15 T 100-200 mg/d, d 1-21, D weekly (40/20 mg cycles 1-4/5+) in 28-d cycles CFZ, twice-weekly, 3 + 3 dose escalation at 36, 45, and 8 Treatment was continued for 8 cycles or until 56 mg/m2, followed by dose expansion. unacceptable toxicity, withdrawal of consent, or CFZ on d 1, 2, 8, 9, 15, 16 of each cycle (20 mg/m2 on progressive disease. d 1 + 2 of cycle 1 and the escalated dose thereafter), Cy 300 mg/m2 on d 1, 8, 15 D on d 1, 8, 15, 22 of each 28-day cycle. CFZ 36 mg/m2 on d 1, 2, 8, 9, 15, 16 in NCT02204241, 9 After completing 9 cycles, patients received 3 dose levels escalated from 45-70 mg/m2 on d 1, 8, 15 in maintenance with CFZ until progression or NCT01857115, and on d 1, 2, 8, 9, 15, 16 in NCT01346787. intolerance. Cy 300 mg/m2 on d 1, 8, 15, D 40 mg 1x/week Median age: 72 (range: 55-85) years, 38 (26%) patients were >75 years old. Median follow-up was 21 months. continued on the next page

1426

haematologica | 2018; 103(9)

Cardiovascular safety in MM continued from the previous page

Study

Phase

N° of patients

CYKLONE61

1-2

Jakubowiak 201251

1-2

Dytfeld 201452

1-2

Regimen

Median prior lines (range)

CFZ was-escalated to 15/20, 20/27, 20/36 and 20/45 mg/m2 to determine the MTD (20/36 mg/m2). CFZ on d 1, 2, 8, 9, 15, 16 Cy 300 mg/m2 on d 1, 8, 15 T 100 mg on d 1–28, D 40 mg on d 1, 8, 15, 22 53 CFZ (20, 27, 36 mg/m2, on d 1, 2, 8, 9, 15, 16 +d 1, 2, 15, 16 after cycle 8) R 25 mg/d, d 1-21, D weekly (40/20 mg cycles 1-4/5+) in 28-d cycles 23 (subset of Dose escalation: elderly patients CFZ 20 mg/m2 for cycle 1 on d 1, 2, 8, 9, 15, 16 of Jakubowiak of 28 d, and then at 20, 27, or 2012(49) 36 mg/m2 for 7 cycles. Maintenance CFZ-R-D (cycles 9-24): CFZ on d 1, 2, 15, 16. R 25 mg on d 1–21, D on d 1, 8, 15, 22 at 40 mg (cycles 1-4), and at 20 mg thereafter.

CARTHADEX62

111

Bringhen 201453

Bringhen 201754

1/2

Korde63

FORTE55

154 309

12 (1-25)

24 (1-24)

Dose escalation: 4 CFZ 20 mg/m2 on d 1, 2 then 27 mg/m2 on d 8, 9, 15, 16 of cycle 1 and on d 1, 2, 8, 9, 15, 16 of cycles 2 to 4. T 200 mg on d 1- 28, D 40 mg on d 1, 8, 15, 22. CFZ was escalated to 20/36 mg/m2 in cohort 2, to 20/45 mg/m2 in cohort 3 and to 20/56 mg/m2 in cohort 4. Dose escalation: 9 CFZ on d 1, 2, 8, 9, 15, 16 (20 mg/m2 on d 1 + 2 of cycle 1 and 36 mg/m2 thereafter). Cy 300 mg/m2 on d 1, 8, 15, D 40 mg on d 1, 8, 15, 22 Maintenance: CFZ 36 mg/m2 on d 1, 2, 15, 16 until progression/ intolerance. Dose escalation: CFZ on d 1, 8, 15 (20 mg/m2 on cycle 1, d 1; subsequent doses were escalated in a standard 3+3 dose-escalation 9 scheme at 45, 56, 70 mg/m2. In phase 2, CFZ 70 mg/m2, with same schedule as in phase 1) Cy 300 mg/m2 on d 1, 8, 15, D 40 mg on d 1, 8, 15, 22. Maintenance: CFZ 70 mg/m2 on d 1, 15 until progression or intolerance. CFZ on d 1, 2, 8, 9, 15, 16 (starting dose, 20 mg/m2 8 on d 1 + 2 of cycle 1; target dose, 36 mg/m2 thereafter). R 25 mg on d 2- 21 of cycle 1 and on d 1- 21 of cycles 2 through 8. D on d 1, 2, 8, 9, 15, 16, 22, 23 (20 mg for cycles 1-4 and 10 mg for cycles 5-8; D was not administered on d 1 of cycle 1). CFZ 20/3 6 mg/m2 on d 1, 2, 8, 9, 15, 16 4 Cy 300 mg/m2 on d 1, 8, 15, D 20 mg on d 1, 2, 8, 9, 15, 16 CFZ 20/36 mg/m2 on d 1, 2, 8, 9, 15, 16 D 20mg on d 1, 2, 8, 9, 15, 16, R 25 mg d 1-21 cycles

Notes

After cycle 4, ASCT-candidates underwent stem cell collection then continued treatment with option of ASCT. Max. planned dose (CFZ 36 mg/m2) was expanded in phase 2. MTD was not established, but based on efficacy and safety, in phase II CFZ dose was 36 mg/m2.13 After 24 cycles, patients continued single-agent maintenance lenalidomide off protocol. Median age: 72 years (range: 65-81). 14 patients received a median of 24 cycles (range: 1-24), 2 at a CFZ dose of 20 mg/m2, 4 at 27 mg/m2, and 17 at 36 mg/m2. Induction was followed by stem cell harvest after Cy priming (2 to 4 mg/m2) and G-CSF. Hereafter patients received high-dose M (200 mg/m2) and ASCT followed by consolidation treatment with 4 cycles of CFZ-T-D in the same schedule except a lower dose of T (50 mg).

ASCT-eligible patients underwent stem cell collection after 4 cycles and continued with treatment. After 8 cycles, all patients with at least stable disease were to receive 2 years of extended dosing with R. The first 4 cycles were followed by high-dose M and ASCT and consolidation with 4 CFZ-Cy-D cycles The first 4 cycles were followed by high-dose M and ASCT and consolidation with 4 CFZ-R-D

CFZ: carfilzomib; R: lenalidomide; D: dexamethasone;V: bortezomib; P: prednisone; Cy: cyclophosphamide; d: day (s), MTD: maximum tolerated dose; M: melphalan; T: thalidomide; ASCT: autologous stem cell transplantation; MM: multiple myeloma.

ASPIRE37 and ENDEAVOR,31 and after analysis of safety data derived from 526 patients enrolled in four phase 2 carfilzomib studies.38 In the ASPIRE study, the incidence of dyspnea was 19.4% considering all grades and 2.8% considering only grade ≥3 events, although the origin (cardiac, infectious, haematologica | 2018; 103(9)

pulmonary) remained unspecified.37 There was also increased hypertension, which was almost double in the carfilzomib-Rd (KRd) group as compared to the Rd group (14.3% versus 6.9% for all grades and 4% versus 2% for grade ≥3 in the KRd and Rd groups, respectively). However, cardiac death rates were similar in the two 1427

S. Bringhen et al.

arms: four deaths with KRd (3 myocardial ischemia, 1 cardiac failure) and four with Rd (1 myocardial ischemia, 3 cardiac failures). The findings of the ENDEAVOR study31 were similar to those of ASPIRE, with a 28% incidence of dyspnea of unspecified origin and 25% of hypertension in patients treated with Kd. In this study, there were five cardiac deaths with Kd and six with Vd. Patients in the Kd group showed a higher incidence of grade ≥3 CVAEs, including hypertension, cardiac failure and dyspnea, while the incidence of ischemic heart disease was similar in the two groups. Venous thromboembolism was more common with Kd than with Vd (deep vein thrombosis: 3.7% for all grades and 0.9% for grade ≥3 in Kd recipients versus 0.9% for all grades and 0.4% for grade ≥3 in Vd recipients; pulmonary embolism: 2.6% for all grades and 1.7% for grade ≥3 in Kd recipients versus 0.9% for all grades and 0.9% for grade ≥3 in Vd recipients). Of interest, a preplanned prospective ENDEAVOR substudy (ECHO study) was performed to evaluate changes from baseline in LVEF, right ventricular function and pulmonary artery systolic pressure via echocardiography.39 Patients with a LVEF ≥40% and no evidence of New York Heart Association class III or IV heart failure, symptomatic ischemia, uncontrolled arrhythmias or recent myocardial ischemia were enrolled. All patients (75 treated with Kd, 76 treated with Vd) were assessed using two-dimensional transthoracic echocardiograms at baseline, every 12 weeks and at the end of treatment. Notable differences between the two treatment groups were that more patients in the Kd group were older than 75 years: 21.3% versus 14.5% in the Vd group) and more had a prior cardiac-related history (26.7% versus 14.5%, respectively). Patients in the Kd arm had a higher incidence of heart failure reported by the treating physician (10.8% versus 4.1%, respectively). A history of cardiac disorders was associated with an elevated but not statistically significant different risk of heart failure. Patients receiving carfilzomib had a higher incidence of hypertension compared to those receiving bortezomib (20.3% versus 8.1%, respectively). Twenty-three patients (15.2%) discontinued treatment because of adverse events; eight due to non-fatal cardiac-related adverse events (Kd: n=6 versus Vd: n=2). The primary endpoint was a reduction in LVEF by 24 weeks. The ECHO data were analyzed in a blinded fashion. Among all patients in both arms, only one (in the Vd group) had a significant reduction in LVEF within the first 24 weeks. Six patients (3 in each group) had significant LVEF reductions at some time during the study, with normal LVEF being recovered in all but one. Fourteen patients (Kd: n=8; Vd: n=6) had clinically meaningful changes in echocardiograms; 79% did not meet the echocardiographic criteria for decreased LVEF. More Kd than Vd recipients had clinical evidence of heart failure (n=4 versus n=0, respectively) or pulmonary hypertension (n=4 versus n=1, respectively) based on investigator assessment. Thus, heart failure and pulmonary hypertension occurred more frequently with Kd than with Vd, although a echocardiographically detectable significant decline in LVEF was low in both treatment arms and occurred with similar frequencies. The substudy found limited utility for serial screening with echocardiograms to mitigate cardiac risks for unselected patients receiving carfilzomib. Another subgroup analysis of the ENDEAVOR study, conducted in 109 Asian patients, showed a toxicity profile similar to that in the general population. In the carfilzomib 1428

arm, hypertension was reported in 26.4% of patients, dyspnea in 17.0%.44 The different incidence of cardiotoxicity in ENDEAVOR and ASPIRE can be explained by comparing both study designs. The first factor is the carfilzomib dose: there was a higher incidence of cardiotoxicity in the ENDEAVOR study, but carfilzomib doses were substantially higher (56 mg/m2 versus 27 mg/m2). The second issue is that nearly 20% of the patients enrolled in the ENDEAVOR trial had a creatinine clearance <50 mL/min (and 6% of them had a creatinine clearance <30 mL/min) and, therefore, had a higher cardiovascular risk, while in ASPIRE only 6.3% of patients had a creatinine clearance between 30 and 50 mL/min.45 Consequently, a side-by-side comparison of both trials is hampered by different patient eligibility criteria. An integrated analysis of patients with advanced MM enrolled in four phase 2 studies showed that 73.6% of patients had a history of cardiovascular events, and 70% had baseline cardiac risks.38 Cardiac failure (including chronic heart failure, pulmonary edema, and decreased LVEF) occurred in 7.2% of patients. The overall mortality rate was the same in patients with or without cardiovascular risks at baseline. CVAEs occurred in 22.1% and 14.3% of patients had hypertension (mainly grade 1–2). Cardiac events leading to treatment discontinuation occurred in 4.4% of patients. Cardiac adverse events did not increase in later treatment cycles, suggesting that there was no cumulative toxicity. Among CVAEs, arrhythmias of any grade were observed in 13.3% of patients (grade ≥3 in 2.3%). Dyspnea was included in respiratory adverse events and appeared in 42.2% of patients (all grades) and was grade ≥3 in 4.9%. Newly diagnosed multiple myeloma Carfilzomib, alone or in combination with other drugs, has also been investigated in newly diagnosed MM patients in several ongoing and completed studies.46 Data derived from these trials are very useful, because they are not biased by possible cardiotoxic effects of previous lines of therapy. The incidence of any grade hypertension was 9-24.7% while that of grade ≥3 was 2-10%. The incidence of any grade dyspnea was 18.1-28.6% while that of grade ≥3 was 3.6-4.8%. The most frequent cardiovascular events were dyspnea and hypertension (Table 3) with no apparent correlation with carfilzomib doses, schedules or other drug associations. The CLARION trial compared carfilzomib, melphalan, and prednisone (KMP) versus bortezomib, melphalan, and prednisone (VMP) in newly diagnosed MM patients ineligible for transplant.47,48 The incidence of grade ≥3 adverse events was 74.7% versus 76.2%, respectively. Adverse events of interest included cardiac failure (all grades: 10.8% versus 4.3%; grade ≥3: 8.2% versus 2.8%), dyspnea (all grades: 18.1% versus 8.5%; grade ≥3: 3.6% versus 0.6%), and hypertension (all grades: 24.7% versus 8.1%; grade ≥3: 10.1% versus 3.6%, respectively). The CHAMPION-2 study evaluated carfilzomib with cyclophosphamide and dexamethasone (KCd).49 One patient in the 45 mg/m2 cohort died due to sudden cardiac arrest, whereas no deaths occurred in the 36 or 56 mg/m2 cohorts. The authors concluded that carfilzomib administered at a dose of 56 mg/m2 twice weekly in combination with cyclophosphamide and dexamethasone showed acceptable toxicity. Likewise, an integrated analysis of haematologica | 2018; 103(9)

Cardiovascular safety in MM

newly diagnosed MM patients enrolled in three phase 1/2 studies evaluated CVAEs in patients treated with KCd.50 Any-grade cardiovascular events were reported in 42% of patients. The risk of CVAEs during treatment with carfilzomib was significantly higher in patients ≥75 years old and the most important risk factor, regardless of age, was hypertension. Grade ≥3 adverse events occurred in 30% of patients. In patients <75 years, the most frequent grade ≥ 3 adverse events were hypertension (9%) and dyspnea

(10%). In patients ≥75 years, the most frequent grade ≥3 adverse events were hypertension and pulmonary edema (both: 13%). Thirty-four percent of patients who developed CVAEs had hypertension at baseline or developed it during treatment compared with 14% of patients who did not have CVAEs. The risk of cardiotoxicity was lower during maintenance, with 22% of patients developing a cardiovascular event of any grade. The most frequent CVAE during maintenance was again hypertension (12%),

Table 3. Incidence of cardiovascular events in patients with newly diagnosed multiple myeloma treated with carfilzomib in phase 1, 2 and 3 studies.

Studies

Hypertension

Cardiac failure

Ischemic heart disease

Dyspnea

Exertional dyspnea

All Grade grades ≥3

Phase 3 studies CLARION47,48 KMP (n=474) 24.7 VMP (n=470) 8.1 60 MYELOMA XI KCRd (n=526) 0.5 CTd (n=1021) 0.3 CRd (n=1021) 1 Phase 1/2 studies Phase 1B CHAMPION 249 K dose: 36 mg/m2 (n=3), 45 mg/m2 (n=3), 56 mg/m2 (n=16) CYKLONE61 KCTd (n=64) 9 Jakubowiak, 201251 KRd (n=53) Dytfeld, 201452 KRd ≥65 years (n=23) Phase 2 studies Bringhen, 201453 KCd (n=58) 9 Bringhen, 201754 wKCd (n=54) 11 Korde63 KRd (n=45 NDMM) FORTE55 KRd (n=309) KCd (n=154)

10.1 3.6

10.8 4.3

8.2 2.8

3.0 1.9

2.1 1.3

18.1 8.5

DVT/PE

Arrhythmia

Aggregated cardiac failure * All Grade All Grade All Grade grades ≥3 grades ≥3 grades ≥3

3.6 0.6

1 0 0

2.5 1.3 1.2

4.8

28.6

4.7

38.1

4.8

5.5

11 (DVT)4(DVT) 6 (PE) 6 (PE) 13

5.5

6.6

3.7

1.8

4.4

3 3

1 0

K: carfilzomib; M: melphalan; P: prednisone;V: bortezomib; C: cyclophosphamide; d: dexamethasone; T: thalidomide; R: lenalidomide; w: weekly; DVT: deep vein thrombosis; PE: pulmonary embolism. *Aggregated cardiac failure includes heart failure, pulmonary edema, and myocardial infarction; NDMM: newly diagnosed multiple myeloma.

Table 4. Main studies conducted on real-life patients treated with carfilzomib-based regimens.

Study

Type of study

N. of patients

Rate of pre-existing CV history

Rate of CVAE

Atrash56 Chari64

R R

130 498

11.5% hospitalized for heart failure 22% had ≥1 CVAE; 5% had ≥1 hospitalization for

Rosenthal65 Dimopoulos66

P P

62 60

54% 84% of non-hospitalized; 92% of hospitalized patients 20% baseline hypertension 28%

8% had cardiac SAE; 32% had hypertension 11.6% had a CVAE

R: retrospective; P: prospective; N.: number; CV: cardiovascular; CVAE: cardiovascular adverse event(s); SAE: serious adverse event(s).

haematologica | 2018; 103(9)

1429

S. Bringhen et al.

regardless of age. Five cardiovascular carfilzomib-related deaths were reported. No difference was observed in relation to different carfilzomib doses and schedules. A phase 1/2 study in newly diagnosed MM patients assessed carfilzomib, lenalidomide and weekly dexamethasone. During induction, grade 3/4 CVAEs included deep venous thrombosis/pulmonary embolism in 9% of cases. Grade 3/4 dyspnea was observed only during phase 1 and within the first three cycles.51 An updated follow-up in a subset of 23 elderly patients (≥65 years) was also performed:52 during induction, non-hematologic grade 3/4 adverse events (>10%) included thromboembolic events (13%). During maintenance, most adverse events were grade 1/2 and did not include CVAEs. This could be partially because patients reaching the maintenance phase have deeper and longer responses to therapy, and tolerate the treatment better. In a multicenter, open-label phase 2 trial of KCd in elderly newly diagnosed MM patients, ineligible for autologous SCT53, cardiac adverse events of any grade occurred during induction in 16% of patients and of grade ≥3 in 7% and included four cardiac events (heart failure, arrhythmia, myocardial ischemia and hypertension in one patient each) and stroke, acute pulmonary edema and pulmonary thromboembolism (also in one patient each). The safety profile was similar in the 15 patients >75 years who received one or more doses of KCd. As expected, no CVAEs occurred during maintenance. Seven patients died while on the study, one due to hypertension (related to carfilzomib) and another due to atrial fibrillation (deemed unrelated to carfilzomib). A multicenter, open-label phase 1/2 trial determined the safety and efficacy of KCd in a weekly schedule in newly diagnosed patients aged ≥65 years who were ineligible for autologous SCT.54 Cardiopulmonary adverse events occurred in 9%. During induction, the incidence of CVAEs of any grade was 24%, while the incidence of CVAEs of grade 3–5 was 9%. Treatment-emergent serious adverse events occurred during induction in 26% of patients and included eight cardiac events (heart failure, pulmonary edema, sudden death and hypertension) and one pulmonary thromboembolism. During maintenance, any grade hypertension occurred in 15% of cases. Grade 3-5 non-hematologic adverse events were rare except for hypertension (10% of cases) and CVAEs (5%: 1 heart failure and 1 myocardial ischemia). Treatment-emergent serious adverse events occurred during maintenance in three patients and included heart failure and myocardial ischemia. The FORTE trial compared KRd versus KCd in transplant-eligible patients with newly diagnosed MM.55 No significant differences were seen in rates of CVAEs (hypertension: 7% versus 6%, cardiac adverse events: 3% versus 5%), with the exception of venous thromboembolism/ pulmonary embolism (8% versus 2%). The slightly lower incidence of CVAEs, as compared to other studies, could be partially explained by the fact that newly diagnosed MM patients eligible for transplantation are fitter than those with relapsed or refractory MM and/or patients ineligible for autologous SCT. Real-life experience Data derived from “real-life” experience with carfilzomib unite more unselected patients than those treated within clinical trials, although they do not add much to 1430

currently available knowledge (Table 4). The most robust data are derived from the study by Atrash et al., in 130 relapsed/refractory patients. In that study, 20% of patients had CVAEs. A history of cardiac events was present in 54% of patients.56 Fifteen (11.5%) were hospitalized for chronic heart failure, serious arrhythmias or pulmonary edema. Of four patients hospitalized for arrhythmia, two had cardiac arrests.

Conclusions Melphalan, cyclophosphamide, and doxorubicin are frequently used in the treatment of MM, especially in the context of high-dose chemotherapy. Alkylating agents and anthracyclines are both characterized by the potential for cardiotoxicity. Cardiac function must be monitored carefully in patients taking these drugs, particularly in the long-term, because of late onset toxicity. Electrocardiography and transthoracic echocardiography are useful for assessing cardiac function before high-dose melphalan and/or anthracycline-based treatment. Immunomodulatory drugs, especially in combination with corticosteroids and chemotherapy, are associated with a high frequency of thromboembolic complications. Prophylaxis with cardio-aspirin, low molecular weight heparin or warfarin is mandatory. The choice of the most appropriate drug depends on the thrombotic risk. If the probability of venous thromboembolism/pulmonary embolism is high, low molecular weight heparin or warfarin should be preferred. Cardiac adverse events are thought to be a class effect of proteasome inhibitors because they have been reported with bortezomib, carfilzomib and ixazomib; however, they appear to be more frequent with carfilzomib. A recent meta-analysis showed that the most frequent CVAEs during treatment with carfilzomib are hypertension (all grades: 12.2%; grade ≥3: 4.3%), heart failure (all grades: 4.1%; grade ≥3: 2.5%) and ischemic heart disease (all grades: 1.8%; grade ≥3: 0.8%).8 The incidence of heart failure increases up to 20-25% in patients ≥75 years of age. Although true carfilzomib-induced cardiac failure is infrequent and usually reversible, MM patients are typically elderly individuals and may frequently have cardiovascular comorbidities, thereby increasing their risk of drugrelated CVAEs.57 As found in the clinical trials described above, age is not the only risk factor for cardiovascular toxicity. Comorbidities, especially pre-existing hypertension or cardiovascular diseases, carfilzomib dose, infusion time, and volume of hydration may increase the risk of CVAEs during carfilzomib treatment.43 However, carfilzomib has been shown to prolong both progression-free and overall survival in MM patients, and maximizing the benefit while reducing cardiovascular risks has become a priority in the management of these patients. At present, there are no strategies to prevent CVAEs that have been validated in prospective studies. The EMN expert panel suggests, based on clinical experience, that cardiovascular risks should be assessed before starting carfilzomib and that measures to correct modifiable risk factors, such as hypertension, high cholesterol levels, hyperglycemia, tobacco use, incorrect diet, should be started. Patients with cardiac risks may benefit from a cardiology review prior to receiving treatment and should be closely monitored for fluid overload. Patients receiving haematologica | 2018; 103(9)

Cardiovascular safety in MM

antihypertensive medications may need drug adjustments to reduce their blood pressure while receiving carfilzomib. Regular clinical surveillance with blood pressure control is recommended before and during treatment. In a sub-study of the ENDEAVOR trial, the rate of hypertension was 25% in the entire cohort and 26.4% in an Asian cohort. These rates are higher than in other studies and suggest that particular attention is needed in patients treated with carfilzomib 56 mg/m2 and in subjects of Asian ethnicity. Serial monitoring of cardiac function via echocardiography39 or cardiac biomarkers such as NT-proBNP are considered of limited value in mitigating the risk of carfil-

References 1. Moreau P, San Miguel J, Ludwig. H, et al, on behalf of the ESMO Guidelines Working Group. Multiple myeloma: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2013;24(6): vi133-vi137. 2. National Cancer Institute. SEER Stat Fact Sheets: Myeloma. Available at: www.seer.cancer.gov/statfacts/html/mulmy .html. Accessed March 2016. 3. Kistler KD, Rajangam K, Faich G, Lanes S. Cardiac event rates in patients with newly diagnosed and relapsed multiple myeloma in US clinical practice. Blood. 2012;102 (21):2916. 4. WHO 2016 Fact sheets-media centre. http://www.who.int/mediacentre/factsheets/fs317/en/. 5. WHO 2016 fact files. http://www.who.int/ features/factfiles/ageing/ageing_facts/en/. 6. Benjamin EJ, Blaha MJ, Chiuve SE, et al; American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics-2017 update: a report from the American Heart Association. Circulation. 2017;135(10):e146-e603. 7. Kistler KD, Kalman J, Sahni G, et al. Incidence and risk of cardiac events in patients with previously treated multiple myeloma versus matched patients without multiple myeloma: an observational, retrospective, cohort study. Clin Lymphoma Myeloma Leuk. 2017;17(2):89-96. 8. Waxman AJ, Clasen S, Hwang WT, et al. Carfilzomib-associated cardiovascular adverse events. A systematic review and meta-analysis. JAMA Oncol. 2018;4(3): e174519. 9. Yeh ETH, Bickford CL. Cardiovascular complications of cancer therapy: incidence, pathogenesis, diagnosis, and management. J Am Coll Cardiol. 2009; 53(24):2231–2247. 10. Zamorano JL, Lancellotti P, Rodriguez Muñoz D, et al. 2016 ESC Position Paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: The Task Force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC). Eur Heart J. 2016;37(36):2768-2801. 11. Bhave M, Akhter N, Rosen ST. Cardiovascular toxicity of biologic agents for cancer therapy. Cancer Network 2014. Available at http://www.cancernetwork.com.

haematologica | 2018; 103(9)

zomib-associated cardiac failure.58 In the event of grade 3 or 4 cardiac events, carfilzomib should be withheld until recovery.59 Carfilzomib may be resumed at the physician’s discretion based on a benefit/risk assessment, although preferably at a reduced dose. In general, the risk-benefit ratio for a drug must be considered in the context of the nature and severity of the disease for which it is used. Dynamic, interdisciplinary cooperation between hematologists and cardiologists is the key to the future studies that are needed to assess and manage cardiovascular safety in patients treated with carfilzomib.

12. Sheppard RJ, Berger J, Sebag IA. Cardiotoxicity of cancer therapeutics: current issues in screening, prevention and therapy. Front Pharmacol. 2013;4:19. 13. Hong RA, Limura T, Sumida KN, Eager RM. Cardio-oncology/onco-cardiology. Clin Cardiol. 2010;33(12):733-737. 14. Palumbo A, Rajkumar SV, Dimopoulos MA, et al; International Myeloma Working Group. Prevention of thalidomide- and lenalidomide-associated thrombosis in myeloma. Leukemia. 2008;22(2):414-423. 15. Dimopoulos MA, Eleutherakis-Papaiakovou V. Adverse effects of thalidomide administration in patients with neoplastic diseases. Am J Med. 2004;117(7):508-515. 16. Carrier M, Le Gal G, Tay J, Wu C, Lee AY. Rates of venous thromboembolism in multiple myeloma patients undergoing immunomodulatory therapy with thalidomide or lenalidomide: a systematic review and meta-analysis. J Thromb Haemost. 2011;9(4):653–663. 17. Terpos E, Kleber M, Engelhardt M, et al. European Myeloma Network guidelines for the management of multiple myeloma-related complications. Haematologica. 2015;100(10):1254-1266. 18. Dimopoulos MA, Leleu X, Palumbo A, et al. Expert panel consensus statement on the optimal use of pomalidomide in relapsed and refractory multiple myeloma. Leukemia. 2014;28(8):1573-1585. 19. Pegourie B, Pernod G, Karlin L, et al. Evaluation of an oral direct anti-Xa anticoagulant, apixaban, for the prevention of venous thromboembolism in patients with myeloma treated with IMiD* compounds: a pilot study (MYELAXAT). J Clin Oncol. 2017;35(Suppl15):8019. 20. Kubiczkova L, Pour L, Sedlarikova L, et al. Proteasome inhibitors-molecular basis and current perspectives in multiple myeloma. J Cell Mol Med. 2014;18(6):947-961. 21. Dou PQ, Zonder JA. Overview of proteasome inhibitor-based anti-cancer therapies: perspective on bortezomib and second generation proteasome inhibitors versus future generation inhibitors of ubiquitin-proteasome system. Curr Cancer Drug Targets. 2014;14(6):517-536. 22. Moreau P, Richardson PG, Cavo M, et al. Proteasome inhibitors in multiple myeloma: 10 years later. Blood. 2012;120(5):947-959. 23. Hasinoff BB, Patel D, Wu X. Molecular mechanism of the cardiotoxicity of the proteasomal targeted drugs bortezomib and carfilzomib. Cardiovasc Toxicol. 2017;17(3): 237-250. 24. Takamatsu H, Yamashita T, Kotani T, et al.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35. 36.

Ischemic heart disease associated with bortezomib treatment combined with dexamethasone in a patient with multiple myeloma. Int J Hematol. 2010;91(5):903906. Nowis D, M czewski M, Mackiewicz U, et al. Cardiotoxicity of the anticancer therapeutic agent bortezomib. Am J Pathol. 2010;176(6):2658-2668. Hacihanefioglu A, Tarkun P, Gonullu E. Acute severe cardiac failure in a myeloma patient due to proteasome inhibitor bortezomib. Int J Hematol. 2008;88(2):219-222. Foley P, Hamilton MS, Leyva F. Myocardial scarring following chemotherapy for multiple myeloma detected using late gadolinium hyperenhancement cardiovascular magnetic resonance. J Cardiovasc Med. 2010;11(5): 386-388. Voortman J, Giaccone G. Severe reversible cardiac failure after bortezomib treatment combined with chemotherapy in a nonsmall cell lung cancer patient: a case report. BMC Cancer. 2006;6:129. Xiao Y, Yin J, Shang Z. Incidence and risk of cardiotoxicity associated with bortezomib in the treatment of cancer: a systematic review and meta-analysis. Plos One. 2014;9(1):e87671. http://www.ema.europa.eu/docs/en_GB/ document_library/EPAR__Summary_for_the_public/human/000539/ WC500048136.pdf Dimopoulos M, Moreau P, Palumbo A, et al; ENDEAVOR Investigators. Carfilzomib and dexamethasone versus bortezomib and dexamethasone for patients with relapsed or refractory multiple myeloma (ENDEAVOR): a randomized, phase 3, open-label, multicentre study. Lancet Oncol. 2016;17(1):2738. Kumar SK, Berdeja JG, Niesvizky R, et al. Safety and tolerability of ixazomib, an oral proteasome inhibitor, in combination with lenalidomide and dexamethasone in patients with previously untreated multiple myeloma: an open-label phase 1/2 study. Lancet Oncol. 2014;15(13):1503-1512. Moreau P, Masszi T, Grzasko N, et al; TOURMALINE-MM1 Study Group. Oral Ixazomib, lenalidomide, and dexamethasone for multiple myeloma. N Engl J Med. 2016;374(17):1621-1634. Jouni H, Aubry MC, Lacy MQ, et al. Ixazomib cardiotoxicity: a possible class effect of proteasome inhibitors. Am J Hematol. 2017;92(2):220-221. Kortuem KM, Stewart AK. Carfilzomib. Blood. 2013;121(6):893-897. Schlafer D, Shah KS, Hall Panjic E, Lonial S.

1431

S. Bringhen et al.

37.

38.

39.

40.

41.

42. 43.

44.

45.

46.

1432

Safety of proteasome inhibitors for treatment of multiple myeloma. Expert Opin Drug Saf. 2017;16(2):167-183. Stewart AK, Rajkumar SV, Dimopoulos MA, et al; ASPIRE Investigators. Carfilzomib, lenalidomide, and dexamethasone for relapsed multiple myeloma. N Engl J Med. 2015;372(2):142-152. Siegel D, Martin T, Nooka A, et al. Integrated safety profile of single-agent carfilzomib: experience from 526 patients enrolled in 4 phase II clinical studies. Haematologica. 2013;98(11):1753-1761. Hájek R, Russell SD, Lyon A, et al. A substudy of the phase 3 ENDEAVOR study: serial echocardiographic assessment of patients with relapsed multiple myeloma (RMM) receiving carfilzomib plus dexamethasone or bortezomib plus dexamethasone. Haematologica. 2016;101(S1):263. Hájek R, Masszi T, Petrucci MT, et al. A randomized phase III study of carfilzomib vs low-dose corticosteroids with optional cyclophosphamide in relapsed and refractory multiple myeloma (FOCUS). Leukemia. 2017;31(1):107-114. Berenson JR, Cartmell A, Bessudo A, et al. CHAMPION-1: a phase 1/2 study of once weekly carfilzomib and dexamethasone for relapsed or refractory multiple myeloma. Blood. 2016;127(26):3360-3368. Delforge M, Ludwig H. How I manage the toxicities of myeloma drugs. Blood. 2017;129(17):2359-2367. Ludwig H, Delforge M, Facon T, et al. Prevention and management of adverse events of novel agents in multiple myeloma: a consensus of the european myeloma network. Leukemia. 2018 May 2. [Epub ahead of print] Lee JH, Huang SY, Liu JH, et al. Outcomes for Asian patients with relapsed multiple myeloma treated with carfilzomib and dexamethasone Vs bortezomib and dexamethasone: a subgroup analysis of the phase 3 ENDEAVOR Study (NCT01568866). Haematologica. 2016;101(S1):549-550. Dimopoulos MA, Siegel D, White DJ, et al. Superior efficacy of carfilzomib and dexamethasone (Kd56) vs bortezomib and dexamethasone (Vd) in multiple myeloma (MM) patients with moderate or serious renal failure: a sobgroup analysis of the phase 3 ENDEAVOR study. Blood. 2017;130(Suppl 1):1845. Sugumar D, Keller J, Vij R. Targeted treat-

47.

48.

49.

50.

51.

52.

53.

54.

55.

ments for multiple myeloma: specific role of carfilzomib. Pharmgenomics Pers Med. 2015;20;8:23-33. Facon T, Lee JH, Moreau P, et al. CLARION: Phase 3 study of carfilzomib, melphalan, prednisone vs bortezomib, melphalan, prednisone in newly diagnosed multiple myeloma. Clin Lymphoma Myeloma Leuk. 2017;17(Suppl):e26-e27. Data available at https://www.amgen.com/ media/news-releases/2016/09/amgenannounces-topline-results-from-phase-3kyprolis-carfilzomib-clarion-study-innewly-diagnosed-multiple-myelomapatients/ Boccia RV, Bessudo A, Agajanian R, et al. A multicenter, open-label, phase 1b study of carfilzomib, cyclophosphamide, and dexamethasone in newly diagnosed multiple myeloma patients (CHAMPION-2). Clin Lymphoma Myeloma Leuk. 2017;17(7):433437. Bringhen S, Petrucci MT, Giuliani N, et al. An integrated analysis of cardio-vascular adverse events of carfilzomib, cyclophosphamide and dexamethasone in elderly newly diagnosed myeloma patients enrolled in 3 phase I/II trials. Blood. 2016;128(22): 3336. Jakubowiak AJ, Dytfeld D, Griffith KA, et al. A phase 1/2 study of carfilzomib in combination with lenalidomide and low-dose dexamethasone as a frontline treatment for multiple myeloma. Blood. 2012;120(9):18011809. Dytfeld D, Jasielec J, Griffith KA, et al. Carfilzomib, lenalidomide, and low-dose dexamethasone in elderly patients with newly diagnosed multiple myeloma. Haematologica. 2014;99:e162-e164. Bringhen S, Petrucci MT, Larocca A, et al. Carfilzomib, cyclophosphamide, and dexamethasone in patients with newly diagnosed multiple myeloma: a multicenter, phase 2 study. Blood. 2014;124(1):63-69. Bringhen S, D’Agostino M, De Paoli L, et al. Phase 1/2 study of weekly carfilzomib, cyclophosphamide, dexamethasone in newly diagnosed transplant-ineligible myeloma. Leukemia. 2018;32(4):979-985. Gay F, Rota Scalabrini D , Belotti A, et al. Carfilzomib-lenalidomide-dexamethasone vs carfilzomib-cyclophosphamide-dexamethasone induction: planned interim analysis of the randomized FORTE trial in newly diagnosed multiple myeloma. J Clin Oncol.

2017;35(Suppl 15):8003. 56. Atrash S, Tullos A, Panozzo S, et al. Cardiac complications in relapsed and refractory multiple myeloma patients treated with carfilzomib. Blood Cancer J. 2015;5:e272. 57. Lenihan DJ, Potluri R, Bhandari H, Ranjan S and Chen C. Evaluation of cardiovascular comorbidities among patients with multiple myeloma in the United States. Blood. 2016;128(4):479-487. 58. Rosenthal A, Luthi J, Belohlavek M, et al. Carfilzomib and the cardiorenal system in myeloma: an endothelial effect? Blood Cancer J. 2016;6(1):e384. 59. Atrash S, Tullos A, Panozzo S, et al. Cardiac complications in relapsed and refractory multiple myeloma patients treated with carfilzomib. Blood Cancer J. 2015;5(1):e272e272. 60. Pawlyn C, Davies F, Cairns D, et al. Quadruplet vs sequential triplet induction Therapy for myeloma patients: results of the MYELOMA XI study. Haematologica. 2017; 102(S2):142. 61. Mikhael JR, Reeder CB, Libby EN, et al. Phase Ib/II trial of CYKLONE (cyclophosphamide, carfilzomib, thalidomide and dexamethasone) for newly diagnosed myeloma. Br J Haematol. 2015;169(2):219-227. 62. Wester R, van der Holt B, Asselbergs E, et al. Phase 2 study of carfilzomib, thalidomide, and low-dose dexamethasone as induction/consolidation in newly diagnosed, transplant eligible patients with multiple myeloma, the carthadex trial. Blood. 2016;128(22):1141. 63 Korde N, Roschewski M, Zingone A, et al. Treatment with carfilzomib-lenalidomidedexamethasone with lenalidomide extension in patients with smoldering or newly diagnosed multiple myeloma. JAMA Oncol. 2015;1(6):746-754. 64. Chari A, Aggarwal S, Mezzi K, et al. Cardiac events in real-world multiple myeloma patients treated with carfilzomib: a retrospective claims database analysis. Blood. 2016;128(22):3319. 65. Rosenthal AC, Luthi J, Belohlavek M, et al. The cardiovascular impact of carfilzomib in multiple myeloma. Blood. 2014;124(21): 4748. 66. Dimopoulos MA, Roussou M, Gavriatopoulou M, et al. Cardiac and renal complications of carfilzomib in patients with multiple myeloma. Blood Adv. 2017;1(7):449-454.

haematologica | 2018; 103(9)

REVIEW ARTICLE

Hereditary hemorrhagic telangiectasia: diagnosis and management from the hematologist’s perspective

Ferrata Storti Foundation

Athena Kritharis,1 Hanny Al-Samkari2 and David J Kuter2

Division of Blood Disorders, Rutgers Cancer Institute of New Jersey, New Brunswick, NJ and 2Hematology Division, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA 1

ABSTRACT

Haematologica 2018 Volume 103(9):1433-1443

ereditary hemorrhagic telangiectasia (HHT), also known as OslerWeber-Rendu syndrome, is an autosomal dominant disorder that causes abnormal blood vessel formation. The diagnosis of hereditary hemorrhagic telangiectasia is clinical, based on the Curaçao criteria. Genetic mutations that have been identified include ENG, ACVRL1/ALK1, and MADH4/SMAD4, among others. Patients with HHT may have telangiectasias and arteriovenous malformations in various organs and suffer from many complications including bleeding, anemia, iron deficiency, and high-output heart failure. Families with the same mutation exhibit considerable phenotypic variation. Optimal treatment is best delivered via a multidisciplinary approach with appropriate diagnosis, screening and local and/or systemic management of lesions. Antiangiogenic agents such as bevacizumab have emerged as a promising systemic therapy in reducing bleeding complications but are not curative. Other pharmacological agents include iron supplementation, antifibrinolytics and hormonal treatment. This review discusses the biology of HHT, management issues that face the practising hematologist, and considerations of future directions in HHT treatment.

Introduction Hereditary hemorrhagic telangiectasia (HHT), also known as Osler-WeberRendu syndrome, is a common autosomal dominant disorder that causes abnormal blood vessel formation.1 The eponym recognizes the 19th century physicians William Osler, Henri Jules Louis Marie Rendu, and Frederick Parkes Weber, who each independently described the disease.2 Clinical sequelae of HHT include mucocutaneous telangiectasias, arteriovenous malformations (AVMs), and bleeding, with consequent iron deficiency anemia. Patients with HHT have been found to have abnormal plasma concentrations of transforming growth factor-beta (TGF-β)3 and vascular endothelial growth factor (VEGF)4 secondary to mutations in ENG, ACVRL1 and MADH4.5 There is considerable inter- and intra-family variation in disease onset and clinical severity, even in cases resulting from an identical mutation. Iron deficiency and associated anemia are frequent complications of the disease due to recurrent epistaxis and/or gastrointestinal bleeding. There are no accepted guidelines on management of patients with HHT beyond supportive measures of iron supplementation, red cell transfusion, and directed treatments to ablate bleeding sites and AVMs. Bevacizumab, a recombinant humanized monoclonal antibody that blocks angiogenesis via VEGF inhibition, appears to be promising in HHT as an intravenous formulation for reducing the frequency and severity of epistaxis and impacting quality of life.6,7 However, data on intranasal bevacizumab have been conflicting, and studies investigating the use of intravenous bevacizumab are limited to case reports and retrospective series.8 Treatment of HHT involves a multidisciplinary approach of specialists in cardiology, pulmonology, hepatology, interventional radiology, ear, nose and throat (ENT), genetics, and hematology. This haematologica | 2018; 103(9)

Correspondence: hal-samkari@mgh.harvard.edu

Received: April 4, 2018. Accepted: May 14, 2018. Pre-published: May 24, 2018.

doi:10.3324/haematol.2018.193003 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1433 ©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.

1433

A. Kritharis et al.

review focuses on the biology of HHT and the management issues that confront the hematologist, as well as proposing a hematology management scheme.

Pathogenesis Pathology Hereditary hemorrhagic telangiectasia is a disease characterized by vascular lesions, including AVMs and telangiectasias. AVMs are abnormal connections that form between arteries and veins without an intermediary capillary system. They can occur anywhere in the body, such as in the central nervous system (CNS), lungs, liver or spine. Vascular malformations may be composed of small (nidi 1-3 cm) or micro (nidi <1 cm) AVMs, pulmonary sacs, or direct high-flow connections. While the terms “telangiectasia” and “arteriovenous malformation” are often used interchangeably, as they both occur from a direct connection between an artery and a vein whilst bypassing the capillary system, they are actually pathologically-distinct terms. Telangiectasias, by definition, occur on mucocutaneous surfaces, such as the skin, gastrointestinal (GI) mucosa, or upper aerodigestive tract. AVMs occur in internal organs, such as the liver, lung, and brain.9 Histological evaluation of AVMs reveals an irregular endothelium, increased collagen and actin, and a convoluted basement membrane.10

• Mutations in MADH4 (which encodes for the SMAD4 protein, a transcription factor that mediates signal transduction in the TGF-β pathway17) result in a juvenile polyposis with HHT syndrome (JP-HHT), described later in this review. Over 80% of HHT patients have identifiable mutations,18,19 leaving approximately 20% who meet clinical diagnostic criteria but do not have definitive mutations. Of those with a pathogenic mutation, 61% have ENG mutations, 37% have ACVRL1 mutations, and 2% have MADH4 mutations;20 very small minorities of patients have pathogenic mutations in other genes, described below. Over 600 different mutations have been uncovered in ENG and ACVRL1 in all exons as well as exon/intron boundaries and splice-sites.21 Frameshift and nonsense mutations appear to be more frequent in ENG. Additional loci associated with HHT have been identified on chromosomes 5q31 (HHT3) and 7q14 (HHT4), but have not been completely characterized.20,22,23 Bone morphogenetic protein 9 (BMP9, also known as growth differentiation factor 2 or GDF2), encoded by BMP9 (also called GDF2), is a ligand for the ACVRL1 gene product ALK1. Consequently, mutations in BMP9/GDF2 result in the clinical manifestations of HHT and are referred to as HHT5. In addition, pathogenic mutations in the RASA1 gene have also been associated with a clinical syndrome consistent with HHT24 as well as other vascular anomalies. Little is known about RASA1-mutated HHT.

Gene mutations Gene mutations that have been described in HHT include ENG, ACVRL1 (also known as ALK1), and MADH4 (also known as SMAD4), as well as other postulated loci (Table 1).11,12 • In 1994, ENG, located on chromosome 9q34 and encoding for the protein endoglin (CD105), was the first gene identified in which mutations resulted in HHT, and so HHT due to ENG mutations is known as HHT type 1 (HHT1).13 Endoglin is a cell-surface glycoprotein that functions as part of the transforming growth factor beta (TGFβ) signaling complex that plays an important role in angiogenesis and vascular remodeling.14,15 • In 1996, defects in the ACVRL1 gene on chromosome 12q13, which encodes for the activin receptor-like kinase 1 (ALK1), were recognized to cause HHT, and defects in this gene result in HHT type 2 (HHT2). Like endoglin, ALK1 is a cell-surface protein that is part of the TGF-β signaling pathway and is important in the regulation of angiogenesis.16

Pathophysiology All three identified causative genes are involved in cell signaling via the TGF-β/BMP signaling pathway, which has roles in cell growth, apoptosis, smooth muscle cell differentiation, and vascular remodeling and maintenance.25 The vasculature normally develops from the capillary system with the activation and growth of endothelial cells, the intercellular junctions between them, and the maturation of the basement membrane.26 Capillaries then develop into larger vessels with the recruitment of smooth muscle cells to the endothelial wall where TGF-β is essential. In the healthy patient, ligands in the extracellular space such as TGF-β, activins and BMPs bind to type I and type II serine/threonine receptors of the cell membrane. TGFβ1/2/3 ligand binds to the type II receptor of the TGF-β signaling cascade (TGFβRII) that becomes phosphorylated and recruits the TGF-β type I receptors ALK1 or ALK5.27 Endoglin is an endothelial specific receptor that associates

Table 1. Classification and genetics of the most common hereditary human telangiectasia (HHT) subtypes.

Disease

Genetic mutation (locus)

Primary visceral manifestations

Function of normal gene product

HHT type 1

ENG (9q34.11)

Pulmonary AVMs Brain AVMs

HHT type 2

ACVRL1 (ALK1;12q13.13)

Combined syndrome of HHT and JP-HHT

MADH4 (SMAD4; 18q21.2)

Liver AVMs Pulmonary hypertension Spinal AVMs Gastrointestinal polyps AVMs Pulmonary hypertension

Membrane glycoprotein receptor on endothelial cells, part of the transforming growth factorbeta (TGF-β) receptor complex Activin receptor-like kinase 1 (ALK1), a cell-surface serine/threonine-protein kinase receptor, part of the TGF-β receptor complex MADH4 encodes SMAD4, a transcription factor acting as a mediator in the TGF-β/BMP pathway signaling

AVM: arteriovenous malformations; JP-HHT: juvenile polyposis-HHT.

1434

haematologica | 2018; 103(9)

HHT for the hematologist

with multiple receptor complexes of the TGF-β receptor complex and also modulates ALK1 and ALK5.27,28 Circulating BMP9 has been demonstrated to bind strongly with endoglin and ALK1 receptors found abundantly in the surface membrane of endothelial cells.29 ALK1 receptors phosphorylate SMAD1/5/8 in the cytoplasm to form the SMAD1/5/8-SMAD4 complex that translocates to the nucleus to promote normal endothelial cell proliferation and smooth muscle migration. In contrast, the ALK5 pathway works through SMAD2/3 to inhibit normal endothelial cell proliferation and smooth muscle migration.27,28,30 The result is contrasting responses that balance endothelial proliferation, angiogenesis and smooth muscle migration. In patients with HHT, mutations in endoglin, ALK1, or one of several other proteins in this pathway alter the normal endothelial response. In HHT1, the ENG mutation leads to reduced endoglin, ALK1 and ALK5 signaling; in HHT2, the ALK1 mutation causes reduced ALK1 signaling alone. Mice with one functioning copy of Eng or Acvrl1 show clinical signs of HHT.31 The haploinsufficiency of these proteins along with a second hit, such as tissue injury, infection or hypoxia, likely cause the focal vascular lesions of HHT1 and HHT2 as reduced levels of endoglin or ALK1 cannot maintain the balance needed for normal

blood vessel formation (recruitment of smooth muscle cells and proliferation of endothelial cells).26,32 Decreased TGF-β transcription normally mediated through this pathway, therefore, disrupts the vascular integrity and smooth muscle differentiation of the endothelium resulting in an abnormal cytoskeleton and fragile small vessels. Vascular endothelial growth factor, an endothelial-specific factor for angiogenesis, is of major interest in diseases of vascular malformation and is elevated in HHT patients.33 VEGF production is stimulated by ALK5 (and SMAD2 through activation of ALK5) and inhibited by ALK1 (and SMAD1 through activation of ALK1).34 Therefore, any mutation along the ALK1 pathway (BMP9, ACVRL1, ENG, MADH4) results in elevation of VEGF through reduced ALK1 pathway signaling. VEGF drives many of the pathogenic manifestations of HHT, as normalizing VEGF has been shown to prevent AVMs in Acvrl1-deficient mice.35 This may be secondary to reduced angiogenic stimuli and reduction of feeding arteries from blocking the VEGF that would normally develop and maintain arteriovenous shunts. Other factors that may contribute to the severity of disease include repeated injury and chronic inflammation in keeping with the two-hit hypothesis and stimulation of the ALK1 signaling pathway. An abnormal endothelium

Figure 1. Molecular pathophysiology of hereditary hemorrhagic telangiectasia (HHT). Physiological signaling via ALK1 and ALK5 receptors (activated via binding BMP9 and TGFβ) results in activation of different SMAD pathways, which converge at SMAD4 resulting in transcription of genes involved in angiogenesis.30 In HHT, mutations perturb signaling through ALK1 via mutations in the ALK1 receptor itself, its ligand BMP9, or its modulator, the glycoprotein membrane receptor endoglin. The result is decreased signaling through ALK1 and increased signaling through ALK5, perturbing normal endothelial proliferation and smooth muscle cell migration. Reduced ALK1 signaling and increased ALK5 signaling also result in higher vascular endothelial growth factor (VEGF) levels, causing increased endothelial proliferation (which may be exacerbated by stress or hypoxia), resulting in arteriovenous malformations (AVMs), telangiectasias, and the manifestations of HHT.

haematologica | 2018; 103(9)

1435

A. Kritharis et al.

may also lead to defective synthesis of von Willebrand Factor (VWF) and prolonged bleeding. There have been reports of families affected by both von Willebrand disease (VWD) and HHT. This poses the question of a potential relationship between the two diseases, which has been studied in published case reports. A potential type IIA VWD mutation (IIe865 to Thr) has been identified in affected families.36 While this is a simplified discussion of complex vascular biology, it illustrates why mutations in ENG, ACVRL1/ALK1, MADH4/SMAD4, and BMP9/GDF2 result in the HHT phenotype. A streamlined schematic summarizing the normal physiological signaling of the TGF-β pathway and the pathophysiology of HHT is shown in Figure 1.

Clinical manifestations Patients with HHT vary in disease severity and bleeding complications. This variability is likely attributed to other genes, inflammation, and the environment that modify the primary genetic defect. Common AVM complications include epistaxis, GI bleeding, iron deficiency, iron deficiency anemia, ischemic and hemorrhagic stroke, brain abscess, high output heart failure, and liver failure. It is suggested that certain mutated genes in HHT may be associated with specific clinical manifestations. ENG mutations may be associated with more pulmonary and brain AVMs; ACVRL1 with more liver AVMs, spinal AVMs, epistaxis and pulmonary hypertension; and MADH4 with juvenile colonic polyposis.44

Pulmonary AVMs Epidemiology and disease course Hereditary hemorrhagic telangiectasia affects approximately 1 in 5000 individuals in North America,37 but the highest prevalence is seen in the Afro-Caribbean regions of the Dutch Antilles and France.38 There is also variability regarding HHT subtype, with type 1 HHT being found more in North America and Europe and type 2 being more common in the Mediterranean and South America. However, these statistics may underestimate the actual disease prevalence as the diagnosis is often missed and some patients may be asymptomatic. HHT exhibits incomplete penetrance and clinical manifestations can vary between patients, even within families with known mutations. Patients may relate a history of epistaxis in childhood, often apparent during adolescence. Mild epistaxis or bleeding tendencies increase with age and telangiectasias may be seen after adolescence, often in adulthood.1 Clinical signs of bleeding become more apparent in adulthood, often after the age of 40 years. Symptoms from anemia may be an initial complaint at presentation from gastrointestinal bleeding, seen in approximately one-third of patients. Patients with mutations of ACVRL1 may present later in life, while those with MADH4 mutations may present earlier in childhood with juvenile colonic polyps and early onset colorectal cancer (at a mean age of 28 years).1,39 As a population, patients with HHT probably have a reduced life expectancy, but this is highly dependent on the severity of disease. Patients without internal organ manifestations (such as hepatic, cerebral or pulmonary AVMs) are expected to have a normal or near-normal lifespan, but approximately 10% of patients may die or become debilitated from vascular complications.40 In a large case-control study, 675 HHT patients were compared with age- and sex-matched healthy controls using a population-based UK primary care database. Patients with HHT were more likely to suffer from cerebral abscess, migraine, ischemic/embolic stroke, heart failure, colon cancer, and the numerous bleeding complications characteristic of the disease. The hazard ratio for death for patients with HHT compared with controls was 2.03 (CI: 1.59-2.60; P<0.0001).41 Life expectancy was seven years shorter in HHT patients in one study, with two mortality peaks, one under 50 years and one between 60-79 years of age.42 Finally, a population study in Denmark demonstrated mortality rates double that of the general population in those under 60 years of age.43 1436

Pulmonary AVMs will develop in at least 50% of HHT patients and are more common in HHT1 than HHT2. Since approximately 70% of pulmonary AVMs are due to HHT, the diagnosis of HHT1 should be considered in all patients with pulmonary AVMs. Migraines are quite frequent in patients with pulmonary AVMs.45 Between 5 and 30% of patients may have pulmonary AVMs that may be asymptomatic or present as hemoptysis, dyspnea, hypoxemia or digital clubbing. Brain abscesses and stroke may occur following “dirty” procedures (e.g. dental cleaning) if bacteria can bypass the pulmonary filtration system via right to left shunting from AVMs.46 Polycythemia may occur if there is significant AV shunting. The locus designated as HHT3 appears to predispose to pulmonary AVM formation.

Liver AVMs Liver AVMs may be seen in up to 70% of patients with HHT. HHT2 appears to be associated with more liver AVMs. Although often asymptomatic, the shunting of blood through these AVMs in the liver can precipitate highoutput heart failure, liver failure, or portal hypertension.

High-output heart failure High-output heart failure can manifest due to large pulmonary AVMs and/or hepatic AVMs.47 High-output failure can be defined by: 1) symptoms of heart failure (such as shortness of breath, fatigue, and exercise intolerance); 2) cardiac output >8 L/min or cardiac index >3.9 L/min/m2; and 3) ejection fraction (EF) >50% and venous oxygen saturation >75%.48 Due to abnormal vascular flow through AVMs of the liver or lung, the vasculature may dilate because of increased high flow and/or decreased resistance. This causes the heart to compensate for the lower blood pressure with an increase in heart rate and output, leading to high-output failure. In these HHT patients, anemia may lead to an increased risk of heart failure due to the stress imposed from tachycardia and increased stroke volume.

Epistaxis Epistaxis will manifest in approximately 50% of patients by the age of ten years. This increases with age such that 95% of all HHT patients eventually develop recurrent epistaxis.49 This will become evident in adulthood with consequent iron deficiency anemia.

Gastrointestinal bleeding When significant, gastrointestinal bleeding affects approximately 20% of patients. GI telangiectasias and haematologica | 2018; 103(9)

HHT for the hematologist

AVMs can involve the large and small intestines as well as the stomach.

Central nervous system manifestations Central nervous system manifestations may affect up to 10% of patients with HHT. Cerebral AVMs can be symptomatic and multiple in number,50 and are often present at birth.51 Neurological involvement may result in epilepsy, transient ischemic attack, stroke, or spinal hemorrhage. In addition to embolic strokes and hemorrhage, CNS infections such as brain abscesses may occur in 1% or more of patients, ranging in severity from mild to lifethreatening. They are likely a result of bacterial seeding or septic emboli from ischemic brain matter or pulmonary AVMs.52,53

Skin telangiectasias Skin telangiectasias can be seen on the fingertips, tongue, face, lip, mucosa, and arms in up to 90% of patients (Figure 2).45 These sites can bleed and can be treated with laser ablation.

Iron deficiency/iron deficiency anemia Iron deficiency/iron deficiency anemia is common in HHT. The underlying cause of iron deficiency in this patient population is the chronic blood loss from telangiectasias (e.g. nasal mucosa or intestinal tract) leading to iron store depletion. Approximately 5% of patients with HHT may have severe hemorrhages from epistaxis and/or intestinal AVMs. This consequently leads to a microcytic or normocytic anemia and symptoms of fatigue. Cardiopulmonary complications as described above can develop.

Other events Though other events are not frequently reported, they include thromboembolic disease, pulmonary hypertension, liver disease, high-risk pregnancies, and spinal events. There is a 1% risk of mortality during pregnancy due to hemorrhage from cerebral or pulmonary AVMs.44 Patients are also affected socially and psychologically due to uncontrolled bleeding episodes. They commonly face difficulties with work, travel, social phobias, isolation, anxiety, and depressive disorders.

Figure 2. Clinical manifestations of telangiectasias. (A) Small red telangiectasias are often seen on the skin of hereditary hemorrhagic telangiectasia patients. (B) Similar lesions may be present on the tongue, lips, or palate.

haematologica | 2018; 103(9)

Juvenile polyposis Juvenile polyposis is a rare association with HHT and results from a germline mutation in MADH4.54 This condition is also an autosomal dominant disorder. Mutations in MADH4 may manifest phenotypically as juvenile polyposis alone, HHT alone, or the combined syndrome of JPHHT.55 The polyposis is best characterized by numerous hamartomatous polyps (i.e. 5-100) that are typically benign, but some patients may develop gastric or colorectal cancer, and so screening is encouraged. Patients with JP-HHT associated with MADH4 mutations are at an increased risk for early colorectal cancer.44 These patients may also have thoracic aorta dilation.

Diagnosis Hereditary hemorrhagic telangiectasia is primarily a clinical diagnosis based on the following Curaçao criteria:44 • spontaneous and recurrent epistaxis

Table 2. Screening and management of hereditary human telangiectasia patients. Anemia • Evaluate for blood transfusion and iron requirements • Monitor ferritin, reticulocytes, hemoglobin • Start oral iron to maintain transferrin saturation >20% and ferritin >50 ng/mL • IV iron: 1 g over multiple infusions Epistaxis • Otolaryngology evaluation • Humidification • Nasal moisture with spray/ointment • Electrocautery or laser therapy • Antifibrinolytics, estrogen or progesterone therapy, surgery, and embolization Gastrointestinal bleeding • Evaluation for telangiectasias and AVMs with upper endoscopy, colonoscopy, capsule endoscopy • Antifibrinolytics, estrogen or progesterone therapy, laser therapy, surgery, and embolization CNS AVM • MRI/MRA brain • >1 cm in diameter: neurosurgical evaluation, embolotherapy, +/- stereotactic radiosurgery Pulmonary AVM • Pulmonary evaluation • Transthoracic echocardiogram with bubble study for screening +/- CT/CTA • If 1+ bubbles on echocardiogram: avoid scuba diving, use IV with filters, antibiotic prophylaxis for procedures (amoxicillin or clindamycin if PCN allergic) • Consider embolization Hepatic AVM • Abdominal ultrasound screening +/- CT/MRI • Consider embolization/ligation, liver transplantation Other • Genetic consultation • Evaluation for other bleeding disorders • Discussion regarding anticoagulation, antiplatelet agents • Pregnancy is considered high risk • Consider assessment for hypercoagulability IV: intravenous; CNS: central nervous system; AVM: arteriovenous malformation; CT: computed tomography; MRI: magnetic resonance imaging; MRA: magentic resonance angiography; CTA: CT angiography; PCN: penicillin.

1437

A. Kritharis et al.

• telangiectasias at characteristic sites • visceral arteriovenous malformations or telangiectasias • a first degree relative with HHT (inheritance is usually autosomal dominant). Patients are classified as follows: 3-4 criteria: definite HHT 2 criteria: probable HHT 0-1 criteria: HHT unlikely. Genetic testing can be performed to inform family members, to increase patient awareness, and can guide more focused preventative screening and in cases of uncertainty. For patients with all 4 features present, the clinical sensitivity of the 5 gene HHT panel (assessing for pathogenic mutations in ENG, ACVRL1, MADH4, RASA1, and BMP9) is approximately 87% or higher.19 Although there has recently been an increase in awareness of HHT, it has been estimated that only 10% of all HHT patients are formally diagnosed; this is because of minimal symptoms or the fact that caregivers are not familiar with the disease and its diagnostic criteria.56

Assessment and management The prevention of future HHT complications is as important as treating the immediate active issues (e.g. bleeding) in caring for patients with HHT. Patients are often asymptomatic from undiagnosed AVMs that can lead to significant morbidity and mortality. Knowledge of a patient’s genetic mutation or family history may help confirm the urgency of certain screening tests over others. As outlined in Table 2, the following are relevant measures for identifying potentially significant AVMs: 1) brain magnetic resonance imaging (MRI)/magnetic resonance angiography (MRA); 2) transthoracic echocardiogram with bubble study, followed by computed tomography (CT) scan as appropriate; 3) colonoscopy/endoscopy/video capsule endoscopy; 4) abdominal doppler ultrasound of the liver, followed by CT scan or MRI as indicated; 5) full ENT evaluation (especially if the patient has epistaxis); and 6) skin evaluation. Hematologic evaluation must also include complete blood count, reticulocyte count, erythrocyte sedimentation rate, iron, total iron binding capacity and ferritin. Ferritin levels alone may not accurately reflect iron stores due to the increased inflammation seen in many HHT patients. Consideration should be given to assessment for inherited thrombophilias prior to using antifibrinolytics for treatment of bleeding associated with HHT. Treatment options are patient-specific and are best grouped by local versus systemic measures in a stepwise approach. There are no standard medical therapies for HHT given the few randomized trials in this field. Management can include supportive care, lesion-specific therapy, and systemic treatment. Lesion-specific therapy may call for involvement from otolaryngology, interventional radiology and neurosurgery.

Management of epistaxis The first step in epistaxis management should always be appropriate patient counseling and use of preventive measures within the home to prevent the nasal mucosa from becoming dry. These may include nasal humidification, use of over-the-counter saline sprays or ointments to keep the nasal mucosa moist, and avoidance of nasal trauma (i.e. from nose blowing and/or nose picking).20 1438

When epistaxis occurs that does not cease within a short period of time at home, nasal packing and direct use of topical agents such as tranexamic acid-soaked gauze in an outpatient clinic or emergency room setting may help curtail bleeding but may also increase trauma to the nasal mucosa. Additional local measures that are commonly employed to control bleeding include laser treatments to the nasal mucosa and septodermoplasty.57 Historically, laser photocoagulation and other interventional procedures have been the cornerstone of therapy,58 although this may begin to shift with effective disease-modifying systemic therapeutics on the horizon, detailed later in this review. Nasal closure57 is an effective but extreme form of therapy that is rarely used. Management of epistaxis with antifibrinolytic agents is another consideration when preventive measures and local or topical treatments fail. Hyperfibrinolysis contributes to the bleeding phenotype in HHT59,60 and antifibrinolytics may work to inhibit fibrinolysis on the telangiectatic wall. By preventing fibrin degradation from plasmin, these agents may act to slow bleeding. Epsilonaminocaproic acid and tranexamic acid can be considered in the care of patients with moderate or severe epistaxis.61 In a randomized, double-blind, placebo-controlled, crossover study of 22 patients, tranexamic acid 1 g 3 times daily resulted in a 54% reduction in nosebleeds while on tranexamic acid as compared with the placebo treatment period, although there was no statistically significant improvement in hemoglobin concentration.62 Apart from its inhibition of plasmin, tranexamic acid may have some effect on the underlying disease process; it appears to increase endoglin and ALK1 levels on the endothelium, selectively stimulating the TGF-β pathway.63 Tranexamic acid may have a higher potency and longer half-life than aminocaproic acid in these patients.63 Dosing can be titrated upward if tolerable to tranexamic acid 650-1300 mg orally 3 times daily or aminocaproic acid 500-2000 mg orally every 4-8 hours.63 Other non-specific hemostatic agents, such as desmopressin or factor replacement products, are not optimal management as HHT is not a disease of coagulation factor deficiency. Antifibrinolytics should be avoided in patients with hypercoagulable conditions and/or prior thrombotic events. While the evidence for its use is limited, N-acetylcysteine dosed 600 mg 3 times daily was modestly effective in reducing epistaxis in HHT patients in a pilot study, with the only statistically significant benefit seen in male patients and those with ENG mutations (HHT1).64

Management of GI bleeding Evidence of GI bleeding or a sharp decline in hematocrit without epistaxis should involve a prompt GI evaluation and an upper and lower endoscopy and, if these do not provide clear results, consideration of video capsule endoscopy. Telangiectasias and AVMs may be visualized in the esophagus, stomach, small intestine and/or colon.65 If accessible, local endoscopic treatment should be attempted. Patients with recurrent bleeding, multiple AVMs, and small bowel AVMs may require additional pharmacological measures. As in the management of epistaxis, antiangiogenic, antifibrinolytic agents and/or other hormonal agents may be considered. Octreotide therapy has also been proposed in reducing transfusion needs66 but is without much supporting data. Management of the anehaematologica | 2018; 103(9)

HHT for the hematologist

mia and iron deficiency that result from this blood loss is addressed below.

Management of pulmonary, hepatic, and CNS AVMs Collaboration with a pulmonologist, hepatologist, gastroenterologist, neurologist, neurosurgeon, and interventional radiologist with experience in treating HHT patients is crucial to the management of AVMs found in the lungs, liver or brain. Screening is, therefore, important early in the diagnosis of these patients. Management will depend on the size of the AVMs, symptoms and location, and may include embolization of a pulmonary AVMs, surgical intervention for a CNS AVM and/or continued surveillance. Angiographic treatment of hepatic AVMs may be helpful in some patients but is often considered a higher risk by interventional radiologists.

Management of iron deficiency anemia The development of anemia can have significant consequences for the patient with HHT. Although oral iron [e.g. ferrous sulfate 325 mg 3 times daily, ferrous asparto glycinate-polysaccharide iron complex 150 mg capsules 1-3 times daily] may be adequate for mildly affected HHT patients, many require intravenous iron such as ferumoxytol, iron sucrose or ferric carboxymaltose. Some patients may require 500-1000 mg of iron a month. Sometimes red blood cell (RBC) transfusion support is needed, but chronic RBC transfusion carries risk of infections and can lead to transfusion reactions and alloimmunization. In some patients, supplementation with erythroid stimulating agents (e.g. epoetin alfa, darbepoetin alfa) may be helpful. A suggested approach to the anemic HHT patient is presented in Figure 3.

Use of hormonal agents Estrogen and progestins (e.g. ethinyl estradiol, norethindrone or mestranol) have been used in HHT patients to

reduce bleeding complications. Mestranol or norethynodrel may help increase nasal squamous epithelium and protect nasal lesions from injury. This hormonal therapy, however, can result in gynecomastia and/or loss of libido in men, weight gain, coronary events, and venous thromboembolism (VTE). Given the age of some patients and the potential side effects of this treatment, it has not been widely used. The overall improvement in hematologic parameters is also questionable. Other hormonal treatment options include danazol 200 mg 3-4 times oral daily, tamoxifen 20 mg oral daily or raloxifene 60 mg oral daily.67 But these are not widely used.

Use of novel systemic anti-angiogenic therapies Anti-VEGF therapies are relatively new for patients with HHT, and their use has been increasing. Thalidomide, used commonly in the management of multiple myeloma, is thought to have both vascular and immunomodulatory effects. Its antiangiogenic activity may be due to the suppression of production of VEGF and basic fibroblast growth factor (bFGF).68 Serum levels of VEGF were found to be decreased after thalidomide treatment in patients with GI bleeding.69 Nasal mucosal biopsies in HHT patients with epistaxis treated with thalidomide demonstrated vessel maturation and improved vessel wall defects.70 Bevacizumab, an anti-VEGF antibody, is a rational therapeutic for HHT as it may reduce excessive angiogenesis (Figure 4). To date, all of the studies describing the use of systemic bevacizumab for the management of HHT have been retrospective cohorts, small case series, or single patient case reports (Table 3). A very recent retrospective study by Iyer et al. describes a large cohort of HHT patients receiving bevacizumab to treat GI bleeding and epistaxis.8 Thirty-four patients were given intravenous bevacizumab according to a standardized protocol, resulting in a statistically significant reduction in epistaxis sever-

Figure 3. Treatment algorithm for iron deficiency anemia in hereditary hemorrhagic telangiectasia (HHT). Oral iron may be attempted first but is typically insufficient in HHT patients with moderate or severe chronic bleeding. In this case, intravenous (IV) iron should be given at regular intervals unless bleeding ceases. When bleeding is so severe that IV iron is insufficient, consideration of antifibinolytics, such as tranexamic acid, is the next step. If this is unsuccessful, a trial of bevacizumab therapy is reasonable. CBC: complete blood count; IV: intravenous; PO: oral administration; TID: 3 times / day; q: every.

haematologica | 2018; 103(9)

1439

A. Kritharis et al.

ity scores and RBC transfusion requirements, although 4 patients developed new-onset or worsened hypertension. Most published studies have used bevacizumab at a dose of 5-10 mg/kg every 2-4 weeks for up to 6 cycles. A lower dose may be sufficient based on pharmacokinetic data showing VEGF suppression at 0.3 mg/kg.71 Adverse effects of bevacizumab may include hypertension, proteinuria, venous thromboembolism, intestinal perforation, and poor wound healing. Interestingly, epistaxis, which is often cited as a side effect in non-HHT patients, has not been a major complication in published studies or in our centerâ&#x20AC;&#x2122;s extensive experience. Bevacizumab may have an impact on high output states in reducing cardiac output. In one study,48 25 patients with severe hepatic vascular AVMs were treated with bevacizumab 5 mg/kg every 14 days for 6 cycles and showed an improvement in cardiac index at three months, reduced epistaxis, and improved

quality of life. Bevacizumab nasal spray has been studied as a treatment for epistaxis. In a randomized phase I study (the ELLIPSE study), 40 patients received a single day treatment of 0.05-0.1 mL of (dose escalated) bevacizumab nasal spray into each nostril for a total dose of 12.5-100 mg.72 Initial results suggested that intranasal treatment was safe but not effective.

Use of anticoagulation in patients with thrombosis Patients who develop thrombotic complications present a difficult therapeutic dilemma given the inherent bleeding of the disease. The low serum iron levels in HHT patients have been associated with elevated factor VIII levels, along with a 2.5-fold increased risk of VTE events.67 In those patients who develop a VTE, therapeutic anticoagulation can be administered. This should be managed with caution and the patient should be screened for pul-

Figura 4. Bevacizumab treatment course in a 71-year-old woman with hereditary hemorrhagic telangiectasia (HHT) and chronic gastrointestinal bleeding and epistaxis. For years, the patient was only able to maintain her hemoglobin with 1-2 units of packed red blood cell (RBC) transfusion weekly plus darbepoetin alfa 300 mcg every other week. After beginning bevacizumab 5 mg/kg at time 0, she became transfusion independent immediately and her hemoglobin normalized within two weeks. Bevacizumab was administered every other week for the first 4 infusions, then monthly as maintenance. Duration of maintenance therapy is patientdependent and optimal dose and duration is not known.

Table 3. Published data using bevacizumab to treat chronic bleeding in hereditary human telangiectasia are patients.

Study

Country

Patient number

Dosing

Duration of efficacy

Effect on epistaxis

U.S.

12 months

Immediate improvement

the Netherlands

12 months

Immediate improvement

Brinkerhoff 201179 Thompson 201480

U.S. U.S.

1 9

10 mg/kg every 2 weeks x 2 cycles then 5 mg/kg every 2 weeks x 2 cycles 5 mg/kg every 2 weeks or 7.5 mg/kg every 2 weeks 5 mg/kg every 2 weeks x 4 cycles 0.125 mg/kg IV every 4 weeks x 6 cycles

Epperla 201681

U.S.

5 mg/kg every 2 weeks x 6 cycles

Guilhem 201782

France

36 treated for bleeding

5 mg/kg every 2 weeks x 6 cycles

U.S.

5 mg/kg every 2 weeks x 4 cycles, with modification of dosing depending on response

Bose 200977 Oosting 200978

Iyer 20188

12 months 6 months

Resolution after 4 cycles Improvement in frequency and severity after 3 cycles on average 12 months Reduced need for nasal cautery procedures 6 months (median) 78% of patients had improved bleeding by physician assessment 6.4 months (median), Significant improvement intermittent treatment in epistaxis severity scores

HHT: hereditary hemorrhagic telangiectasia; IV: intravenous.

1440

haematologica | 2018; 103(9)

HHT for the hematologist

monary and cerebral AVMs that may increase their bleeding risk.

Clinical trials and future directions There are several ongoing clinical trials studying new therapies for HHT (Online Supplementary Table S1). HHT is relatively unique in the family of rare bleeding disorders in that several off-the-shelf therapeutics, such as bevacizumab, and the immunomodulatory agents (IMiDs) currently being used are rational targeted therapies that may be highly effective. The majority of studies are currently investigating the use of bevacizumab via different routes of administration (submucosal, topical or intravenous). In a murine model of HHT, four anti-angiogenic agents were studied for their impact on AVM formation.73 Sorafenib (a dual Raf kinase/VEGF receptor inhibitor with additional tyrosine kinase targets) and a pazopanib analog (pazopanib is a multi-target tyrosine kinase inhibitor with anti-VEGF receptor properties) were beneficial in improving anemia from bleeding from the GI tract more than from mucocutaneous lesions in the upper aerodigestive tract. A phase II study is being conducted to examine the efficacy of increasing doses of pazopanib, from 50 mg to 400 mg daily, in reducing epistaxis and improving anemia. Tacrolimus, a calcineurin inhibitor used principally as an immunosuppressive therapy, may have a therapeutic role in HHT. Ruiz et al. identified tacrolimus as an activator of the ALK1-SMAD1/5/8 pathway, improving defects caused by ALK1 loss.74 Their data in human embryonic vascular endothelial cells demonstrated that tacrolimus activated

References 1. McDonald J, Bayrak-Toydemir P, Pyeritz RE. Hereditary hemorrhagic telangiectasia: an overview of diagnosis, management, and pathogenesis. Genet Med. 2011;13(7):607616. 2. Fuchizaki U, Miyamori H, Kitagawa S, Kaneko S, Kobayashi K. Hereditary haemorrhagic telangiectasia (Rendu-Osler-Weber disease). Lancet. 2003;362(9394):1490-1494. 3. Letarte M, McDonald ML, Li C, et al. Reduced endothelial secretion and plasma levels of transforming growth factor-beta1 in patients with hereditary hemorrhagic telangiectasia type 1. Cardiovasc Res. 2005;68(1):155-164. 4. Sadick H, Riedel F, Naim R, et al. Patients with hereditary hemorrhagic telangiectasia have increased plasma levels of vascular endothelial growth factor and transforming growth factor-beta1 as well as high ALK1 tissue expression. Haematologica. 2005;90(6):818-828. 5. Gallione CJ, Richards JA, Letteboer T, et al. SMAD4 mutations found in unselected HHT patients. J Med Genet. 2006;43(10):793-797. 6. Karnezis TT, Davidson TM. Efficacy of intranasal bevacizumab (Avastin) treatment in patients with hereditary hemorrhagic telangiectasia associated epistaxis. Laryngoscope. 2011;121(3):636-638.

haematologica | 2018; 103(9)

ALK1 HHT mutants unresponsive to BMP9, and inhibited Akt and p38 stimulation by VEGF (normally a major driver of angiogenesis). In a mouse model of HHT, hypervascularization and AVMs were reduced in number by treatment with tacrolimus. Tacrolimus may, therefore, represent yet another off-the-shelf pharmacological option of potential therapeutic benefit in HHT patients. Lastly, the aforementioned IMiDs are promising. In comparison with thalidomide and lenalidomide, pomalidomide may be a superior potential therapeutic option due to its efficacy and reduced toxicity (such as less peripheral neuropathy and cytopenias). Interim results from a phase I study of pomalidomide in HHT patients have been reported in which its use was associated with reduced bleeding outcomes in a small cohort of patients.75 Larger studies are needed to better evaluate the efficacy of this and other IMiDs in the management of bleeding in HHT. Future directions in HHT may look to evaluate other antiangiogenic agents and other targets of the vascular endothelium. In patients with Heyde syndrome, acquired VWF syndrome occurs due to the loss of large molecular multimers of VWF from high shear stress.76 The reduced level of VWF observed in a small case series of patients with HHT36 raises the question of whether VWF replacement may reduce bleeding, as patients with ineffective or low VWF cannot effectively clot. Better understanding of the role of acquired VWF deficiency in the pathogenesis of angiodysplasia in Heyde syndrome may prove useful in the development of novel HHT therapies. In conclusion, HHT is a rare but poorly recognized genetic bleeding disorder that demands greater attention in order to develop targeted and rational management strategies that are both safe and cost-effective.

7. Riss D, Burian M, Wolf A, Kranebitter V, Kaider A, Arnoldner C. Intranasal submucosal bevacizumab for epistaxis in hereditary hemorrhagic telangiectasia: A double blind, randomized, placebo controlled trial. Head Neck. 2015;37(6):783-787. 8. Iyer VN, Apala DR, Pannu BS, et al. Intravenous Bevacizumab for Refractory Hereditary Hemorrhagic TelangiectasiaRelated Epistaxis and Gastrointestinal Bleeding. Mayo Clin Proc. 2018;93(2):155166. 9. Olitsky SE. Hereditary hemorrhagic telangiectasia: diagnosis and management. Am Fam Physician. 2010;82(7):785-790. 10. Duncan BW, Kneebone JM, Chi EY, et al. A detailed histologic analysis of pulmonary arteriovenous malformations in children with cyanotic congenital heart disease. J Thorac Cardiovasc. 1999;117(5):931-938. 11. Prigoda NL, Savas S, Abdalla SA, et al. Hereditary haemorrhagic telangiectasia: mutation detection, test sensitivity and novel mutations. J Med Genet. 2006;43(9):722-728. 12. Bossler AD, Richards J, George C, Godmilow L, Ganguly A. Novel mutations in ENG and ACVRL1 identified in a series of 200 individuals undergoing clinical genetic testing for hereditary hemorrhagic telangiectasia (HHT): correlation of genotype with phenotype. Hum Mutat. 2006;27(7):667675.

13. Klaus DJ, Gallione CJ, Kara A, et al. Novel missense and frameshift mutations in the activin receptor-like kinase-1 gene in hereditary hemorrhagic telangiectasia. Hum Mutat. 1998;12(2):137. 14. Guerrero-Esteo M, Sanchez-Elsner T, Letamendia A, Bernabeu C. Extracellular and cytoplasmic domains of endoglin interact with the transforming growth factorbeta receptors I and II. J Biol Chem. 2002;277(32):29197-29209. 15. Li DY, Sorensen LK, Brooke BS, et al. Defective angiogenesis in mice lacking endoglin. Science. 1999;284(5419):15341537. 16. Oh SP, Seki T, Goss KA, et al. Activin receptor-like kinase 1 modulates transforming growth factor-beta 1 signaling in the regulation of angiogenesis. Proc Natl Acad Sci USA. 2000;97(6):2626-2631. 17. Massague J. TGF-beta signal transduction. Annu Rev Biochem. 1998;67:753-791. 18. Abdalla SA, Letarte M. Hereditary haemorrhagic telangiectasia: current views on genetics and mechanisms of disease. J Med Genet. 2006;43(2):97-110. 19. Richards-Yutz J, Grant K, Chao EC, Walther SE, Ganguly A. Update on molecular diagnosis of hereditary hemorrhagic telangiectasia. Hum Genet. 2010;128(1):61-77. 20. Kuhnel T, Wirsching K, Wohlgemuth W, Chavan A, Evert K, Vielsmeier V. Hereditary Hemorrhagic Telangiectasia. Otolaryngol

1441

A. Kritharis et al. Clin North Am. 2018;51(1):237-254. 21. Albiñana V, Zafra MP, Colau J, et al. Mutation affecting the proximal promoter of Endoglin as the origin of hereditary hemorrhagic telangiectasia type 1. BMC Med Genet. 2017;18(1):20. 22. Cole SG, Begbie ME, Wallace GM, Shovlin CL. A new locus for hereditary haemorrhagic telangiectasia (HHT3) maps to chromosome 5. J Med Genet. 2005;42(7):577-582. 23. Bayrak-Toydemir P, McDonald J, Akarsu N, et al. A fourth locus for hereditary hemorrhagic telangiectasia maps to chromosome 7. Am J Med Genet A. 2006;140(20):21552162. 24. Hernandez F, Huether R, Carter L, et al. Mutations in RASA1 and GDF2 identified in patients with clinical features of hereditary hemorrhagic telangiectasia. Hum Genome Var. 2015;2:15040. 25. Shovlin CL. Hereditary haemorrhagic telangiectasia: pathophysiology, diagnosis and treatment. Blood Rev. 2010;24(6):203219. 26. Goumans M-J, Liu Z, Ten Dijke P. TGF-β signaling in vascular biology and dysfunction. Cell Res. 2009;19(1):116-127. 27. Fernández-L A, Sanz-Rodriguez F, Blanco FJ, Bernabéu C, Botella LM. Hereditary hemorrhagic telangiectasia, a vascular dysplasia affecting the TGF- signaling pathway. Clin Med Res. 2006;4(1):66-78. 28. Pomeraniec L, Hector-Greene M, Ehrlich M, Blobe GC, Henis YI. Regulation of TGFreceptor hetero-oligomerization and signaling by endoglin. Mol Biol Cell. 2015;26(17):3117-3127. 29. Scharpfenecker M, van Dinther M, Liu Z, et al. BMP-9 signals via ALK1 and inhibits bFGF-induced endothelial cell proliferation and VEGF-stimulated angiogenesis. J Cell Sci. 2007;120(6):964-972. 30. Cunha SI, Magnusson PU, Dejana E, Lampugnani MG. Deregulated TGFbeta/BMP Signaling in Vascular Malformations. Circ Res. 2017;121(8):981999. 31. Tual-Chalot S, Oh P, Arthur HM. Mouse Models of Hereditary Haemorrhagic Telangiectasia: Recent Advances and Future Challenges. Front Genet. 2015;6:25. 32. Garrido-Martín EM, Blanco FJ, Roquè M, et al. Vascular Injury Triggers Krüppel-Like Factor 6 (KLF6) Mobilization and Cooperation with Sp1 to Promote Endothelial Activation through Upregulation of the Activin Receptor-Like Kinase 1 (ALK1) Gene. Circ Res. 2012;112(1):113-127. 33. Cirulli A, Liso A, D’Ovidio F, et al. Vascular endothelial growth factor serum levels are elevated in patients with hereditary hemorrhagic telangiectasia. Acta Haematol. 2003;110(1):29-32. 34. Shao ES, Lin L, Yao Y, Bostrom KI. Expression of vascular endothelial growth factor is coordinately regulated by the activin-like kinase receptors 1 and 5 in endothelial cells. Blood. 2009;114(10):21972206. 35. Han C, Choe S-w, Kim YH, et al. VEGF neutralization can prevent and normalize arteriovenous malformations in an animal model for hereditary hemorrhagic telangiectasia 2. Angiogenesis. 2014;17(4):823-830. 36. Iannuzzi MC, Hidaka N, Boehnke M, et al. Analysis of the relationship of von Willebrand disease (vWD) and hereditary hemorrhagic telangiectasia and identification of a potential type IIA vWD mutation (IIe865 to Thr). Am J Hum Genet.

1442

1991;48(4):757-763. 37. Marchuk DA. Genetic abnormalities in hereditary hemorrhagic telangiectasia. Curr Opin Hematol. 1998;5(5):332-338. 38. Westermann CJ, Rosina AF, de Vries V, Coteau PAd. The prevalence and manifestations of hereditary hemorrhagic telangiectasia in the Afro Caribbean population of the Netherlands Antilles: A family screening. Am J of Med Genet. 2003;116(4):324-328. 39. Williams J-CB, Hamilton JK, Shiller M, Fischer L, Deprisco G, Boland CR. Combined juvenile polyposis and hereditary hemorrhagic telangiectasia. Proc (Bayl Univ Med Cent). 2012;25(4):360-364. 40. Baert A. Vascular Embolotherapy: A Comprehensive Approach, Volume 1: General Principles, Chest, Abdomen, and Great Vessels: Springer Science & Business Media, 2006. 41. Donaldson JW, McKeever TM, Hall IP, Hubbard RB, Fogarty AW. Complications and mortality in hereditary hemorrhagic telangiectasia: A population-based study. Neurology. 2015;84(18):1886-1893. 42. Sabba C, Pasculli G, Suppressa P, et al. Life expectancy in patients with hereditary haemorrhagic telangiectasia. QJM. 2006;99(5):327-334. 43. Kjeldsen AD, Vase P, Green A. [Hereditary hemorrhagic telangiectasia. A populationbased study on prevalence and mortality among Danish HHT patients]. Ugeskr Laeger. 2000;162(25):3597-3601. 44. Shovlin CL, Guttmacher AE, Buscarini E, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu Osler Weber syndrome). Am J Med Genet. 2000;91(1):6667. 45. Garg N, Khunger M, Gupta A, Kumar N. Optimal management of hereditary hemorrhagic telangiectasia. J Blood Med. 2014;5:191-206. 46. Brydon HL, Akinwunmi J, Selway R, UlHaq I. Brain abscesses associated with pulmonary arteriovenous malformations. Br J Neurosurg. 1999;13(3):265-269. 47. Cho D, Kim S, Kim M, et al. Two cases of high output heart failure caused by hereditary hemorrhagic telangiectasia. Korean Circ J. 2012;42(12):861-865. 48. Dupuis-Girod S, Ginon I, Saurin J-C, et al. Bevacizumab in patients with hereditary hemorrhagic telangiectasia and severe hepatic vascular malformations and high cardiac output. JAMA. 2012;307(9):948-955. 49. OS AA, Friedman CM, White RI Jr. The natural history of epistaxis in hereditary hemorrhagic telangiectasia. Laryngoscope. 1991;101(9):977-980. 50. Jessurun G, Kamphuis D, Van der Zande F, Nossent J. Cerebral arteriovenous malformations in the Netherlands Antilles: high prevalence of hereditary hemorrhagic telangiectasia-related single and multiple cerebral arteriovenous malformations. Clin Neurol Neurosurg. 1993;95(3):193-198. 51. Morgan T, McDonald J, Anderson C, et al. Intracranial hemorrhage in infants and children with hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu syndrome). Pediatrics. 2002;109(1):E12. 52. Press OW, Ramsey PG. Central nervous system infections associated with hereditary hemorrhagic telangiectasia. Am J Med. 1984;77(1):86-92. 53. Dong SL, Reynolds SF, Steiner IP. Brain abscess in patients with hereditary hemorrhagic telangiectasia: case report and literature review. J Emerg Med. 2001;20(3):247251.

54. Gallione CJ, Repetto GM, Legius E, et al. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4). Lancet. 2004;363(9412):852-859. 55. Jelsig AM, Torring PM, Kjeldsen AD, et al. JP-HHT phenotype in Danish patients with SMAD4 mutations. Clin Genet. 2016;90 (1):55-62. 56. Pierucci P, Lenato GM, Suppressa P, et al. A long diagnostic delay in patients with hereditary haemorrhagic telangiectasia: a questionnaire-based retrospective study. Orphanet J Rare Dis. 2012;7(1):33. 57. Harvey RJ, Kanagalingam J, Lund VJ. The impact of septodermoplasty and potassiumtitanyl-phosphate (KTP) laser therapy in the treatment of hereditary hemorrhagic telangiectasia-related epistaxis. Am J Rhinol. 2008;22(2):182-187. 58. Reh DD, Yin LX, Laaeq K, Merlo CA. A new endoscopic staging system for hereditary hemorrhagic telangiectasia. Int Forum Allergy Rhinol; 2014: Wiley Online Library; 2014. p. 635-639. 59. Kwaan HC, Silverman S. Fibrinolytic activity in lesions of hereditary hemorrhagic telangiectasia. Arch Dermatol. 1973;107(4): 571-573. 60. Watanabe M, Hanawa S, Morishima T. Fibrinolytic activity in cutaneous lesions of hereditary hemorrhagic telangiectasia. Jpn J Dermatol B. 1985;95(1):11. 61. Zaffar N, Ravichakaravarthy T, Faughnan ME, Shehata N. The use of anti-fibrinolytic agents in patients with HHT: a retrospective survey. Ann Hematol. 2015;94(1):145152. 62. Geisthoff UW, Seyfert UT, Kubler M, Bieg B, Plinkert PK, Konig J. Treatment of epistaxis in hereditary hemorrhagic telangiectasia with tranexamic acid - a double-blind placebo-controlled cross-over phase IIIB study. Thromb Res. 2014;134(3):565-571. 63. Fernandez-L A, Garrido-Martin EM, SanzRodriguez F, et al. Therapeutic action of tranexamic acid in hereditary haemorrhagic telangiectasia (HHT): Regulation ofALK1/endoglin pathway in endothelial cells. Thromb Haemost. 2007;97(2):254-262. 64. de Gussem EM, Snijder RJ, Disch FJ, Zanen P, Westermann CJ, Mager JJ. The effect of Nacetylcysteine on epistaxis and quality of life in patients with HHT: a pilot study. Rhinology. 2009;47(1):85-88. 65. Longacre AV, Gross CP, Gallitelli M, Henderson KJ, White Jr RI, Proctor DD. Diagnosis and management of gastrointestinal bleeding in patients with hereditary hemorrhagic telangiectasia. Am J Gastroenterol. 2003;98(1):59-65. 66. Nardone G, Rocco A, Balzano T, Budillon G. The efficacy of octreotide therapy in chronic bleeding due to vascular abnormalities of the gastrointestinal tract. Aliment Pharmacol Ther. 1999;13(11):1429-1436. 67. Livesey JA, Manning RA, Meek JH, et al. Low serum iron levels are associated with elevated plasma levels of coagulation factor VIII and pulmonary emboli/deep venous thromboses in replicate cohorts of patients with hereditary haemorrhagic telangiectasia. Thorax. 2012;67(4):328-333. 68. Peng HL, Yi YF, Zhou SK, Xie SS, Zhang GS. Thalidomide Effects in Patients with Hereditary Hemorrhagic Telangiectasia During Therapeutic Treatment and in FliEGFP Transgenic Zebrafish Model. Chin Med J (Engl). 2015;128(22):3050-3054. 69. Bauditz J, Schachschal G, Wedel S, Lochs H. Thalidomide for treatment of severe

haematologica | 2018; 103(9)

HHT for the hematologist

70.

71.

72.

73.

intestinal bleeding. Gut. 2004;53(4):609612. Lebrin F, Srun S, Raymond K, et al. Thalidomide stimulates vessel maturation and reduces epistaxis in individuals with hereditary hemorrhagic telangiectasia. Nat Med. 2010;16(4):420-428. Gordon M, Margolin K, Talpaz M, et al. Phase I safety and pharmacokinetic study of recombinant human anti-vascular endothelial growth factor in patients with advanced cancer. J Clin Oncol. 2001;19(3):843-850. Dupuis-Girod S, Ambrun A, Decullier E, et al. ELLIPSE Study: a Phase 1 study evaluating the tolerance of bevacizumab nasal spray in the treatment of epistaxis in hereditary hemorrhagic telangiectasia. Mabs. 2014;6(3):794-799. Kim YH, Kim MJ, Choe SW, Sprecher D, Lee Y, P Oh S. Selective effects of oral antiangiogenic tyrosine kinase inhibitors on an animal model of hereditary hemorrhagic telangiec-

haematologica | 2018; 103(9)

74.

75.

76.

77. 78.

tasia. J Thromb Haemost. 2017;15(6):10951102. Ruiz S, Chandakkar P, Zhao H, et al. Tacrolimus rescues the signaling and gene expression signature of endothelial ALK1 loss-of-function and improves HHT vascular pathology. Hum Mol Genet. 2017;26(24): 4786-4798. Samour M, Saygin, C., Abdallah, R., Kundu, S., McCrae, K.R. Pomalidomide in Hereditary Hemorrhagic Telangiectasia: Interim Results of a Phase I Study. Blood. 2016;128(22):210. Warkentin TE, Moore JC, Morgan DG. Gastrointestinal angiodysplasia and aortic stenosis. N Engl J Med. 2002;347(11):858859. Bose P, Holter JL, Selby GB. Bevacizumab in hereditary hemorrhagic telangiectasia. N Engl J Med. 2009;360(20):2143-2144. Oosting S, Nagengast W, de Vries E. More on bevacizumab in hereditary hemorrhagic

79.

80.

81.

82.

telangiectasia. N Engl J Med. 2009;361(9): 931; author reply 931-932. Brinkerhoff BT, Poetker DM, Choong NW. Long-term therapy with bevacizumab in hereditary hemorrhagic telangiectasia. N Engl J Med. 2011;364(7):688-689. Thompson AB, Ross DA, Berard P, FigueroaBodine J, Livada N, Richer SL. Very low dose bevacizumab for the treatment of epistaxis in patients with hereditary hemorrhagic telangiectasia. Allergy Rhinol (Providence). 2014;5(2):91-95. Epperla N, Kapke JT, Karafin M, Friedman KD, Foy P. Effect of systemic bevacizumab in severe hereditary hemorrhagic telangiectasia associated with bleeding. Am J Hematol. 2016;91(6):E313-314. Guilhem A, Fargeton AE, Simon AC, et al. Intra-venous bevacizumab in hereditary hemorrhagic telangiectasia (HHT): A retrospective study of 46 patients. PLoS One. 2017;12(11):e0188943.

1443

ARTICLE

Hematopoiesis

Ferrata Storti Foundation

Transient inhibition of NF-κB signaling enhances ex vivo propagation of human hematopoietic stem cells

Mehrnaz Safaee Talkhoncheh, Agatheeswaran Subramaniam, Mattias Magnusson, Praveen Kumar, Jonas Larsson and Aurélie Baudet

Haematologica 2018 Volume 103(9):1444-1450

Division of Molecular Medicine and Gene Therapy, Lund Stem Cell Center, Lund University, Sweden

ABSTRACT

espite extensive studies, defining culture conditions in which hematopoietic stem cells can be expanded ex vivo has been challenging. Here we show that chemical inhibition of the NF-κB signaling pathway leads to a significant improvement of hematopoietic stem cell function from ex vivo cultured human umbilical cord blood derived CD34+ cells. We found a distinct peak of activation of the NF-κB pathway shortly after cells were put in culture, and consequently inhibition of the pathway was both necessary and sufficient during the first 24 hours of culture where it reduced the levels of several pro-inflammatory cytokines. Taken together, NF-κB pathway inhibition facilitates propagation of hematopoietic stem cells in culture and may complement other strategies for hematopoietic stem cell expansion by relieving stress signals that are induced as an immediate response to culture initiation.

Introduction Correspondence: jonas.larsson@med.lu.se

Received: January 12, 2018. Accepted: May 17, 2018. Pre-published: June 7, 2018. doi:10.3324/haematol.2018.188466 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/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.

1444

Umbilical cord blood (CB) has emerged as a promising source of hematopoietic stem and progenitor cells (HSPCs) for transplantation. Unfortunately, the use of CB grafts is mainly restricted to pediatric transplantation as the number of HSPCs per unit is usually too low to allow the infusion of the minimal cell dose required for successful transplantation in adults.1-3 Potentially, the ex vivo expansion of CBderived HSPCs prior to transplantation could extend the use of CB transplantation to adult patients.4 Successful HSPC expansion would further facilitate the development of more advanced cell therapies for hematologic diseases, including gene therapy applications.5 Hematopoietic stem cell self-renewal is regulated by a combination of positivenegative feedback signaling.6 An incomplete understanding of this complex regulatory mechanism and how it would fit in a culture system has limited successful HSC ex vivo expansion. Despite the well-studied role of positive signals such as growth factors on HSC self-renewal, several studies highlighted the importance of inhibitory signals in restricting HSC self-renewal and function ex vivo. It has been suggested that, when under culture, HSCs undergo an immediate and transient “culture shock” that compromises their maintenance and function. This is due to upregulation of negative regulators of HSC expansion such as tumor necrosis factor (TNF signaling) and members of the aryl hydrocarbon receptor (AhR) signaling pathway.7,8 Furthermore, the secretion of inhibitory signaling factors, mainly from more differentiated cells, constitutes a major restriction to all HSC culture systems.8 To further explore this notion, our laboratory has performed high-throughput forward RNAi screens and identified negative regulators of ex vivo expansion of human HSPCs, including the cohesin family of genes, and p38 (MAPK14).9,10 Through these studies, we also identified shRNAs which display a remarkable ability to expand phenotypically-defined HSPCs in culture but whose gene target has not been validated. We reasoned that, although it would be very difficult to identify the gene targets responsible for the phenotype, the molecular characterization of cells targeted by such off-target shRNAs could provide general clues about the context under which HSPCs can be propagated ex vivo. Here, we report that the expression signature of cells showing enhanced expansion from one specific off-target shRNA (sh758) haematologica | 2018; 103(9)

NF-κB inhibition promotes human HSCs in culture

involved the downmodulation of genes involved in NF-κB signaling pathway. We found that pharmacological inhibition of the NF-κB signaling leads to a significant improvement in HSC function from ex vivo cultured CB-derived CD34+ cells, as assessed by transplantation to NSG mice. The effect of NF-κB pathway inhibition was most critical early during the culture where it reduced the levels of several pro-inflammatory cytokines induced as an immediate response to culture initiation.

Statistical analysis Statistical significance was calculated using a two-tailed Student t-test with GraphPad Prism software unless otherwise stated. Statistical significance in the figures are indicated: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Error bars indicate Standard Error of the Mean (SEM) unless otherwise stated and n represents the number of independent experiments. Methods describing microarray analysis, CFC assays, western blot and cell cycle analysis can be found in the Online Supplementary Appendix.

Methods Results shRNA experiments The RNAi screening strategy has been thoroughly described previously.9,10 The target sequence for the candidate shRNA sh758 is GATATGCAAGTCTGTGAATTT. CD34+ cells were transduced with a pLKO1-GFP lentiviral vector harboring either sh758 or control (scrambled) shRNA and subsequently cultured for several weeks according to previously described protocol.9,10

Cord blood CD34+ isolation and culture Umbilical CB samples were collected from full-term deliveries at maternity wards of Lund, Malmö and Helsingborg Hospitals. CB unit collection, mononuclear cell isolation, and CD34+ cell enrichment and culture were carried out as previously described.10 IKKβ inhibitors, PF184 and TPCA1 (Tocris Bioscience), kept in DMSO, were added at a final concentration of 400nM. Control wells were supplemented with DMSO at a matching concentration. Cultures were kept at 37°C and 5% CO2 and the medium (including inhibitors) was refreshed after four days.

Flow cytometry and cell sorting For cell surface marker staining, cells were collected, washed once with PBS supplemented with 2% FCS (FACS buffer). Cells were incubated with anti CD34 (#343516581), CD90 (#3281145E10) and EPCR (#351906) (BioLegend) antibodies for 30 minutes (min) at 4°C, and washed once with cold FACS buffer. For cell sorting, CD34+ cells were quickly thawed and stained for CD34, CD38 (#345806), CD45RA (#560362) (BD Bioscience) and CD90 following the same procedure as above. When specified, cells were stained with the Annexin V Apoptosis Detection Kit, according to the manufacturer’s protocol (BD Bioscience). All data were collected on FACS Canto II or LSRII analyzer (Becton Dickinson), and analyzed with FlowJo software. Cells were sorted on a FACS Aria II or III (Becton Dickinson).

Human engraftment assay All experiments with mice were reviewed and conducted under approved protocol from the Lund/Malmö Local Ethical Committee. NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ mice (NSG; Jackson Laboratory) were sublethally irradiated (300 cGy) before transplantation. Fresh cells or the cultured equivalent of 30,000 input CD34+ cells were injected intravenously into 10-12-week old NSG mice. Human cell contribution in peripheral blood (PB) and bone marrow (BM) of NSG was assessed 16 weeks post transplantation.

The NF-κB pathway modulates ex vivo cultured human HSPCs From RNAi-based screens conducted in our laboratory aimed at identifying novel modifiers of HSPC expansion,10,11 we have identified several off-target hits: shRNAs that display profound effects on HSPC expansion but do not affect the expression of their predicted target. One such shRNA, TRC00000758 (sh758), predicted to specifically knock-down (CLK1), induced an especially strong proliferative advantage of CB CD34+ cells, as the transduced, GFP+ population eventually dominated the culture (Figure 1A). Moreover, when cultured for seven days, CD34+CD38-CD90+CD45RA- cells expressing sh758 maintained a substantially higher percentage of immature CD34hiCD90+ cells, which is known to contain the engraftable HSCs (Figure 1B). Given the dramatic increase of CD34hiCD90+ cells induced by sh758, we reasoned that the molecular signature of cells expressing sh758 could provide valuable clues about the genes and pathways that modulate HSC expansion ex vivo. We performed global transcriptional profiling of CD34hiCD90+ cells expressing either sh758 or a control shRNA (shCTL) after seven days of culture. The gene expression profiles revealed a downmodulation of cellular stress-related genes and pathways, such as MAPK14 and genes involved in NF-κB signaling, in sh758 transduced cells (Figure 1C). Of note, the expression levels of CLK1, as well as other predicted targets of sh758 (based on sequence homology), were not repressed by the shRNA. We have previously reported on the role of MAPK14 (p38) in HSPC expansion10 and therefore decided to further investigate the role of NF-κB signaling in this context. We first addressed whether the NF-κB pathway is activated upon ex vivo culture of CB CD34+ cells and found that the key regulator of NF-κB signaling, IkBa,12 was readily detected and also showed an increased expression during culture (Figure 1D). We further measured the phosphorylated form of the IκBa protein (p-IκBa), as an indicator of NF-κB activity, at different time points during culture of CB CD34+ cells. We observed that the p-IκBa signal was detected throughout the culture and with particularly high levels during the first 24 hours (Figure 1D). Collectively, this suggests that the NF-κB pathway is activated in cultured HSPCs and may negatively influence their expansion.

Cytokine secretion and Bioplex assay Supernatants were collected from duplicate samples after six hours treatment of CB CD34+ cells with TPCA1 or STF. The secreted cytokines in the supernatants were measured by using human 27-plex panel (M500KCAF0Y, Bio-Rad) in the Bio-Rad Luminex instrument. Samples were prepared and analyzed as per the manufacturer’s protocol. haematologica | 2018; 103(9)

NF-κB pathway inhibition mediated by targeting IKKβ increases expansion of CD34+CD90+ cells ex vivo To further explore the role of NF-κB signaling in cultured HSPCs, we next blocked the pathway pharmacologically. As the down-regulated genes detected in the gene expression profiling (Figure 1C) are acting through canon1445

M.S. Talkhoncheh et al. A

ical NF-κB signaling, we decided to target IKKβ which is a central molecule in integrating upstream signals in the canonical pathway.13 Thus, we treated CD34+ cells with two specific inhibitors of the IKKβ subunit: PF184 and TPCA1.14,15 IκBa is the canonical target of IKKβ and both inhibitors induced a strong reduction of the phosphorylated-to-total IκBa ratio after four hours treatment (Figure 2A). After seven days of culture, we observed a marked increase in both the frequency and number of CD34+CD90+ cells (Figure 2B). However, the total cell number remained unaffected (Figure 2C), suggesting that the increase in CD34+CD90+ cells number is not due to a generally higher proliferation rate of cells treated with NF-κB inhibitors. Since NF-κB has well described effects on apoptosis16 and proliferation,17 we assayed cell cycle and apoptosis parameters in CD34+ cells treated with PF184 and TPCA1 in an attempt to understand the basis for the increased numbers of CD34+CD90+ cells. However, we could not detect any significant difference in cell cycle status (Figure 2D) or apoptotic cell levels (Figure 2E) of CD34+90+ cells.

NF-κB pathway inhibition enhances the ex vivo propagation of transplantable HSCs To assess in more detail the functional consequences of growing HSPCs in the presence of NF-κB inhibition, we cultured CD34+ cells for seven days with or without IKKβ inhibitors, and then assayed the cells for colony forming ability in methylcellulose medium. Cells cultured in the presence of inhibitors produced a significantly higher number of colonies compared to the control, and also gen1446

Figure 1. The sh758-dependent ex vivo expansion effect is associated with a downmodulation of the NFκB pathway. (A) Growth of GFP+ cells was assessed by FACS and normalized to day 2-transduction efficiency. (B) The percentage of CD34+90+ cells in the GFP+ fraction was quantified by FACS seven days after transduction of CD34+CD38-CD90+CD45RAcells (n=2). (C) Gene expression data mining showed modulation of members of NF-κB signaling in sh758 CD34+CD90+ expressing cells. (D) The expression of total and phosphorylated ΙκΒa was quantified by western blot in CD34+ cells at time culture indicated points. Actin expression is shown as loading control.

erated a higher proportion of primitive colonies with a mixed phenotype (Figure 3A), indicating improved functional integrity of the cultured progenitor cells. Next, to test the in vivo regenerative capacity of HSCs cultured with IKKβ inhibitors, we performed transplantation experiments. Fresh CB CD34+ (30,000) cells or 7-day cultured equivalents of 30,000 CB CD34+ were transplanted into sub-lethally irradiated NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mice (Figure 3B). Cells treated with PF184 and TPCA1 yielded 53% and 64% human engraftment, respectively, while control cells generated lower (38%) human engraftment in bone marrow 16 weeks post transplantation (Figure 3C). The engrafted cells contributed to both myeloid and lymphoid lineages (Figure 3D and E). Finally, the frequency of the HSC enriched CD34+CD38population among the engrafted human cells was similar to that of control cells (Online Supplementary Figure S1), demonstrating an overall higher long-term engraftment of also the most immature cells from cultures with PF184 and TPCA1. Although IKKβ inhibitor treated cells showed slightly lower long-term engraftment compared to noncultured cells, our data demonstrate a strong benefit of NF-κB inhibition in preserving functional HSPCs during ex vivo expansion cultures.

NF-κB inhibition is most beneficial during an early time window where it down-regulates several pro-inflammatory cytokines We had previously observed a peak in p-IκBa levels during the first 24 hours of culture (Figure 1D), and therefore next assessed if there is a critical time window for addihaematologica | 2018; 103(9)

NF-κB inhibition promotes human HSCs in culture

Figure 2. Targeting of the NF-κB regulator IKKβ increases the number of human hematopoietic stem and progenitor cells (HSPCs) in vitro without affecting cell proliferation and survival. (A) The protein level of total and phosphorylated IκBa were quantified in CD34+ cells after 4-hour treatment with STF/PF/TP in culture. Actin expression is shown as loading control (n=2). (B) FACS plots and graph show cell surface expression of CD34 and CD90 and number of CD34+CD90+ cells respectively at day 7 (n=4). (C) Graph displays total live (7AAD-) cell number at day 7 (n=4). (D) Cell cycle distribution, and (E) apoptosis (7AAD-/ AnnexinV+) of CD34+CD90+ cells were assessed by FACS at days 4 and 7, respectively (n=3). p-IκBa: phosphorylated IκBa; t-IκBa: total IκBa; PF: PF184; TP: TPCA1.

tion of IKKβ inhibitors. We found that delaying IKKβ inhibition for 24 hours abolished the beneficial effects and resulted in lower numbers of CD34+CD90+ cells after five days compared to cultures where the inhibitor was added directly. Similarly, treating cells for only the first 24 hours with IKKβ inhibitor resulted in similarly elevated numbers of CD34+CD90+ cells compared to 5-day treatment (Figure 4A). Collectively, these findings suggest that NF-κB activation is most detrimental during the first 24 hours of culture and define a time window when NF-κB inhibition is both necessary and sufficient to execute protective effects on HSPCs. In order to further characterize the potential of NF-κB inhibition in improving expansion protocols, we compared TPCA1 to two other known expansion molecules for human HSCs: SR1 and UM171.18,19 We evaluated the output of CD34+90+ cells as well as CD34+CD90+EPCR+ cells, as EPCR expression recently has been associated with the regenerative activity of ex vivo cultured HSPCs in the context of UM171 mediated expansion.20 All three compounds showed a clear increase in both CD34+CD90+ and CD34+CD90+EPCR+ cell numbers compared to control cultures (Figure 4B and C, and Online Supplementary Figure S2). However, TPCA1 treatment resulted in larger numbers of these cell populations compared to SR1, while UM171 showed a significant increase compared to both TPCA1 and SR1 (Figure 4B and C). Interestingly, the combination of the IKKβ inhibitor with either SR1 and UM171 led to a marked increase in CD34+90+ and CD34+CD90+EPCR+ cell numbers, compared to each compound alone (Figure 4B and C), indicating that targeting NF-κB signaling may enhance HSPC expansion protocols driven by other compounds. NF-κB proteins are considered key players in inflammation and immunity. However, they also play an important haematologica | 2018; 103(9)

role in other processes such as cell growth, survival and development.13 To further identify which NF-κB associated functions were targeted in our cultured HSPCs, we subjected bulk CD34+ and sorted CD34+CD90+ cells to global gene expression profiling using microarrays (Figure 4D and E, Online Supplementary Table S1 and Online Supplementary Figure S3). We used 6-hour culture time for the transcriptome analysis as this time point showed a clear increase in NF-κB activity (Figure 1D). Although relatively few transcriptional changes were detected within the more heterogeneous bulk of CD34+ cells (Online Supplementary Figure S3), 51 annotated genes were modulated in the CD34+CD90+ population upon exposure to the IKKβ inhibitor (Figure 4D and E, and Online Supplementary Table S1). Out of these, all coding genes (23 genes) were down-regulated and mainly related to inflammatory processes (15 genes), for example, IL1a, TNF-a, GM-CSF (CSF2), MCP-1 (CCL2), TCA3 (CCL1), and MIP-1β (CCL4) (Online Supplementary Table S1). Using Bioplex bead arrays, we confirmed the reduction of several proinflammatory factors, including TNF-a, IL-6, IP-10 (CXCL10) and IL-8 in supernatants collected from cultures of CB CD34+ cell after a 6-hour exposure to the IKKβ inhibitor (Figure 4F and G). Taken together, these findings indicate that the beneficial effect of IKKβ inhibitors on cultured human HSPCs is associated with the downregulation of NF-κB-dependent pro-inflammatory molecules.

Discussion In this study, we demonstrate that the NF-κB signaling has a negative impact on ex vivo cultured human CB HSPCs, and that pharmacological inhibition of the pathway improves their propagation and regenerative poten1447

M.S. Talkhoncheh et al.

Figure 3. Targeting of the NFκB regulator IKKβ improves the in vitro and in vivo function of cultured cord blood (CB)-derived hematopoietic stem and progenitor cells (HSPCs). (A) Colonies per 300 input cells were CD34+ scored after 14 days (n=3). (B) Schematic representation of in vivo transplantation experiment. (C) Human engraftment was assessed by quantifying the percentage of human CD45+ cells in the bone marrow (BM) of NSG recipients four months post transplantation (data from 2 independent experiments). (D) Lineage distribution was quantified by FACS analysis of human B cells (CD19), T cells (CD3), and myeloid (CD33/CD15) in CD45+ cells of the BM of recipients. (E) Representative FACS plots for gating strategy of lineage reconstitution. PF: PF184; TP: TPCA1.

tial after culture. Our findings further showed that this is associated with inhibition of pro-inflammatory signals in CD34+CD90+ cells during the early phases of the culture. NF-κB signaling can be activated in both immune and non-immune tissues by various extracellular signals, such as reactive oxygen species,21 pro-inflammatory cytokines such as interleukin-1,22 and members of the TNF superfamily.23 Cellular responses largely depend on the cell type, include positive or negative modulation of proliferation and apoptosis,16 differentiation,24 development,13,25 and are mediated by secretion of a large variety of signals21 and direct expression control of cell cycle/apoptosis mediators.16 In hematopoiesis, knockout mouse models of different NF-κB subunits have revealed defects in HSPCs function, as well as impaired immune response, lymphopoiesis, granulocytosis, and splenomegaly.26,27 In human settings, overexpression of p65 or a constitutively active form of IKKβ did not influence growth and differentiation of CB-derived CD34+ cells.28 However, the actual role of NF-κB signaling and the consequences of inhibiting the pathway in normal HSPCs had not previously been addressed. We initially identified the NF-κB pathway as part of a 1448

molecular signature associated with sh758, an shRNA that dramatically expands phenotypically defined HSPCs. However, even if chemical inhibition of IKKβ, the integrating kinase of NF-κB signaling, did lead to a robust and consistent increase in CD34+CD90+ cells, it could not reproduce the full extent of the sh758 phenotype. Specifically, the effect of IKKβ inhibition was limited to a short time window during the first 24 hours of culture. This suggests that the profound and persistent expansion of CD34+CD90+ cells induced by sh758 is only partly mediated by reduced activity of the NF-κB pathway, and that other genes and pathways must be involved as well. One strong candidate is MAPK14 (p38), which was also down-regulated in sh758-transduced cells. We have previously shown that p38 inhibition enhances the stem cell output from cultured HSPCs. Despite the well described function of NF-κB in regulating proliferation and apoptosis,29 we did not observe any significant changes in cell cycle, or apoptotic status of CD34+CD90+ cells upon NF-κB pathway inhibition. Additionally, NF-κB signaling has been linked to ROS production,21 which must be tightly regulated to maintain HSPC function.30 However, intracellular ROS levels were haematologica | 2018; 103(9)

NF-κB inhibition promotes human HSCs in culture

Figure 4. The benefit of NF-κB pathway inhibition is limited to a short time window, and results in downregulation of inflammatory signaling. (A) Number of CD34+CD90+ cells was assessed after five days by FACS, when TP/STF was added directly or 24 hours after culture start or removed after 24 hours (n=3). (B) Graph shows number of CD34+CD90+ cells after treatment with STF, TP, SR1, UM171 and combination of TP with either SR1 or UM171 at day 6 (representative data from 2 experiments). (C) Graph shows number of CD34+CD90+EPCR+ cells after treatment with STF, TP, SR1, UM171 and combination of TP with either SR1 or UM171 at day 6 (representative data from 2 experiments). (D and E) CD34+ cells were treated with or without TP for six hours and subjected to transcriptome analysis by microarrays after sorting for CD34+90+ population. Volcano plot and GO classification of 2-fold modulated (down-regulated) genes are shown for each condition. See also Online Supplementary Figure S3 and Online Supplementary Table S1. (F) The log ratio (TP/STF) of factors present in supernatant after six hours of culture of CB-derived CD34+ cells is shown (n=3). (G) Each graph displays individual data point for significantly modulated cytokines (IL6, TNFa, IL8, and IP10), extracted from the experiments presented in (F). The stars in (F) indicate that those factors were not detectable. PF: PF184; TP: TPCA1.

not significantly altered during the early phase of the culture (data not shown), when the effects of NF-κB pathway inhibition were seen to enhance HSPC activity. This suggests that the preserved stem cell output of cultured HSPCs upon NF-κB pathway inhibition is not related to ROS modulation. Based on the IκBa phosphorylation pattern, we showed that there is a critical 6-24-hour time frame necessary and sufficient for NF-κB pathway inhibition to benefit the propagation of the CD34+CD90+ population. In agreement with this, previous studies also reported on an immediate variation in gene expression profile of inhibitory HSPC regulators as the cultured cells experience an initial culture shock.7,8 In fact, we showed downregulation of several pro-inflammatory genes such as MCP-1 (CCL2), TCA3 (CCL1), IL1a, TNF-a and MIP-1β (CCL4) upon NF-κB inhibition, which are known to inhibit HSPC function.31 Although these pro-inflammatory cytokines are essential for a proper defense mechanism,24 they have been shown haematologica | 2018; 103(9)

to negatively influence self-renewal and lineage commitment of HSCs.32,33 Therefore, considering the lack of effect on cell proliferation, the increased HSC output after NF-κB pathway inhibition may rather be associated with preserved stem cell integrity by restricting differentiation. It is likely that the most immature HSPC are particularly sensitive to the early culture shock as they showed the most profound transcriptional changes upon NF-κB pathway inhibition. It could be argued that pro-inflammatory stress could be relieved by simply exchanging the media similar to the fed-batch system.8 However, during the critical first 24 hours, the cell density is very low and a media exchange system is not likely to have an impact during this short time frame. It is possible that restricting inhibitory signals in expansion cultures have a bi-phasic pattern; one pulse that is induced immediately upon culture initiation and one that accumulates over time and that is dependent on cell density and media changes. The latter can be relieved 1449

M.S. Talkhoncheh et al.

by fed-batch systems, but the acute effect may rather be intrinsically triggered and require pharmacological targeting, in this case NF-κB pathway inhibition, to be reduced. Our finding that targeting NF-κB signaling further enhances HSPC expansion driven by other compounds suggests that this should be further exploited in strategies aimed at HSC expansion for clinical benefit. Acknowledgments The authors are thankful to the technical staff and manage-

References 1. Brunstein CG, Gutman JA, Weisdorf DJ, et al. Allogeneic hematopoietic cell transplantation for hematologic malignancy: relative risks and benefits of double umbilical cord blood. Blood. 2010;116(22):4693-4699. 2. Wagner JE, Gluckman E. Umbilical cord blood transplantation: the first 20 years. Semin Hematol. 2010;47(1):3-12. 3. Ballen KK, Gluckman E, Broxmeyer HE. Umbilical cord blood transplantation: the first 25 years and beyond. Blood. 2013;122(4):491-498. 4. Broxmeyer HE. Enhancing the efficacy of engraftment of cord blood for hematopoietic cell transplantation. Transfus Apher Sci. 2016;54(3):364-372. 5. Farahbakhshian E, Verstegen MM, Visser TP, et al. Angiopoietin-like protein 3 promotes preservation of stemness during ex vivo expansion of murine hematopoietic stem cells. PLoS One. 2014;9(8):e105642. 6. Kirouac DC, Ito C, Csaszar E, et al. Dynamic interaction networks in a hierarchically organized tissue. Mol Syst Biol. 2010;6:417. 7. Magnusson M, Sierra MI, Sasidharan R, et al. Expansion on stromal cells preserves the undifferentiated state of human hematopoietic stem cells despite compromised reconstitution ability. PLoS One. 2013;8(1): e53912. 8. Csaszar E, Kirouac DC, Yu M, et al. Rapid expansion of human hematopoietic stem cells by automated control of inhibitory feedback signaling. Cell Stem Cell. 2012;10 (2):218-229. 9. Galeev R, Baudet A, Kumar P, et al. Genome-wide RNAi Screen Identifies Cohesin Genes as Modifiers of Renewal and Differentiation in Human HSCs. Cell Rep. 2016;14(12):2988-3000. 10. Baudet A, Karlsson C, Safaee Talkhoncheh M, et al. RNAi screen identifies MAPK14 as a druggable suppressor of human hematopoietic stem cell expansion. Blood. 2012;119(26):6255-6258. 11. Ali N, Karlsson C, Aspling M, et al. Forward

1450

12. 13. 14.

15.

16. 17. 18.

19.

20.

21. 22.

ment at the Division of Molecular Medicine and Gene Therapy, The Lund Stem Cell Center FACS and Vector cores, as well as the Animal facility at the Biomedical Center. This work was funded by grants from the Swedish Research Council, the Swedish Cancer Foundation, the Swedish Pediatric Cancer Foundation and the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement n. 648894) to JL. The work was further supported by the HematoLinné and StemTherapy programs at Lund University.

RNAi screens in primary human hematopoietic stem/progenitor cells. Blood. 2009;113(16): 3690-3695. Hayden MS, Ghosh S. Shared principles in NF-kappaB signaling. Cell. 2008;132(3):344362. Oeckinghaus A, Hayden MS, Ghosh S. Crosstalk in NF-kappaB signaling pathways. Nat Immunol. 2011;12(8):695-708. Mora E, Guglielmotti A, Biondi G, SassoneCorsi P. Bindarit: an anti-inflammatory small molecule that modulates the NFkappaB pathway. Cell Cycle. 2012;11(1):159-169. Podolin PL, Callahan JF, Bolognese BJ, et al. Attenuation of murine collagen-induced arthritis by a novel, potent, selective small molecule inhibitor of IkappaB Kinase 2, TPCA-1 (2-[(aminocarbonyl)amino]-5-(4fluorophenyl)-3-thiophenecarboxamide), occurs via reduction of proinflammatory cytokines and antigen-induced T cell Proliferation. J Pharmacol Exp Ther. 2005;312(1):373-381. Barkett M, Gilmore TD. Control of apoptosis by Rel/NF-kappaB transcription factors. Oncogene. 1999;18(49):6910-6924. Naugler WE, Karin M. NF-kappaB and cancer-identifying targets and mechanisms. Curr Opin Genet Dev. 2008;18(1):19-26. Boitano AE, Wang J, Romeo R, et al. Aryl hydrocarbon receptor antagonists promote the expansion of human hematopoietic stem cells. Science. 2010;329(5997):13451348. Fares I, Chagraoui J, Gareau Y, et al. Cord blood expansion. Pyrimidoindole derivatives are agonists of human hematopoietic stem cell self-renewal. Science. 2014;345(6203):1509-1512. Fares I, Chagraoui J, Lehnertz B, et al. EPCR expression marks UM171-expanded CD34(+) cord blood stem cells. Blood. 2017;129(25):3344-3351. Morgan MJ, Liu ZG. Crosstalk of reactive oxygen species and NF-kappaB signaling. Cell Res. 2011;21(1):103-115. Stylianou E, O'Neill LA, Rawlinson L, et al. Interleukin 1 induces NF-kappa B through its type I but not its type II receptor in lym-

23. 24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

phocytes. J Biol Chem. 1992;267(22):1583615841. Baud V, Karin M. Signal transduction by tumor necrosis factor and its relatives. Trends Cell Biol. 2001;11(9):372-377. Espin-Palazon R, Stachura DL, Campbell CA, et al. Proinflammatory signaling regulates hematopoietic stem cell emergence. Cell. 2014;159(5):1070-1085. Oeckinghaus A, Ghosh S. The NF-kappaB family of transcription factors and its regulation. Cold Spring Harb Perspect Biol. 2009;1(4):a000034. Stein SJ, Baldwin AS. Deletion of the NFkappaB subunit p65/RelA in the hematopoietic compartment leads to defects in hematopoietic stem cell function. Blood. 2013;121(25):5015-5024. Nakata S, Matsumura I, Tanaka H, et al. NFkappaB family proteins participate in multiple steps of hematopoiesis through elimination of reactive oxygen species. J Biol Chem. 2004;279(53):55578-55586. Schepers H, Eggen BJ, Schuringa JJ, Vellenga E. Constitutive activation of NF-kappa B is not sufficient to disturb normal steady-state hematopoiesis. Haematologica. 2006;91(12): 1710-1711. Radhakrishnan SK, Kamalakaran S. Proapoptotic role of NF-kappaB: implications for cancer therapy. Biochim Biophys Acta. 2006;1766(1):53-62. Ito K, Hirao A, Arai F, et al. Regulation of oxidative stress by ATM is required for selfrenewal of haematopoietic stem cells. Nature. 2004;431(7011):997-1002. Baldridge MT, King KY, Goodell MA. Inflammatory signals regulate hematopoietic stem cells. Trends Immunol. 2011;32(2): 57-65. Dybedal I, Bryder D, Fossum A, Rusten LS, Jacobsen SE. Tumor necrosis factor (TNF)mediated activation of the p55 TNF receptor negatively regulates maintenance of cycling reconstituting human hematopoietic stem cells. Blood. 2001;98(6):1782-1791. Pietras EM. Inflammation: a key regulator of hematopoietic stem cell fate in health and disease. Blood. 2017;130(15):1693-1698.

haematologica | 2018; 103(9)

ARTICLE

Bone Marrow Failure

Hematopoietic stem cell loss and hematopoietic failure in severe aplastic anemia is driven by macrophages and aberrant podoplanin expression

Ferrata Storti Foundation

Amanda McCabe,1* Julianne N.P. Smith,1** Angelica Costello,1 Jackson Maloney,1 Divya Katikaneni1 and Katherine C. MacNamara1 1

Department for Immunology and Microbial Disease, Albany Medical College, NY, USA

AM and JNPS contributed equally to this work. *Current address: Boston Children’s Hospital, Division of Hematology/Oncology, Harvard Medical School, Karp Family Research Lab, Boston, MA, USA. **Current address: Case Western Reserve University, Department of Medicine, Wolstein Research Building, Cleveland, OH, USA

Haematologica 2018 Volume 103(9):1451-1461

ABSTRACT

evere aplastic anemia (SAA) results from profound hematopoietic stem cell loss. T cells and interferon gamma (IFNγ) have long been associated with SAA, yet the underlying mechanisms driving hematopoietic stem cell loss remain unknown. Using a mouse model of SAA, we demonstrate that IFNγ-dependent hematopoietic stem cell loss required macrophages. IFNγ was necessary for bone marrow macrophage persistence, despite loss of other myeloid cells and hematopoietic stem cells. Depleting macrophages or abrogating IFNγ signaling specifically in macrophages did not impair T-cell activation or IFNγ production in the bone marrow but rescued hematopoietic stem cells and reduced mortality. Thus, macrophages are not required for induction of IFNγ in SAA and rather act as sensors of IFNγ. Macrophage depletion rescued thrombocytopenia, increased bone marrow megakaryocytes, preserved platelet-primed stem cells, and increased the platelet-repopulating capacity of transplanted hematopoietic stem cells. In addition to the hematopoietic effects, SAA induced loss of nonhematopoietic stromal populations, including podoplanin-positive stromal cells. However, a subset of podoplanin-positive macrophages was increased during disease, and blockade of podoplanin in mice was sufficient to rescue disease. Our data further our understanding of disease pathogenesis, demonstrating a novel role for macrophages as sensors of IFNγ, thus illustrating an important role for the microenvironment in the pathogenesis of SAA.

Correspondence: macnamk@amc.edu

Received: January 30, 2018. Accepted: May 14, 2018. Pre-published: May 17, 2018. doi:10.3324/haematol.2018.189449

Introduction Severe aplastic anemia (SAA) is a rare, lethal bone marrow (BM) failure disease that can be inherited or acquired. The most effective treatment for SAA is BM transplantation but disease management also includes immunosuppressive therapy (IST). Not all patients are good transplant candidates, however, and IST responsiveness varies. Therefore more specific treatments are necessary.1,2 Chemical-induced toxicity, myeloablation, and lymphocyte infusion-based BM destruction have been used to model SAA in mice and define factors critical for initiating disease.1,2 SAA can be acquired as a result of radiation, toxic drug exposure, or infection. Acquired forms are often immune-mediated,3 thus the lymphocyteinfusion model is clinically relevant. Sublethal irradiation and subsequent lymphocyte or bulk splenocyte transfer elicits pancytopenia and death within 2-3 weeks.4 Importantly, disease progression and IST treatment responses are similar to those in SAA patients.5 T cells promote hematopoietic stem cell (HSC) loss during SAA through a “bystander effect” involving inflammatory cytokines, including IFNγ.6-9 IFNγ negatively regulates HSC function, and it was first observed over thirty years ago that SAA patients show elevated IFNγ levels.10 Despite this knowledge, the haematologica | 2018; 103(9)

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1451 ©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.

1451

A. McCabe et al.

underlying mechanisms whereby IFNγ drives SAA are still unknown. Thrombocytopenia causes substantial morbidity and mortality in SAA patients.11 Megakaryocytes (Mks) not only produce platelets via thrombopoiesis, they serve as critical niches for HSCs.12,13 Thrombopoiesis is regulated by soluble factors, vascular integrity, and extracellular matrix composition, and requires adequate numbers and location of Mks.14 Platelet-biased HSCs, including HSCs that highly express CD41, have been observed in settings of inflammation and aging, where increased platelet output may be necessary to maintain vascular function.15,16 Inflammation can impact megakaryopoiesis and thrombopoiesis, though it is unclear whether this process is modulated in SAA. Bone marrow macrophages (Mfs) support stromal niche cell function at steady-state,17-19 however, little is known about the impact of inflammation on BM Mfs. Herein, we demonstrate that Mfs are essential for IFNγ-dependent HSC loss in murine SAA. IFNγ signaling in Mfs was necessary for the selective maintenance of BM resident Mfs, whereas all other myeloid cells were diminished. Targeting Mfs during SAA, via depletion or blocking their ability to respond to IFNγ, rescued HSCs and markedly improved survival. We demonstrate a key role for BM Mfs in sensing IFNγ, and our findings suggest that dysfunctional megakaryopoiesis and thrombopoiesis underlie hematopoietic collapse in SAA.

Bone-associated cell analysis Bones were crushed using a mortar and pestle in HBSS, washed and twice digested in 2 mg/mL collagenase type 2 and 1X trypsin enzyme solution at 37°C (with rocking, 30 min).

Complete blood cell count Complete blood counts (CBCs) were determined using an automated hematology analyzer (Cell Dyn 3700, Abott Laboratories).

Tissue preparation for protein quantification Bone marrow cell lysates were homogenized with a pestle in a buffer containing IGEPAL CA-630 and proteinase inhibitors for protein analysis.

Transplantation 150 HSCs (Lin– cKit+ CD48- CD150+) were sorted from PBS- and clod-lip-treated ACTB-tdTomato F1 mice 8 d.p.s.t. and transplanted separately into lethally irradiated F1 recipients with 2.5x105 protective F1 whole BM cells.

Macrophage depletion and antagonist delivery

250 μL of PBS- or clodronate-liposomes (ClodronateLiposomes. com) was administered intravenous (i.v.) at 1 d.p.s.t. (day 7 analysis) or 1 and 7 d.p.s.t. (day 15 analysis). Anti-podoplanin (PDPN) antibody (clone 8.1.1) or Syrian Hamster IgG (both from BioXcell) was administered i.v. 3, 7, and 10 d.p.s.t. at 125 μg/dose.

Platelet analysis Methods

Fluorescently-tagged anti-Gp1bβ antibodies (Emfret Analytics) were administered to mice 5 or 10 d.p.s.t., according to the manufacturer’s instructions.

Mice C57BL/6 (Hb/b) and BALB/c (Hd/d) mice were from Taconic (Albany, NY, USA). C57BL/6-TG(UBC-GFP)30Scha/J mice and ACTB-tdTomato mice were from Jackson Laboratory (Bar Harbor, ME, USA). MIIG (Mf-insensitive to IFNγ) mice20 were a gift from Dr. Michael Jordan. Hybrid B6 F1 (Hb/d) were generated by crossing C57BL/6 with BALB/c mice. To generate MIIG F1 (Hb/d) mice, MIIG (C57BL/6 background) and BALB/c mice were crossed. Hybrid F1 progeny were screened by PCR to identify mice carrying the MIIG transgene, and MIIG-negative mice were included as littermate controls (LC). Mice were bred and housed in the Animal Research Facility at Albany Medical College (AMC) under microisolator conditions. Protocols were approved by the AMC Institutional Animal Care and Use Committee.

Histology Sternums were fixed in 10% buffered formalin, decalcified in 14% EDTA, and paraffin-embedded. Megakaryocytes were stained with anti-rat GP1bβ (Emfret) and nuclear fast red (Poly Scientific R&D) counterstaining. Images were taken on an Olympus SC30 light microscope using CellSens software.

Statistical analysis Analysis was performed using Prism software. Two-tailed Student t-test was used to compare between indicated groups, unless otherwise reported.

Results SAA induction Age- (6-8 weeks) and sex-matched B6 F1 mice were sublethally irradiated (300 RADs) using a 137Cs source four hours prior to intraperitoneal (i.p.) transfer of 5x107 C57BL/6 splenocytes.4 Mice were euthanized by CO2 inhalation at the indicated day post splenocyte transfer (d.p.s.t.). For survival studies, mice were examined twice daily and humanely euthanized upon 20% loss of initial body weight or if found moribund. Surviving mice were euthanized approximately ten days after the last mouse succumbed to disease.

Cell preparation and flow cytometry Bone marrow was flushed from femurs and tibias, and spleens were homogenized. After red blood cell (RBC) lysis single-cell suspensions were plated and stained. Surface-stained cells underwent nuclear or cytoplasmic permeabilization (BD Pharmingen) prior to T-bet (4B10) and IFNγ staining, respectively. Data were collected on an LSR II (BD Biosciences) and analyzed using FlowJo software (TreeStar, Ashland, OR, USA). 1452

BM macrophages are maintained during SAA Hematopoietic stem cell loss and BM destruction are key features of SAA and are associated with cytokine production by T cells.6-8 It is unclear, however, if inflammation depletes HSCs directly or through the microenvironment. To examine resident Mfs in SAA, we used an established murine model of BM failure involving histocompatibility mismatched recipients and splenocyte infusion.4 SAA was induced via sublethal irradiation of F1 hybrids (C57BL/6 x BALB/c), followed by adoptive transfer of C57BL/6 splenocytes (Online Supplementary Figure S1A). Significant cytopenias were observed in SAA mice at 8 and 15 d.p.s.t. compared to radiation controls (Rad) (Online Supplementary Figure S1B-D). Thrombocytopenia was evident 8 d.p.s.t in control and SAA mice, and progressed in SAA mice (Online Supplementary Figure S1D). SAA-associated cytopenias coincided with BM hypocellularity (Figure 1A) and HSC loss (Online Supplementary Figure S1E-G). haematologica | 2018; 103(9)

MΦ-mediated BM failure

Monocytes (CD11b+ Ly6Chi) and Mfs (defined as CD11blo/– Mfs: F4/80+ VCAM1+ CD169+ CD11blo/– SSClo and CD11b+ Mfs: F4/80+ VCAM1+ CD169+ CD11b+Ly6Cint SSClo; Online Supplementary Figure S2), were increased by frequency 8 d.p.s.t., relative to Rad mice (Figure 1C). At 15 d.p.s.t., CD11b+ Mfs, monocytes, and neutrophils were reduced by frequency whereas CD11blo/- Mf frequencies were significantly increased (Figure 1D). Despite severe BM hypocellularity at 15 d.p.s.t., CD11blo/– Mf numbers remained stable (Figure 1F). Thus, BM CD11blo/– Mfs per-

sist despite SAA-associated cytopenias, myeloid cell loss, and HSC loss in SAA. These data suggest that CD11blo/– Mfs do not require hematopoietic input during SAA, potentially due to their long lifespan or ability to selfrenew.

IFNγ-dependent increase in BM macrophages during SAA drives HSC loss and thrombocytopenia Interferon-γ mediates SAA pathology5,9 and maintains BM Mfs during infection,21 thus we next addressed

Figure. 1. Bone marrow (BM) macrophages are maintained during aplastic anemia. (A and B) Hematoxylin and eosin-stained BM in radiation control (Rad) (top) and severe aplastic anemia (SAA) (bottom) mice on days 8 and 15. Scale bar=100μm. Frequencies (C and D) and numbers (E and F) of CD11blo/- Mfs, CD11b+ Mfs, monocytes, and neutrophils in radiation control (open bars) and SAA (filled bars) mice on days 8 and 15. Data represent one experiment repeated at least twice, n=3-6 mice/group. Mean±Standard Error of Mean is shown. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

haematologica | 2018; 103(9)

1453

A. McCabe et al.

MIIG mice exhibit increased CD41hi HSCs and megakaryocytes during SAA

whether IFNγ was necessary for preservation of Mfs during SAA. Using transgenic mice in which Mf-lineage cells are insensitive to IFNγ (referred to as MIIG mice) due to a CD68-driven dominant-negative IFNγ receptor,20 we noted improved cellularity upon induction of SAA (Figure 2A and B). MIIG and littermate control (LC) mice showed no significant difference in response to sublethal radiation; however, CD11blo/- and CD11b+ Mfs were significantly reduced in MIIG mice relative to LC counterparts 8 d.p.s.t, when HSC loss is first noted (Figure 2C). Thus, we define a novel role for IFNγ in maintaining and increasing BM Mfs during SAA. Based on our prior findings in bacterial infection, where Mfs drive IFNγ-induced HSC depletion,21 we predicted that the BM HSC pool would be preserved in MIIG mice during SAA. Indeed, HSCs were preserved in MIIG mice, relative to LCs (Figure 2D and E), demonstrating IFNγsensing by Mf-lineage cells drives HSC loss in SAA. Anemia was slightly, but significantly, ameliorated (Figure 2F and G) whereas thrombocytopenia was strikingly rescued in MIIG relative to LC controls (Figure 2H). In fact, platelet levels were higher in MIIG mice with SAA than in radiation-control mice. The robust platelet rescue suggested that IFNγ-stimulated Mfs contribute specifically to thrombocytopenia in SAA.

Inflammation-induced megakaryopoiesis reportedly relies on the emergence of a CD41hi stem-like Mk progenitor cell type (SL-MkP) within the phenotypic HSC gate.15 SL-MkPs self-renew and rapidly produce Mks and platelets, while CD41lo/int HSCs contain multi-lineage potential.15,16 CD41 expression increased robustly on HSCs in MIIG relative to LC mice at day 15 p.s.t. (Figure 3A), suggesting that IFNγ-sensing Mfs limit CD41hi HSC emergence in response to SAA-induced inflammation. MIIG and LC radiation-control mice exhibited similar numbers of CD41lo/int and CD41hi at days 8 and 15 p.s.t. (Figure 3B). In SAA conditions, however, MIIG mice exhibited significantly more CD41hi HSCs (Figure 3C), and increased CD41lo/int on day 15 p.s.t. than LC. Consistent with increased phenotypic SL-MkPs we observed increased BM Mks in MIIG SAA mice, relative to LC (Figure 3D and E). Our data indicate that IFNγ signaling in Mfs during SAA is associated with rapid loss of both CD41lo/int and CD41hi HSCs, which correlates with Mk depletion and severe thrombocytopenia. Moreover, SAAinduced mortality was significantly reduced in MIIG compared to LC mice (Figure 3F). Thus, Mfs are key sensors of IFNγ, and our data strongly suggest that Mfs drive disease and death by reducing platelet-biased CD41hi HSCs.

Figure 2. IFNγ sensing by macrophages is required for bone marrow (BM) macrophage maintenance and hematopoietic stem cell (HSC) loss in aplastic anemia. (A) Severe aplastic anemia (SAA) was induced in MIIG and littermate control (LC) F1 hybrids. (B) Hematoxylin and eosin-stained BM of LC and MIIG mice 15 days post induction. Scale bar=50 μm. (C) Frequencies and absolute numbers of CD11blo/- Mfs and CD11b+ Mfs in LC (D) and MIIG (▲) mice 8 days post induction. Shading represents ranges of each Mf population in radiation control LC and MIIG mice. (D) CD150 and CD48 expression on BM Lin- c-Kit+ (LK+) cells. Numbers represent mean HSC (LK+ CD150+ CD48–) frequency±Standard Error of Mean (SEM). (E) HSC numbers in MIIG (D) and LC (▲) mice 8 days post induction. Shading represents ranges of radiation control LC and MIIG mice. (F) Red blood cells (RBCs), (F) hemoglobin, and (H) platelets in the blood 15 days post induction. Shading represents ranges of radiation control LC and MIIG mice. Data represent one experiment repeated at least two times, n=5-7 mice/group. Mean±SEM is shown. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

1454

haematologica | 2018; 103(9)

MΦ-mediated BM failure

Clodronate-liposomes specifically deplete macrophages, increase CD41hi HSCs and platelets, and rescue survival during SAA To test the impact of Mf depletion on SAA pathogenesis, we administered clodronate-encapsulated liposomes (clod-lip) to mice one day after SAA induction. BM Mfs were significantly and specifically reduced 8 d.p.s.t. (Figure 4A). Monocytes and neutrophils are also phagocytic and may be transiently depleted; however, they were quickly replaced and no sustained depletion was observed with clod-lip. Mf depletion correlated with improved cellularity at day 15 (Figure 4B and C), increased total HSCs (Online Supplementary Figure S3A), and increased CD41hi HSCs at 8 and 15 p.s.t (Figure 4D), thus supporting the idea that Mfs negatively regulate HSCs during SAA. Macrophage-colony stimulating factor (M-CSF) is critical for tissue Mf survival and self-renewal,22,23 and similar to clod-lip administration, M-CSFR antagonism significantly increased HSCs during SAA (Online Supplementary Figure S3B). Similar to MIIG SAA mice, CD41hi HSCs correlated with increased circulating platelets, significantly increased BM Mks, and improved survival (Figure 4E-G). HSCs were increased in both MIIG SAA and Mf-depleted SAA mice, while more downstream progenitors, including shortterm HSCs and multipotent progenitors (MPPs), were more variable (Online Supplementary Figure S4A). Consistent with improved thrombocytopenia in both

models, a similar and significant increase in megakaryocyte progenitors was observed in MIIG SAA mice and Mf-depleted SAA mice (Online Supplementary Figure S4B). A similar rate of platelet removal from circulation was observed in PBS- and clod-lip-treated SAA mice (Figure 4H and I). Thus, improved platelet counts were not due to loss of consumption by Mfs. The increase in BM Mks and significantly reduced mortality in Mf-depleted mice compared to PBS-lip-treated controls demonstrate Mfs drive SAA mortality, possibly via their ability to restrict phenotypically-defined platelet-biased HSCs. Thus, HSC loss and thrombocytopenia is dependent on Mfs and Mf growth factors in SAA.

T-cell responses are not impaired in clod-lip-treated and MIIG mice Macrophages may drive HSC loss by enhancing activated T-cell infiltration into the BM; therefore, we tracked donor T cells by inducing SAA with splenocytes from UBC-GFP mice (C57BL/6 background). Expression of the T-helper 1 transcription factor T-bet, which is expressed in T cells of SAA patients and increases IFNγ gene transcription,7 was not diminished in T cells from Mf-depleted mice (Online Supplementary Figure S5A). IFNγ protein levels (Online Supplementary Figure S5B) and IFNγ-secreting donor T cells (Online Supplementary Figure S5C-E) in the BM of SAA mice were also unaffected by Mf depletion. T-

Figure 3. MIIG mice exhibit increased CD41hi hematopoietic stem cells (HSC), increased megakaryocytes in the bone marrow (BM), and reduced mortality during aplastic anemia. (A) CD41 expression on BM HSCs in radiation control (Rad) (top) and severe aplastic anemia (SAA) (bottom) MIIG and LC mice 15 days post-splenocyte transfer (d.p.s.t.). CD41 mean fluorescence intensity (MFI) on HSCs is shown on the plots and gates represent CD41lo/int and CD41hi HSCs. (B and C) CD41lo/int and CD41hi HSC numbers in MIIG (▲) versus LC (D) mice on days 8 (mean is shown) and 15 (median is shown, and a Mann-Whitney test was used to compare between groups). *P<0.05, **P<0.01. (D) Gp1bβ staining in BM of Rad and SAA MIIG (▲) and LC (r) mice 15 d.p.s.t. Scale bar=100 μm. (E) GP1bβ+ megakaryocytes per 100 mm2 of sternal BM. Mean±Standard Error of Mean (SEM) is shown. **P<0.01. (F) Kaplan-Meier survival curve for radiation control (Rad; LC and MIIG mice; □, n=4) and SAA MIIG (▲; n=9) and LC (D; n=11) mice. Log-rank (Mantel-Cox) test was used to compare between groups. *P<0.05.

haematologica | 2018; 103(9)

1455

A. McCabe et al.

bet+ donor T-cell numbers, IFNγ protein levels, and IFNγ+ donor T cells were also comparable between LC and MIIG mice in SAA (Online Supplementary Figure S5F-J). Though donor T lymphocytes are necessary for disease initiation, HSC loss during SAA occurred independently of the direct effects of T-cell-derived IFNγ. Rather, HSC loss occurred through Mf-dependent sensing of T-cell-derived IFNγ. Interferon-γ acts on Mfs to promote inflammation during disease contributing to M1 polarization of Mfs,24,25 and inflammation can impact HSC pool size and function.26,27 However, we observed similar levels of inflammatory factors previously associated with SAA, including TNFa, IL-6, and IL-1β, in the BM of MIIG and LC mice during SAA (Online Supplementary Figure S6A). Furthermore, Mf depletion did not alter TNFa, IL-6, and IL-1β levels (Online Supplementary Figure S6B), demonstrating that HSC loss is not due to induction of a broad inflammatory response during SAA. Despite the similar inflammatory milieu, purified Mfs from LC and MIIG SAA mice exhibited differential expression of genes associated with M1 and M2 polarization, programs associated with inflammatory and wound healing, respectively.28 In MIIG-derived CD11b+

Mf depletion increases functional platelet-biased HSCs To examine Mf-dependent regulation of HSC function and lineage bias in SAA, we transplanted HSCs sorted from the BM of PBS- or clod-lip-treated SAA mice 8 d.p.s.t. (Figure 5A). To our knowledge, HSC function and lineage bias have not previously been assayed in models of SAA, likely due to the severity of BM hypocellularity. HSCs exposed to Mfs during SAA showed little repopulating activity, indicating that exposure to secondary stress severely compromised their function, whereas HSCs from clod-lip-treated SAA mice exhibited platelet-, myeloid-, and lymphoid-repopulating capacity (Figure 5B). Thus our data demonstrate that in SAA, Mfs reduce HSC function and impair platelet output.

Mfs, we observed a reduction in inducible nitric oxide synthesis (Nos2) expression, whereas induction of arginase 1 was seen in CD11blo/- Mfs (Online Supplementary Figure S6C). Elevated Nos2 expression and increased nitric oxide concentrations are associated with disease in SAA patients,29 thus IFNγ signaling in Mfs may drive pathological M1 polarization in SAA.

Figure 4. Clodronate-liposomes specifically deplete macrophages and rescue hematopoietic stem cell numbers during aplastic anemia. (A) Myeloid bone marrow (BM) cell numbers in PBS- (○) or clodronate-loaded (Clod; ●) liposome-treated severe aplastic anemia (SAA) mice. (B) BM cellularity in PBS- (○) or clodronate-loaded (Clod; ●) liposome-treated SAA mice on days 8 and 15 post-splenocyte transfer (p.s.t.) is shown. (C) Hematoxylin and eosin-stained BM from PBS- or clod-lip-treated radiation control (top) and SAA (bottom) mice 15 days post splenocyte transfer (d.p.s.t.). Scale bar=50 μm. (D) CD41lo/int and CD41hi HSC numbers 8 and 15 days d.p.s.t. *P<0.05. (E) Platelets in the blood 15 d.p.s.t. ***P<0.001. (F) GP1bβ+ megakaryocytes per 100mm2 of BM 15 d.p.s.t. Mean±Standard Error of Mean is shown. *P<0.05. (G) Kaplan-Meier survival curve for mice with SAA treated with PBS- (○; n=8) or clod-lip (●; n=9) 1 and 7 d.p.s.t. Log-rank (Mantel-Cox) test was used to compare between groups. **P<0.01. (H) Schematic showing administration of anti-Gp1bβ antibodies to mice 5 or 10 d.p.s.t. (I) Labeled platelets were measured in the blood over time. Two-way ANOVA was used to compare between groups. P<0.0001 and P=0.0007 for SAA-PBS versus radiation control (Rad) at days 5-9 p.s.t. and days 10-12 p.s.t., respectively; P<0.0001 for SAA-Clod versus Rad at days 5-9 p.s.t. and days 10-12 p.s.t.; P=0.004 and P=0.7 for SAA-PBS versus SAA-Clod at days 5-9 p.s.t. and days 10-12 p.s.t., respectively.

1456

haematologica | 2018; 103(9)

MΦ-mediated BM failure

Aberrant podoplanin expression during SAA drives HSC loss, thrombocytopenia, and death Macrophages negatively regulated both HSCs and Mks, and we questioned whether Mfs also regulate nonhematopoietic BM stromal cells, known to be impaired in SAA patients.30-32 In contrast to radiation alone, SAA reduced osteoblastic and endothelial cells (Online Supplementary Figure S7A and B). Because SAA was associated with severe thrombocytopenia we examined expression of podoplanin (PDPN), recently identified in the BM and shown to increase platelet production.33 We noted a striking loss in PDPN+ stromal cells in SAA (Online Supplementary Figure S7C and D). At the same time, however, we observed induction of PDPN on hematopoietic cells that appeared to be entirely restricted to Mfs and a majority were CD11blo/- Mfs (Figure 6A and B, and Online Supplementary Figure S8A). We found no change in PDPN expression among other hematopoietic or nonhematopoietic stromal cells (data not shown). In addition, we observed increased podoplanin (gp38) transcripts specifically in the CD11blo/- Mfs, relative to CD11b+ Mfs, T cells, and neutrophils in SAA mice (Figure 6C). PDPN+ Mfs were reduced in MIIG mice relative to controls, though PDPN MFI was unchanged (Online Supplementary

Figure S8B and C). This suggests that IFNγ increased numbers of PDPN+BM Mfs, rather than directly regulating PDPN expression, during SAA. To determine if aberrant PDPN expression on Mfs mediated pathology during SAA, we administered an antiPDPN monoclonal antibody during SAA. PDPN blockade significantly increased CD41lo/int and CD41hi HSCs and resulted in a preservation of BM cellularity compared to isotype control treatment (Figure 7A and B). Administration of an anti-PDPN antibody did not rescue HSCs by reducing or impairing T-cell activation because similar numbers of T-bet+ CD4 and CD8 T cells and IFNγ levels were observed in the BM of anti-PDPN and controltreated mice during SAA (Online Supplementary Figure S9A and B). Consistent with improved HSC numbers PDPN blockade rescued thrombocytopenia, increased BM Mks, and increased survival in SAA (Figure 7C-E). Anti-PDPN antibody conferred significant protection, whereas isotype-control antibody-treated mice had a median survival of only 16.5 days and died between days 12 and 19 induction of SAA. PDPN-CLEC-2 interaction was reported to induce RANTES,33 which can support platelet production, thus our finding that PDPN blockade improved thrombocy-

Figure 5. Macrophage depletion increases plateletbiased hematopoietic stem cells (HSC) in severe aplastic anemia (SAA). (A) HSC function in SAA was assessed by transplantation of HSCs from PBS- (™) or clod-lip (Clod;˜)-treated TdTomato+ F1 mice 8 days post-splenocyte transfer (d.p.s.t.). (B) Peripheral blood was analyzed for reconstitution at indicated time points. P<0.0001 for platelets, P=0.002 for myeloid, P=0.06 for lymphoid. Each transplantation data set represents one experiment, n=4-5 recipient mice per group. Two-way ANOVA was used to compare between groups.

haematologica | 2018; 103(9)

1457

A. McCabe et al. A

Figure 6. Macrophages exhibit increased expression of podoplanin (PDPN) during severe aplastic anemia (SAA). (A) PDPN expression in bone marrow (BM) cells (top) 8 days post-splenocyte transfer (d.p.s.t.). CD11b expression among PDPN+ F4/80+ cells (bottom). (B) PDPN+ F4/80+ MΦ numbers in healthy (●), radiation control (■), and SAA (+Rad +Splenocytes; □) mice 8 d.p.s.t. (C) Gp38 expression in sort-purified BM populations, relative to β-actin and normalized to expression in neutrophils. Data represent data pooled from 3 independent experiments n=5-10 mice/group.

topenia was somewhat surprising. Consistent with a role for PDPN in RANTES production we observed increased RANTES during SAA (Online Supplementary Figure S9C), where we also observe increased PDPN+ Mfs. PDPN blockade did not impact BM RANTES in SAA, likely because the antibody clone (8.1.1) does not interfere with CLEC-2 binding in vivo or in vitro (Online Supplementary Figure S9D-F).34 Thus, PDPN-dependent HSC loss and hematopoietic failure occurs via a RANTES-independent mechanism. Administration of anti-PDPN antibody induced a specific decrease in CD11blo/- Mfs whereas CD11b+ Mf numbers were not significantly different (Figure 7F). This suggests PDPN signaling may be important for CD11blo/- Mf survival during SAA. It also demonstrates that selective reduction of CD11blo/- Mfs is associated with improved survival during SAA. PDPN can bind and activate ezrin, radixin, and moesin family proteins to promote cytoskeletal reorganization and contractility of fibroblastic reticular cells in lymph nodes.34 Microenvironmental stiffness can reduce physical support for HSCs and Mks,35 and we observed reduced expression of a-smooth muscle actin (aSMA), a marker of contractile stress fibers,36 by PDPN+ BM Mfs at day 8 p.s.t. upon anti-PDPN treatment (Online Supplementary Figure S10A and B). We also noted a striking increase in expression of arginase-1, a marker of M2-polarized Mfs, in both CD11blo/- and CD11b+ Mfs upon antiPDPN treatment (Online Supplementary Figure S10C). Thus, CD11blo/- Mfs aberrantly express PDPN in the BM during SAA, correlating with hematopoietic failure. Future studies are warranted to determine the precise impact of PDPN-expressing Mfs on the microenvironment and whether stiffness and Mf-activation state contribute to SAA pathology.

Discussion Hematopoietic stem cell loss and BM destruction are key features of SAA, and are associated with cytokine production by T cells.6-8 It is still unclear, however, if inflammation depletes HSCs directly or does so through the microenvironment. Findings from SAA patient BM suggest that stromal support of hematopoietic cells is signifi1458

cantly reduced.30-32,37,38 In a mouse model of SAA, we observed reduced stromal cells, but, at the same time, BM Mfs were maintained. CD11blo Mfs exhibited a unique survival advantage in SAA that correlated with their expression of PDPN. Consistent with our findings, SAA patient BM exhibited CD169+ Mfs persistence despite significant reductions in nearly all other hematopoietic cell types.31 Our findings reveal that, rather than direct IFNγmediated HSC depletion, IFNγ signaling in Mfs promotes HSC loss during SAA. IFNγ and Mfs limit CD41hi HSCs during disease, thus contributing to severe thrombocytopenia and mortality in SAA (Figure 7G). To the best of our knowledge, this is the first in vivo study addressing the mechanistic role of the BM microenvironment in HSC loss and disease progression during SAA. HSCs reportedly undergo apoptosis during SAA,39 yet studies in models of infection suggest that excessive differentiation and reduced self-renewal contribute to IFNγdependent HSC depletion.40,41 We previously identified an IFNγ-dependent increase in monopoiesis during Ehrlichia muris infection, which occurred at the expense of HSCs.4244 Monocytes are increased early in SAA, prior to their ultimate loss, supporting the idea that increased IFNγ-driven HSC differentiation contributes to HSC loss. It is also possible that increased apoptosis in SAA is a product of enhanced differentiating divisions that render HSCs more susceptible to inflammatory stress and/or cell death. Aberrant immune cell function, specifically T-cell activation and homing to the BM, is associated with SAA.7 Since IFNγ primes Mfs for activation,25 and Mfs produce cytokines and present antigen to T cells, we predicted that IFNγ signaling in Mfs increase T-cell activation. SAA progression is mitigated in MIIG mice, however, despite similar numbers of activated and IFNγ-secreting donor T cells in the BM. Cytokines associated with SAA (TNFa, IL-1β, and IFNγ) were also similarly induced. Thus, resident Mfs do not appear to drive disease through their capacity to present antigen to T cells or general inflammatory disposition. Mf polarization can contribute to disease through exaggerated inflammation and wound healing responses.28 During SAA, differential expression of M1-associated Nos2 was observed between the MIIG model and antiPDPN treatment and between CD11b+ and CD11blo/- Mfs, indicating functional differences between these two Mf haematologica | 2018; 103(9)

MΦ-mediated BM failure

Figure 7. Podoplanin (PDPN) antagonism rescues hematopoietic stem cells (HSC), circulating platelet levels, and mortality. (A) CD41lo/int. and CD41hi HSC numbers in anti-PDPN (■; a-PDPN) or isotype control (□)-treated mice 8 and 14 days post-splenocyte transfer (d.p.s.t.). (B) Hematoxylin and eosin-stained bone marrow (BM) from isotype control- or a-PDPN-treated mice 14 d.p.s.t. Scale bar=50 μm. (C) Platelets in the blood of isotype control- (□) and α-PDPN (■)-treated mice 14 d.p.s.t. (D) GP1bβ+ megakaryocytes per 100mm2 of BM 14 d.p.s.t. Data represent one experiment repeated at least twice, n=4-18 mice/group. Mean±Standard Error of Mean is shown. **P<0.01, ***P<0.001. (E) Kaplan-Meier survival curve for a-PDPN- (■) or isotype control (□)-treated severe aplastic anemia (SAA) mice. Data represent one experiment, n=8 mice/group. Log-rank (Mantel-Cox) test was used to compare between groups. (F) CD11blo/- and CD11b+ MΦ populations were enumerated 8 d.p.s.t. (G) Schematic summarizing the steady state role(s) for MΦs in the BM (left) and the impact of IFNγ on the BM microenvironment and resulting HSC loss in SAA (right): increased PDPN-expressing MΦs that drive reduced Mks, impaired platelet production, and correlate with reduced stromal.

populations. Of note, we observed increased expression of M2-associated arginase1 in CD11blo/- Mfs from MIIG mice and after anti-PDPN treatment. Mechanistically, a role for M1 activation in SAA pathogenesis may differ between the MIIG and anti-PDPN models; however, enhanced M2 activity correlates with protection in SAA. Macrophages participate in immune-mediated thrombocytopenia (ITP), where increased platelet clearance drives thrombocytopenia and reduced platelet production.45 Mf clearance of platelets does not appear to drive thrombocytopenia in SAA, however, as clearance rates were similar in Mf−depleted and control mice. Our findings are consistent with and add to previous reports of Mf−dependent impairment of megakaryopoiesis and haematologica | 2018; 103(9)

platelet production at steady state.46,47 Additionally, macrophage-colony stimulating factor M-CSF, a factor that increases Mf self-renewal, transiently causes thrombocytopenia.47 Our study builds upon these findings and identify Mfs as key sensors or target cells of IFNγ. Acute inflammation increases CD41hi HSCs or SLMkPs15 and we observed the emergence of a CD41hi HSC population in SAA that is accompanied by increased Mk numbers and platelet recovery when Mfs are depleted or unresponsive to IFNγ. While CD41hi HSC-derived Mks may support sustained platelet production during SAA, another intriguing possibility is that Mk-lineage cells protect HSCs through a positive feedback loop. Mk-derived factors promote HSC quiescence and protect 1459

A. McCabe et al. HSC from myeloablative injury.12,13 Therefore, CD41hi HSCs may represent more committed progenitors that augment HSC-protective niches during SAA. It is currently unclear if Mk preservation is necessary for or predictive of HSC rescue. However, our finding that HSC loss precedes Mk loss would argue in favor of a decline in megakaryopoiesis as a result of reduced CD41hi HSCs. We identify a unique population of PDPN-expressing Mfs that restrict the HSC compartment and contribute to thrombocytopenia during SAA. PDPN regulates contractility and migration of lymphatic endothelial cells, FRCs, and tumor cells.48 Thus PDPN signaling in Mfs may influence hematopoiesis in a similar manner by regulating BM stiffness or migration within BM niches. Because increased or decreased matrix stiffness impairs proplatelet extension by Mks in vitro, it is possible that SAA-induced stromal stiffness restricts thrombopoiesis.45 PDPN expression is low and confined mainly to stromal cells in the BM at steady state and even upon radiation injury. PDPN is specifically increased on Mfs in the BM during SAA, and anti-PDPN antibody specifically reduced CD11blo/- Mfs, but not CD11b+ Mfs, during SAA. Our data demonstrate a direct correlation between CD11blo/- Mfs and SAA pathogenesis. Future studies to test the impact of PDPN expression on Mf migration and survival during SAA are warranted. Podoplanin-expressing Mfs are increased in SAA, consistent with previous reports that PDPN is expressed on inflammatory Mfs in response to IFNγ during infection.49 PDPN is a CLEC-2 ligand that triggers downstream signaling in CLEC2-expressing cells. However CLEC-2/PDPN ligation also elicits bidirectional signaling, eliciting RANTES production from PDPN-expressing cells.33,34 We noted increased RANTES in SAA, relative to radiation controls. Thus, our observation that PDPN blockade did

References 1. Chen J. Animal models for acquired bone marrow failure syndromes. Clin Med Res. 2005;3(2):102-108. 2. Scheinberg P, Chen J. Aplastic anemia: what have we learned from animal models and from the clinic. Semin Hematol. 2013; 50(2):156-164. 3. Young NS, Maciejewski J. The pathophysiology of acquired aplastic anemia. N Engl J Med. 1997;336(19):1365-1372. 4. Roderick JE, Gonzalez-Perez G, Kuksin CA, et al. Therapeutic targeting of NOTCH signaling ameliorates immune-mediated bone marrow failure of aplastic anemia. J Exp Med. 2013;210(7):1311-1329. 5. Bloom ML, Wolk AG, Simon-Stoos KL, Bard JS, Chen J, Young NS. A mouse model of lymphocyte infusion-induced bone marrow failure. Exp Hematol. 2004; 32(12):1163-1172. 6. Lu J, Basu A, Melenhorst JJ, Young NS, Brown KE. Analysis of T-cell repertoire in hepatitis-associated aplastic anemia. Blood. 2004;103(12):4588-4593. 7. Solomou EE, Keyvanfar K, Young NS. T-bet, a Th1 transcription factor, is up-regulated in T cells from patients with aplastic anemia. Blood. 2006;107(10):3983-3991. 8. Zoumbos NC, Ferris WO, Hsu SM, et al.

1460

10.

11.

12.

13.

14.

not impact SAA-induced RANTES was somewhat surprising. However, our data and a previous report demonstrate that clone 8.1.1 does not interfere with CLEC-2 binding.34 Thus, PDPN blockade with clone 8.1.1 during SAA likely does not interfere with CLEC-2-driven RANTES production. Clone 8.1.1 may interfere with CLEC-2 binding,50 though very high concentrations of antibody were needed, and it is unlikely that this can be achieved in vivo. Our observations demonstrate protection independently of CLEC-2; however, future studies are necessary to test the involvement of the CLEC-2-PDPN axis and define the precise action of PDPN in the BM during SAA. Interferon-γ impacts SAA disease progression, yet neutralization of IFNγ is not currently a treatment option during this disease. Our finding that IFNγ maintains Mfs provides the rationale for targeting Mfs during BM failure. We reveal that Mfs errantly express PDPN during SAA and antagonizing PDPN signaling rescues HSCs and enhances platelet output, thus revealing a novel circuit in the microenvironment during BM failure. Understanding the mechanisms whereby PDPN expression in Mfs regulates HSC function and platelet production may reveal novel treatment options for SAA. Acknowledgments We would like to acknowledge Kathleen Curran and Candace Ross at the New York State Department of Health Wadsworth Center (Albany, NY) for running CBCs on blood samples. We would like to thank Dr. Livingston Van De Water for helpful discussion. Funding This work was supported by R01 GM105949 to KCM, an Aplastic Anemia and MDS International Foundation grant to KCM, and BM160071 (DOD-BMFRP-IDA) to KCM.

Analysis of lymphocyte subsets in patients with aplastic anaemia. Br J Haematol. 1984;58(1):95-105. Chen J, Lipovsky K, Ellison FM, Calado RT, Young NS. Bystander destruction of hematopoietic progenitor and stem cells in a mouse model of infusion-induced bone marrow failure. Blood. 2004;104(6):16711678. Zoumbos NC, Gascon P, Djeu JY, Young NS. Interferon is a mediator of hematopoietic suppression in aplastic anemia in vitro and possibly in vivo. Proc Natl Acad Sci USA. 1985;82(1):188-192. Townsley DM, Desmond R, Dunbar CE, Young NS. Pathophysiology and management of thrombocytopenia in bone marrow failure: possible clinical applications of TPO receptor agonists in aplastic anemia and myelodysplastic syndromes. Int J Hematol. 2013;98(1):48-55. Zhao M, Perry JM, Marshall H, et al. Megakaryocytes maintain homeostatic quiescence and promote post-injury regeneration of hematopoietic stem cells. Nat Med. 2014;20(11):1321-1326. Bruns I, Lucas D, Pinho S, et al. Megakaryocytes regulate hematopoietic stem cell quiescence through CXCL4 secretion. Nat Med. 2014;20(11):1315-1320. Avecilla ST, Hattori K, Heissig B, et al.

15.

16.

17.

18.

19.

Chemokine-mediated interaction of hematopoietic progenitors with the bone marrow vascular niche is required for thrombopoiesis. Nat Med. 2004;10(1):6471. Haas S, Hansson J, Klimmeck D, et al. Inflammation-Induced Emergency Megakaryopoiesis Driven by Hematopoietic Stem Cell-like Megakaryocyte Progenitors. Cell Stem Cell. 2015;17(4):422-434. Gekas C, Graf T. CD41 expression marks myeloid-biased adult hematopoietic stem cells and increases with age. Blood. 2013;121(22):4463-4472. Chang MK, Raggatt LJ, Alexander KA, et al. Osteal tissue macrophages are intercalated throughout human and mouse bone lining tissues and regulate osteoblast function in vitro and in vivo. J Immunol. 2008;181(2): 1232-1244. 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. 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.

haematologica | 2018; 103(9)

MÎŚ-mediated BM failure 20. Lykens JE, Terrell CE, Zoller EE, et al. Mice with a selective impairment of IFN-gamma signaling in macrophage lineage cells demonstrate the critical role of IFN-gammaactivated macrophages for the control of protozoan parasitic infections in vivo. J Immunol. 2010;184(2):877-885. 21. McCabe A, Zhang Y, Thai V, Jones M, Jordan MB, MacNamara KC. MacrophageLineage Cells Negatively Regulate the Hematopoietic Stem Cell Pool in Response to Interferon Gamma at Steady State and During Infection. Stem Cells. 2015;33(7):2294-2305. 22. 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. 23. Hume DA, MacDonald KP. Therapeutic applications of macrophage colony-stimulating factor-1 (CSF-1) and antagonists of CSF-1 receptor (CSF-1R) signaling. Blood. 2012;119(8):1810-1820. 24. Fujiwara N, Kobayashi K. Macrophages in inflammation. Curr Drug Targets Inflamm Allergy. 2005;4(3):281-286. 25. Mosser DM. The many faces of macrophage activation. J Leukoc Biol. 2003;73(2):209-212. 26. King KY, Goodell MA. Inflammatory modulation of HSCs: viewing the HSC as a foundation for the immune response. Nat Rev Immunol. 2011;11(10):685-692. 27. Schuettpelz LG, Link DC. Regulation of hematopoietic stem cell activity by inflammation. Front Immunol. 2013;4:204. 28. Zhou D, Huang C, Lin Z, et al. Macrophage polarization and function with emphasis on the evolving roles of coordinated regulation of cellular signaling pathways. Cell Signal. 2014;26(2):192-197. 29. Chung IJ, Lee JJ, Nam CE, et al. Increased inducible nitric oxide synthase expression and nitric oxide concentration in patients with aplastic anemia. Ann Hematol. 2003;82(2):104-108. 30. Chatterjee S, Dutta RK, Basak P, et al. Alteration in marrow stromal microenvironment and apoptosis mechanisms involved in aplastic anemia: an animal model to study the possible disease pathology. Stem Cells

haematologica | 2018; 103(9)

Int. 2010;2010:932354. 31. Park M, Park CJ, Jang S, et al. Reduced expression of osteonectin and increased natural killer cells may contribute to the pathophysiology of aplastic anemia. Appl Immunohistochem Mol Morphol. 2015;23(2):139-145. 32. Wu L, Mo W, Zhang Y, et al. Impairment of hematopoietic stem cell niches in patients with aplastic anemia. Int J Hematol. 2015;102(6):645-653. 33. Tamura S, Suzuki-Inoue K, Tsukiji N, et al. Podoplanin-positive periarteriolar stromal cells promote megakaryocyte growth and proplatelet formation in mice by CLEC-2. Blood. 2016;127(13):1701-1710. 34. Astarita JL, Cremasco V, Fu J, et al. The CLEC-2-podoplanin axis controls the contractility of fibroblastic reticular cells and lymph node microarchitecture. Nat Immunol. 2015;16(1):75-84. 35. Shin JW, Swift J, Spinler KR, Discher DE. Myosin-II inhibition and soft 2D matrix maximize multinucleation and cellular projections typical of platelet-producing megakaryocytes. Proc Natl Acad Sci USA. 2011;108(28):11458-11463. 36. Talele NP, Fradette J, Davies JE, Kapus A, Hinz B. Expression of alpha-Smooth Muscle Actin Determines the Fate of Mesenchymal Stromal Cells. Stem Cell Reports. 2015;4(6):1016-1030. 37. Chen J, Brandt JS, Ellison FM, Calado RT, Young NS. Defective stromal cell function in a mouse model of infusion-induced bone marrow failure. Exp Hematol. 2005;33(8):901-908. 38. Juneja HS, Lee S, Gardner FH. Human longterm bone marrow cultures in aplastic anemia. Int J Cell Cloning. 1989;7(2):129-135. 39. Philpott NJ, Scopes J, Marsh JC, GordonSmith EC, Gibson FM. Increased apoptosis in aplastic anemia bone marrow progenitor cells: possible pathophysiologic significance. Exp Hematol. 1995;23(14):1642-1648. 40. Matatall KA, Jeong M, Chen S, et al. Chronic Infection Depletes Hematopoietic Stem Cells through Stress-Induced Terminal Differentiation. Cell Rep. 2016;17(10):25842595. 41. de Bruin AM, Demirel O, Hooibrink B,

42.

43.

44.

45.

46.

47.

48.

49.

50.

Brandts CH, Nolte MA. Interferon-gamma impairs proliferation of hematopoietic stem cells in mice. Blood. 2013;121(18):35783585. MacNamara KC, Oduro K, Martin O, et al. Infection-induced myelopoiesis during intracellular bacterial infection is critically dependent upon IFN-gamma signaling. J Immunol. 2011;186(2):1032-1043. Zhang Y, Jones M, McCabe A, Winslow GM, Avram D, MacNamara KC. MyD88 signaling in CD4 T cells promotes IFNgamma production and hematopoietic progenitor cell expansion in response to intracellular bacterial infection. J Immunol. 2013;190(9):4725-4735. MacNamara KC, Jones M, Martin O, Winslow GM. Transient activation of hematopoietic stem and progenitor cells by IFNgamma during acute bacterial infection. PLoS One. 2011;6(12):e28669. Ballem PJ, Segal GM, Stratton JR, Gernsheimer T, Adamson JW, Slichter SJ. Mechanisms of thrombocytopenia in chronic autoimmune thrombocytopenic purpura. Evidence of both impaired platelet production and increased platelet clearance. J Clin Invest. 1987;80(1):33-40. Alves-Rosa F, Vermeulen M, Cabrera J, et al. Macrophage depletion following liposomalencapsulated clodronate (LIP-CLOD) injection enhances megakaryocytopoietic and thrombopoietic activities in mice. Br J Haematol. 2003;121(1):130-138. Baker GR, Levin J. Transient thrombocytopenia produced by administration of macrophage colony-stimulating factor: investigations of the mechanism. Blood. 1998;91(1):89-99. Astarita JL, Acton SE, Turley SJ. Podoplanin: emerging functions in development, the immune system, and cancer. Front Immunol. 2012;3:283. Hitchcock JR, Cook CN, Bobat S, et al. Inflammation drives thrombosis after Salmonella infection via CLEC-2 on platelets. J Clin Invest. 2015;125(12):44294446. Rayes J, Lax S, Wichaiyo S, et al. The podoplanin-CLEC-2 axis inhibits inflammation in sepsis. Nat Commun. 2017;8(1):2239.

1461

ARTICLE

Myelodysplastic Syndromes

Ferrata Storti Foundation

Haematologica 2018 Volume 103(9):1462-1471

Transforming growth factor β1-mediated functional inhibition of mesenchymal stromal cells in myelodysplastic syndromes and acute myeloid leukemia

Stefanie Geyh,1* Manuel Rodríguez-Paredes,1,2* Paul Jäger,1 Annemarie Koch,1 Felix Bormann,2 Julian Gutekunst,2 Christoph Zilkens,3 Ulrich Germing,1 Guido Kobbe,1 Frank Lyko,2 Rainer Haas1 and Thomas Schroeder1

Department of Hematology, Oncology and Clinical Immunology, University of Duesseldorf, Medical Faculty; 2Division of Epigenetics, DKFZ-ZMBH Alliance, German Cancer Research Center, Heidelberg and 3Department of Orthopedic Surgery, University of Duesseldorf, Medical Faculty, Germany 1

*SG and MR-P contributed equally to this work.

ABSTRACT

Correspondence: thomas.schroeder@med.uni-duesseldorf.de

Received: December 19, 2017. Accepted: May 14, 2018. Pre-published: May 17, 2018.

doi:10.3324/haematol.2017.186734 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1462

esenchymal stromal cells are involved in the pathogenesis of myelodysplastic syndromes and acute myeloid leukemia, but the underlying mechanisms are incompletely understood. To further characterize the pathological phenotype we performed RNA sequencing of mesenchymal stromal cells from patients with myelodysplastic syndromes and acute myeloid leukemia and found a specific molecular signature of genes commonly deregulated in these disorders. Pathway analysis showed a strong enrichment of genes related to osteogenesis, senescence, inflammation and inhibitory cytokines, thereby reflecting the structural and functional deficits of mesenchymal stromal cells in myelodysplastic syndromes and acute myeloid leukemia on a molecular level. Further analysis identified transforming growth factor β1 as the most probable extrinsic trigger factor for this altered gene expression. Following exposure to transforming growth factor β1, healthy mesenchymal stromal cells developed functional deficits and adopted a phenotype reminiscent of that observed in patient-derived stromal cells. These suppressive effects of transforming growth factor β1 on stromal cell functionality were abrogated by SD-208, an established inhibitor of transforming growth factor β receptor signaling. Blockade of transforming growth factor β signaling by SD-208 also restored the osteogenic differentiation capacity of patient-derived stromal cells, thus confirming the role of transforming growth factor β1 in the bone marrow microenvironment of patients with myelodysplastic syndromes and acute myeloid leukemia. Our findings establish transforming growth factor β1 as a relevant trigger causing functional inhibition of mesenchymal stromal cells in myelodysplastic syndromes and acute myeloid leukemia and identify SD-208 as a candidate to revert these effects.

©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.

1462

Introduction Myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) are blood stem cell disorders that are characterized by hematopoietic insufficiency causing relevant morbidity and mortality. For a long time both entities have been considered to be hematopoietic-cell autonomous diseases originating from the accumulation of genetic and epigenetic alterations within the hematopoietic stem and progenitor cell (HSPC) population.1,2 Recently, it has become apparent that these myeloid disorders do not exclusively arise from HSPC but also involve the haematologica | 2018; 103(9)

TGFβ1-mediated inhibition of MDS- and AML-derived MSC

entire bone marrow microenvironment and, in particular, mesenchymal stromal cells (MSC).3,4 In this context, it was demonstrated in several mouse models that genetic perturbation of mesenchymal cells can induce MDS and promote the emergence of AML.5,6 Vice versa, malignant myeloid cells can also act on niche elements such as MSC leading to a self-reinforcing mechanism supporting leukemic cells at the expense of normal hematopoiesis.7-9 Investigating the pathogenic role of MSC in humans, we and others have recently shown that primary MSC from the bone marrow of MDS and AML patients are structurally, genetically and functionally altered.4,10-15 For instance, we found that these MSC have impaired growth capacity and a decreased ability to undergo osteogenic differentiation, accompanied by a specific methylation signature. Along with impaired stromal support of healthy HSPC this suggests that MSC alterations contribute substantially to the inadequate hematopoiesis in MDS and AML.12,13 Furthermore, these findings imply significantly deregulated bi-directional crosstalk between malignant myeloid cells and MSC. In the current analysis, we used discovery-based strategies, such as RNA sequencing, followed by candidate-testing approaches. We identified transforming growth factor β (TGFβ) signaling as a relevant mechanism responsible for the functional inhibition of MSC in MDS and AML and showed that this was pharmacologically reversible upon TGFβ blockage.

Methods

with the gene sets contained in the Molecular Signature Database or with self-made gene lists. The Signal2Noise metric and 1000 gene set-based permutations were applied to all analyses. Ingenuity pathway analysis software (Qiagen, Hilden, Germany) was used to predict the potential extrinsic factors capable of generating the different RNA sequencing data sets.

Cell culture conditions and reagents Healthy MSC were cultured in Dulbecco modified Eagle medium low glucose supplemented with 30% fetal bovine serum and 1% penicillin/streptomycin/L-glutamine (all from Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) in the presence of TGFβ1 (10 ng/mL, PeproTech, Rocky Hill, NJ, USA) and/or SD-208 (0.5 μM, Tocris/R&D Systems, MN, USA), an established small molecule pyridopyrimidine inhibitor of the TGFβ receptor I kinase.21-23 The inhibitor was diluted in dimethyl sulfoxide (10 mM). The concentration of TGFβ1 used in our experiments is within the range of TGFβ1 levels previously detected in patients with MDS24,25 and has demonstrated inhibitory effects on healthy and MDS-derived hematopoietic cells.26-28 Furthermore, it was previously shown that TGFβ1 at this concentration mediates pathophysiologically relevant effects in other myeloid malignancies such as AML, chronic myeloid leukemia and myeloproliferative syndromes.9,29 Depending on the experimental setting, incubation time ranged from 3 to 28 days corresponding to the duration of the culture of MSC investigated by RNA sequencing. Detailed information is given in the legends of the respective figures. Subsequently, to investigate the effects of TGFβ1 on the functionality of cells, pre-incubated MSC underwent the phenotypic and functional analyses described below.

Patients, healthy controls and cell preparation Bone marrow samples were obtained from a total of 28 patients with newly-diagnosed MDS or AML (median age 59 years; range, 25-89 years) and 16 age- and sex-matched healthy controls (median age 68 years; range, 39-86 years; P=0.14). Details of the patients’ characteristics are given in Online Supplementary Table S1. The study was approved by our local institutional review board (approval number: 4777) and all individuals gave written informed consent to participation in it. MSC were derived from the mononuclear cell fraction of the specimens and cultured as described previously.12,13 All experiments were carried out using MSC derived from passages 3–4. Furthermore, CD34+ HSPC were obtained from the bone marrow of healthy controls by density gradient separation and subsequent immunomagnetic selection (Miltenyi Biotec, Bergisch Gladbach, Germany) as described elsewhere.16

RNA sequencing Transcriptome sequencing libraries were prepared from isolated total RNA using the TruSeq RNA sample preparation kit (Illumina, San Diego, CA, USA), and single-read 50 bp sequencing was performed in a HiSeq-2000 device (Illumina). Reads were then trimmed by removing stretches of bases with a quality score <30 at their ends, and subsequently mapped using Tophat2.0.6 against the hg19 assembly of the human genome.17 Finally, differential expression was quantified using DESeq2 and Cuffdiff 2.0, and subjected to diverse testing corrections.18,19 Genes with a q-value <0.05 were considered differentially expressed. Principal component analysis plots and heat maps were created in R using the FactoMineR and pheatmap packages, respectively. For gene set enrichment analysis (GSEA, Broad Institute, Boston, MA, USA),20 the fragments per kilobase of transcript per million mapped reads (FPKM) values obtained from Cuffdiff 2.0 for the different samples were compared either haematologica | 2018; 103(9)

Phenotypic and functional characterization of mesenchymal stromal cells The morphology and growth properties of primary and preincubated MSC were characterized by light microscopy. For quantification of growth potential, absolute cell numbers and cumulative population doubling were determined as described elsewhere.12,13 Furthermore, surface expression of established stromal cell markers CD73, CD90 and CD105,30 hematopoietic antigens CD34 and CD45 as well as Jagged1 were measured by flow cytometry using a FACSCalibur (BD Biosciences, Heidelberg, Germany) and FCS Express V3 software (De Novo Software, Los Angeles, CA, USA) for data analysis. Antibody specifications are provided in Online Supplementary Table S2. Osteogenic differentiation capacity was investigated and visualized by Alizarin red staining as reported elsewhere.12,13 In addition, mRNA expression of markers of osteogenesis, osterix and osteocalcin, and other candidate genes was measured by quantitative polymerase chain reaction (PCR) on a StepOne Plus Realtime PCR Cycler (Applied Biosystems, Life Technologies, Carlsbad, CA, USA) as described before.12,13

Long-term culture-initiating cell assay After pre-incubation with TGFβ1 or its antagonist, SD-208, 0.8x106-1.2x106 MSC were cultured on 96-well plates (Costar, Corning, USA) and irradiated with 30 Gray using Gulmay RS225 X-ray equipment. Subsequently, 6x103 healthy CD34+ cells were plated on these MSC feeder layers and then further processed using the same conditions and reagents as in our previous work.12,13

Statistical analysis Statistical analyses were performed using Prism 5.01 (GraphPad Software Inc., La Jolla, USA). For inter-individual 1463

S. Geyh et al.

comparisons the two-sided unpaired Student t-test was employed, while for intra-individual analyses the Wilcoxon signed rank test was used. For all experiments means and the standard error of mean (SEM) are given. Statistical significance was established at Pâ&#x2030;¤0.05.

Data access RNA sequencing expression data have been stored in the Gene Expression Omnibus database (GSE107490).

Results RNA sequencing analysis of myelodysplastic syndrome- and acute myeloid leukemia-derived mesenchymal stromal cells reveals a specific gene expression profile reflecting their phenotypic and functional deficits We have previously shown that MSC from MDS and AML patients display common functional deficits and

DNA methylation changes.12,13 In order to determine whether these similarities extend to the gene expression level and contribute to the altered phenotype, we performed RNA sequencing. Sequencing data were obtained for MSC from three healthy donors (all male, median age 68 years) and nine patients with myeloid malignancies [all male, median age 60 years, P=1.0; 3 MDS patients (n. 4-6) with refractory cytopenia with multilineage dysplasia (RCMD), 3 MDS patients (n. 9-11) with refractory anemia with excess blasts (RAEB), and 3 AML patients (n. 9-11)]. Data analysis revealed 1673 genes that were significantly (q<0.05) differentially expressed, relative to controls, in the MSC from the three subentities (Online Supplementary Figure S1, Online Supplementary Table S3). GSEA confirmed a significant overlap of the deregulated genes with those involved in osteogenesis or cellular senescence, two processes that we previously found to be deregulated in MDS- and AML-related MSC (Figure 1A and Online Supplementary Tables S4 and S5).12,13 Pairwise comparisons between healthy samples and each of the different disease

Figure 1. RNA sequencing analysis of mesenchymal stromal cells from patients with myeloid malignancies and from healthy controls. (A) GSEA plots showing the specific deregulation of gene sets associated with osteogenesis and cellular senescence in MSC from patients with MDS and AML. For both plots, the normalized enriched score (NES), false discovery rate (FDR) and P-values are given. (B) Venn diagram representing the number of genes differentially expressed (q<0.05) in MSC from patients with RCMD, RAEB or AML with respect to healthy MSC, as well as those genes deregulated in common in the three different myeloid entities. (C) Gene ontology analysis of the 112 genes deregulated in common in RCMD-, RAEB- and AML-derived MSC. (D) Examples of GSEA plots demonstrating an inflammatory reaction in MSC from patients with MDS or AML. For both plots, NES, FDR and P-values are given. (E) Heat map illustrating the expression changes of the set of cytokines found to be statistically significantly (q<0.05) deregulated in the MSC from MDS and AML patients. A total of 173 cytokines involved in the regulation of hematopoiesis, some of which previously described in MDS,10 were analyzed. FC: fold-change in expression.

1464

haematologica | 2018; 103(9)

TGFβ1-mediated inhibition of MDS- and AML-derived MSC

conditions defined a set of 112 deregulated genes (q<0.05) in common, mostly related to general developmental processes and osteogenesis (Figure 1B,C, Online Supplementary Figure S2, and Online Supplementary Tables S3 and S6). Further analysis revealed a remarkable abundance of gene signatures associated with inflammatory responses (Figure 1D) as well as altered expression of a variety of cytokines (Figure 1E and Online Supplementary Figure S3). Overall, the expression of 18 cytokines was significantly altered (11 upregulated, 7 downregulated) with several of them known to be involved in the regulation of hematopoiesis. Collectively, these results are consistent with the malignant phenotypes previously reported in MDS- and AML-related MSC, and further suggest a related pathomechanism for the MSC from the three entities.

RNA sequencing analysis suggests that transforming growth factor β1 signaling is a common cause of abnormal gene expression patterns in myelodysplastic syndrome- and acute myeloid leukemia-derived mesenchymal stromal cells Our previous work suggested that the phenotypic abnormalities observed in AML-derived MSC may be triggered by an extrinsic factor.13 We hypothesized that the conserved aberrant gene expression patterns are caused by an extrinsic factor. To identify this molecule, we used ingenuity pathway analysis,29 which enables the prediction of upstream regulators for a given RNA sequencing data set. This computational approach identified the inhibitory cytokines tumor necrosis factor a (TNFa) and TGFβ1 as the two most probable (P values of 2.17x10-11 or lower) secreted factors for the induction of

the aberrant gene expression patterns in the MSC derived from RCMD, RAEB and AML patients (Figure 2A). These two molecules also appeared among the most probable secreted factors in analyses of RNA sequencing data sets from pairwise comparisons (Online Supplementary Figure S4). To analyze whether only one of these two molecules could be sufficient to induce the gene expression deregulation observed in the MSC from the patients, we used GSEA. The results showed that our gene set was significantly enriched in the TGFβ1 signature but not in the TNFa signature (Figure 2B). Our RNA sequencing experiments therefore suggest that increased TGFβ1 signaling in the bone marrow may lead to the aberrant gene expression patterns observed in the MSC from these myeloid malignancies and, ultimately, to their functional inhibition.

Transforming growth factor β1 induces functional deficits in healthy mesenchymal stromal cells recapitulating the phenotype of these cells in myelodysplastic syndrome and acute myeloid leukemia

Having identified TGFβ1 as a candidate factor for the induction of the observed phenotypic alterations and functional deficits of MSC in MDS and AML, we experimentally addressed this hypothesis using an in vitro culture system. For this purpose, we cultured healthy MSC in the presence or absence of TGFβ1 to investigate whether healthy MSC adopt a phenotype that is similar to the phenotype of patient-derived MSC. Furthermore, in this experimental system we also used SD-208, which specifically abrogates the signaling downstream of the TGFβ receptor (Online Supplementary Figure S5).

haematologica | 2018; 103(9)

Figure 2. RNA sequencing identified transforming growth factor β1 as the most probable secreted upstream regulator inducing the aberrant gene expression patterns. (A) Ingenuity pathway analysis29 predicted that TNFα and TGFβ1 are the most probable secreted upstream regulators of the gene expression aberrations observed in the MSC from MDS and AML patients. (B) GSEA confirmed that overactivation of TGFβ1 signaling, but not TNFa, in the bone marrow caused the aberrant gene expression patterns. For both plots, the normalized enriched score (NES), false discovery rate (FDR) and P-values are given.

1465

S. Geyh et al.

As a first read-out, we quantified the proliferative potential and the growth capacities of MSC. TGFβ1 significantly inhibited the proliferation and growth properties of healthy MSC, as indicated by the absolute cell number and cumulative population doubling of the cells treated in this way being similar to those observed in primary patient-derived MSC (Figure 3A,B and Online Supplementary Figure S6). This suppressive effect also persisted when healthy MSC were primed with TGFβ1 for 3 days and subsequently cultured in control media for another 3 days (Figure 3A,B). In line with these findings, microscopic analysis of healthy MSC exposed to TGFβ1 revealed a disorganized cellular architecture, which differed sharply from the fibroblastoid feeder layer of the control MSC (Figure 3C). The inhibitory effects of TGFβ1 on growth and proliferation of MSC were abrogated by the TGFβ receptor I inhibitor SD-208 (Figure 3A-C). In subsequent experiments we investigated the influence of TGFβ1 on osteogenic differentiation, which has been previously shown to be significantly impaired in MDS- and AML-derived MSC.12,13 Using established culture conditions and staining methods12,13 we found that TGFβ1 significantly impaired osteogenic differentiation capacity (Figure 4A-B, Online Supplementary Figure S7). In

agreement with this result, mRNA expression of the bone formation marker osteocalcin was significantly reduced after exposure to TGFβ1 (Figure 4D). Similar to its effect on MSC growth, SD-208 also abolished the TGFβ1-mediated suppression of osteogenic differentiation (Figure 4AC). RNA sequencing of healthy MSC after incubation with TGFβ1 revealed a specific gene expression pattern substantially overlapping with the expression profile of primary patient-derived MSC (Figure 4E, Online Supplementary Figure S8 and Online Supplementary Table S7). In this regard, we decided to check whether the expression of PITX2, HOXB6 and TBX15, three genes physiologically involved in cell differentiation and skeletal morphogenesis, and which we had previously linked to the impaired osteogenesis in MDS and AML,12,13 could also be affected in the TGFβ1-treated healthy MSC. Although only PITX2 showed a statistically significant change in the RNA sequencing experiment (fold change=3.57, q<0.001), the dysregulation of the three genes could be further verified by quantitative real-time PCR (Online Supplementary Figure S9). SD-208 again reverted the TGFβ1-induced gene expression changes, including those for the three marker genes. (Figure 4F and Online Supplementary Figure S9).

Figure 3.Transforming growth factor β1 impairs growth properties of healthy mesenchymal stromal cells. (A) Bar charts showing the absolute MSC numbers after 3 days pre-incubation of 2x105 MSC with the respective factors (10 ng/mL). HC: healthy control; DMSO: dimethylsulfoxide; w/o TGFβ1: healthy MSC were treated with TGFβ1 for 3 days, then the medium was changed and the MSC were cultured for 3 additional days without TGFβ1. (B) Bar charts depicting cumulative population doubling (CPD). (C) Representative micrographs showing the morphology of healthy MSC after incubation in the presence of the respective compounds. Scale bars represent 100 μm. For all experiments results are expressed as mean ± SEM. Asterisks indicate P-values **P<0.01, ***P<0.001.

1466

haematologica | 2018; 103(9)

TGFβ1-mediated inhibition of MDS- and AML-derived MSC

Transforming growth factor β1 impairs stromal hematopoietic support function We have previously described the deregulation of Jagged1, Angiopoietin-1 and Kit-ligand, three signaling molecules physiologically involved in the regulation of HSPC, in patient-derived MSC.12,13 We now examined whether TGFβ1 can also induce alterations of these factors and of three candidate cytokines (CCL26, SPP1 and LIF) that were shown to be deregulated in primary patient-derived MSC. Indeed, quantitative PCR and flow cytometry detected an expression pattern (Figure 5A,B, Online Supplementary Figures S10 and S11) that is congruent with the expression previously detected in primary MDS- and AML-derived MSC.12,13 Again, these effects were reversible by adding SD-208 (Figure 5A,B). In light of the profound molecular alterations that were induced by TGFβ1 in healthy MSC, we were further interested in whether TGFβ1 also suppressed stromal hematopoietic support. We therefore used the established long-term culture-initiating cell (LTC-IC) assay and cultured healthy CD34+ HSPC on human MSC feeder layers which had previously been incubated with TGFβ1 or with TGFβ1 together with SD-208 for 28 days. As indicated in Figure 5C, TGFβ1 significantly inhibited the ability of healthy MSC to support CD34+ HSPC in the LTCIC assay, an effect that could be reversed by inhibiting TGFβ1 signaling through SD-208. In summary, TGFβ1

induces a phenotype and functional deficits recapitulating those observed in primary MSC obtained from patients with MDS or AML. Our data thus strongly support the hypothesis derived from RNA sequencing data analysis, namely that TGFβ1 contributes significantly to the functional inhibition of MSC in these two myeloid malignancies.

Inhibition of transforming growth factor β1 signaling restores osteogenic differentiation and hematopoietic support capacity of myelodysplastic syndromeand acute myeloid leukemia-derived mesenchymal stromal cells

In a final set of experiments we directly addressed the impact of TGFβ1 on MSC functionality in human MDS and AML. To abrogate the potential influence of TGFβ1 on MSC functions we again used SD-208. This inhibitor had already been demonstrated to abolish TGFβ signaling in the above-mentioned experiments, as well as previously in MDS-derived CD34+ cells.12,13 Given the substantial inhibition of osteogenic differentiation in primary patient-derived MSC, we focused on the potential effects of TGFβ1 on osteogenesis. For this purpose we cultured MSC, freshly isolated from the bone marrow of patients with MDS and AML, in the presence of SD-208 and documented that osteogenic differentiation capacity could be significantly rescued compared to that occurring in the

Figure 4. Transforming growth factor β1 suppresses the osteogenic differentiation capacity of healthy mesenchymal stromal cells and induces a specific gene expression profile. Healthy MSC (n=6) were pre-incubated with TGFβ1 and/or SD-208 for up to 28 days. Medium was changed every 3 days and supplemented with TGFβ1 at a concentration ranging from 5 ng/mL to 10 ng/mL and/or SD-208 (0.25 μM to 0.5 μM). Subsequently, osteogenic differentiation was induced for 14 days and visualized by Alizarin red staining as described previously.12,13 (A) Overview of a representative experiment. (B) Representative micrographs of healthy MSC after exposure to the respective factors with scale bars indicating 100 μm. (C) For the purpose of quantification, osteogenic differentiation capacity was graded according to microscopic analysis of staining intensity as follows: 0 = absent; 1 = weak; 2 = moderate; 3 = intensive as previously described.12 (D) Messenger RNA expression of osteocalcin was measured by quantitative real-time PCR analysis of healthy MSC (n=5) after 3 days of incubation. HC: healthy control; DMSO: dimethylsulfoxide; w/o TGFβ1: healthy MSC were treated with TGFβ1 for 3 days, then the medium was changed and the MSC were cultured for 3 additional days without TGFβ1. For all experiments results are expressed as mean ± SEM. Asterisks indicate P-values *P<0.05, **P<0.01. (E) After a 28-day incubation period healthy MSC (n=2) were subjected to RNA sequencing analysis. The Venn diagram illustrates a substantial overlap of 312 genes deregulated in both the TGFβ1-treated MSC and the patientderived MSC. (F) Bar charts demonstrate that most of the differential gene expression provoked by TGFβ1 in healthy MSC (1812 genes, q<0.05) is abrogated by SD208 (only 58 genes remained differentially expressed, q<0.05).

haematologica | 2018; 103(9)

1467

S. Geyh et al.

control condition of patient-derived MSC cultured in the absence of SD-208 (Figure 6A-C). In addition, treatment of MDS- and AML-derived MSC with SD-208 substantially increased these cells’ support of healthy CD34+ cells, as reflected by their 2.1-fold higher LTC-IC frequency in comparison to that of untreated MDS- and AMLderived MSC (Online Supplementary Figure S12, P=0.078). These findings clearly indicate that TGFβ1 inhibits osteogenic differentiation and hematopoietic support capacity of MSC in human MDS and AML and that this effect can be pharmacologically reverted by SD-208.

Discussion In our previous work we demonstrated that MSC from patients with MDS and AML exhibit impaired growth and osteogenic differentiation capacity as well as altered expression of hematopoietic signaling molecules such as Angiopoietin-1, Kit-ligand and Jagged1. Comparable with recent results from Bhagat et al.,31 these structural and functional deficits were associated with a specific methylation pattern and resulted in impaired stromal support for HSPC.12,13 These findings were compatible with confirmatory results published by other groups in the last

years.4,11,14,15,32,33 and directly linked alterations of MSC to the pathophysiology of MDS and AML. To elucidate the underlying molecular mechanisms, in this study we performed RNA sequencing of MSC from patients with early MDS, advanced MDS and AML. Our data analysis defined a specific common molecular profile which clearly separated MSC of these myeloid diseases from those of healthy controls. Pathway analyses further identified an enrichment of genes involved in general developmental processes, cellular senescence and in particular genes essential for skeletal morphogenesis, including the three candidate genes PITX2, HOXB6 and TBX15 that are commonly deregulated in MDS- and AMLderived MSC.12,13 On the molecular level these results provide an excellent explanation for the phenotypic and functional alterations of MDS- and AML-derived MSC that were previously reported by our group and by others.11-14 Additional findings suggested an ongoing stromal response to an inflammatory environment. This is in line with the results from two other groups who also reported an inflammatory stress response in MSC derived from patients with low-risk MDS.10,14 Our data now expand this finding to MSC from patients with advanced MDS and AML, and suggest that the pro-inflammatory signal-

Figure 5. Impaired hematopoietic support capacities of healthy mesenchymal stromal cells exposed to transforming growth factor β1. Healthy MSC (n=5) were pre-incubated with TGFβ1 and/or SD-208 for up to 28 days. Medium was changed every 3 days and supplemented with TGFβ1 and/or SD-208 at a concentration ranging from 5 ng/mL to 10 ng/mL (SD-208: from 0.25 μM to 0.5 μM). Cells were harvested on days 7, 14, 21, and 28. (A) Bar charts illustrate protein expression of Jagged1 as measured by flow cytometry at the given time points. The mean fluorescence intensity43 is shown. HC: healthy control; DMSO: dimethylsulfoxide. (B) mRNA expression of Angiopoietin-1 (Angpt1) and Kit-ligand (Kitlg) was measured by quantitative real-time PCR after 7 days of incubation. (C) Bar charts showing LTCIC frequencies of healthy CD34+ HSPC cultured on healthy MSC which had been previously incubated with TGFβ1 and/or SD-208 for 28 days. For all experiments results are expressed as mean ± SEM. Asterisks indicate P-values *P<0.05, **P<0.01, ***P<0.001.

1468

haematologica | 2018; 103(9)

TGFβ1-mediated inhibition of MDS- and AML-derived MSC

ing within the bone marrow microenvironment influences the stromal compartment in addition to HSPC, thereby contributing to insufficient hematopoiesis. Ingenuity pathway analysis and GSEA predicted that the inhibitory cytokine TGFβ1 (but not TGFβ2, data not shown) is an upstream regulator of the observed functional and molecular alterations in MDS- and AML-derived MSC. Again, this confirms and expands the results from Chen et al., who recently identified signatures related to TGFβ signaling in primary MSC from patients with lowrisk MDS.10 This molecular similarity between our culture-expanded MSC and the non-expanded, FACS-sorted MSC reported by Chen et al. also indicates that the most relevant pathophysiological mechanisms seem to be well preserved even after culture. Apart from its pathophysiological relevance this also has important methodological implications. The use of non-expanded FACS-sorted MSC does not allow subsequent direct functional testing of candidate genes because of the low number of cells, but also requires culture or cell lines.10 In contrast, we were able to test the effects of TGFβ1 on stromal cell functionality in the same culture system previously used to isolate and characterize MDS- and AML-derived MSC. This methodological issue should ideally be further investigated by a direct comparison of matched FACS-sorted,

unexpanded and culture-propagated MSC from the same individual patients. We did not detect any relevant effects on MSC functions after exposure to TNFa (data not shown). Considering the limited clinical efficacy of TNFa blockade by etanercept and infliximab in MDS patients,34-36 we believe that TNFa plays, at most, a minor role. In contrast, exposure to TGFβ1 suppressed growth and osteogenic differentiation of healthy MSC and induced changes in the expression of candidate genes (PITX2, HOXB6, TBX15, Kit-ligand, Angiopoietin-1, and Jagged1). Together with impaired stromal hematopoietic support, the phenotype of healthy MSC therefore resembles that of primary MDS- and AML-derived MSC after exposure to TGFβ1.12,13 The inhibitory effect of TGFβ1 on normal HSPC is well documented.37 In addition, overactivation of TGFβ signaling in CD34+ HSPC has been shown to be an important mechanism mediating hematopoietic insufficiency in MDS.23,38,39 Growth differentiation factor-11, another ligand of the TGFβ superfamily, inhibits maturation of erythroid precursors and elevated levels of this factor have been found in patients with MDS, implicating it in the development of anemia.40 In addition to suppression of HSPC and erythroid progenitors, our data from RNA sequencing and functional experiments show for

C Figure 6. Blockage of transforming growth factor β1 signaling by SD-208 restores the osteogenic differentiation capacity of myelodysplastic syndrome- and acute myeloid leukemia-derived mesenchymal stromal cells. Osteogenic differentiation of MDS- and AMLderived MSC was induced and visualized by Alizarin red staining. SD-208 (0.5 μM) was added to the osteogenic medium during this procedure. (A) Overview of a representative experiment investigating MDS-derived MSC (n=5, MDS patients n. 1, 2, 7, 12, 13). Representative micrographs of MDS-derived MSC after exposure to SD-208 with scale bars indicating 100 μm. (B) Overview of a representative experiment investigating AML-derived MSC (n=4, AML patients n. 1, 5, 12, 13). Representative micrographs of AML-derived MSC after exposure to SD-208 with scale bars indicating 100 μm. (C) Differences in osteogenic differentiation between patientderived MSC (black bars, MDS n=5; AML n=4) either native or treated with SD-208 (white bar) were quantified by a previously described score.12 In detail, differentiation was graded according to microscopic analysis of staining intensity as follows: 0 = absent; 1 = weak; 2 = moderate; 3 = intensive. For all experiments results are expressed as mean ± SEM. Asterisks indicate Pvalues **P<0.01, ***P<0.001.

haematologica | 2018; 103(9)

1469

S. Geyh et al.

the first time that activated TGFβ signaling contributes significantly to ineffective hematopoiesis in MDS and AML via functional inhibition of MSC. Envisaging targeted stromal therapies, it is of particular interest that we could pharmacologically revert the suppressive effects of TGFβ1 on healthy MSC in vitro using SD-208. This small molecule pyridopyrimidine inhibitor of TGFβ receptor I abrogates the signaling downstream of TGFβ receptor I, independently of the specific ligand. SD208 has been shown to stimulate hematopoiesis in a mouse model of bone marrow failure as well as when MDS HSPC were cultured in vitro.23 In agreement with these findings, SD-208 was also able to restore the osteogenic differentiation capacity and, with a clear trend, the hematopoietic support capacity of MDS- and AMLderived MSC in our analysis. Targeting TGFβ signaling to improve hematopoiesis is the aim of two ligand-trapping approaches with luspatercept and sotatercept, currently under clinical investigation in MDS.40,41 Of particular interest, the effects of the latter are not mediated via CD34+ HSPC but rather via the stromal compartment,41 thus further supporting the role of TGFβ-mediated stromal inhibition in MDS and AML. At this point it remains unclear whether TGFβ1-mediated functional inhibition of MSC in MDS and AML is a cell-intrinsic mechanism or is extrinsically mediated by other TGFβ1-producing cells within the bone marrow microenvironment. While Chen et al. recently detected overexpression of TGFβ1 in FACS-sorted unexpanded MSC of patients exclusively with low-risk MDS, we did not find overexpression of TGFβ1 in cultureexpanded MSC derived from patients with early or advanced MDS or AML (data not shown). On the other hand, we did find mRNA overexpression of TGFβ1 in CD34+ HSPC of patients with MDS or AML by performing quantitative PCR as well as by analyzing data from

References 1. Shastri A, Will B, Steidl U, Verma A. Stem and progenitor cell alterations in myelodysplastic syndromes. Blood. 2017;129(12): 1586-1594. 2. Thomas D, Majeti R. Biology and relevance of human acute myeloid leukemia stem cells. Blood. 2017;129(12):1577-1585. 3. Medyouf H. The microenvironment in human myeloid malignancies: emerging concepts and therapeutic implications. Blood. 2017;129(12):1617-1626. 4. Schroeder T, Geyh S, Germing U, Haas R. Mesenchymal stromal cells in myeloid malignancies. Blood Res. 2016;51(4):225232. 5. Kode A, Manavalan JS, Mosialou I, et al. Leukaemogenesis induced by an activating beta-catenin mutation in osteoblasts. Nature. 2014;506(7487):240-244. 6. Raaijmakers MH, Mukherjee S, Guo S, et al. Bone progenitor dysfunction induces myelodysplasia and secondary leukaemia. Nature. 2010;464(7290):852-857. 7. Frisch BJ, Ashton JM, Xing L, et al. Functional inhibition of osteoblastic cells in an in vivo mouse model of myeloid leukemia. Blood. 2012;119(2):540-550.

1470

173 AML samples in comparison to 967 samples from 28 different healthy tissues included in The Cancer Genome Atlas database (Online Supplementary Figures S13 and S14). This suggests that TGFβ1 released from CD34+ HSPC may contribute, at least in part, to the inhibition of MSC in MDS and AML. In further support of an extrinsic mechanism, it was recently shown that myeloid-derived suppressor cells also overproduce TGFβ1 and thereby suppress hematopoiesis in MDS.42 Overall, these data imply that different cells may be involved as sources of TGFβ1 and TGFβ1-related inhibition of hematopoiesis in MDS and AML. Indeed, since there may even be variations between early and advanced MDS as well as between MDS and AML this issue needs to be addressed in future studies. In conclusion, our RNA sequencing analysis revealed a specific molecular signature of MDS- and AML-derived MSC which reflected and explained these cells’ structural and functional deficits. By subsequent functional testing we confirmed that TGFβ1 is a relevant trigger for the molecular alterations and functional inhibition of MSC in MDS and AML. Our results also further support targeting of this signal pathway as a promising treatment approach. Acknowledgments This work was supported by a grant from the Deutsche Forschungsgemeinschaft (SCHR 1470/1-1), the Research Committee of the Medical Faculty, Heinrich-Heine-University, Duesseldorf, Germany and the Leukämie Lymphom Liga e. V., Duesseldorf, Germany to TS. We would like to thank Dennis Sohn from the Laboratory of Molecular Radio-oncology, Clinic and Polyclinic for Radiation Therapy and Radio-oncology for technical assistance regarding LTC-IC assays.

8. Hanoun M, Zhang D, Mizoguchi T, et al. Acute myelogenous leukemia-induced sympathetic neuropathy promotes malignancy in an altered hematopoietic stem cell niche. Cell Stem Cell. 2014;15(3):365-375. 9. Schepers K, Pietras EM, Reynaud D, et al. Myeloproliferative neoplasia remodels the endosteal bone marrow niche into a selfreinforcing leukemic niche. Cell Stem Cell. 2013;13(3):285-299. 10. Chen S, Zambetti NA, Bindels EM, et al. Massive parallel RNA sequencing of highly purified mesenchymal elements in low-risk MDS reveals tissue-context-dependent activation of inflammatory programs. Leukemia. 2016;30(9):1938-1942. 11. Ferrer RA, Wobus M, List C, et al. Mesenchymal stromal cells from patients with myelodyplastic syndrome display distinct functional alterations that are modulated by lenalidomide. Haematologica. 2013;98(11):1677-1685. 12. Geyh S, Oz S, Cadeddu RP, et al. Insufficient stromal support in MDS results from molecular and functional deficits of mesenchymal stromal cells. Leukemia. 2013;27(9):18411851. 13. Geyh S, Rodriguez-Paredes M, Jager P, et al. Functional inhibition of mesenchymal stromal cells in acute myeloid leukemia.

Leukemia. 2016;30(3):683-691. 14. Medyouf H, Mossner M, Jann JC, et al. Myelodysplastic cells in patients reprogram mesenchymal stromal cells to establish a transplantable stem cell niche disease unit. Cell Stem Cell. 2014;14(6):824-837. 15. von der Heide EK, Neumann M, Vosberg S, et al. Molecular alterations in bone marrow mesenchymal stromal cells derived from acute myeloid leukemia patients. Leukemia. 2017;31(5):1069-1078. 16. Schroeder T, Czibere A, Zohren F, et al. Meningioma 1 gene is differentially expressed in CD34 positive cells from bone marrow of patients with myelodysplastic syndromes with the highest expression in refractory anemia with excess of blasts and secondary acute myeloid leukemia. Leuk Lymphoma. 2009;50(6):1043-1046. 17. Kim D, Pertea G, Trapnell C, et al. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. 2013;14(4):R36. 18. 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. 19. Trapnell C, Hendrickson DG, Sauvageau M, et al. Differential analysis of gene regulation at transcript resolution with RNA-seq. Nat

haematologica | 2018; 103(9)

TGFÎ˛1-mediated inhibition of MDS- and AML-derived MSC

Biotechnol. 2013;31(1):46-53. 20. Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA. 2005;102(43):15545-15550. 21. Bruns I, Cadeddu RP, Brueckmann I, et al. Multiple myeloma-related deregulation of bone marrow-derived CD34(+) hematopoietic stem and progenitor cells. Blood. 2012;120(13):2620-2630. 22. Hayashi T, Hideshima T, Nguyen AN, et al. Transforming growth factor beta receptor I kinase inhibitor down-regulates cytokine secretion and multiple myeloma cell growth in the bone marrow microenvironment. Clin Cancer Res. 2004;10(22):7540-7546. 23. Zhou L, Nguyen AN, Sohal D, et al. Inhibition of the TGF-beta receptor I kinase promotes hematopoiesis in MDS. Blood. 2008;112(8):3434-3443. 24. Akiyama T, Matsunaga T, Terui T, et al. Involvement of transforming growth factorbeta and thrombopoietin in the pathogenesis of myelodysplastic syndrome with myelofibrosis. Leukemia. 2005;19(9):15581566. 25. Zorat F, Shetty V, Dutt D, et al. The clinical and biological effects of thalidomide in patients with myelodysplastic syndromes. Br J Haematol. 2001;115(4):881-894. 26. Blank U, Karlsson S. TGF-beta signaling in the control of hematopoietic stem cells. Blood. 2015;125(23):3542-3550. 27. Garbe A, Spyridonidis A, Mobest D, et al. Transforming growth factor-beta 1 delays formation of granulocyte-macrophage colony-forming cells, but spares more primitive progenitors during ex vivo expansion of CD34+ haemopoietic progenitor cells. Br J

haematologica | 2018; 103(9)

Haematol. 1997;99(4):951-958. 28. Sitnicka E, Ruscetti FW, Priestley GV, Wolf NS, Bartelmez SH. Transforming growth factor beta 1 directly and reversibly inhibits the initial cell divisions of long-term repopulating hematopoietic stem cells. Blood. 1996;88(1):82-88. 29. Krause DS, Fulzele K, Catic A, et al. Differential regulation of myeloid leukemias by the bone marrow microenvironment. Nat Med. 2013;19(11):1513-1517. 30. Dominici M, Le Blanc K, Mueller I, et al. Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy. 2006;8 (4):315-317. 31. Bhagat TD, Chen S, Bartenstein M, et al. Epigenetically aberrant stroma in MDS propagates disease via Wnt/beta-catenin activation. Cancer Res. 2017;77(18):48464857. 32. Chandran P, Le Y, Li Y, et al. Mesenchymal stromal cells from patients with acute myeloid leukemia have altered capacity to expand differentiated hematopoietic progenitors. Leuk Res. 2015;39(4):486-493. 33. Kim JA, Shim JS, Lee GY, et al. Microenvironmental remodeling as a parameter and prognostic factor of heterogeneous leukemogenesis in acute myelogenous leukemia. Cancer Res. 2015;75(11): 2222-2231. 34. Bachegowda L, Gligich O, Mantzaris I, et al. Signal transduction inhibitors in treatment of myelodysplastic syndromes. J Hematol Oncol. 2013;6:50. 35. Baron F, Suciu S, Amadori S, et al. Value of infliximab (RemicadeÂŽ) in patients with low-risk myelodysplastic syndrome: final

36.

37.

38.

39.

40.

41.

42.

43.

results of a randomized phase II trial (EORTC trial 06023) of the EORTC Leukemia Group. Haematologica. 2012;97 (4):529-533. Deeg HJ, Gotlib J, Beckham C, et al. Soluble TNF receptor fusion protein (etanercept) for the treatment of myelodysplastic syndrome: a pilot study. Leukemia. 2002;16(2):162-164. Isufi I, Seetharam M, Zhou L, et al. Transforming growth factor-beta signaling in normal and malignant hematopoiesis. J Interferon Cytokine Res. 2007;27(7):543552. Bhagat TD, Zhou L, Sokol L, et al. miR-21 mediates hematopoietic suppression in MDS by activating TGF-beta signaling. Blood. 2013;121(15):2875-2881. Zhou L, McMahon C, Bhagat T, et al. Reduced SMAD7 leads to overactivation of TGF-beta signaling in MDS that can be reversed by a specific inhibitor of TGF-beta receptor I kinase. Cancer Res. 2011;71(3): 955-963. Suragani RN, Cadena SM, Cawley SM, et al. Transforming growth factor-beta superfamily ligand trap ACE-536 corrects anemia by promoting late-stage erythropoiesis. Nat Med. 2014;20(4):408-414. Iancu-Rubin C, Mosoyan G, Wang J, et al. Stromal cell-mediated inhibition of erythropoiesis can be attenuated by Sotatercept (ACE-011), an activin receptor type II ligand trap. Exp Hematol. 2013;41(2):155-166.e17. Chen X, Eksioglu EA, Zhou J, et al. Induction of myelodysplasia by myeloidderived suppressor cells. J Clin Invest. 2013;123(11):4595-4611. Dohner H, Weisdorf DJ, Bloomfield CD. Acute myeloid leukemia. N Engl J Med. 2015;373(12):1136-1152.

1471

ARTICLE

Acute Myeloid Leukemia

Ferrata Storti Foundation

Pharmacological inhibition of dihydroorotate dehydrogenase induces apoptosis and differentiation in acute myeloid leukemia cells

Dang Wu,1,# Wanyan Wang,1,# Wuyan Chen,2 Fulin Lian,2 Li Lang,1 Ying Huang,3 Yechun Xu,2 Naixia Zhang,2 Yinbin Chen,4 Mingyao Liu,4 Ruth Nussinov,5,6 Feixiong Cheng,7,8,9,10 Weiqiang Lu4 and Jin Huang1

Shanghai Key Laboratory of New Drug Design, School of Pharmacy, East China University of Science and Technology, China; 2CAS Key Laboratory of Receptor Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences (CAS), China; 3Guangdong Institute for Drug Control, Guangzhou, China; 4Shanghai Key Laboratory of Regulatory Biology, Institute of Biomedical Sciences and School of Life Sciences, East China Normal University, China; 5Cancer and Inflammation Program, Leidos Biomedical Research, Inc., Frederick National Laboratory for Cancer Research, National Cancer Institute at Frederick, MD, USA; 6Department of Human Molecular Genetics and Biochemistry, Sackler School of Medicine, Tel Aviv University, Israel; 7Center for Complex Networks Research and Department of Physics, Northeastern University, Boston, MA, USA; 8Center for Cancer Systems Biology and Department of Cancer Biology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA; 9Genomic Medicine Institute, Lerner Research Institute, Cleveland Clinic, OH, USA and 10Department of Molecular Medicine, Cleveland Clinic Lerner College of Medicine, Case Western Reserve University, OH, USA 1

Haematologica 2018 Volume 103(9):1472-1483

DW and WW contributed equally to this work.

ABSTRACT

Correspondence: huangjin@ecust.edu.cn or wqlu@bio.ecnu.edu.cn or chengf@ccf.org Received: January 10, 2018. Accepted: May 30, 2018. Pre-published: June 7, 2018.

doi:10.3324/haematol.2018.188185 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1472 ÂŠ2018 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.

1472

cute myeloid leukemia is a disorder characterized by abnormal differentiation of myeloid cells and a clonal proliferation derived from primitive hematopoietic stem cells. Interventions that overcome myeloid differentiation have been shown to be a promising therapeutic strategy for acute myeloid leukemia. In this study, we demonstrate that CRISPR/Cas9-mediated knockout of dihydroorotate dehydrogenase leads to apoptosis and normal differentiation of acute myeloid leukemia cells, indicating that dihydroorotate dehydrogenase is a potential differentiation regulator and a therapeutic target in acute myeloid leukemia. By screening a library of natural products, we identified a novel dihydroorotate dehydrogenase inhibitor, isobavachalcone, derived from the traditional Chinese medicine Psoralea corylifolia. Using enzymatic analysis, thermal shift assay, pull down, nuclear magnetic resonance, and isothermal titration calorimetry experiments, we demonstrate that isobavachalcone inhibits human dihydroorotate dehydrogenase directly, and triggers apoptosis and differentiation of acute myeloid leukemia cells. Oral administration of isobavachalcone suppresses subcutaneous HL60 xenograft tumor growth without obvious toxicity. Importantly, our results suggest that a combination of isobavachalcone and adriamycin prolonged survival in an intravenous HL60 leukemia model. In summary, this study demonstrates that isobavachalcone triggers apoptosis and differentiation of acute myeloid leukemia cells via pharmacological inhibition of human dihydroorotate dehydrogenase, offering a potential therapeutic strategy for acute myeloid leukemia.

Introduction Acute myeloid leukemia (AML) is a malignant disorder characterized by differentiation and abnormal growth of hematopoietic stem or progenitor cells.1 AML is typically associated with a rapid onset of symptoms attributed to bone marrow haematologica | 2018; 103(9)

Inhibition of DHODH induces differentiation in AML

failure, and may be fatal within weeks or months without treatment.1 AML is the most common acute leukemia, with an estimated incidence of about 19,000 cases in 2018 in the United States (USA) and a 5-year relative survival rate of approximately 60% in a population of children in the USA, based on data from 2007 to 2013.2 The major therapeutic paradigm for AML in the past several decades has been either chemotherapy with an anthracycline/cytarabine combination or allogeneic stem cell transplantation.3 Although traditional chemotherapy induces remission in de novo AML patients, only 20-30% of patients survive disease free in the long-term because of the high toxicity of chemotherapeutics, subsequent relapses, and development of drug resistance.4,5 On the other hand, although targeted therapies, such as the isocitrate dehydrogenase 2 (IDH2) inhibitor enasidenib and the FMS-like tyrosine kinase 3 (FLT3) inhibitor midostaurin, were recently approved by the USA Food and Drug Administration for AML treatment,6 only relatively few AML patients with actionable mutations of IDH2 or FLT3 will benefit from these drugs.7-9 Development of innovative therapeutic agents is, therefore, a pressing need to improve the clinical efficacy and quality of life for AML patients. The human dihydroorotate dehydrogenase (DHODH) enzyme belongs to the class 2 DHODH family. It is anchored at the inner mitochondrial membrane.10 As an essential enzyme that catalyzes dihydroorotate to orotic acid, DHODH plays a critical role in the de novo pyrimidine biosynthesis of DNA and RNA.11 Rapidly proliferating cells, such as cancer cells and lymphocytes, mainly depend on de novo pyrimidine biosynthesis to support their growth rate, indicating that this enzyme is a potential target in the treatment of cancer and autoimmune diseases.10 A previous study suggested that DHODH is required for rapid proliferation of tumor cells, playing an important role in tumorigenesis and tumor development.12 Using a unique Homeobox A9-driven leukemia model, Sykes et al. recently found that DHODH is a novel metabolic target in differentiation therapy of AML.13,14 Their pioneer work offers a potential differentiation treatment strategy for patients with AML.13,14 Several DHODH inhibitors, such as brequinar, have already been evaluated in various clinical cancer trials, but severe adverse reactions limit their clinical application.15,16 In this study, we show that CRISPR-Cas9-mediated knockout of DHODH greatly impairs growth, increases apoptosis, and induces differentiation of two AML cell lines, HL60 and THP-1, indicating once again that DHODH is a potential therapeutic target. We identified a novel, direct inhibitor of DHODH, isobavachalcone, by screening a library of natural products. We demonstrate that isobavachalcone effectively triggers apoptosis and induces differentiation in AML cells via direct inhibition of DHODH. Furthermore, our results suggest that isobavachalcone, alone or in combination with adriamycin, significantly prolongs survival in an intravenous HL60 leukemia model.

complied with the protocol approved by the Animal Care and Use Committee at East China University of Science and Technology.

Knockout of dihydroorotate dehydrogenase in HL60 and THP-1 cells The guide RNA sequences targeting DHODH were designed and cloned into a LentiCRISPRv2 construct (Addgene, #52961).17 This construct along with psPAX2 (Addgene, #12260) and pMD2.G (Addgene, #12259) helper constructs were co-transfected into HEK-293T cells using Lipofectamine 2000 reagent (Invitrogen, NY, USA) to produce lentiviral supernatants. Viral production was subsequently concentrated 60X by ultracentrifugation. Cells were infected with lentiviral supplemented with polybrene (8 μg/mL) in 24-well plates and infected cells were selected in medium containing puromycin (0.8 μg/mL). The knockout efficiency of single guide RNA (sgRNA) was determined by western blot analysis. The sgRNA targeting exon 2 of human DHODH were listed as follows: sgRNA1: 5’-TTCTTCGACATTGCCGTCGA-3’; sgRNA2: 5’ACAAGGTCCCAAAGACAG-3’.

Cell apoptosis assay Cells were seeded into six-well plates and incubated with the indicated concentrations of compounds. The apoptosis assays were performed using an AnnexinV-FITC Apoptosis Detection kit (eBioscience, MA, USA) according to the instructions.18 The apoptotic cells were analyzed using a BD FACS Calibur flow cytometer (BD Biosciences, NJ, USA).

Differentiation marker analysis Cultured cells were harvested and washed with phosphatebuffered saline on ice, then resuspended in FACS buffer (phosphate-buffered saline, pH 7.4, supplemented with 1 mM ethylenediaminetetraacetic acid and 2% fetal bovine serum).19 Antibodies of differentiation markers (CD11b, CD14, CD33 and CD34) were added and incubated for 1 h at 37°C in the dark. Flow cytometer data were collected and analyzed on a BD FACS Caliber using Cell Quest software (BD Biosciences, NJ, USA).

Wright-Giemsa staining Cells were harvested, washed with phosphate-buffered saline and fixed with 95% ice-cold methanol for 30 min at 4°C.14 The cells were then seeded on the slide and allowed to dry in the air. Next, the cells were stained with Wright-Giemsa for 5 min and rinsed in deionized water. Finally, coverslips were fixed with Permount prior to microscopy (Nikon, Tokyo, Japan).

Statistical analysis All values are expressed as the mean ± standard deviation of at least three independent experiments. GraphPad Prism 5.0 software (GraphPad software, CA, USA) was used for the statistical analysis. Comparisons between two groups were analyzed using the two-tailed Student t-test. For multiple comparisons, one-way ANOVA followed by Tukey multiple comparison tests were performed. P values < 0.05 are considered statistically significant.

Results Methods The Online Supplement contains detailed information on the experimental materials and methods. All animal care and experimental procedures in this study haematologica | 2018; 103(9)

Dihydroorotate dehydrogenase overexpression is associated with poor prognosis in acute myeloid leukemia We examined the relationship between DHODH expression and overall survival in AML patients. In the 1473

D. Wu et al.

Kaplan-Meier survival analyses (see Online Supplementary Methods), we found that high expression of DHODH was significantly correlated with poor prognosis in patients with cytogenetically normal AML based on data from a previous microarray study20 (P=2.5 × 10-3) (Figure 1A), suggesting a potential clinical role of human DHODH in AML.

Dihydroorotate dehydrogenase is required for maintenance of acute myeloid leukemia cancer cell malignancy We next examined the levels of DHODH expression in a panel of human cancer cell lines including AML. We found that DHODH expression was higher in AML than in other cancer cell lines (Figure 1B), consistent with the results of bioinformatics analysis (Online Supplementary Figure S1) using large-scale cancer cell lines from the Cancer Cell Line Encyclopedia database.21 We next silenced DHODH completely by a CRISPR/Cas9 knockout system in both HL60 and THP-1 cells. A substantial knockout of the DHODH protein was observed in the knockout groups compared with the control group (Figure 1C and Online Supplementary Figure S2A). Notably, DHODH knockout impaired the growth of HL60 and THP-1 cells (Figure 1D and Online Supplementary Figure S2B). Knockout of DHODH caused an increase of HL60 cell apoptosis from 1.34±0.21% to 23.47±1.23% (sgRNA1) or 26.18±0.84% (sgRNA2), compared to the control (Figure 1E). Similar cell apoptosis induction was observed in THP-1 (from 4.72±0.41% to 19.93±1.74% (sgRNA1) or 21.79±1.32% (sgRNA2) (Online Supplementary Figure S2C). Western blot analysis further showed significant upregulation of three apoptosis-related markers (cleaved PARP, cleaved caspase-3 and cleaved caspase-9) in HL60 and THP-1 cells, thus revealing an apoptotic mechanism (Figure 1F and Online Supplementary Figure S2D). We observed that DHODH-knockout significantly increased the expression of CD11b and CD14 (differentiation markers of myeloid cells), whereas it had no effect on CD33 and CD34, in either HL60 (Figure 1G) or THP-1 cells (Online Supplementary Figure S2E), suggesting that DHODH-knockout induces myeloid differentiation of AML cells. The best-known MYC protein family member, MYC, a crucial myeloid cell differentiation modulator, is frequently overexpressed in AML.22 Bioinformatics analysis revealed that DHODH is highly co-expressed with MYC in AML patients (Online Supplementary Figure S3) according to RNA-sequencing data from The Cancer Genome Atlas.23 We found that DHODH knockout significantly reduced the expression of MYC protein in HL60 and THP-1 cells (Figure 1H and Online Supplementary Figure S2F). p21 is transcriptionally suppressed by MYC in cancer cells.24 Notably, MYC loss induced by DHODH inhibition accompanied an elevation of protein expression of p21 in HL60 and THP-1 (Figure 1H and Online Supplementary Figure S2F). Taken together, abrogation of DHODH activity in AML markedly alleviated malignant characteristics, indicating that DHODH is a potential therapeutic target in AML.

Isobavachalcone is a novel dihydroorotate dehydrogenase inhibitor Through screening an in-house natural product library using the 2,6-dichloroindophenol assay (see Online 1474

Supplementary Methods), we found that, at the concentration of 10 μM, isobavachalcone, a chalcone derived from traditional Chinese medicine Psoralea corylifolia, showed the greatest inhibitory activity on recombinant human DHODH protein among 337 natural products (Figure 2A). The chemical structure of isobavachalcone is presented in Figure 2B. Specifically, isobavachalcone showed a half maximal inhibitory concentration (IC50) value of 0.13 μM on DHODH, which is approximately 2-fold stronger than that of leflunomide, a Food and Drug Administrationapproved DHODH inhibitor for the treatment of rheumatoid arthritis (Figure 2C).25 To further examine the direct interaction between isobavachalcone and DHODH, we first performed a thermal shift assay, a commonly used assay to evaluate ligand-protein interaction.26 Figure 2D reveals that isobavachalcone significantly stabilized DHODH protein with an over 14°C melting temperature (DTm) increase in the presence of a 10-fold molar excess of DHODH (Figure 2D), suggesting a direct interaction between isobavachalcone and DHODH. Furthermore, we observed a dose-dependent attenuation of signal in the Carr-Purcell-Meiboom-Gill nuclear magnetic resonance (NMR) spectra, confirming that DHODH influences the state of isobavachalcone (Figure 2E).27 In addition, isobavachalcone bound to DHODH with a KD value of 1.33 μM (Figure 2F), according to an isothermal titration calorimetry experiment, which is consistent with results of the thermal shift assay and NMR experiments. We next performed kinetic analysis of isobavachalcone against human DHODH using a Lineweaver-Burk plot. We found that isobavachalcone is a competitive inhibitor against coenzyme Q0 and uncompetitive for the substrate dihydroorotate (Online Supplementary Figure S4). Figure 2G reveals that isobavachalcone occupies the “ubiquinone channel”, a well-known ligand binding pocket of DHODH (PDB ID: 4YLW),28 as determined from molecular docking simulation (see Online Supplementary Methods). Specifically, several hydrophobic contacts are involved in the binding between isobavachalcone and amino acid residues of DHODH, including Tyr38, Gln47, Ala55, Ala59, Leu67, Phe98, Val143, Thr360 and Pro364. In brief, we identified a direct, high potential DHODH inhibitor, isobavachalcone, which could be a therapeutic agent for AML.

Isobavachalcone inhibits the proliferation of acute myeloid leukemia cells via inhibition of dihydroorotate dehydrogenase We next investigated DHODH protein expression in four human AML cell lines: HL60, THP-1, U937 and MOLM-13. Among these four AML cell lines, HL60 has a high level of DHODH protein and is sensitive to isobavachalcone (Online Supplementary Figure S5A,B). Figure 3A,B shows that isobavachalcone significantly suppresses HL60 and THP-1 cell proliferation in a concentration-dependent and time-dependent manner. The cellcounting assay further revealed isobavachalcone concentration- and time-dependent suppression of HL60 cell growth (Online Supplementary Figure S5C). Importantly, Figure 3C illustrates that both recombinant DHODH and endogenous DHODH in HL60 cells bind directly bind to isobavachalcone-conjugated-Sepharose 4B beads, but not to Sepharose 4B beads alone. Isobavachalcone stabilized DHODH in HL60 cells in a cellular thermal shift assay (Figure 3D), suggesting that isobavachalcone inhibits DHODH directly in AML cells. Notably, knockout of haematologica | 2018; 103(9)

Inhibition of DHODH induces differentiation in AML

DHODH markedly reduced the sensitivity of HL60 cells to isobavachalcone (Figure 3E). Taken together, these findings indicate that isobavachalcone suppresses HL60 cell growth through direct inhibition of DHODH. After treatment with increasing concentrations of isobavachalcone for 72 h, the percentage of apoptotic cells

increased in a dose-dependent manner (Figure 3F,G). To further investigate the mechanisms underlying isobavachalcone-induced apoptosis in HL60 cells, we performed a western blot assay to detect the protein marker of apoptosis. Figure 3H reveals that several protein markers of apoptosis (cleaved caspase-9, cleaved caspase-3 and cleaved

F A

Figure 1. Dihydroorotate dehydrogenase is required for acute myeloid leukemia cells to maintain their malignant characteristics.. (A) Kaplan-Meier survival curves for AML patients divided by level of DHODH expression. The P-value of the Kaplan-Meier survival analysis was determined using a log-rank test (see the Online Supplementary Methods). (B) Western blot analysis of the expression levels of DHODH in different cancer types. SiHa: cervical carcinoma; H1299, A549 and H446: lung carcinoma; HL60 and THP1: AML; Jurkat: acute T-cell leukemia; HepG2: hepatic carcinoma; U251: glioma. (C) Knockout of DHODH in HL60 cells was analyzed by western blot. (D) Knockout of DHODH impaired the growth of HL60 cells. Cell viability was evaluated by MTS assay at 24 h intervals up to 96 h in three independent experiments. The graph represents the means Âą SD. The Student t-test was performed, **P<0.01. (E and F) DHODH knockout resulted in apoptosis of HL60 cells. Cell apoptosis was analyzed by flow cytometry and the expression levels of apoptosis-related proteins in HL60 cells was detected by western blot at 96 h after infection. (G) Flow cytometry demonstrated upregulation of cell surface markers CD14 and CD11b after knockout of DHODH in HL60 cells whereas there was no effect on CD33 and CD34 expression. The cells were measured at 96 h after infection. Data represent the mean Âą SD of three independent experiments. (H) Knockout of DHODH resulted in reduced expression of MYC protein and upregulated expression of p21 in HL60 cells.

haematologica | 2018; 103(9)

1475

D. Wu et al.

PARP) are increased after isobavachalcone treatment. Given that DHODH represented the rate-limiting step of de novo pyrimidine biosynthesis in the endogenous synthesis of uridine monophosphate, we wondered whether uridine could rescue isobavachalcone-induced apoptosis. Figure 3F-H shows that uridine alone did not affect cell apoptosis, whereas it reversed apoptosis induced by isobavachalcone treatment. Additionally, we examined the effect of isobavachalcone on cell morphology using fluorescence microscopy. Chromatin condensation was observed with

Hochest 33258 staining after treatment with 30 μM isobavachalcone for 48 h (Figure 3I). Similar apoptotic events following isobavachalcone treatment were also observed in THP-1 cells (Online Supplementary Figure S6). It can be concluded that isobavachalcone induces apoptosis of HL60 cells by inhibiting DHODH activity.

Isobavachalcone triggers differentiation by inhibiting dihydroorotate dehydrogenase We then investigated whether isobavachalcone triggers

E F

Figure 2. A natural product, isobavachalcone, is a newly identified direct dihydroorotate dehydrogenase inhibitor. (A) Graphical presentation of screening results of 337 compounds tested at a concentration of 10 μM in a DHODH enzymatic assay. Each dot represents one compound. (B) Chemical structure of isobavachalcone. (C) Dose-response curves of isobavachalcone and leflunomide in the DHODH enzymatic assay. (D) A thermofluor assay shows that isobavachalcone robustly stabilizes DHODH and produces a thermal shift over 14°C (ratio 1:10). (E) NMR measurement of direct binding between isobavachalcone and DHODH. Carr-PurcellMeiboom-Gill NMR spectra for isobavachalcone (red), isobavachalcone in the presence of DHODH at 2.5 μM (green) and 5 μM (cyan). (F) Isothermal titration calorimetry of isobavachalcone binding to DHODH. Binding curves were fitted as a single binding event. (G) Computational docking analysis of the binding mode of isobavachalcone with DHODH. The structure is shown as a ribbon diagram and the isobavachalcone molecule (left) is presented as a sphere model based on PDB ID: 4YLW. The amino acid residues surrounding isobavachalcone (yellow sticks, right) are represented by slate sticks. Figure 1G was generated by PyMOL software (https://www.pymol.org/). IBC: isobavachalocone; LEF: leflunomide.

1476

haematologica | 2018; 103(9)

Inhibition of DHODH induces differentiation in AML

AML cell differentiation. We measured the level of expression four myeloid differentiation markers, CD14, CD11b, CD33 and CD34, by flow cytometry. Of note, we found that isobavachalcone increased the expression of CD14 and CD11b (Figure 4A,B), but had no effect on the expression of CD33 and CD34 (Online Supplementary Figure S7A) in HL60 cells. As DHODH is involved in the

intracellular synthesis of uridine,29 we observed that isobavachalcone treatment lead to the depletion of uridine in HL60 cells (Online Supplementary Figure S8A). The uridine rescue experiment demonstrates that the endogenous cellular pyrimidine deficiency induced by DHODH inhibition is crucial for the differentiation of AML cells (Figure 4A,B and Online Supplementary Figure S7B).

D H

Figure 3. Isobavachalcone shows anti-proliferative activity against acute myeloid leukemia cells. (A) HL60 and THP1 cells were treated with increasing concentrations of isobavachalcone for 48 h, and cell viability was measured by MTS assay. (B) The time-response curve of 30 ÎźM isobavachalcone on cell viability of HL60 and THP1 cells. (C) Recombinant DHODH protein or HL60 cell lysate was incubated with control or isobavachalcone-conjugated Sepharose 4B beads. Proteins bound to the beads were analyzed by western blot. (D) Cellular thermal shift assay shows that isobavachalcone stabilizes and targets DHODH in intact HL60 cells. Cells were incubated with isobavachalcone for 12 h and the assay was performed. (E) The IC50 value of isobavachalcone against DHODH-knockout HL60 cells. (F) HL60 cells were treated with isobavachalcone at the indicated concentrations for 72 h. Cell apoptosis was detected by flow cytometry using staining with annexin V, fluorescein isothiocyanate (FITC) and propidium iodine (PI). (G) The quantitative data of cell apoptosis in (F). (H) Changes in apoptosis-related proteins after treatment with isobavachalcone or uridine for 72 h. (I) Representative images of Hoechst 33258-stained cells were analyzed by fluorescence microscopy in HL60 cells treated with 30 ÎźM isobavachalcone for 48 h. Red arrows indicate apoptotic cells. IBC: isobavachalocone.

haematologica | 2018; 103(9)

1477

D. Wu et al.

Wright–Giemsa staining shows that isobavachalcone strikingly reduced the nuclear cytoplasmic ratio (N: C ratio) compared to the control (Figure 4C), suggesting monocytic differentiation. MYC is a main regulator of AML cell differentiation, and DHODH knockout reduces

the protein level of MYC in HL60 cells (Figure 1H). As expected, after isobavachalcone treatment, there was a significant concentration-dependent loss of MYC expression accompanied by a significant up-regulation of p21 (Figure 4D). Furthermore, MYC protein level was down-

Figure 4. Isobavachalcone induces HL60 cell differentiation. (A) HL60 cells were treated with isobavachalcone for 72 h and CD14 expression was detected by flow cytometry analysis. Right: quantification data of CD14 expression in HL60 cells. (B) HL60 cells were treated with isobavachalcone for 72 h and CD11b expression was detected by flow cytometry analysis. Right: quantification data of CD11b expression in HL60 cells. (C) Morphological changes associated with differentiation of HL60 cells were evidenced by Wright-Giemsa staining in the presence of 30 μM isobavachalcone. (D) HL60 cells were incubated with different concentrations of isobavachalcone for 24 h. After incubation, western blot assay was performed to examine the expression levels of MYC and p21. (E) The expression of MYC was analyzed by western blot 1.5, 3, 6 and 12 h after treatment with 30 μM isobavachalcone. (F) HL60 cells were treated with 30 μM isobavachalcone for 3 h with or without 10 μM MG132. The levels of MYC expression were subsequently examined by western blot analysis. (G) 293T cells were transfected with the MYC-luc reporter plasmid together with pRSVluc plasmid (as an internal control) and incubated with different concentrations of isobavachalcone for 24 h. (H) HL60 cells were treated with 30 μM isobavachalcone for 24 h, then MYC and p21 gene levels in HL60 cells were examined by reverse transcriptase polymerase chain reaction. Data represent mean ± SD of three independent experiments. A Student t-test was performed, *P<0.05, **P<0.01, ***P<0.001. IBC: isobavachalocone.

1478

haematologica | 2018; 103(9)

Inhibition of DHODH induces differentiation in AML

regulated by isobavachalcone in a time-dependent manner (Figure 4E). MYC has been reported to be an unstable protein that is degraded rapidly via the proteasome pathway.30 As shown in Figure 4F, isobavachalcone induced significant degradation of MYC, and treatment with MG132 (a pro-

teasome inhibitor) blocked MYC deregulation. Moreover, extended isobavachalcone treatment (24 h) reduced MYC gene transcription in a MYC reporter assay (Figure 4G). Figure 4H further confirms down-regulation of MYC and p21 gene expression following isobavachalcone treatment.

Figure 5. Isobavachalcone suppresses tumor growth in a subcutaneous HL60 xenograft mouse model. (A) Measurements of tumor volume in an HL60 xenograft mouse model treated with vehicle or the indicated dosages of compounds for 18 days. Changes in mean tumor sizes compared with the control group. Bars represent mean Âą SD for eight animals in each group. (B) Effect of isobavachalcone on tumor weights. (C) The body weight of mice was measured every 3 days. (D) Histological morphology of tumor tissue, stained with hematoxylin and eosin, from the different groups. (E) Images of the major organ tissues, stained with hematoxylin and eosin, from the different groups of animals. (F) DHODH protein expression in tumors was examined at the time the animals were sacrificed. (G) The enzyme activity of DHODH in xenograft tumor tissues treated with vehicle, isobavachalcone or leflunomide was measured by fluorescence assay. (H) Western blot analysis of the changes of apoptosis-associated proteins at day 18 of HL60 xenograft tumors treated with vehicle, isobavachalcone or leflunomide. (I) Expression levels of MYC and p21 in tumors treated with vehicle, isobavachalcone, or leflunomide were estimated by western blot. Data represent means Âą SD. A one-way ANOVA test was performed, **P<0.01 and ***P<0.001. IBC: isobavachalocone; LEF: leflunomide.

haematologica | 2018; 103(9)

1479

D. Wu et al.

Previous studies have shown that O-linked N-acetylglucosamine transferase (OGT) is directly involved in the regulation of MYC expression.31, 32 As expected, our results further showed that isobavachalcone treatment led to a significant reduction of OGT (Online Supplementary Figure S8B). Taken together, these findings indicate that the inhibition of DHODH by isobavachalcone induces MYC degradation-dependent differentiation of AML cells.

isobavachalcone exhibits antitumor efficacy by inhibiting dihydroorotate dehydrogenase activity in vivo Figure 5A shows that isobavachalcone significantly suppressed tumor growth (37.81 ± 4.32% and 78.91 ± 9.73%, 15 and 30 mg/kg isobavachalcone, respectively),

compared with the control group, in a subcutaneous HL60 xenograft mouse model. Similarly, the weight of tumors in the groups of animals treated with isobavachalcone was significantly reduced (Figure 5B) by 37.65 ± 3.74% and 74.41 ± 8.47%, respectively. No obvious body weight loss or deaths were observed in any of the groups of mice (Figure 5C), suggesting that isobavachalcone has a low toxicity in vivo. Hematoxylin and eosin staining analysis further supports the potent tumor suppression exerted by isobavachalcone on the HL60 xenograft model (Figure 5D) without significant damage to the main organs of the mice treated with this compound (Figure 5E). We then examined the expression level and activity of DHODH protein in vivo. Western

Figure 6. The combination of isobavachalcone and adriamycin shows synergistic antileukemic effects in vitro and in vivo. (A-D) Synergistic effects of the isobavachalcone and adriamycin combination on AML cells. AML cell lines (HL60, THP1, U937 and MOLM-13) were treated with several increasing concentrations of isobavachalcone and adriamycin alone or in combination for 48 h. The combination index (CI) calculation was performed using CalcuSyn software (Version 2.1; Biosoft). Drug combinations with a CI<1 are considered to be synergistic. (E) Synergistic effects of isobavachalcone and adriamycin combination therapy in an intravenous HL60 leukemia model. Mice with established tumors (4 per group) were divided into four groups and treated with vehicle, isobavachalcone, adriamycin or a combination of isobavachalcone and adriamycin. The P value was determined using the log-rank test, P=0.0003 for the survival analysis (E). (F) Leukemia cells isolated from the isobavachalcone-treated group exhibit morphological features of differentiation. ADR: adriamycin; IBC: isobavachalocone.

1480

haematologica | 2018; 103(9)

Inhibition of DHODH induces differentiation in AML

blot revealed that DHODH expression levels were not affected in isobavachalcone-treated tumors (Figure 5F). Interestingly, DHODH activity decreased in isobavachalcone-treated tumors compared with vehicle-treated tumors, as determined by a cellular DHODH enzyme assay based on the use of a 4-trifluoromethyl-benzamidoxime fluorogenic reagent (Figure 5G). Hence, we hypothesized that isobavachalcone suppresses tumor growth by inhibiting DHODH enzymatic activity in vivo. We checked how isobavachalcone affects the expression levels of three apoptosis-related markers (cleaved caspase-9, cleaved caspase-3, and cleaved PARP) and two differentiation-related proteins (MYC and p21) in xenograft tumors. As expected, cleaved caspase-9, cleaved caspase-3 and cleaved PARP and p21 levels were noticeably elevated in the isobavachalcone-treated groups, while MYC protein level decreased (Figure 5H), consistent with the in vitro analysis (Figures 3H and 4D). Notably, isobavachalcone shows greater efficacy than leflunomide (Figure 5A-H).

The combination of isobavachalcone and adriamycin shows synergistic antileukemic effects in vitro and in vivo Adriamycin is a widely used chemotherapy drug for treatment of AML in the clinic. However, as monotherapy, its therapeutic efficacy is limited by problems such as acquired resistance.9 We investigated whether isobavachalcone sensitizes AML to adriamycin. We first determined the cytotoxicity of adriamycin in four human AML cell lines: HL60, THP-1, U937 and MOLM-13 (Online Supplementary Figure S9). We then examined the efficacy of a combination of isobavachalcone and adriamycin across the four AML cell lines and found that the isobavachalcone and adriamycin combination led to a cooperative suppression of AML cell growth (Figure 6AD). We further investigated the potency of combinational therapy using a disseminated HL60 model of AML. As shown in Figure 6E, mice treated with the isobavachalcone and adriamycin combination had a significantly longer survival compared with that of animals in the other groups. Of note, treatment with isobavachalcone led to AML differentiation in vivo, as evidenced by Wright–Giemsa staining (Figure 6F). We also investigated the effect of isobavachalcone on adriamycin-resistant HL60/adriamycin cells in vitro. As expected, the HL60/adriamycin cell line showed greater resistance to adriamycin (IC50 = 38.77±0.81 μM), (Online Supplementary Figure S10A) compared to HL60 cells (IC50 = 0.36±0.05 μM), (Online Supplementary Figure S9A). We found that isobavachalcone markedly suppressed HL60/adriamycin cell growth in a concentration-dependent manner (Online Supplementary Figure S10B). Online Supplementary Figure S10C shows that the combination of isobavachalcone and adriamycin had an enhanced antiproliferative effect in HL60/adriamycin cells compared with the effect of isobavachalcone or adriamycin alone in vitro. The values of the combination index, which were all <0.8, suggest a synergistic antitumor effect between isobavachalcone and adriamycin.33 Notably, the group co-administered isobavachalcone and adriamycin exhibited markedly coordinative anti-tumor activity compared with monotherapy and control groups (P<0.001, one-way ANOVA) (Online Supplementary Figure S10D,E). No obvious changes were observed in the animals’ body haematologica | 2018; 103(9)

weight, indicating that the combined therapy was well tolerated (Online Supplementary Figure S10F). In summary, isobavachalcone combined with adriamycin effectively suppresses the growth of adriamycin-resistant AML cells in vitro and in vivo, offering a potential drug combination strategy for AML therapy.

Discussion Differentiation therapy is inspired by the observation that hormones and cytokines can induce differentiation ex vivo; it can, therefore, be a powerful way of irreversibly altering the phenotype of malignant cells.34 The high cure rates of acute promyelocytic leukemia by a combination of retinoic acid and arsenic underscore the success of differentiation therapy.34 However, approximately 90% of patients with AML other than acute promyelocytic leukemia do not benefit from the combination of retinoic and arsenic. New differentiation therapy strategies are urgently needed to improve the clinical outcome of these patients. In this study, we demonstrated that DHODH is a potent regulator of AML cell growth, apoptosis and differentiation. Specifically, genetic knockout or pharmacological inhibition of DHODH overcomes a differentiation blockade of myeloid cells by promoting MYC degradation. Via systematic screening of an in-house natural product library, we identified isobavachalcone as an effective, direct DHODH inhibitor. By targeting DHODH, isobavachalcone suppressed tumor growth, overcoming the differentiation blockade of AML in vitro and in vivo, without a significant toxic profile, thus making it a potential therapeutic agent for AML differentiation. MYC is a pro-oncogenic transcription factor that contributes to tumorigenesis and tumor progression of human cancers including leukemia.35,36 MYC levels are correlated with tumor cell progression and differentiation and myeloid cell differentiation is dependent on the suppression of this transcription factor.37 An earlier study revealed that inhibition of DHODH abrogates transcriptional elongation of the MYC gene in melanoma.38 The DHODH inhibitor leflunomide can reduce MYC expression and consequently reduce proliferation of human melanoma cells.38 However, the ramifications of DHODH inhibition on MYC in AML are still unclear. In our experiments, the level of MYC expression was markedly lowered after the silencing of DHODH, either by knockout of DHODH or by the introduction of the DHODH inhibitor. Isobavachalcone induced MYC degradation in a proteasome-dependent manner. In addition, extended isobavachalcone treatment inhibited MYC transcriptional activity in a luciferase reporter assay. Sykes and colleagues reported that inhibition of DHODH overcame differentiation blockade in AML; however, the precise pathway affected by DHODH inhibition is still not understood.14 We suggest that inhibition of DHODH through down-regulation of MYC induces AML cell differentiation. It was previously reported that isobavachalcone, a naturally occurring chalcone, exhibited anticancer activity in several types of malignancies. For instance, isobavachalcone showed anti-cancer activities in a two-stage mouse skin cancer model.39 Yang et al. reported that isobavachalcone impaired the growth and increased apoptosis of the ovarian carcinoma cell line OVCAR-8 and prostate cancer 1481

D. Wu et al.

cell line PC3.40 Isobavachalcone has been reported to inhibit AKT1 kinase in a dose-dependent manner in vitro with an IC50 value of 32.90 μM.39 However, the anti-cancer properties and mechanisms of isobavachalcone are not fully understood. As described in this paper, we identified isobavachalcone as a potent, direct human DHODH inhibitor, and systematically validated it by, for example, enzymatic and isothermal titration calorimetry assays, thermal shift, and NMR. To the best of our knowledge, this is the first report of isobavachalcone’s potent differentiation-inducing activity in AML and antileukemic effect in mouse xenograft models. However, other potential off-target effects caused by isobavachalcone in AML cells remain to be determined in the future. Taken together, these results provide compelling evidence of isobavachalcone’s potent anti-leukemia activity by interfering with the biosynthetic pathway of pyrimidine nucleotides through suppression of DHODH catalytic activity. Several major breakthroughs have been made recently in the diagnosis and therapy of AML; however resistance is still a daunting barrier.9,23 It was reported that known DHODH inhibitors, leflunomide and A771726 can increase the sensitivity of cells to adriamycin in triplenegative breast cancer.41 However, the biological consequences and clinical benefits of these agents in AML remain unclear. In this study, we demonstrated that the combination of the DHODH inhibitor isobavachalcone and adriamycin effectively suppressed the growth of AML cells and prolonged survival in xenograft models of AML without obvious toxicity, offering a promising therapeutic strategy for AML. The detailed molecular mechanism of the synergistic effect of this combination of drugs remains to be studied further. For example, PTEN is a well-known tumor suppressor gene and loss of PTEN is associated with chemoresistance in multiple types of cancers.42 Recently, Deepti et al. revealed that PTEN-deficient

References 1. Khwaja A, Bjorkholm M, Gale RE, et al. Acute myeloid leukaemia. Nat Rev Dis Primers. 2016;2:16010. 2. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018;68(1):7-30. 3. Roboz GJ. Current treatment of acute myeloid leukemia. Curr Opin Oncol. 2012;24(6):711-719. 4. Tallman MS, Gilliland DG, Rowe JM. Drug therapy for acute myeloid leukemia. Blood. 2005;106(4):1154-1163. 5. Pollyea DA, Gutman JA, Gore L, Smith CA, Jordan CT. Targeting acute myeloid leukemia stem cells: a review and principles for the development of clinical trials. Haematologica. 2014;99(8):1277-1284. 6. Wei AH, Tiong IS. Midostaurin, enasidenib, CPX-351, gemtuzumab ozogomycin and venetoclax bring new hope to AML. Blood. 2017;130(23):2469-2474. 7. Buckley SA, Kirtane K, Walter RB, Lee SJ, Lyman GH. Patient-reported outcomes in acute myeloid leukemia: where are we now? Blood Rev. 2018;32(1):81-87. 8. Kadia TM, Ravandi F, Cortes J, Kantarjian H.

1482

10.

11.

12.

13.

14.

cancers are heavily dependent on an intact pathway for de novo pyrimidine synthesis, which makes such tumors vulnerable to DHODH inhibition.43 Thus, dual targeting of DHODH and mTOR/AKT may offer a strategy for further overcoming chemoresistance in AML patients with PTEN alterations. In summary, we demonstrated that isobavachalcone triggers apoptosis and differentiation of AML cells via pharmacological inhibition of human DHODH, laying the groundwork for future AML treatment strategies. Importantly, administration of isobavachalcone enhanced the anti-cancer efficiency of adriamycin in a xenograft model of human AML. The comprehensive preclinical findings presented here suggest that isobavachalcone is a potential therapeutic agent for AML with a novel molecular mechanism, and that further development of the drug for use in the clinic is warranted. Acknowledgments Funds: this work was supported by the National Natural Science Foundation of China (81773775), Shanghai Committee of Science and Technology (15431902000), and State Key Laboratory for Chemistry and Molecular Engineering of Medicinal Resources (Guangxi Normal University, CMEMR2017-B01). This work was also supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Number K99HL138272 to FC. The project was funded in whole or in part with Federal funds from the Frederick National Laboratory for Cancer Research, National Institutes of Health, under contract HHSN261200800001E. This research was also supported [in part] by the Intramural Research Program of the NIH, Frederick National Laboratory, Center for Cancer Research. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the US Government.

Toward individualized therapy in acute myeloid leukemia: a contemporary review. JAMA Oncol. 2015;1(6):820-828. Kavanagh S, Murphy T, Law A, et al. Emerging therapies for acute myeloid leukemia: translating biology into the clinic. JCI Insight. 2017;2(18). pii: 95679. Diao Y, Lu W, Jin H, et al. Discovery of diverse human dihydroorotate dehydrogenase inhibitors as immunosuppressive agents by structure-based virtual screening. J Med Chem. 2012;55(19):8341-8349. Liu S, Neidhardt EA, Grossman TH, Ocain T, Clardy J. Structures of human dihydroorotate dehydrogenase in complex with antiproliferative agents. Structure. 2000;8 (1):25-33. Löffler M, Fairbanks LD, Zameitat E, Marinaki AM, Simmonds HA. Pyrimidine pathways in health and disease. Trends Mol Med. 2005;11(9):430-437. Janzer A, Sykes D, Gradl S, et al. Abstract 3086: Inhibitors of the enzyme dihydroorotate dehydrogenase, overcome the differentiation blockade in acute myeloid leukemia. Cancer Res. 2017;77(13):3086-3086. Sykes D, Kfoury Y, Mercier F, et al. Inhibition of dihydroorotate dehydrogenase overcomes

15.

16.

17.

18.

19.

20.

differentiation blockade in acute myeloid leukemia. Cell. 2016;167(1):171-186. Rd BH, Raymond E, Awada A, et al. Pharmacokinetic and phase I studies of brequinar (DUP 785; NSC 368390) in combination with cisplatin in patients with advanced malignancies. Invest New Drug. 1998;16(1): 19-27. Pally C, Smith D, Jaffee B, et al. Side effects of brequinar and brequinar analogues, in combination with cyclosporine, in the rat. Toxicology. 1998;127(1-3):207-222. Wang J, Hu K, Guo J, et al. Suppression of KRas-mutant cancer through the combined inhibition of KRAS with PLK1 and ROCK. Nat Commun. 2016;7:11363. Lu W, Cheng F, Yan W, et al. Selective targeting p53WT lung cancer cells harboring homozygous p53 Arg72 by an inhibitor of CypA. Oncogene. 2017;31(10):4719-4731. Fiskus W, Sharma S, Shah B, et al. Highly effective combination of LSD1 (KDM1A) antagonist and pan-histone deacetylase inhibitor against human AML cells. Leukemia. 2014;28(11):2155-2164. Metzeler KH, Hummel M, Bloomfield CD, et al. An 86-probe-set gene-expression signa-

haematologica | 2018; 103(9)

Inhibition of DHODH induces differentiation in AML

21.

22.

23.

24.

25.

26.

ture predicts survival in cytogenetically normal acute myeloid leukemia. Blood. 2008;112(10):4193-4201. Barretina J, Caponigro G, Stransky N, et al. The Cancer Cell Line Encyclopedia enables predictive modeling of anticancer drug sensitivity. Nature. 2012;483(7391):603-607. Wu S, Cetinkaya C, Munozalonso MJ, et al. Myc represses differentiation-induced p21 CIP1 expression via Miz-1-dependent interaction with the p21 core promoter. Oncogene. 2003;22(3):351-360. Cancer Genome Atlas Research N, Ley TJ, Miller C, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):20592074. Seoane J, Le HV, Massagué J. Myc suppression of the p21 (Cip1) Cdk inhibitor influences the outcome of the p53 response to DNA damage. Nature. 2002;419(6908):729734. Smolen JS, Kalden JR, Scott DL, et al. Efficacy and safety of leflunomide compared with placebo and sulphasalazine in active rheumatoid arthritis: a double-blind, randomised, multicentre trial. Lancet. 1999;353 (9149):259-266. Grøftehauge MK, Hajizadeh NR, Swann MJ, Pohl E. Protein–ligand interactions investigated by thermal shift assays (TSA) and dual polarization interferometry (DPI). Acta Crystallogr D Biol Crystallogr. 2015;71 (1):36-44.

haematologica | 2018; 103(9)

27. Guo Z-Q, Zheng T, Chen B, et al. Smallmolecule targeting of E3 ligase adaptor SPOP in kidney cancer. Cancer Cell. 2016;30(3):474-484. 28. K Vyas V, Ghate M. Recent developments in the medicinal chemistry and therapeutic potential of dihydroorotate dehydrogenase (DHODH) inhibitors. Mini Rev Med Chem. 2011;11(12):1039-1055. 29. Sun Y, Hess JL. Targeting the pyrimidine synthesis pathway for differentiation therapy of acute myelogenous leukemia. Transl Cancer Res. 2017;6(1):S109-S111. 30. Dang CV. MYC on the path to cancer. Cell. 2012;149(1):22-35. 31. Itkonen H M, Minner S, Guldvik I J, et al. OGlcNAc transferase integrates metabolic pathways to regulate the stability of c-MYC in human prostate cancer cells. Cancer Res. 2013;73(16):5277-5287. 32. Jóźwiak P, Forma E, Bry M, et al. OGlcNAcylation and metabolic reprograming in cancer. Front Endocrinol. 2014;5(5):145. 33. Chou T-C. Drug combination studies and their synergy quantification using the ChouTalalay method. Cancer Res. 2010;70(2): 440-446. 34. de Thé H. Differentiation therapy revisited. Nat Rev Cancer. 2018;18(2):117-127. 35. Vita M, Henriksson M. The Myc oncoprotein as a therapeutic target for human cancer. Semin Cancer Biol. 2006;16(4):318-330. 36. Nesbit CE, Tersak JM, Prochownik EV. MYC oncogenes and human neoplastic dis-

ease. Oncogene. 1999;18(19):3004-3016. 37. Yu Z-Y, Xiao H, Wang L-M, et al. Natural product vibsanine A induces differentiation of myeloid leukemia cells through PKC activation. Cancer Res. 2016;76(9):2698-2709. 38. White RM, Cech J, Ratanasirintrawoot S, et al. DHODH modulates transcriptional elongation in the neural crest and melanoma. Nature. 2011;471(7339):518-522. 39. Akihisa T, Tokuda H, Hasegawa D, et al. Chalcones and other compounds from the exudates of Angelica keiskei and their cancer chemopreventive effects. J Nat Prod. 2006; 69(1):38-42. 40. Jing H, Zhou X, Dong X, et al. Abrogation of Akt signaling by isobavachalcone contributes to its anti-proliferative effects towards human cancer cells. Cancer Lett. 2010;294(2):167-177. 41. Brown KK, Spinelli JB, Asara J, Toker A. Adaptive reprogramming of de novo pyrimidine synthesis is a metabolic vulnerability in triple-negative breast cancer. Cancer Discov. 2017;7(4):391-399. 42. Cully M, You H, Levine AJ, Mak TW. Beyond PTEN mutations: the PI3K pathway as an integrator of multiple inputs during tumorigenesis. Nat Rev Cancer. 2006;6(3): 184-192. 43. Mathur D, Stratikopoulos E, Ozturk S, et al. PTEN regulates glutamine flux to pyrimidine synthesis and sensitivity to dihydroorotate dehydrogenase inhibition. Cancer Discov. 2017;7(4):380-390.

1483

ARTICLE

Acute Myeloid Leukemia

Ferrata Storti Foundation

Haematologica 2018 Volume 103(9):1484-1492

Clofarabine, high-dose cytarabine and liposomal daunorubicin in pediatric relapsed/refractory acute myeloid leukemia: a phase IB study Natasha K.A. van Eijkelenburg,1,2,3* Mareike Rasche,4* Essam Ghazaly,5 Michael N. Dworzak,6 Thomas Klingebiel,7 Claudia Rossig,8 Guy Leverger,9 Jan Stary,10 Eveline S.J.M. De Bont,11 Dana A. Chitu,12 Yves Bertrand,13 Benoit Brethon,14 Brigitte Strahm,15 Inge M. van der Sluis,1,3 Gertjan J.L. Kaspers,2,16,17 Dirk Reinhardt3,17 and C. Michel Zwaan1,2,3

Department of Pediatric Oncology/Hematology, Erasmus MC-Sophia Children's Hospital, Rotterdam, the Netherlands; 2Department of Pediatric Oncology, Princess Máxima Center for Pediatric Oncology, Utrecht, the Netherlands; 3European Consortium for Innovative Therapies for Children with Cancer (ITCC), Villejuif, France; 4Department of Pediatric Oncology, University Children's Hospital, Essen, Germany; 5Centre for Haemato-Oncology, Barts Cancer Institute, Queen Mary University of London, UK; 6 Children’s Cancer Research Institute and St. Anna Children’s Hospital, Department of Pediatrics, Medical University of Vienna, Austria; 7Pediatric Hematology/Oncology, Johann Wolfgang Goethe University, Frankfurt, Germany; 8Pediatric Hematology and Oncology, University Children’s Hospital, Münster, Germany; 9Department of Pediatric Hematology and Oncology, AP-HP, GH HUEP, Trousseau Hospital, Paris, France; 10 Department of Pediatric Hematology and Oncology, 2nd Faculty of Medicine, Charles University Prague, University Hospital Motol, Czech Republic; 11Department of Pediatric Oncology, University Medical Center Groningen, University of Groningen, the Netherlands; 12Clinical Trial Center, Department of Hematology, Erasmus Medical Center, Rotterdam, the Netherlands; 13Pediatric Hematology Department, IHOP and Claude Bernard University, Lyon, France; 14Department of Pediatric Hematology, Robert Debré Hospital, Paris, France; 15Division of Pediatric Hematology and Oncology, Department of Pediatrics and Adolescent Medicine, University of Freiburg, Germany; 16 Department of Pediatric Oncology, VU University Medical Center, Amsterdam, the Netherlands and 17I-BFM-AML committee, Kiel, Germany 1

Correspondence: c.m.zwaan@erasmusmc.nl

Received: December 24, 2017. Accepted: May 16, 2018. Pre-published: May 17, 2018. doi:10.3324/haematol.2017.187153 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1484 ©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.

1484

*NKAE and MR contributed equally to this work as first authors.

ABSTRACT

urvival in children with relapsed/refractory acute myeloid leukemia is unsatisfactory. Treatment consists of one course of fludarabine, cytarabine and liposomal daunorubicin, followed by fludarabine and cytarabine and stem-cell transplantation. Study ITCC 020/I-BFM 2009-02 aimed to identify the recommended phase II dose of clofarabine replacing fludarabine in the abovementioned combination regimen (3+3 design). Escalating dose levels of clofarabine (20-40 mg/m2/day x 5 days) and liposomal daunorubicin (40-80 mg/m2/day) were administered with cytarabine (2 g/m2/day x 5 days). Liposomal DNR was given on day 1, 3 and 5 only. The cohort at the recommended phase II dose was expanded to make a preliminary assessment of anti-leukemic activity. Thirty-four children were enrolled: refractory 1st (n=11), early 1st (n=15), ≥2nd relapse (n=8). Dose level 3 (30 mg/m2 clofarabine; 60 mg/m2 liposomal daunorubicin) appeared to be safe only in patients without subclinical fungal infections. Infectious complications were dose-limiting. The recommended phase II dose was 40 mg/m2 clofarabine with 60 mg/m2 liposomal daunorubicin. Side-effects mainly consisted of infections. The overall response rate was 68% in 31 response evaluable patients, and 80% at the recommended phase II dose (n=10); 22 patients proceeded to stem cell transplantation. The 2year probability of event-free survival (pEFS) was 26.5±7.6 and probability of survival (pOS) 32.4±8.0%. In the 21 responding patients, the 2-year pEFS was 42.9±10.8 and pOS 47.6±10.9%. Clofarabine expohaematologica | 2018; 103(9)

Clofarabine in pediatric AML

sure in plasma was not significantly different from that in single-agent studies. In conclusion, clofarabine was well tolerated and showed high response rates in relapsed/refractory pediatric acute myeloid leukemia. Patients with (sub)clinical fungal infections should be treated with caution. Clofarabine has been taken forward in the Berlin-Frankfurt-Münster study for newly diagnosed acute myeloid leukemia. The Study ITCC-020 was registered as EUDRA-CT 2009-009457-13; Dutch Trial Registry number 1880.

Introduction Despite enhanced cure rates for pediatric acute myeloid leukemia (AML), relapsed patients still suffer from poor clinical outcome,1 especially those who relapse within one year of diagnosis.2 Recently, a randomized phase III study in relapsed/refractory AML revealed an improved early treatment response when liposomal daunorubicin (DNX) was added to the FLAG (fludarabine, cytarabine and granulocyte-colony stimulating factor (G-CSF)) regimen.2 Although this did not translate into a survival benefit [4year probability of survival (pOS) of 38±3%], in Europe, the combination of FLAG and DNX (FLAG-DNX) followed by stem cell transplantation (SCT) was considered the standard treatment for children with AML in first relapse.2,3 For several reasons, including data obtained in adults, subsequently G-CSF priming was deleted from the FLAG regimen4,5 and the current standard is FLA-DNX. Intensified relapse therapy increased survival over time,6 but insufficient cure rates in the 70% range (pOS),3 and potential long-term toxicities such as anthracycline cardiomyopathy when cumulative dosages exceed 300-400 mg/m2,7,8 mean that novel chemotherapy combinations must be developed. Clofarabine, a structural hybrid of fludarabine and cladribine, was developed to enhance efficacy and stability of the drug, while reducing the formation of toxic compounds like 2-F adenine compared to previous nucleoside analogs.9,10 Inhibition of DNA polymerase and ribonucleotide reductase as well as induction of apoptosis,10 and enhanced accumulation of cytarabine may contribute to the drug's high antitumor activity.11 The first pediatric phase I study of clofarabine identified a maximum tolerated dose (MTD) of 52 mg/m2/day, with reversible hepatotoxicity and skin rash as dose-limiting toxicities.12 Based on a phase II study, clofarabine was approved in 2004 for relapsed pediatric ALL.13 However, in relapsed pediatric AML, activity of clofarabine was not confirmed, with a response rate of only 26%, mainly consisting of partial responses, probably due to the inclusion of heavily pre-treated patients.13 In contrast, in adult AML, clofarabine showed antileukemic activity in several early phase studies.14-16 Randomized data showed that, even though remission rates were improved in untreated older patients with AML and high-risk myelodysplastic syndrome (MDS), no survival benefit of clofarabine (20 mg/m2/d for 5 days) over low-dose cytarabine was shown.17 Moreover, another randomized study comparing cytarabine with clofarabine (20 mg/m2/d for 5 days) in induction courses I and II in elderly AML also failed to show a survival benefit.18 However, a recent study by the HOVON-group did show a survival advantage for patients randomized to clofarabine in intermediate-risk AML subsets, albeit at a very low dose of 10 mg/m2/d for five days added to induction courses I and II.19 haematologica | 2018; 103(9)

In pediatric ALL, clofarabine combination therapy was developed to overcome resistance, e.g. combinations with cyclophosphamide, or etoposide and cyclophosphamide, or topotecan, vinorelbine and thiopeta.20-23 In children with relapsed/refractory AML, clofarabine (52 mg/m2/d for 5 days) with cytarabine (1 g/m2/d for 5 days) resulted in a 3-year pOS of 46%±27% in responders.24 In the CLOUD study, 9 children with relapsed/refractory AML were treated with clofarabine (30 mg/m2/d for 5 days) and liposomal daunorubicin (60 mg/m2, days 1, 3 and 5); 33% obtained complete remission (CR) and were subsequently transplanted.25 In this study, we aimed to combine clofarabine with high-dose cytarabine and liposomal daunorubicin as in the FLA-DNX regimen, replacing fludarabine with clofarabine assuming that this may have greater anti-leukemic potential when tolerable. The treatment schedule was based on an adapted ‘Faderl regimen’ developed in adult AML.26,27

Methods Study ITCC-020 (EUDRA-CT 2009-009457-13; Dutch Trial Registry number 1880) was an investigator-initiated open-label phase IB dose-escalation study sponsored by Erasmus MC, Rotterdam, the Netherlands. Patients were enrolled in 15 centers in 5 countries within the Innovative Therapies for Children with Cancer (ITCC) consortium.

Study design The primary objective was to establish the MTD and recommended phase II dose (RP2D) of clofarabine in combination with cytarabine and DNX in relapsed/refractory pediatric AML. Doselimiting toxicities were evaluated in the first course only. Secondary objectives included tabulation of additional safety and tolerability data across both treatment courses, preliminary estimation of response, event-free survival (EFS), and overall survival (OS), and the pharmacokinetics (PK) of clofarabine in this combination [serum and cerebrospinal fluid (CSF)]. A classical 3+3 design was used with dose escalation to MTD, after which the cohort was expanded (n=10) at the RP2D. Separate expansion cohorts were planned for dose level (DL) 1-4 and for DL5 (in early 1 first relapse only), respectively, when considered safe.

Patient eligibility Pediatric patients below 19 years of age with early 1st relapse (within 12 months from initial diagnosis), refractory 1st relapse (≥ 20% blasts in the bone marrow after the first course of standard re-induction therapy), or those with at least a second relapsed AML were eligible. Only patients with early 1st relapse without prior SCT were eligible for DL5. Inclusion and exclusion criteria are described in Online Supplementary Methods S1, and included recovery from prior organ toxicity. An amendment in November 2011 excluded patients with evidence for subclinical fungal infections using high-resolution computed tomography (CT) scan of 1485

N.K.A. van Eijkelenburg et al.

the thorax, and elevated serum levels of galactomannan. All patients/parents had to provide written informed consent according to local law and regulations, after the institutional review boards of the participating institutes approved the study. The study was conducted in accordance with good clinical practice guidelines and the Declaration of Helsinki.

Treatment Clofarabine and DNX dosages were escalated from 20-40 mg/m2/day at days 1-5, and 40-80 mg/m2/day at days 1, 3 and 5, respectively, whereas the dose of cytarabine was fixed at 2 g/m2/d for days 1-5 (Table 1, detailed information in Online Supplementary Methods S2). DL5 (DNX 80 mg/m2/day at days 1, 3 and 5) was considered as a separate cohort with restricted inclusion criteria. This cohort was added as up-front pediatric AML protocols use DNX at this dose level rather than 60 mg/m2/d28 (Figure 1).

Safety and efficacy evaluations Adverse events (AEs) were graded according to the Common Terminology Criteria for Adverse Events v.3.0. Dose Limiting Toxicities (DLT) were defined as grade 3 or 4 non-hematologic AEs and hematologic AEs lasting longer than 42 days, limited to the first course, and at least possibly drug-related, with some exceptions (definitions and efficacy evaluation in Online Supplementary Methods S3 and Table S1).

Statistical analysis Dosing and efficacy were analyzed using descriptive statistics. Survival estimates were computed by Kaplan-Meier. The database lock for this analysis was set at 11th February 2018. All analyses were performed with Stata v.13.1. Further information concerning pharmacokinetics and statistical analysis is available in Online Supplementary Methods S4, Methods S5 and Table S2.

Table 1. Dose levels of clofarabine, liposomal daunorubicin and cytarabine.

Age â&#x2030;Ľ 1 year (Age < 1 year) Dose level -1

Dose level 1 (starting dose) Dose level 2 Dose level 3A Dose level 3B* Dose level 4 Dose level 5**

Clofarabine

DNX

Ara-C

15 mg/m2/d x 5 d (0.5 mg/kg/d x5 d) 20 mg/m2/d x 5 d (0.7 mg/kg/d x 5 d) 30 mg/m2/d x 5 d (1.0 mg/kg/d x 5 d) 30 mg/m2/d x 5 d (1.0 mg/kg/d x 5 d) 30 mg/m2/d x 5 d (1.0 mg/kg/d x 5 d) 40 mg/m2/day x 5 d (1.3 mg/kg/day x 5 d) 40 mg/m2/day x 5 d

40 mg/m2/day 1-3-5 (1.3 mg/kg/d 1-3-5) 40 mg/m2/d 1-3-5 (1.3 mg/kg/d 1-3-5) 40 mg/m2/d 1-3-5 (1.3 mg/kg/d 1-3-5) 60 mg/m2/d 1-3-5 (2.0 mg/kg/d 1-3-5) 60 mg/m2/d 1-3-5 (2.0 mg/kg/d 1-3-5) 60 mg/m2/d 1-3-5 (2.0 mg/kg/d 1-3-5) 80 mg/m2/d 1-3-5

2 gr/m2/d x 5 d (70 mg/kg/d x 5 d) 2 gr/m2/d x 5 d (70 mg/kg/d x 5 d) 2 gr/m2/d x 5 d (70 mg/kg/d x 5 d) 2 gr/m2/d x 5 d (70 mg/kg/d x 5 d) 2 gr/m2/d x 5 d (70 mg/kg/d x 5 d) 2 gr/m2/d x 5 d (70 mg/kg/d x 5 d) 2 gr/m2/d x 5 d

Ara-C: cytarabine; DNX:liposomal daunorubicin. *Dose level 3B: cohort 3 was repeated after an amendment implementing screening for subclinical fungal infections. **Dose level 5: this cohort was open for patients with early 1st relapse of acute myeloid leukemia without prior stem cell transplantation only. Dosages in brackets are for children below 1 year of age or below 10 kg body weight.

Figure 1. Treatment schedule. Clofarabine was administered intravenously (IV) in 2 hours (h) (days 1-5); liposomal daunorubicin (DNX) in 1 h (days 1, 3, 5); DNX in 1 h (days 1, 3, 5), starting 30 minutes after the end of clofarabine; cytarabine was administered (IV) in 3 h (days 1-5), starting 3 h after the end of clofarabine. Intrathecal therapy was administered at day 6 with cytarabine for prophylaxis, or triple therapy (cytarabine and methotrexate and prednisolone) with age-adjusted dosages in case of central nervous system involvement. G-CSF: granulocyte-colony stimulating factor.

1486

haematologica | 2018; 103(9)

Clofarabine in pediatric AML

Results

patients, prior cumulative dose of anthracycline was calculated using the equivalent formula, as previously described.29 With respect to the missing equivalent doses, prior amsacrine and prior liposomal daunorubicin were not included in the calculation. Prior anthracycline dosages are heterogenous in this cohort (range 80-686 mg/m2). Detailed patients' characteristics are provided in Table 2. A total of 46 CLARA-DNX cycles were administered: 12 patients received two cycles (1 in DL3A; 2 in DL3B; 6 in DL4; 3 in DL5). In 31 patients, cycle 1 was given according to the schedule provided in the protocol. Deviations concerned intrathecal treatment (delayed or deleted in one patient each), and in one patient cytarabine was halted at day 3 due to an allergic reaction. The median absolute dose of clofarabine in cycle one was 160 mg (range 38-380 mg): 165 mg (range 38-456) for DNX, and 9.37 g (range 1.26-19 g) for cytarabine.

Patients and treatment Between 10th May 2010 and 28th May 2014, 34 AML patients (early first relapse, n=15; refractory first relapse, n=11; ≥2nd relapse, n=8) were recruited. The median age was 8.3 years (range 1.0-19.6 years); the median WBC 4.6x109/L (range 0.3-326x109/L), and 41% of patients were female. We had conclusive cyto(genetic) data from initial diagnosis in 30 of 34 patients, and from time point of inclusion in this study in 23 patients. None of these karyotypes included good risk characteristics such as t(8;21)(q22;q22) or inv(16)(p13q22)/t(16;16)(p13;q22). Eighteen patients had been pre-treated with FLAG-DNX reinduction according to protocol AML-BFM 2001/01.2 Twelve patients had received a prior SCT (4 in CR1; 7 in CR2; and 1 unknown), including one patient with 2 prior SCTs. For 32 of 34

Table 2. Patients’ baseline characteristics.

Characteristic

Age, years Median Range Sex Male Female FAB at initial diagnosis M0 M1 M2 M4 M5 M6 Non-classified Disease status Early 1st rel Refractory 1st rel ≥ 2nd rel WBC at inclusion (x109/L) Median Range <10 ≥10 Prior SCT NO SCT SCT in 1st CR SCT in 2nd CR Unknown Pre-treatment with FLA-DNX Yes No

ALL patients (N=34) N. (%)

DL1 (N=4) N.

DL2 (N=3) N.

DL3A (N=6) N.

DL3B (N=6) N.

DL4 (N=10) N.

DL5 (N=5) N.

8.3 1.0-19.6

7.1 2.8-8.8

14.2 1.3-15.6

12.9 1.0-17.6

4.6 2.5-19.6

10.2 1.4-18.8

2.4 1.3-16.0

20 (58.8) 14 (41.2)

3 1

3 -

2 4

4 2

6 4

2 3

4 3 6 4 14 1 2

1 1 1 1

1 1 1 -

1 2 3 -

2 3 1

2 2 1 5 -

2 1 2 -

15 11 8

2 2

1 1 1

4 2

3 2 1

6 2 2

5 -

4.6 0.3-326 27 6

3 1-5 4 -

6 1-127 2 1

3 1-326 4 2

4 0-34 4 2

3 0-30 6 2

4 1-6 5 -

22 4 7 1

2 2 -

1 1 1 -

4 1 1 -

2 2 1 1

8 2 -

5 -

18 16

4 -

1 2

6 -

3 3

4 6

WBC: white blood cells; FAB: French-American-British; SCT: stem cell transplantation; FLA: fludarabine, cytarabine (Ara-C); DNX: liposomal daunorubicin; CR: complete remission; DL: dose-level; N: number of patients; rel: relapse.

haematologica | 2018; 103(9)

1487

N.K.A. van Eijkelenburg et al.

Safety and tolerability Initial dose-escalation was halted at DL3 because of DLTs (3 Grade 3 pulmonary fungal infections (aspergillosis); 1 Grade 3 non-fungal pulmonary infection). An amendment was issued to repeat cohort 3 after adding screening for subclinical fungal infections, and only one DLT in 6 patients was noted (Grade 3 pulmonary candida infection), hence escalation to DL4 was pursued. In DL4 only one DLT was noted (Grade 4 septicemia). As DL4 was considered safe, this dose-level was expanded to 10 patients. Subsequently, DL5 was opened for patients with early 1st relapse without prior SCT. At DL5, 2 of 5 patients experienced DLT (1 Grade 3 pseudomonas aeruginosa cellulitis and 1 Grade 3 Gram-negative septicemia), and DL5 was closed therafter (Table 3). Of note, the use of prophylactic antibacterial, antifungals, and antiviral agents was recommended according to each institutionâ&#x20AC;&#x2122;s guidelines. Twenty-eight of 34 patients received anti-fungal prophylaxis with azoles (either itraconazole, voriconazole or fluconazole); 17 of 34 patients were on prophylactic treatment with amphotericin B. Hence 11 of 34 patients received both azole prophylaxis as well as amphotericin B. In total, 34 SAEs were reported, mostly consisting of febrile neutropenia (n=18), documented infections (n=9), or gastrointestinal SAEs (n=3) (Online Supplementary Table S3). Non-hematologic AEs related to the first cycle of

study treatment are summarized in Table 3. Overall, the most common treatment-related AEs were gastrointestinal, pain and infection. These non-hematologic AEs were mild (Grade 1-2) in most patients. In 2 patients, acute renal failure was reported, one in combination with a tumor lysis syndrome. Another patient had a capillary leak syndrome with ascites and reduced diuresis. Hematologic AEs occurred frequently and mostly concerned Grade 3-4 myelosuppression, not resulting in DLTs (data not shown). Overall, 21 patients died: 2 as the result of an AE (one multi-organ failure secondary to febrile neutropenia, 1 hypoxia and respiratory and cardiovascular failure); 15 deaths were due to progressive leukemia following other treatment attempts in some but not all patients. The remaining 4 deaths were: SCT procedure related (n=2), lung toxicity after allogeneic transplantation with cytomegalovirus (CMV) reactivation (n=1), and invasive aspergillosis (n=1).

Efficacy Of 31 evaluable patients, the ORR was 68% after the 1st cycle of treatment, including 5 (16%) patients with CR, 15 (48%) patients with CR with incomplete blood count recovery (CRi), and 1 patient with a partial response (PR) (3%). Overall response rate (ORR) differed according to disease phase: 87% in 15 patients with early 1st relapse;

Table 3. Non-hematologic adverse events in the first treatment cycle.

AEs AE term

Total

Grade 1-2 n (%)

All grades, n (%) Gastrointestinal 114 (24) Pain 62(13) Infection 61(13) Metabolic/laboratory 46(10) Skin/dermatological 34(7) Constitutional symptoms 2(5) Pulmonary/upper respiratory 19(4) Cardiac problems including arrhythmia 13(3) Neurology (or ocular/visual) 10(7) Allergy/immunology 8(2) Others (renal/genitourinary/vascular/syndromes) 5(1)

DLTs Dose level

Grade 3-4 n (%)

AEs

Patients

AEs

Patients

96(29) 54(17) 9(3) 37(11) 31(9) 22(7) 13(4) 8(2) 14(3) 7(2) -

27(79) 16(47) 8(24) 8(24) 24(71) 11(32) 9(27) 7(21) 11(32) 6(18) -

18(13) 8(6) 52(36) 9(6) 3(2) 2(1) 6(4) 5(3) 3(2) 1(<1) 5(3)

14(41) 7(21) 27(79) 5(15) 3(9) 2(6) 3(9) 4(12) 3(9) 1(3) 4(12)

Total patients per dose level (n)

Patients with DLT (n)

Details about DLTs

Dose level 1 Dose level 2 Dose level 3A

4 3 6

1 0 4

Dose level 3B

Dose level 4 Dose level 5

10 5

1 2

Gr 3 Streptococcal infection None Gr 3 Pulmonary infection including 3 fungal infections/ aspergillosis Gr 3 Pulmonary infection: fungal infection / candida albicans Gr 3 Sepsis Gr 3 Cellulitis infection (pseudomonas) Gr 3 Sepsis

AE: adverse events; DLT: dose-limiting toxicities; Gr: grade; n: number.

1488

haematologica | 2018; 103(9)

Clofarabine in pediatric AML

27% in 11 patients with refractory 1st relapse; and 50% in 8 patients with ≥2nd relapse (Table 4). Of the 18 patients pre-treated with FLA-DNX (refractory 1st relapse, n=11; second relapse n=7), 6 patients (5 CR/1 CRi) responded (33%). Three responders (1 CR/2 CRi) were seen in the group of refractory 1st relapse patients. In the expansion cohort at DL4 (n=10), the overall response rate was 80% (2 CR, 5 Cri, 1 PR) (Online Supplementary Table S4). In 9 of the 18 patients with FAB M4 and M5 AML an objective response was seen; this was equally divided between both groups (FAB M4 and M5). Of the 13 patients with other FAB classifications, 9 objective responses were seen (69%, P M4/M5 vs. other FAB 0.284). Following clofarabine, 22 patients underwent SCT (17 MUD, 1 matched family donor, 1 HLA identical sibling and 2 haplo-identical, 1 unknown). Of the 21 clofarabine

responding patients, 18 subsequently underwent SCT. Overall, 9 of the 22 patients who were transplanted died (6 MUD, 1 matched family donor, 1 haplo-identical, 1 unknown); this included 2 patients with SCT-procedurerelated deaths and one death due to lung toxicity with CMV reactivation. The remaining patients died due to progressive leukemia. At time of database lock, 10 of 34 patients were alive with a median survival time of 56 months (range 32.3-78.4 months). The 1-year pEFS was 35.3±8.2% and the pOS 50.0±8.6% (Figure 2A and B); the 2-year pEFS was 26.5±7.6% and the 2-year pOS 32.4±8.0%. The 1-year pEFS was 57.1±10.8% in the 21 responding patients, and the pOS was 71.4±9.9%; the 2year pEFS was 42.9±10.8% and the pOS 47.6±10.9% (data not shown). Most events in the 21 responders were due to relapse (Figure 2C). In the 10 patients treated at the

Figure 2. Survival estimates after clofarabine combination chemotherapy reinduction. (A) Overall survival of all patients. (B) Event-free survival of all patients. (C) Cumulative Incidence of relapse of the 21 responding patients. (D) Overall survival of all patients at dose level (DL) 4. (E) Event-free survival of all patients at DL4. N: number; F: females; F: failure (event).

haematologica | 2018; 103(9)

1489

N.K.A. van Eijkelenburg et al.

RP2D at DL4, both 1-year pEFS and pOS were 60.0±15.5%, while 2-year pEFS was 50.0±15.8% and 2year pOS 60.0±15.5% (Figure 2D and E).

Pharmacokinetic analysis Blood samples were available from the first 19 patients, after which collection of PK data was halted after interim analysis. Comparison of normalized plasma concentrations at day 5 pre-dose and 24 hours (h) after the last infusion of clofarabine showed similar results with a median concentration of 0.12 and 0.10 ng/mL/mg, indicating a steady state plasma concentration. The median AUC at day 1 was 28.0 ng/mL/mg/h (range 6.0 to 401.2) and 44.0 ng/mL/mg/h (range 19.4 to 135.9) at day 5. Clofarabine t1/2 was identical at day 1 and day 5 (average value of 1.5 h) (Online Supplementary Figure S3A). Clofarabine levels in CSF (n=11) ranged from 0.3 ng/mL to 3.2ng/mL [median CSF penetration (CSF conc/plasma conc) was 32.9% (range 8.0-66.5%)]. These data are in contrast with the previously reported low clofarabine median penetration of 5% (range 3-26%) into the CSF in non-human primates.30 Bonate et al.31 have fit clofarabine single-agent PK data derived from 3 clinical trials (i.e. the ID99-383 study

recruiting pediatric hematologic malignancies, the CLO212 study in pediatric ALL, and the CLO-222 study in pediatric refractory AML) using a non-parametric LOESS fit to the observed dose normalized concentration-time plot. Our clofarabine PK data fitted well with the Bonate et al. 2004 PK model (Figure 3B), confirming that the current combination did not alter clofarabine PK. Furthermore, our PK data in combination were not significantly different from our data with single-agent clofarabine in relapsed ALL patients,32 suggesting that clofarabine has no major drug-drug interaction with either cytarabine or daunorubicin which can affect clofarabine PK.

Discussion In this phase IB study, the RP2D of clofarabine in combination with cytarabine and DNX was established at 40 mg/m2/day for five consecutive days in combination with 60 mg/m2 DNX at days 1, 3 and 5 and cytarabine 2 gram/m2 for five days in patients with no clinical evidence of subclinical fungal infections at time of treatment. This regimen resulted in a high ORR of 68% in 31 response evaluable patients, and 80% at the RP2D (n=10). In general, the combination of clofarabine, liposomal

1490

Figure 3. Clofarabine pharmacokinetics and Ppasma concentrations. (A) Clofarabine plasma concentrations normalized to infused dose (ng/mL/mg of infused dose) as measured by liquid chromatography mass spectrometry (LC-MS)/MS. Each line represents plasma concentrations for a single patient before receiving clofarabine infusion (Pre-dose), 2 (T2), 5 (T5) and 24 (T24) hours (h) after starting of clofarabine infusion at day (d)1 and d5 of the first treatment cycle. Samples from 2 patients (ns. 0708 and 0717) were available from two different treatment cycles. (B) A scatter plot for clofarabine plasma concentrations as measured by LC-MS/MS (in color) overlaid on the pharmacokinetic model developed by Bonate et al.31 (in gray) for single agent clofarabine in 3 previous clinical studies (ID99383, CLO-212 and CLO-222) fitted using non-parametric LOESS fit.

haematologica | 2018; 103(9)

Clofarabine in pediatric AML

daunorubicin and high-dose cytarabine was well-tolerated in our cohort of patients. Most observed adverse events were as expected (febrile neutropenia and infections, gastrointestinal symptoms, dermatological manifestations and pain). Our data on infections are comparable to the single agent study in pediatric relapsed AML (67% of patients experienced ≥grade 3 infections).13 The highest dose level, however, appeared not to be tolerable, again due to infectious complications. It might be that the higher anthracycline dose of 80 mg/m2 DNX can be tolerated in newly diagnosed patients, but this was not assessed in our study. The RP2D of clofarabine (40 mg/m2) was higher compared to the CLOUD study (30 mg/m2/d for 5 days),25 but lower than clofarabine (52 mg/m2/d for 5 days) in combination with 1 gram/m2 of Ara-C without anthracyclines.24 The ORR of 68% observed with this combination regimen is extremely encouraging. In addition, this response rate is higher compared than the single-agent study13 and the CLOUD study,25 and is for instance also higher than in our prior study with single-agent Mylotarg.33 Patients with early 1st relapse (n=15) responded in 93%, whereas in the AML 2001/01 study only 70% (treated with FLAG-DNX) and 54% (treated with FLAG) responded, although a different definition for response was used (≤ 20% blasts in BM after cycle 1).2 The number of responding patients (3 of 10) who were refractory to FLA-DNX chemotherapy given directly before clofarabine is also interesting, although it cannot be excluded that this was the effect of repeated chemotherapy. This was also observed in the AML 2001/01 study where 20% of patients responded after the second FLAG course who were not in CR after the first course.2 However, we have to consider that, in our cohort, we observed a low response rate among the refractory 1st relapse patients (CR 2 of 10, Cri 1 of 10). Our response and survival data are in accordance with the data reported for the COG AAML0523 study.24 Of interest, 90% of the patients in this study (that combined cytarabine 1 g/m2 with clofarabine 52 mg/m2) were in 1st relapse, while the remaining 10% of patients had refractory disease. The combination of cytarabine and clofarabine in the COG study resulted in an ORR of 48% in 48 evaluable patients, and 21 of 23 responders underwent SCT. The overall survival rate at 3 years was 46% for respon-

ders.24 In our study, 10 of 34 patients were still alive at last follow up, with a 2-year pOS 32±8%. This was even higher in responding patients: 2-year pOS 48±11%. A phase III, randomized, double-blind, placebo-controlled trial was recently published in 320 adults over 55 years of age with relapsed/refractory AML, comparing cytarabine 1 g/m2 plus clofarabine 40 mg/m2 versus cytarabine with placebo.27 This study showed significantly improved response rates and enhanced EFS in patients treated with clofarabine and cytarabine compared to placebo and cytarabine. However, no significant impact on survival was achieved, probably related to the higher incidence of mortality in the clofarabine arm.27 This is in line with the randomized study performed by Burnett et al.18 However, the recent study from Löwenberg et al.19 did show that intermediate risk AML patients in the clofarabine arm benefitted compared to cytarabine in the standard arm, including a survival benefit. Children can usually tolerate higher dosages of chemotherapy, certainly when compared to elderly patients, and the dosages used in the adult studies were very low compared to the regimen that we tested here. Better salvage regimens in relapsed pediatric AML, as summarized in various papers,34,35 including ours, have contributed to better survival rates of patients with pediatric AML in the last decade, together with improved supportive care measures.3 Allogeneic SCT is considered the standard treatment in relapsed pediatric AML in Europe after re-induction with salvage chemotherapy. In our study, 22 patients underwent SCT following the clofarabine combination regimen. No particular toxicities (e.g. veno-occlusive disease) (data not shown) were noted during the SCT procedure, and pre-treatment with this chemotherapy combination did not preclude a successful SCT procedure. Pharmacokinetic analysis showed no difference between data obtained with single agent clofarabine in prior studies versus clofarabine in combination in this study. Moreover, we demonstrated that cerebrospinal fluid (CSF) penetration was limited in CSF samples taken approximately 24 h after the last clofarabine infusion. To the best of our knowledge, this study is the first to study CSF penetration of clofarabine. The results of this trial indicate high efficacy of the com-

Table 4. Response after cycle 1 by disease status and after prior treatment with FLA-DNX.

All patients All patients Pre-treated n=34 n=18 Morphological response Evaluable CR CRi PR NEL SD PD Treatment failure

31 (100%) 5 (16%) 15 (48%) 1 (3%) 0 6 (19%) 3 (10%) 1 (3%)

15 (100%) 1 (7%) 5 (33%) 0 5 (33%) 3 (20%) 1 (7%)

Early 1st relapse Refractory 1st relapse All patients Pre-treated All patients Pre-treated n=15 n=0 n=11 n=11 15 (100%) 4 (27%) 9 (60%) 1 (7%)

10 (100%) 1 (10%) 2 (20%)

1 (7%)

3 (30%) 3 (30%) 1 (10%)

≥ 2nd relapse All patients Pre-treated n=8 n=7 6 (100%)

5 (100%)

4 (67%)

3 (60%)

2 (33%)

2 (40%)

FLA: fludarabine, cytarabine (Ara-C); CR: complete remission; DNX: liposomal daunorubicin; n: number; CRi: morphological complete remission with incomplete blood count recovery; NEL: no evidence of leukemia; PR: partial response; SD: stable disease; PD: progressive disease.

haematologica | 2018; 103(9)

1491

N.K.A. van Eijkelenburg et al.

bination of clofarabine, liposomal daunorubicin and highdose cytarabine, while having an acceptable toxicity profile, even in heavily pre-treated patients. Of interest, clofarabine has been taken forward in front-line treatment in the AML-BFM 2012 study (EudraCT: 2013-000018-39) as an induction randomization, albeit low-dose cytarabine is used rather than high-dose.

References 1. Gorman MF, Ji L, Ko RH, et al. Outcome for children treated for relapsed or refractory acute myelogenous leukemia (rAML): a Therapeutic Advances in Childhood Leukemia (TACL) Consortium study. Pediatr Blood Cancer. 2010;55(3):421-429. 2. Kaspers GJ, Zimmermann M, Reinhardt D, et al. Improved outcome in pediatric relapsed acute myeloid leukemia: results of a randomized trial on liposomal daunorubicin by the International BFM Study Group. J Clin Oncol. 2013;31(5):599-607. 3. Zwaan CM, Kolb EA, Reinhardt D, et al. Collaborative Efforts Driving Progress in Pediatric Acute Myeloid Leukemia. J Clin Oncol. 2015;33(27):2949-2962. 4. Milligan DW, Wheatley K, Littlewood T, Craig JI, Burnett AK, Group NHOCS. Fludarabine and cytosine are less effective than standard ADE chemotherapy in highrisk acute myeloid leukemia, and addition of G-CSF and ATRA are not beneficial: results of the MRC AML-HR randomized trial. Blood. 2006;107(12):4614-4622. 5. Ehlers S, Herbst C, Zimmermann M, et al. Granulocyte colony-stimulating factor (GCSF) treatment of childhood acute myeloid leukemias that overexpress the differentiation-defective G-CSF receptor isoform IV is associated with a higher incidence of relapse. J Clin Oncol. 2010;28(15):25912597. 6. Sander A, Zimmermann M, Dworzak M, et al. Consequent and intensified relapse therapy improved survival in pediatric AML: results of relapse treatment in 379 patients of three consecutive AML-BFM trials. Leukemia. 2010;24(8):1422-1428. 7. Kremer LC, van der Pal HJ, Offringa M, van Dalen EC, Voute PA. Frequency and risk factors of subclinical cardiotoxicity after anthracycline therapy in children: a systematic review. Ann Oncol. 2002;13(6):819829. 8. Lipshultz SE, Colan SD, Gelber RD, PerezAtayde AR, Sallan SE, Sanders SP. Late cardiac effects of doxorubicin therapy for acute lymphoblastic leukemia in childhood. N Engl J Med. 1991;324(12):808-815. 9. Pui CH, Jeha S, Kirkpatrick P. Clofarabine. Nat Rev Drug Discov. 2005;4(5):369-370. 10. Bonate PL, Arthaud L, Cantrell WR Jr, Stephenson K, Secrist JA 3rd, Weitman S. Discovery and development of clofarabine: a nucleoside analogue for treating cancer. Nat Rev Drug Discov. 2006;5(10):855-863. 11. Faderl S, Gandhi V, O'Brien S, et al. Results of a phase 1-2 study of clofarabine in combination with cytarabine (ara-C) in relapsed and refractory acute leukemias. Blood. 2005;105(3):940-947. 12. Jeha S, Gandhi V, Chan KW, et al.

1492

13.

14. 15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Acknowledgments We are grateful to all physicians and study staff who took care of the patients and contributed data for this study. We thank Dr. Simon Joel for the useful advice on the pharmacokinetic section of this study. Satianand Ramnarain was the trial-manager responsible for this study. The study was performed with financial support from Sanofi and by the KiKa-foundation grant number 26.

Clofarabine, a novel nucleoside analog, is active in pediatric patients with advanced leukemia. Blood. 2004;103(3):784-789. Jeha S, Razzouk B, Rytting M, et al. Phase II study of clofarabine in pediatric patients with refractory or relapsed acute myeloid leukemia. J Clin Oncol. 2009;27(26):43924397. Tiley S, Claxton D. Clofarabine in the treatment of acute myeloid leukemia in older adults. Ther Adv Hematol. 2013;4(1):5-13. Kantarjian HM, Erba HP, Claxton D, et al. Phase II study of clofarabine monotherapy in previously untreated older adults with acute myeloid leukemia and unfavorable prognostic factors. J Clin Oncol. 2010; 28(4):549-555. Burnett AK, Russell NH, Kell J, et al. European development of clofarabine as treatment for older patients with acute myeloid leukemia considered unsuitable for intensive chemotherapy. J Clin Oncol. 2010;28(14):2389-2395. Burnett AK, Russell NH, Hunter AE, et al. Clofarabine doubles the response rate in older patients with acute myeloid leukemia but does not improve survival. Blood. 2013;122(8):1384-1394. Burnett AK, Russell NH, Hills RK, et al. A comparison of clofarabine with ara-C, each in combination with daunorubicin as induction treatment in older patients with acute myeloid leukaemia. Leukemia. 2017; 31(2):310-317. Lowenberg B, Pabst T, Maertens J, et al. Therapeutic value of clofarabine in younger and middle-aged (18-65 years) adults with newly diagnosed AML. Blood. 2017;129(12):1636-1645. Abd Elmoneim A, Gore L, Ricklis RM, et al. Phase I dose-escalation trial of clofarabine followed by escalating doses of fractionated cyclophosphamide in children with relapsed or refractory acute leukemias. Pediatr Blood Cancer. 2012;59(7):12521258. Hijiya N, Gaynon P, Barry E, et al. A multicenter phase I study of clofarabine, etoposide and cyclophosphamide in combination in pediatric patients with refractory or relapsed acute leukemia. Leukemia. 2009; 23(12):2259-2264. Hijiya N, Thomson B, Isakoff MS, et al. Phase 2 trial of clofarabine in combination with etoposide and cyclophosphamide in pediatric patients with refractory or relapsed acute lymphoblastic leukemia. Blood. 2011;118(23):6043-6049. Shukla N, Kobos R, Renaud T, Steinherz LJ, Steinherz PG. Phase II trial of clofarabine with topotecan, vinorelbine, and thiotepa in pediatric patients with relapsed or refractory acute leukemia. Pediatr Blood Cancer. 2014;61(3):431-435. Cooper TM, Alonzo TA, Gerbing RB, et al.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34. 35.

AAML0523: a report from the Children's Oncology Group on the efficacy of clofarabine in combination with cytarabine in pediatric patients with recurrent acute myeloid leukemia. Cancer. 2014; 120(16):2482-2489. Kearns P, Graham NJ, Cummins B, et al. Phase I study of clofarabine and liposomal daunorubicin in childhood acute myeloid leukemia [abstract]. J Clin Oncol. 2011; 29(Suppl 15):Abstract 9521. Faderl S, Verstovsek S, Cortes J, et al. Clofarabine and cytarabine combination as induction therapy for acute myeloid leukemia (AML) in patients 50 years of age or older. Blood. 2006;108(1):45-51. Faderl S, Wetzler M, Rizzieri D, et al. Clofarabine plus cytarabine compared with cytarabine alone in older patients with relapsed or refractory acute myelogenous leukemia: results from the CLASSIC I Trial. J Clin Oncol. 2012;30(20):2492-2499. Creutzig U, Zimmermann M, Bourquin JP, et al. Randomized trial comparing liposomal daunorubicin with idarubicin as induction for pediatric acute myeloid leukemia: results from Study AML-BFM 2004. Blood. 2013;122(1):37-43. Creutzig U, Diekamp S, Zimmerman M, Reinhardt D. Longitudinal evaluation of early and late anthracycline carciotoxicity in children with AML. Pediatr Blood Cancer. 2007;48(7):651-662. Berg SL, Bonate PL, Nuchtern JG, et al. Plasma and cerebrospinal fluid pharmacokinetics of clofarabine in nonhuman primates. Clin Cancer Res. 2005;11(16):59815983. Bonate PL, Craig A, Gaynon P, et al. Population pharmacokinetics of clofarabine, a second-generation nucleoside analog, in pediatric patients with acute leukemia. J Clin Pharmacol. 2004;44(11):1309-1322. Joel S, Ghazaly E, Smith H, Kearns P, Saha V. The plasma and intracellular pharmacokinetics of clofarabine in pediatric leukemia patients. Cancer Res. 2007;67(Suppl 9):Abstract 1566. Zwaan CM, Reinhardt D, Zimmerman M, et al. Salvage treatment for children with refractory first or second relapse of acute myeloid leukaemia with gemtuzumab ozogamicin: results of a phase II study. Br J Haematol. 2010;148(5):768-776. Kaspers GJ. How I treat paediatric relapsed acute myeloid leukaemia. Br J Haematol. 2014;166(5):636-645. Creutzig U, Zimmermann M, Dworzak MN, et al. The prognostic significance of early treatment response in pediatric relapsed acute myeloid leukemia: results of the international study Relapsed AML 2001/01. Haematologica. 2014;99(9):14721478.

haematologica | 2018; 103(9)

ARTICLE

Acute Lymphoblastic Leukemia

Investigating chemoresistance to improve sensitivity of childhood T-cell acute lymphoblastic leukemia to parthenolide

Ferrata Storti Foundation

Benjamin C. Ede,1 Rafal R Asmaro,1 John P. Moppett,1,2 Paraskevi Diamanti1,3 and Allison Blair1,3

1 School of Cellular and Molecular Medicine, University of Bristol; 2Bristol Royal Hospital for Children and 3Bristol Institute for Transfusion Sciences, NHS Blood and Transplant, UK

Haematologica 2018 Volume 103(9):1493-1501

ABSTRACT

urrent therapies for childhood T-cell acute lymphoblastic leukemia have increased survival rates to above 85% in developed countries. Unfortunately, some patients fail to respond to therapy and many suffer from serious side effects, highlighting the need to investigate other agents to treat this disease. Parthenolide, a nuclear factor kappa (κ)B inhibitor and reactive oxygen species inducer, has been shown to have excellent anti-cancer activity in pediatric leukemia xenografts, with minimal effects on normal hemopoietic cells. However, some leukemia initiating cell populations remain resistant to parthenolide. This study examined mechanisms for this resistance, including protective effects conferred by bone marrow stromal components. T-cell acute leukemia cells co-cultured with mesenchymal stem cells demonstrated significantly enhanced survival against parthenolide (73±11%) compared to cells treated without mesenchymal stem cell support (11±9%). Direct cell contact between mesenchymal cells and leukemia cells was not required to afford protection from parthenolide. Mesenchymal stem cells released thiols and protected leukemia cells from reactive oxygen species stress, which is associated with parthenolide cytotoxicity. Blocking cystine uptake by mesenchymal stem cells, using a small molecule inhibitor, prevented thiol release and significantly reduced leukemia cell resistance to parthenolide. These data indicate it may be possible to achieve greater toxicity to childhood T-cell acute lymphoblastic leukemia by combining parthenolide with inhibitors of cystine uptake.

Correspondence: allison.blair@bristol.ac.uk

Received: December 18, 2017. Accepted: May 10, 2018. Pre-published: May 17, 2018. doi:10.3324/haematol.2017.186700

Introduction The introduction of contemporary therapies for childhood T-cell acute lymphoblastic leukemia (T-ALL) has resulted in remission rates that are closer to that of B-cell precursor (BCP) ALL but survival rates remain lower and 15-20% of children with T-ALL die from relapsed/refractory disease.1 Patients with high-risk disease or those who relapse often receive more intensive treatment, making them more susceptible to toxicity and long-term secondary complications.2 This highlights the need to investigate other agents to treat this disease. It has been demonstrated that numerous cancers generate high levels of reactive oxygen species (ROS) compared to healthy tissue counterparts, where ROS levels are normally maintained in a tightly controlled manner.3 In T-ALL, ROS levels have been shown to be heightened, and this can inactivate phosphatase and the tensin homolog (PTEN), promoting leukemia cell survival.4 In human T-ALL, ROS levels are restrained by downregulation of protein kinase c theta (PKCθ) caused by NOTCH-1, a commonly activated mutation in T-ALL.5 However, if ROS stress levels are pushed above a certain threshold, cell death is forced to occur.3 Therefore, ROS promoting drugs may be an effective way of targeting cancer cells. Parthenolide (PTL) has been previously shown by ourselves and others to be a promising therapeutic agent for blood cancers.6-8 Importantly, it has limited effects haematologica | 2018; 103(9)

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1493 ©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.

1493

B.C. Ede et al. Table 1. Patients’ sample characteristics.

Patient N. 1 2 3 4 5 6 7 8 9 10

Karyotype

Sex

Age (years)

Disease status at biopsy

MRD risk statusa

46XY t(1;14) 46XY t(8;14) t(11;14), add(17) Runx1 rearr 46XY add (4), add (9) del (4), del (9) add (7), add (9), add (14)

M M M M M M M M M M

7 3 17 3 2 3 6 14 9 4

Diagnosis Diagnosis Diagnosis Diagnosis Relapse Diagnosis Diagnosis Diagnosis Diagnosis Diagnosis

Risk Risk High Low High N/A Low High Low Low

All patients treated on UKALL2011 protocol, except patients 3, 5 and 8 who were treated on UKALL 2003. aMinimal residual disease (MRD) risk status at Day 29, on respective treatment protocols. N: number; M: male.

on normal cells at the doses required to kill cancer cells. PTL can target cancer cells via numerous mechanisms, such as inhibition of nuclear factor (κ)B, p53 activation and ROS stress.6,7 However, the mechanism of PTL toxicity to T-ALL has not been defined. Parthenolide has been shown to be very effective against childhood T-ALL in vivo, with elimination of the disease and restoration of murine hemopoiesis in NOD.Cg-PrkdcscidIl2rtm1Wjl/SzJ (NSG) mice.8 However, in mice engrafted with different leukemia initiating cell populations from 2 of 9 T-ALL cases, disease progression was delayed rather than eliminated, indicating variable sensitivity of certain subpopulations to PTL. Reasons for the differences in sensitivity may be due to the effect of the in vivo microenvironment. Bone marrow (BM) stromal cells release cysteine for uptake by chronic lymphocytic leukemia (CLL) cells, driving anti-oxidative glutathione synthesis, which provides protection against ROS generating chemotherapeutic agents, such as fludarabine and oxiplatin.9 Mesenchymal stem cells (MSC) are key constituents of the BM microenvironment and have been shown to enhance protection against certain drugs in T-ALL cell lines10 and primary samples from patients with B-ALL, acute myeloid leukemia (AML) and CLL.9,11-13 Co-culture of T-ALL cell lines with MSC enhanced resistance to the anthracycline idarubicin.10 However, the role of ROS in stromal cell mediated protection in childhood ALL has not been reported. As we had previously reported resistance to PTL in T-ALL cases, in this study the cytotoxic and ROS inducing effects of the drug on primary T-ALL cells in the presence of MSC were examined to increase our understanding of PTL resistance.

Methods T-ALL and normal samples Bone marrow samples from 10 children, aged 2-17 years (median 5 years), diagnosed with T-ALL at presentation or relapse were collected with informed consent and approval of University Hospitals Bristol NHS Trust and London Brent Research Ethics Committee (Table 1). Mononuclear cells (MNC) were separated via density gradient centrifugation using FicollHypaque (Sigma-Aldrich, Gillingham, UK). MNC were sus1494

pended in 90% fetal calf serum (FCS, Thermo Scientific, Paisley, UK) and 10% dimethyl sulfoxide (DMSO, Origen Biomedical, Solihull, UK) and stored in liquid nitrogen prior to use. Samples from patients with a range of karyotypic abnormalities, diagnostic age and minimal residual disease (MRD) status were investigated. Bone marrow from a consenting healthy donor was used as a source of MSC. See the Online Supplementary Appendix for full details of MSC expansion and characterization.

Cytotoxicity assays T-cell acute lymphoblastic leukemia cells were plated in duplicate (for each drug concentration tested) at 1.2x105 cells/mL in RPMI 1640 medium (Sigma-Aldrich) containing 20% FCS, 1% Lglut and 1% Pen/Strep, hereafter referred to as suspension medium. Drugs used for assays were: PTL (Enzo Life Sciences, Exeter, UK) at 1-10 μM, N-acetyl cysteine (NAC, Sigma-Aldrich) at 15 mM and 30 μM, and sulfasalazine (SSZ, Sigma-Aldrich) at 300 µM, all prepared in suspension medium. For co-culture experiments, 5x104 MSC/mL were seeded per well and left to adhere for 24 hours (h). MSC medium was removed and replaced with suspension media containing 1x105 T-ALL cells/mL. T-ALL cells were left to settle for 1 h and then treated with 10 μM PTL with or without 300 μM SSZ, for 24 h. After treatment, non-adherent cells were removed and stained with annexin-V conjugated to fluorescein isothiocyante (Miltenyi Biotec) for 10 minutes (min). Cells were washed and stained with propidium iodide (PI, Miltenyi Biotec) prior to flow cytometric analysis. See Online Supplementary Figure S1 for details. For transwell separation experiments, 1x105 MSC/mL were seeded per well in MSC media and left to adhere for 24 h. MSC medium was removed and replaced with suspension medium. T-ALL cells were seeded at 2x105 cells/mL onto Costar Transwell inserts (0.4 μM pore size, Corning Life Sciences, Ewloe, UK) above the adherent MSC. T-ALL cells were left to settle for 1 h and then treated with PTL with or without SSZ for 24 h. After treatment, cells in transwell inserts were removed and viability was assessed by flow cytometry, as above.

Reactive oxygen species detection

Cells were treated with 5 μM of the redox sensitive probe 5(and-6)-chloromethyl-2',7'-dichlorodihydrofluorescein diacetate (CM-H2DCFDA, Thermo Fisher Scientific) for 30 min at 37°C. Cells were treated with PTL, then immediately analyzed haematologica | 2018; 103(9)

Improving T-ALL sensitivity to parthenolide

C Figure 1. Parthenolide (PTL) induces changes in reactive oxygen species (ROS) and reduced glutathione (rGSH) levels can be reversed by N-acetyl cysteine (NAC). (A) T-cell acute lymphoblastic leukemia (T-ALL) samples (patients 1, 3, 5, and 9) were pre-incubated with the ROS sensitive probe CMH2DCFDA and ROS accumulation was measured in live cells over 90 minutes by flow cytometry in response to 10 μM PTL or with PTL+15 mM NAC. Fold change in median fluorescence intensity (MFI) compared to untreated cells is shown; data represent mean ± Standard Deviation (SD). Asterisks represent significant differences between PTL and PTL+15 mM NAC. (B) Dose response curve of patient samples 2, 5, 8, and 10 that had been treated with PTL only (1-10 μM), PTL+30 μM NAC, or PTL+15mM NAC; data represent mean±SD. Top row of asterisks represent significant differences between PTL alone versus PTL+15mM NAC; bottom row of asterisks represent significant differences between PTL alone versus PTL+30 μM NAC. (C) GSH levels were measured in T-ALL samples (patients 1-9) after treatment with PTL (10 μM) for 1 hour and compared with those in untreated cells. Symbols represent the mean % change from untreated cells in replicate samples. Each symbol represents an individual patient (see Online Supplementary Figure S1 for patient symbol key). Lines represent median and interquartile range. Results were analyzed by two-way ANOVA (A and B) or paired t-test (C). *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001.

over a 90 min period by staining with PI and measuring the median fluorescence intensity (MFI) of the ROS probe by flow cytometry.

Reduced glutathione detection The level of reduced glutathione (rGSH) was detected using the GSH/GSSG-Glo luminescence assay using a plate reader (GloMax, both Promega, Chilworth, UK). Levels of rGSH were quantified in 2x104 live T-ALL cells either untreated or treated with 10 μM PTL for 1 h.

Thiol detection The thiol concentration in media was detected using a 5,5′dithiobis-(2-nitrobenzoic) (DTNB) assay. MSC were plated at 5x104 cells/mL per well in MSC medium and allowed to adhere for 24 h. MSC medium was replaced with suspension medium and left for 24 h. Subsequently, 25 μL of MSC conditioned suspension medium was removed and mixed with 100 μL TRIS buffer (0.1M) and 25 μL DTNB (2.5mM, Sigma-Aldrich) solution. The reaction was left for 10 min at room temperature, then the absorbance read at 412nm. The concentration of thiols was quantified using a linear regression equation of the standard curve generated from the absorbance of known standards of cysteine.

Cysteine transporter, xCT, knockdown in MSC Higher thiol levels are a feature of cysteine, which is usually synthesized using a specialized amino acid transport system, known as the xc- system. To investigate whether knockdown of the cysteine transporter, xCT, in MSC affected thiol production small interfering RNAs (siRNAs) for xCT and a non-coding haematologica | 2018; 103(9)

scramble control were transfected into MSC with Lipofectamine 3000 (all Thermo Fisher Scientific) using 20nM siRNA. MSC were plated at 5x104 cells/mL per well in MSC media, without Pen/Strep, containing siRNA for 24 h. MSC media was completely removed and replaced with DMEM containing 10% FCS for 24 h. To assess xCT knockdown, mRNA and protein were harvested using a PARIS kit (Thermo Fisher Scientific). (See Online Supplementary Appendix for details.)

Statistical analysis Full details of statistical analyses are provided in the Online Supplementary Appendix.

Results Cytotoxicity is linked to changes in ROS and GSH Eight of 10 samples responded to treatment with PTL, with a median half-maximal inhibitory concentration (IC50) of 7.6 μM (range 2.6-10.0 μM). The highest dose of PTL reduced the viability of samples to an average of 30±18% compared to untreated cells (Online Supplementary Figure S2A). Two cases (patients 3 and 9) were relatively resistant to PTL with viabilities remaining above 55% at the highest dose tested (Online Supplementary Figure S2B). To measure ROS levels, cells from 4 samples were pre-incubated with the ROS sensitive probe CM H2DCFDA and then treated with PTL. After 30 min, the level of ROS stress was significantly higher in PTL treated cells and it was 5-fold higher than untreated cells after 90 min (P≤0.02) (Figure 1A). Addition 1495

B.C. Ede et al. A

of the antioxidant NAC to PTL treated cells caused a complete block in PTL induced ROS stress over the 90-min period (P≤0.02) (Figure 1A). NAC (15 mM) reversed PTL induced cytotoxicity, even at the highest dose (10 μM), where T-ALL cells retained a viability of 110±22% compared to 6±3% in cells treated with PTL alone (P≤0.0001) (Figure 1B). A lower dose of NAC, 30 μM, also significantly reduced PTL cytotoxicity (P≤0.01), increasing the IC50 from 2.3 μM to 5.7 μM. As a further measure of oxidative changes, the levels of the anti-oxidative molecule rGSH were measured in PTL treated cells from samples 1-9. The levels of rGSH detected after 1 h were significantly lower in PTL treated cells with a median of 12% (range 0-58%) compared to untreated cells (P=0.02) (Figure 1C). rGSH levels were higher in the PTL resistant cases (patients 3 and 9).

MSC protection from PTL and ROS stress Mesenchymal stem cells generated from normal BM cells were highly positive for MSC associated markers; CD73 (99±0.3%), CD105 (90±3%), CD90 (99±0.2%) with low expression of hemopoietic cell markers CD45 (0.3%±0.2%) and CD34 (0.6±0.5%). MSC protected TALL cells from the cytotoxic effects of PTL (78% median survival, range 57-84) in co-culture compared to 8% (range 1496

Figure 2. Mesenchymal stem cells (MSC) protect T-cell acute lymphoblastic leukemia (TALL) cells from parthenolide (PTL)-induced cell death, reactive oxygen species (ROS) stress and decreased reduced glutathione (rGSH) levels. (A) The viability of T-ALL samples (patients 1, 2, 5, 6, and 10) after 24 hours (h) of PTL treatment (10 μM) with or without MSC in direct contact and in transwells. (B) ROS levels in patient samples 2, 5, 6, 9, and 10 after 1 h PTL treatment (10 μM) with or without MSC. (C) The concentration of rGSH in patient samples 2, 5, 6, 9, and 10 after 1 h PTL treatment (10 μM) with or without MSC. Symbols represent the average value in replicate samples. Each symbol represents an individual patient. Lines represent median and interquartile range. (D) Increase in thiol concentration in media from MSC culture after 24 h compared to blank media (n=6). Thiol concentrations were derived from the standard curve of known cysteine levels. Results were analyzed by one-way ANOVA (AC) or paired t-test (D). *P≤0.05, **P≤0.01, ***P≤0.001.

2-26%) without MSC (P=0.002) (Figure 2A). When TALL cells were seeded onto transwell inserts to prevent direct cell contact with MSC, median survival following treatment (58%, range 20-71%) was significantly higher compared to cells treated without MSC support (P=0.03). However, T-ALL survival in transwell cultures was lower than cells in direct contact with MSC (P=0.02) (Figure 2A). To ascertain whether this MSC conferred resistance to PTL is retained on removal from the supportive environment, T-ALL cells that had been pre-conditioned with MSC for 24 h were treated with PTL without further MSC support. Pre-conditioned cells showed some evidence for enhanced survival compared to cells without conditioning with an increase in IC50 from 2.6 to 3.3 μM. The increase in PTL resistance in MSC conditioned cells was modest but significant (P=0.04) (Online Supplementary Figure S3). As ROS levels are associated with PTL cytotoxicity, the ability of MSC to modulate ROS levels was investigated. T-ALL cells conditioned with MSC had significantly lower levels of ROS stress, (ROS MFI 234, range 67-380), compared to those without MSC conditioning (313, range 156506; P=0.02) (Figure 2B). When T-ALL cells were treated with PTL, ROS levels were significantly increased in both MSC conditioned (P=0.04) and unconditioned (P=0.01) cells. However, the ROS levels following PTL treatment haematologica | 2018; 103(9)

Improving T-ALL sensitivity to parthenolide

C Figure 3. Combination of parthenolide (PTL) with sulfasalazine (SSZ) overcomes protective effects of mesenchymal stem cells (MSC). (A) Viability of acute lymphoblastic leukemia (ALL) samples (patients 2, 5, 6, 9, and 10) treated with PTL (10 μM), SSZ (300 μM) or a combination of both agents for 24 hours (h) with or without MSC. (B) Viability of samples (patients 1, 2, 5, 6, and 10) following 24-h exposure to PTL (10 μM) or a combination of PTL with SSZ (300 μM) separated from MSC by transwell inserts. Symbols represent the average viability in replicate samples. Each symbol represents an individual patient. Lines represent median and interquartile range. (C) Thiol concentration in media following incubation with MSC with or without SSZ treatment (n=4). Data represent mean±Standard Deviation. Results were analyzed by one-way ANOVA (A) or paired t-test (B and C). *P≤0.05, **P≤0.01.

were significantly lower in MSC conditioned cells (353, range 200-617) compared to those cultured without MSC (573, range 321-796; P=0.05) (Figure 2B). To examine whether these observed protective effects were specific to MSC, T-ALL cells were co-cultured on BM stromal cells and fibroblasts. PTL toxicity was reduced in co-cultures containing stromal cells and fibroblasts but not to the same extent as cultures containing MSC (Online Supplementary Figure S4A). T-ALL cells conditioned with stroma had similar ROS levels to those conditioned with MSC. ROS levels in cells cultured on fibroblasts were similar to those observed in unsupported cultures (Online Supplementary Figure S4B). Parthenolide is known to lower rGSH levels in leukemia cells, therefore the ability of MSC to modulate this effect was examined. T-ALL cells conditioned with MSC had significantly higher levels of rGSH (median 131nM, range 108-135nM) compared to those without conditioning (82nM, range 60-99nM; P=0.03). Following PTL treatment, rGSH levels were significantly decreased in both MSC conditioned (P=0.04) and unconditioned (P=0.0003) cells, indicating reduced anti–oxidant activity. However, rGSH levels were still higher in MSC conditioned cells (median 87nM, range 66-117nM) compared to those without MSC conditioning (26nM, range 7-35nM; P=0.01) (Figure 2C). As cysteine is crucial for GSH synthesis in leukemia, it is possible that MSC release cysteine into the surrounding media, thereby moderating the anti-oxidant effects of PTL. A thiol detection assay was used, since cysteine contains a free thiol group, to investigate cysteine release. Thiol concentrations were significantly higher in media removed from MSC cultures (91±17 μM) compared to haematologica | 2018; 103(9)

background media levels (62±17 μM; P=0.0002), representing a 1.5-fold increase over background (Figure 2D).

Blocking thiol release overcomes protective effect The thiol functional group plays a major role in intracellular anti-oxidant defenses. Cysteine residues eliminate ROS, usually by converting them to 2H2O or H2O and O2, and reduce oxidized protein thiols.14 SSZ, an inhibitor of the cystine uptake antiporter xc-, was used to inhibit cystine uptake and block subsequent cysteine generation and release in MSC, to determine what effect this would have on the protection conferred by MSC. The combination of PTL and SSZ significantly reduced the median viability of T-ALL cells with MSC conditioning to 38% (range 1142%) compared to PTL (71%, range 59-95%; P=0.002) or SSZ alone (88%, range 69-109%; P=0.04) (Figure 3A). The treatment combination reduced viabilities to levels comparable to those observed in cells treated with both agents without MSC conditioning (16%, range 5-24%). In the absence of MSC, SSZ had low cytotoxicity (92% median survival, range 77-102%) and when combined with PTL there was no significant difference in toxicity compared to PTL alone (P=0.9). Addition of PTL+ SSZ to T-ALL cells conditioned with BM stroma or fibroblasts also reduced leukemia cell survival, although to a lesser extent (Online Supplementary Figure S4C). Interestingly, the effects of combining these agents were more remarkable in samples that were relatively resistant to PTL. Fiftysix percent of cells from patient 9 survived treatment with PTL, but viability was reduced to only 6% when PTL was combined with SSZ. Leukemia burden in NSG mice engrafted with this sample was significantly reduced when treated with PTL and SSZ compared to placebos 1497

B.C. Ede et al. A

C Figure 4. xCT knockdown in mesenchymal stem cells (MSC) overcomes MSC mediated resistance to parthenolide (PTL). Viability of T-cell acute lymphoblastic leukemia (T-ALL) samples (patients 5, 6, 8-10) treated with 10 μM PTL for 24 hours (h) in the presence of MSC pre-treated with xCT or scramble control siRNA, in direct contact (A) or samples (1, 2, 6, 8, and 10) in transwell inserts (B). Symbols represent the average viability in duplicate samples. Each symbol represents an individual patient. Lines represent median and interquartile range. (C) Thiol concentration in media following 24-h incubation with MSC pre-treated with xCT or scramble control siRNA compared to background media. Data represent mean±Standard Deviation. (n=4). Results were analyzed by one-way ANOVA (A) or paired t-tests (B and C). *P≤0.05, **P≤0.01.

(P<0.0001) (Online Supplementary Figure S5). Using SSZ alone and in combination with PTL had minimal effects on normal hemopoietic cells (Online Supplementary Figure S6). In transwell experiments, using SSZ with PTL abrogated the protective effects of MSC support, reducing T-ALL survival from 58% (range 20-71%) to 14% (range 9-43%; P=0.01) (Figure 3B). Media harvested from cultures of SSZ treated MSC contained thiol levels similar to background media (53±19 μM vs. 61±10 μM, respectively). In contrast, media from untreated MSC contained significantly higher thiol levels (90±11μM; P=0.005) compared to media from SSZ treated cells, representing a 1.5-fold increase over background media (Figure 3C). As a second approach to block the cystine uptake antiporter xc-, and thereby prevent cysteine release from MSC, xCT was targeted with siRNAs. Relative gene expression levels of xCT were significantly lower in MSC 48 h after transfection with either siRNA 1 (17±9%) or siRNA 2 (9±4%) compared to scramble control siRNA (P≤0.0001) (Online Supplementary Figure S7A). Protein available from xCT siRNA-1 showed a significant reduction in xCT expression (3±0.1%) relative to scramble control levels quantified by western blotting densitometry (P=0.002) (Online Supplementary Figure S7B). The viability of PTL treated T-ALL cells, co-cultured with MSC treated with xCT siRNA-1 (27%, range 3-28%) or -2 (35%, range 24-42%), was significantly lower than cells co-cultured with scramble control MSC (49%, range 46-72%; P<0.05) (Figure 4A). As siRNA-1 provided the best reduction in MSC protection, it was selected for further experiments. When transwell inserts were used to separate xCT knockdown MSC from T-ALL cells, the median viability after PTL treatment was significantly 1498

lower (13%, range 7-30%) compared to scramble control MSC cells (24%, range 15-74%; P=0.05) (Figure 4B). Scramble control MSC media had a significantly higher thiol concentration (86±2 μM) compared to media harvested from xCT knockdown MSC (46±1 μM; P=0.002) (Figure 4C). Media from xCT knockdown MSC did not show any increase in thiol levels above background media (53±1 μM). To confirm that the observed effects of interfering with xCT were a result of blocking cystine uptake, and not a result of toxicity, the viability and morphology of treated MSC were assessed. Treatment with PTL, SSZ or both agents in combination had no effect on the viability and confluency of MSC (P≥0.37) (Figure 5A) nor did siRNA treatment (P≥0.82) (Figure 5B and Online Supplementary Figure S8). This was also confirmed by flow cytometry (Figure 5C).

Discussion To date, PTL is the only drug that has been shown to be capable of completely eradicating childhood ALL in NSG xenografts, as a single agent.8 Most studies, using such models, report reduction in leukemia burden but levels often increase on cessation of treatment. Consequently, there is much interest in the application of PTL for cancer therapy. Several groups are developing strategies to improve the bioavailability of PTL, without having detrimental effects on its pharmacokinetic properties.15-19 PTL can be successfully sequestered into nanoscopic vectors which can achieve equivalent toxicity to unmodified PTL.1618 Nanoparticle PTL formulations can be used in vivo at 40fold lower doses with 20-fold lower administration frehaematologica | 2018; 103(9)

Improving T-ALL sensitivity to parthenolide

C Figure 5. Effects of parthenolide (PTL) and xCT modifying agents on mesenchymal stem cell (MSC) viability. (A) Average live MSC counts from 3 fields (x10 magnification) following treatment with PTL (10 μM), sulfasalazine (SSZ) (300 μM), or a combination of both agents measured by fluorescence microscopy. (B) Average live MSC counts from 3 fields (x10 magnification) following treatment with PTL (10 μM), pre-treated with xCT, or scramble control siRNA. (C) Viability of MSC following treatment with PTL (10 μM) and SSZ (300 μM), or PTL after pre-treatment with xCT siRNA, measured by flow cytometry. Treatment with 70% ethanol served as a positive control for inducing toxicity. Data represent mean±Standard Deviation (n=3). Live cell counts by fluorescence microscopy were calculated using the ImageJ particle analysis software. Results were analyzed by one-way ANOVA (A-C). ***P≤0.001.

quency than standard PTL.16 Some of these delivery systems are inexpensive and formulations are readily scalable.18 Consequently, they should facilitate the use of PTL at clinically relevant doses. However, the efficacy of many therapeutic drugs may be compromised by the host BM microenvironment. Therapy-induced niches, which protect leukemia cells against standard first-line induction agents, have been described in ALL.12,20 In the present study, we investigated whether resistance to PTL, reported in a minority of T-ALL cases,8 may be conferred by BM-derived MSC. In vitro PTL treatment reduced viability of T-ALL cells to less than 30%, confirming previous results in a separate cohort of pediatric cases.8 While there were differences in the responses of individual patient samples, there was no correlation between PTL cytotoxicity with karyotype or MRD risk status, which may be a result of the heterogeneous nature of this disease. PTL was shown to increase ROS stress and lower rGSH levels in T-ALL. The antioxidative compound NAC blocked ROS upregulation and diminished PTL cytotoxicity, suggesting that PTL toxicity is, at least in part, related to ROS stress in T-ALL. These findings concur with reports of increased ROS stress in primary AML and CLL cell lines following PTL treatment.6,21,22 The protective effect provided by NAC may be attributed to the role of cysteine as the rate limiting amino acid in rGSH production.23 Reduction of cystine to cysteine and subsequent supply to leukemia cells is crucial for GSH synthesis in these cells. GSH levels elevate the anti-oxidative capacity of cells, which may provide protection against PTL. In AML and breast cancer, populations of cells endure lower levels of ROS stress and these cells are more resistant to therapy.24,25 It is unclear whether ROS is actively eliminated by rGSH or if PTL directly interacts with rGSH, leading to a reduction in anti-oxidative capacity and increased ROS accumulation. One function of rGSH is to detoxify cells from reactive electrophiles,26 which PTL contains in the haematologica | 2018; 103(9)

form of an a,β-unsaturated carbonyl group. PTL can directly interfere with rGSH synthesis in AML, causing depletion of rGSH, allowing ROS levels to increase.21 Whether PTL directly blocked rGSH synthesis was beyond the scope of this study. However, it is evident that ROS levels increased and rGSH levels dropped, putting the cells under higher levels of stress, which is likely to drive apoptosis. As these data indicate that ROS levels may be linked to PTL cytotoxicity, it is possible that the BM microenvironment provides resistance to drug activity by protecting leukemia cells from ROS stress. CLL cells had decreased sensitivity to the ROS inducing agents fludarabine and oxiplatin when co-cultured with BM stromal cells.9 Chemoprotection was derived from the generation and release of cysteine by stromal cells for uptake by CLL cells. The BM microenvironment can also play a role in leukemia drug resistance, mediated by a diverse set of mechanisms, both by direct cell adhesion11,13,27 and/or soluble factor release.12,20,28,29 In this study, the protective effects of MSC against PTL toxicity were related to decreases in ROS stress and preserved rGSH levels, suggesting that MSC increased the anti-oxidative stress capacity of T-ALL cells and thereby conferred resistance to PTL. To eliminate variability, MSC were generated from a single normal BM donor, allowing direct comparison of the effects on the patients' samples. This may have introduced bias but our results are consistent with those reported in CLL.9 These data also suggest that protection is provided by one or more secretable factors, as preventing direct contact with MSC still provided TALL cells with protection against PTL. However, protection was not to the same extent as observed in direct contact, suggesting involvement of other factors. Mesenchymal stem cell conditioned media contained significantly higher levels of thiols, an important feature of the anti-oxidative compound cysteine, compared to unconditioned media. A key mechanism by which normal cells synthesize cysteine is via intake of extracellular cystine 1499

B.C. Ede et al.

Figure 6. Proposed mechanism of action of parthenolide (PTL) and the protective effect provided by mesenchymal stem cells (MSC). PTL causes apoptosis by increasing reactive oxygen species (ROS) stress and decreasing reduced glutathione (rGSH) resulting in death of T-cell acute lymphoblastic leukemia (T-ALL) cells (A). MSC express high levels of the cystine glutamate antiporter xc-, facilitated by the antiporter protein xCT. Extracellular cystine is taken up by MSC and reduced into cysteine. Cysteine is released into the extracellular space for uptake by T-ALL cells, blocking ROS induction and preventing apoptosis (B). Blocking the xCT system using sulfasalazine (SSZ) or siRNA reduces protective effects provided by MSC (C).

using a specialized amino acid transport mechanism known as the xc– system.14 The system specifically mediates cystine and glutamate exchange across the cell membrane and is comprised of two subunits: the light chain transporter subunit SLC7A11 (or xCT) and the heavy subunit SLC3A2 (4F2hc).30,31 Cysteine is unstable outside of cells and is rapidly oxidized to form a disulphide bridge with another cysteine amino acid, forming cystine. However, upon entry into the cell, cystine enters a stronger reducing environment and is converted into cysteine. Blocking xc– activity with a small molecule inhibitor, SSZ,32 or by knockdown of the xc– light chain component xCT, prevented thiol release and significantly reduced MSC protection to levels close to those observed in cells without MSC support. In addition, leukemia cell survival in a PTL resistant case was significantly decreased when SSZ or xCT siRNA were used to block xc– activity. Furthermore, PTL and SSZ treatment in vivo resulted in significantly reduced leukemia burden in engrafted NSG mice compared to controls. Similar results have been reported in BCP-ALL, where cell viability was diminished in a subset of patients following treatment with cysteine dioxygenase, which catalyzes conversion of cysteine into cysteine sulfinic acid, thereby bypassing GSH synthesis.33 Together, these findings indicate that MSC release cysteine, which confers a survival advantage on leukemia cells. The functional contribution of thiols was further confirmed by the demonstration that T-ALL cells exposed to the thiol containing compound NAC, at an equivalent thiol concentration to that in MSC conditioned media, were more resistant to PTL. The fact that there was still a small protective effect against PTL after interfering with the xc- system, indicates that cysteine release may not be the sole protective mechanism in these experiments. A recent study found that ALL cells release extracellular vesicles that can be taken up by MSC, causing a shift to glycolysis metabolism. The switch to glycolysis led to an increased release of the metabolite lactate, which might be used as an additional source of energy by leukemia cells and confer chemoresistance.34 While MSC are a fundamental component of the BM niche, and contribute to its formation in vivo, other stromal 1500

cells may have a role in leukemia maintenance and protection from therapeutic agents. We demonstrated that BM stromal cells and fibroblasts also conferred protection against PTL. ALL cells themselves can alter the endosteal and vascular compartments of the niche12 and induce apoptosis of osteoblast cells.17 The remodelled niche is dynamically transient and on exposure to chemotherapeutic agents, ALL cells can release CCL3 and cytokines TGF-β1 and GDF15, conferring chemoresistance by activating the TGF-β signaling pathway.12 It may be possible to overcome this resistance by disrupting the interactions between leukemia cells and the BM environment. This is the first report demonstrating MSC provide a protective effect to T-ALL cells against PTL. Moreover, we have shown that targeting the xc- system can overcome this effect (see overview in Figure 6), adding to the evidence that targeting xc– enhances the efficacy of anti-cancer agents in leukemias9 and in several solid cancer models.35-38 MSC viability was unaffected by PTL and targeting xc-, so damage to this important element of the BM environment39,40 should be minimal. A logical progression of this work will be determining whether the effects of PTL, or indeed current chemotherapeutic agents, in vivo can be enhanced using an xc- inhibitor like SSZ. Furthermore, with developments in nanoscopic drug delivery vectors, it may be possible to use dual loaded (PTL+SSZ) nanovectors to overcome selective patient resistance towards one drug and the chemoprotective effects of the leukemia microenvironment in vivo. Acknowledgments The authors would like to thank Dr Jeremy Hancock, Mr Paul Virgo and staff of Bristol Genetics Laboratory, Southmead Hospital for excellent technical assistance. We also thank Elinor Curnow, Statistics and Clinical Studies, NHS Blood and Transplant, consultants and oncology staff at Bristol Royal Hospital for Children. We are grateful to the patients and their families who gave permission for their cells to be used for research. This work was supported by a generous donation from Mr Richard Cunningham and by grants from the Department of Health and NHS Blood and Transplant. haematologica | 2018; 103(9)

Improving T-ALL sensitivity to parthenolide

References 1. Schrappe M, Hunger SP, Pui CH, et al. Outcomes after induction failure in childhood acute lymphoblastic leukemia. N Engl J Med. 2012;366(15):1371-1381. 2. Hunger SP, Mullighan CG. Acute Lymphoblastic Leukemia in Children. N Engl J Med. 2015;373(16):1541-1552. 3. Trachootham D, Alexandre J, Huang P. Targeting cancer cells by ROS-mediated mechanisms: a radical therapeutic approach? Nature Rev Drug Discov. 2009;8(7):579-591. 4. Silva A, Yunes JA, Cardoso BA, et al. PTEN posttranslational inactivation and hyperactivation of the PI3K/Akt pathway sustain primary T cell leukemia viability. J Clin Invest. 2008;118(11):3762-3774. 5. Giambra V, Jenkins CR, Wang H, et al. NOTCH1 promotes T cell leukemia-initiating activity by RUNX-mediated regulation of PKC-theta and reactive oxygen species. Nat Med. 2012;18(11):1693-1698. 6. Guzman ML, Rossi R, Karnischky L, et al. The sesquiterpene lactone parthenolide induces apoptosis of human acute myelogenous leukemia stem and progenitor cells. Blood. 2005;105(11):4163-4169. 7. Steele AJ, Jones DT, Ganeshaguru K, et al. The sesquiterpene lactone parthenolide induces selective apoptosis of B-chronic lymphocytic leukemia cells in vitro. Leukemia. 2006;20(6):1073-1079. 8. Diamanti P, Cox CV, Moppett JP, Blair A. Parthenolide eliminates leukemia-initiating cell populations and improves survival in xenografts of childhood acute lymphoblastic leukemia. Blood. 2013;121(8):1384-1393. 9. Zhang W, Trachootham D, Liu J, et al. Stromal control of cystine metabolism promotes cancer cell survival in chronic lymphocytic leukaemia. Nat Cell Biol. 2012;14(3):276-286. 10. Wu KN, Zhao YM, He Y, et al. Rapamycin interacts synergistically with idarubicin to induce T-leukemia cell apoptosis in vitro and in a mesenchymal stem cell simulated drugresistant microenvironment via Akt/mammalian target of rapamycin and extracellular signal-related kinase signaling pathways. Leuk Lymphoma. 2014;55(3): 668-676. 11. Jacamo R, Chen Y, Wang Z, et al. Reciprocal leukemia-stroma VCAM-1/VLA-4-dependent activation of NF-kappaB mediates chemoresistance. Blood. 2014;123(17):26912702. 12. Duan CW, Shi J, Chen J, et al. Leukemia propagating cells rebuild an evolving niche in response to therapy. Cancer Cell. 2014;25(6):778-793. 13. Xia B, Tian C, Guo S, et al. c-Myc plays part in drug resistance mediated by bone marrow stromal cells in acute myeloid leukemia. Leuk Res. 2015;39(1):92-99. 14. Lewerenz J, Hewett SJ, Huang Y, et al. The

haematologica | 2018; 103(9)

15.

16.

17.

18.

19.

20.

21.

22.

23. 24.

25.

26. 27.

28.

cystine/glutamate antiporter system x(c)(-) in health and disease: from molecular mechanisms to novel therapeutic opportunities. Antioxid Redox Signal. 2013;18(5):522-555. Baranello MP, Bauer L, Jordan CT, Benoit DSW. Micelle Delivery of Parthenolide to Acute Myeloid Leukemia Cells. Cell Mol Bioneng. 2015;8(3):455-470. Zong H, Sen S, Zhang G, et al. In vivo targeting of leukemia stem cells by directing parthenolide-loaded nanoparticles to the bone marrow niche. Leukemia. 2016;30(7): 1582-1586. Deller RC, Diamanti P, Morrison G, et al. Functionalized Triblock Copolymer Vectors for the Treatment of Acute Lymphoblastic Leukemia. Mol Pharm. 2017;14(3):722-732. Ridolfo R, Ede B, Diamanti P, et al. Drug loaded nanovectors via direct hydration as a new platform for cancer therapeutics. Small. 2018 Jul 12:e1703774. doi: 10.1002/ smll.201703774. [Epub ahead of print] Tyagi V, Alwaseem H, O'Dwyer KM, et al. Chemoenzymatic synthesis and antileukemic activity of novel C9- and C14functionalized parthenolide analogs. Bioorg Med Chem. 2016;24(17):3876-3886. Hawkins ED, Duarte D, Akinduro O, et al. T-cell acute leukaemia exhibits dynamic interactions with bone marrow microenvironments. Nature. 2016;538(7626):518-522. Pei S, Minhajuddin M, Callahan KP, et al. Targeting aberrant glutathione metabolism to eradicate human acute myelogenous leukemia cells. J Biol Chem. 2013;288 (47):33542-33558. Zunino SJ, Ducore JM, Storms DH. Parthenolide induces significant apoptosis and production of reactive oxygen species in high-risk pre-B leukemia cells. Cancer Lett. 2007;254(1):119-127. Lu SC. Glutathione synthesis. Biochim Biophys Acta. 2013;1830(5):3143-3153. Lagadinou ED, Sach A, Callahan K, et al. BCL-2 inhibition targets oxidative phosphorylation and selectively eradicates quiescent human leukemia stem cells. Cell Stem Cell. 2013;12(3):329-341. Diehn M, Cho RW, Lobo NA, et al. Association of reactive oxygen species levels and radioresistance in cancer stem cells. Nature. 2009;458(7239):780-783. Hayes JD, Flanagan JU, Jowsey IR. Glutathione transferases. Annu Rev Pharmacol Toxicol. 2005;45:51-88. Winter SS, Sweatman JJ, Lawrence MB, Rhoades TH, Hart AL, Larson RS. Enhanced T-lineage acute lymphoblastic leukaemia cell survival on bone marrow stroma requires involvement of LFA-1 and ICAM-1. Br J Haematol. 2001;115(4):862-871. Iwamoto S, Mihara K, Downing JR, Pui CH, Campana D. Mesenchymal cells regulate the response of acute lymphoblastic leukemia cells to asparaginase. J Clin Invest.

2007;117(4):1049-1057. 29. Laranjeira AB, de Vasconcellos JF, Sodek L, et al. IGFBP7 participates in the reciprocal interaction between acute lymphoblastic leukemia and BM stromal cells and in leukemia resistance to asparaginase. Leukemia. 2012;26(5):1001-1011. 30. Sato H, Tamba M, Ishii T, Bannai S. Cloning and expression of a plasma membrane cystine/glutamate exchange transporter composed of two distinct proteins. J Biol Chem. 1999;274(17):11455-11458. 31. Sato H, Tamba M, Kuriyama-Matsumura K, Okuno S, Bannai S. Molecular cloning and expression of human xCT, the light chain of amino acid transport system xc. Antioxid Redox Signal. 2000;2(4):665-671. 32. Aoki Y, Watanabe T, Saito Y, et al. Identification of CD34+ and CD34leukemia-initiating cells in MLL-rearranged human acute lymphoblastic leukemia. Blood. 2015;125(6):967-980. 33. Boutter J, Huang Y, Marovca B, et al. Imagebased RNA interference screening reveals an individual dependence of acute lymphoblastic leukemia on stromal cysteine support. Oncotarget. 2014;5(22):11501-11512. 34. Johnson SM, Dempsey C, Chadwick A, et al. Metabolic reprogramming of bone marrow stromal cells by leukemic extracellular vesicles in acute lymphoblastic leukemia. Blood. 2016;128(3):453-456. 35. Yoshikawa M, Tsuchihashi K, Ishimoto T, et al. xCT inhibition depletes CD44v-expressing tumor cells that are resistant to EGFRtargeted therapy in head and neck squamous cell carcinoma. Cancer Res. 2013;73(6):1855-1866. 36. Balza E, Castellani P, Delfino L, Truini M, Rubartelli A. The pharmacologic inhibition of the xcâ&#x20AC;&#x201C; antioxidant system improves the antitumor efficacy of COX inhibitors in the in vivo model of 3-MCA tumorigenesis. Carcinogenesis. 2013;34(3):620-626. 37. Lanzardo S, Conti L, Rooke R, et al. Immunotargeting of Antigen xCT Attenuates Stem-like Cell Behavior and Metastatic Progression in Breast Cancer. Cancer Res. 2016;76(1):62-72. 38. Liu R, Blower PE, Pham AN, et al. Cystineglutamate transporter SLC7A11 mediates resistance to geldanamycin but not to 17(allylamino)-17-demethoxygeldanamycin. Mol Pharmacol. 2007;72(6):1637-1646. 39. Le Blanc K, Samuelsson H, Gustafsson B, et al. Transplantation of mesenchymal stem cells to enhance engraftment of hematopoietic stem cells. Leukemia. 2007;21(8):17331738. 40. Koc ON, Gerson SL, Cooper BW, et al. Rapid hematopoietic recovery after coinfusion of autologous-blood stem cells and culture-expanded marrow mesenchymal stem cells in advanced breast cancer patients receiving high-dose chemotherapy. J Clin Oncol. 2000;18(2):307-316.

1501

ARTICLE

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

Haematologica 2018 Volume 103(9):1502-1510

Sustained efficacy and detailed clinical follow-up of first-line ibrutinib treatment in older patients with chronic lymphocytic leukemia: extended phase 3 results from RESONATE-2

Paul M. Barr,1 Tadeusz Robak,2 Carolyn Owen,3 Alessandra Tedeschi,4 Osnat Bairey,5,6 Nancy L. Bartlett,7 Jan A. Burger,8 Peter Hillmen,9 Steven Coutre,10 Stephen Devereux,11 Sebastian Grosicki,12 Helen McCarthy,13 Jianyong Li,14 David Simpson,15 Fritz Offner,16 Carol Moreno,17 Cathy Zhou,18 Lori Styles,18 Danelle James,18 Thomas J. Kipps19 and Paolo Ghia20

University of Rochester, NY, USA; 2Medical University of Lodz, Poland; 3Tom Baker Cancer Centre, Calgary, AB, Canada; 4ASST Grande Ospedale Metropolitano Niguarda, Milan, Italy; 5Rabin Medical Center, Beilinson Campus, Petah Tikva, Israel; 6Sackler Faculty of Medicine, Tel Aviv University, Israel; 7Washington University School of Medicine, St. Louis, MO, USA; 8University of Texas MD Anderson Cancer Center, Houston, TX, USA; 9The Leeds Teaching Hospitals, St. James Institute of Oncology, UK; 10 Stanford University School of Medicine, CA, USA; 11Kings College Hospital, NHS Foundation Trust, London, UK; 12School of Public Health, Silesian Medical University, Katowice, Poland; 13Royal Bournemouth Hospital, UK; 14Jiangsu Province Hospital, Nanjing, China; 15North Shore Hospital, Auckland, New Zealand; 16Universitair Ziekenhuis Gent, Belgium; 17Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; 18 Pharmacyclics, LLC, an AbbVie Company, Sunnyvale, CA, USA; 19University of California, San Diego, Moores Cancer Center, La Jolla, CA, USA and 20Università Vita-Salute San Raffaele and IRCCS Istituto Scientifico San Raffaele, Milan, Italy 1

ABSTRACT

Correspondence: Paul_Barr@URMC.Rochester.edu

Received: February 28, 2018. Accepted: June 4, 2018. Pre-published: June 7, 2018. doi:10.3324/haematol.2018.192328 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1502 ©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.

1502

esults of RESONATE-2 (PCYC-1115/1116) supported approval of ibrutinib for first-line treatment of chronic lymphocytic leukemia. Extended analysis of RESONATE-2 was conducted to determine long-term efficacy and safety of ibrutinib in older patients with chronic lymphocytic leukemia. A total of 269 patients aged ≥65 years with previously untreated chronic lymphocytic leukemia without del(17p) were randomized 1:1 to ibrutinib (n=136) or chlorambucil (n=133) on days 1 and 15 of a 28-day cycle for 12 cycles. Median ibrutinib treatment duration was 28.5 months. Ibrutinib significantly prolonged progression-free survival versus chlorambucil (median, not reached vs. 15 months; hazard ratio, 0.12; 95% confidence interval, 0.07-0.20; P<0.0001). The 24-month progression-free survival was 89% with ibrutinib (97% and 89% in patients with del[11q] and unmutated immunoglobulin heavy chain variable region gene, respectively). Progression-free survival rates at 24 months were also similar regardless of age (<75 years [88%], ≥75 years [89%]). Overall response rate was 92% (125/136). Rate of complete response increased substantially from 7% at 12 months to 18% with extended follow up. Greater quality of life improvements occurred with ibrutinib versus chlorambucil in Functional Assessment of Chronic Illness Therapy-Fatigue (P=0.0013). The most frequent grade ≥3 adverse events were neutropenia (12%), anemia (7%), and hypertension (5%). Rate of discontinuations due to adverse events was 12%. Results demonstrated that firstline ibrutinib for elderly patients with chronic lymphocytic leukemia provides sustained response and progression-free survival benefits over chemotherapy, with depth of response improving over time without new toxicity concerns. This trial was registered at clinicaltrials.gov identifier 01722487 and 01724346. haematologica | 2018; 103(9)

Extended Phase 3 Results From RESONATE-2

Introduction Chronic lymphocytic leukemia (CLL) is the most common leukemia in Western countries and is increasing in prevalence with the prolonged survival observed with introduction of novel combinations and targeted treatments such as ibrutinib.1 With a median age at diagnosis of 71 years,1 management of this predominately older population is controversial given that frequent comorbidities often preclude aggressive therapy. Randomized studies have provided disparate results in older compared with younger patients.2,3 Less intensive approaches, such as chlorambucil, provide limited response durability. While the addition of anti-CD20 antibodies has improved outcomes achieved with single-agent chlorambucil, administration of these intravenous agents has associated toxicity, and response durations remain limited.4 The first-in-class, oral, once-daily, Bruton’s tyrosine kinase (BTK) inhibitor ibrutinib targets signaling via the B-cell receptor cascade, critical to survival of malignant lymphocytes.5-9 Ibrutinib demonstrated tolerability, a high rate of objective responses, and prolongation of progression-free survival and overall survival in patients with relapsed/refractory CLL.10 Early-phase studies demonstrated responses of up to 84% in previously untreated patients, with complete response (CR) rates of up to 23% and up to 3 years of median follow up.11,12 This small cohort suggested that single-agent ibrutinib might provide durable efficacy in first-line treatment of patients with CLL while avoiding toxicity inherent to cytotoxic or other infused regimens. RESONATE-2 was an international phase 3 study designed to definitively evaluate first-line ibrutinib treatment in older patients who often had baseline frailties against a standard chemotherapeutic agent, chlorambucil.13 Primary results demonstrated an 84% reduction in the risk of death at a median follow up of 18 months for ibrutinib compared with chlorambucil. Based on these findings, ibrutinib received approval in the United States, Europe, and other regions for the first-line treatment of patients with CLL, and allows for treatment without chemotherapy.14,15 A detailed analysis of overall survival (OS) with longer follow up and adjustment for the impact of treatment crossover was previously reported.16 A separate data cut was subsequently performed after this detailed OS analysis to evaluate additional outcomes after long-term follow up. Herein, we present the extended analysis of additional outcomes from RESONATE-2 including quality-of-life (QOL) measures that may help guide appropriate use of ibrutinib for previously untreated patients.

Methods Study design and population Eligible patients for RESONATE-2 (PCYC-1115/1116; clinicaltrials.gov identifier 01722487/01724346) had previously untreated CLL or SLL with active disease and were ≥65 years. Patients ≤70 years of age must have had a comorbidity that precluded treatment with fludarabine-cyclophosphamide-rituximab. Eligible patients had an Eastern Cooperative Oncology Group (ECOG) performance status ≤2, absolute neutrophil count ≥1000 cells/mm3, platelet count ≥50,000/mm3, and adequate liver and kidney function. Those with del(17p) CLL were excluded. haematologica | 2018; 103(9)

This study was conducted according to principles of the Declaration of Helsinki and the International Conference on Harmonisation Guidelines for Good Clinical Practice and approved by the institutional review boards of participating institutions. All patients provided written informed consent. Patients were randomly assigned in a 1:1 ratio to treatment with oral ibrutinib, 420 mg once daily until disease progression or chlorambucil, 0.5 mg/kg (increased up to 0.8 mg/kg based on tolerability) on days 1 and 15 of a 28-day cycle for 12 cycles. Patients from the chlorambucil treatment arm with independent review committee (IRC)-confirmed disease progression were eligible to cross over to second-line treatment with ibrutinib at the investigator’s discretion.

End points and assessments End points included progression-free survival (PFS, defined as time from randomization to progression or death, whichever occurs earlier), overall survival (OS), overall response rate (ORR), improvement in hematologic variables, patient-reported healthrelated QOL, and safety. Disease progression and response was determined by investigator. QOL was assessed using the Functional Assessment of Chronic Illness Therapy (FACIT)Fatigue questionnaires. Safety assessments included adverse events (AEs) and laboratory parameters. Non-hematologic AEs were graded using Common Terminology Criteria for Adverse Events, v4.03. Hematologic AEs were graded using International Workshop on CLL criteria.17

Statistical analyses PFS and OS were analyzed using Kaplan-Meier estimates and a 2-sided log-rank test stratified by the randomization factors. Sensitivity analyses were performed to adjust for the impact of crossover on OS as previously described.16 ORR was analyzed with the Cochran-Mantel-Haenszel c2 test, stratified by the randomization factors. QOL analyses were based on the proportion of patients with clinically meaningful changes in scores from baseline (≥3 points for FACIT-Fatigue). Additional QOL analyses used time-dependent mixed-models repeated measures analysis.

Results Patients There were 269 patients randomly assigned to ibrutinib (n=136) or chlorambucil (n=133) monotherapy in the RESONATE-2 study (Online Supplementary Figure S1).13 Patient characteristics were well balanced across treatment arms, as previously reported (Table 1).13 The median patient age was 73 years on the ibrutinib arm and 72 years on the chlorambucil arm. In the ibrutinib treatment arm, of those evaluated, 22% (29/130) had del(11q), and 48% (58/121) had unmutated IGHV. Patients initiated ibrutinib treatment for active disease per iwCLL criteria, most commonly manifesting as marrow failure (progressive anemia or thrombocytopenia [40%]), progressive or symptomatic lymphadenopathy (40%) or splenomegaly (26%), with many patients having more than one indication for treatment including disease symptoms such as fatigue or night sweats (Table 1). Although 32% of patients had substantial fatigue when entering study, only 5% were started on CLL treatment solely for fatigue that was considered to have interfered with work or usual activities. With a median follow up for this extended analysis of 29 months (maximum, 36 months), 107 patients (79%) remain on first-line ibrutinib. 1503

P.M. Barr et al.

Survival outcomes Ibrutinib resulted in significantly longer PFS compared with chlorambucil (median, PFS not reached vs. 15.0 months; Figure 1A). There was an 88% reduction in risk of PFS events (progression or death; hazard ratio [HR], 0.12; 95% CI, 0.07-0.20; P<0.0001) for patients randomized to ibrutinib. PFS at 24 months was 89% with ibrutinib versus 34% with chlorambucil. This rate was relatively stable with ibrutinib with an 18-month PFS of 94%. Ibrutinib consistently demonstrated significant improvements in PFS for patients in all subgroups including those considered high risk (Figure 2). In patients treated with ibrutinib, only 1 patient with del(11q) has had disease progression, and the rates of 24-month PFS were 97% and 86% for those with or without del(11q), respectively (Figure 1B). No significant difference was observed in the PFS of patients with unmutated versus mutated IGHV (24month PFS, 90% and 89%, respectively; Figure 1C). PFS benefits were consistent across additional subgroups of patients, including those with advanced disease (Rai stage 3 or 4) or bulky disease (Figure 2). PFS and OS rates were also similar regardless of age (24-month PFS, <75 years [88%], ≥75 years [89%]; OS, <75 years [94%], ≥75 years [96%]; Figure S2). With longer follow up and despite patient crossover, ibrutinib continues to demonstrate an OS benefit compared with chlorambucil (HR, 0.43; 95% CI, 0.21-0.86; P=0.0145; Online Supplementary Figure S3 and Table S1), with a 24-month OS of 95% for ibrutinib vs. 84% for chlorambucil (Online Supplementary Figure S3).

Responses for ibrutinib-treated patients With a maximum of 36 months of follow up, the ORR with ibrutinib treatment was 92% (Table 2). Eighteen percent of patients achieved CR, which improved from 7% at 12 months and 15% at 24 months (Figure 3). Comparable ORR and CR rates were also observed in high-risk subgroups, including those with del(11q) (ORR, 100%; CR rate, 14%) or unmutated IGHV (ORR, 95%; CR rate, 21%).

Disease burden and symptoms The vast majority of ibrutinib-treated patients experienced substantial reduction in lymphadenopathy and splenomegaly at the time of the primary analysis which was much greater than observed with chlorambucil. A ≥50% reduction in the lymph node sum of the product of longest diameter (SPD) occurred in 95% of patients treated with ibrutinib versus 40% of those treated with chlorambucil, with complete resolution in lymphadenopathy in 42% versus 7%, respectively (Online Supplementary Figure S4A,B). Reduction in splenomegaly by ≥50% occurred in 95% with ibrutinib versus 52% with chlorambucil, with complete resolution in splenomegaly in 56% versus 22%, respectively (Online Supplementary Figure S4C,D). Ibrutinib also resulted in higher rates of improvements in disease symptoms including weight loss, fatigue, and night sweats, which were indications for therapy in many patients.

Table 1. RESONATE-2 reasons for initiation of treatment and baseline patient characteristics.1

Ibrutinib (n=136) Baseline Characteristic Median age (range), y 73 (65-89) ≥ 75 y, (%) 46 (34) Male, n (%) 88 (65) ECOG performance status, n (%) 0 60 (44) 1 65 (48) 2 11 (8) Rai stage III or IV, n (%) 60 (44) Bulky disease ≥ 5 cm, n (%) 54 (40) a Hierarchical Classification , n (%) Del(11q) 29/130 (22) Trisomy 12 20/117 (17) Del(13q) 25/112 (22) None of above 38/112 (34) b IGHV status , n/N (%) Mutated 40/121 (33) Unmutated 58/121 (48) Unclassifiablec 23/121 (19) Patients meeting criteria for active disease, n (%) Progressive marrow failure 54 (40) Lymphadenopathy 55 (40) Splenomegaly 36 (26) Progressive lymphocytosis 23 (17) Autoimmune anemia and/or 3 (2) thrombocytopenia Any documented constitutional 64 (47) symptoms Unintentional weight loss 14 (10) (>10% within 6 months) Significant fatigue 44 (32) Fever 4 (3) Night sweats 32 (24)

Chlorambucil (n=133) 72 (65-90) 47 (35) 81 (61) 54 (41) 67 (50) 12 (9) 62 (47) 40 (30) 25/121 (21) 23/108 (21) 32/108 (30) 28/108 (26) 42/127 (33) 60/127 (47) 25/127 (20) 49 (37) 44 (33) 44 (33) 28 (21) 5 (4) 56 (42) 16 (12) 29 (22) 3 (2) 35 (26)

ECOG: Eastern Cooperative Oncology Group; FISH: fluorescence in situ hybridization; IGHV: immunoglobulin heavy-chain variable-region gene. aPatients with missing results were excluded (Del(11): n=6 for ibrutinib, n=12 for chlorambucil; Trisomy 12: n=19 for ibrutinib, n=25 for chlorambucil; Del(13q): n=24 for ibrutinib, n=25 for chlorambucil; None of the above: n=24 for ibrutinib, n=25 for chlorambucil). bPatients with missing results were excluded (n=15 for ibrutinib, n=6 for chlorambucil). c Unclassifiable includes patients with polyclonal IGHV status if no specific IGHV subfamily member was dominant (>50% of all reads) and samples with no amplification.

Fatigue, although this was not statistically significant (86/136 [63%] vs. 71/133 [53%]; odds ratio, 1.50; 95% CI, 0.92-2.45; P=0.1013).

Safety and tolerability of ibrutinib Patient-reported QOL Greater improvements in QOL occurred with ibrutinib versus chlorambucil in FACIT-Fatigue (P=0.0013) by repeated measure analyses (Online Supplementary Figure S5). Clinically meaningful improvements occurred more frequently with ibrutinib versus chlorambucil in FACIT1504

Median treatment duration with ibrutinib was 28.5 months (range, 0.7-35.9 months). Most patients continue ibrutinib treatment, with 83% (112/135) receiving ibrutinib continuously for durations exceeding 2 years. The most frequent AEs with ibrutinib with extended follow up were similar to the primary report13: diarrhea, fatigue, haematologica | 2018; 103(9)

Extended Phase 3 Results From RESONATE-2

PFS

PFS in the presence or absence of Del(11q)

HR (95% CI) P value

PFS with mutated vs. unmutated IGHV

Figure 1. PFS for the intent-to-treat population. Survival analyses from randomization until event or censored at last follow up using the Kaplan-Meier method. Vertical ticks indicate censored patients. PFS: progression-free survival.

cough, anemia, and nausea (Online Supplementary Table S2). Grade ≥3 AEs were generally observed more frequently during the first 12 months of ibrutinib therapy and generally decreased over time (Figure 4). Rates of grade ≥3 cytopenias decreased over time from 8.1%, 5.9%, and 2.2% during the first year of treatment to 0%, 1%, and 0% during the third year of treatment for neutropenia, anemia, and thrombocytopenia, respectively. Several AEs of clinical interest were characterized in greater detail (Table 3). Diarrhea generally occurred early in treatment (median, 26 days) and was completely reversible in 95% of patients within a median of 6 days. Visual disturbances (blurred vision or reduced visual acuity) were grade 1 or 2, with 57% of these completely resolving within a median of 37.5 days after onset. Hypertension occurred at a median of 187 days, with improvements reported at a median of 14 days after onset. Arthralgia was observed at a median of 135 days and was generally reversible (78% complete, 4% partial) within a haematologica | 2018; 103(9)

median duration of approximately 3 weeks. Atrial fibrillation was observed throughout treatment follow up, with 4% of patients experiencing a grade 3 event. Symptoms of atrial fibrillation quickly resolved in the majority of patients (57% complete, 7% partial resolution) within a median of 3 days. Nine patients (7%) experienced a major hemorrhage occurring at a median of 310 days. Of these, 3 patients were reported to have active treatment with concomitant medications that impact platelets or coagulation (aspirin, low molecular weight heparin, and naproxen, respectively) including a traumatic hematoma, postprocedural hematoma, and hematuria, all of which were grade 3 in severity and did not result in study drug discontinuation. Grade ≥3 infection occurred in 23% of patients at a median of 138 days, including 2 that were fatal (Klebsiella infection and septic shock). Grade ≥3 infections were observed most frequently in the first year of treatment and decreased thereafter (Figure 4). There were no cases of pneumocystis pneumonia or multifocal leukoen1505

P.M. Barr et al.

cephalopathy reported. Serious AEs over the 3 years of follow up occurring in more than 2 ibrutinib-treated patients included pneumonia (11; 8%), atrial fibrillation (6; 4%), urinary tract infection (5; 4%), basal cell carcinoma (5; 4%), hyponatremia (5; 4%), pleural effusion (4; 3%), hypertension (3; 2%), and anemia (3; 2%). Eighteen patients (13%) required dose reductions and 16 patients (12%) discontinued first-line ibrutinib because of AEs. AEs leading to discontinuation in more than 1 patient included infection (n=5), hemorrhage (n=3), atrial fibrillation (n=2), and rash (n=2). Treatment-limiting toxicity including both reductions and discontinuations due to AEs decreased over time with ibrutinib (Figure 4).

Concomitant medications Concomitant medications were collected throughout the duration of ibrutinib treatment (median, 28.5 months) and chlorambucil (median, 7.1 months). Despite longer

follow up recording of the use of these agents in the ibrutinib arm versus the chlorambucil arm, the rate of neutrophil growth factor use and platelet and red blood cell transfusion was higher in the chlorambucil arm. Intravenous immunoglobulin was administered to 4% of ibrutinib-treated patients versus 2% of those randomized to chlorambucil. Anticoagulants and/or antiplatelet agents were frequently used during study therapy (56% and 54% of patients treated with ibrutinib and chlorambucil, respectively; Online Supplementary Table S3), including anticoagulants in 21% of the ibrutinib-treated patients.

Outcomes following ibrutinib discontinuation With up to 3 years follow up, out of 136 patients, only 4 patients discontinued ibrutinib primarily due to disease progression; 1 had unmutated IGHV and none were reported to have del(11q). Two of these 4 patients remain alive. Of the 16 patients who discontinued ibrutinib because of AEs; 13 (81%) are alive with a median of 13

Sex

Figure 2. PFS subgroup analysis.

Table 2. Response rates in ibrutinib-treated patients.

Best response, n (%) ORR CR/CRi nPR PR PR-L

All patients (n=136)

With del(11q) (n=29)

Without del(11q) (n=101)

Mutated IGHV (n=40)

Unmutated IGHV (n=58)

125 (92) 25 (18) 1 (1) 97 (71) 2 (1)

29 (100) 4 (14) 0 25 (86) 0

91 (90) 20 (20) 1 (1) 68 (67) 2 (2)

35 (88) 8 (20) 3 (8) 26 (65) 0

55 (95) 12 (21) 0 43 (74) 0

Cri: complete response with incomplete blood-count recovery; nPR: nodular partial response (defined according to the International Workshop on Chronic Lymphocytic Leukemia criteria for response16 as a complete response with lymphoid nodules in the bone marrow); PR: partial response; PR-L: partial response with lymphocytosis.

1506

haematologica | 2018; 103(9)

Extended Phase 3 Results From RESONATE-2

months follow up after ibrutinib discontinuation and 3 have died (Online Supplementary Table S4). Ten of the patients had PR as best response, 4 patients discontinued after not responding to ibrutinib and 2 patients discontinued prior to response evaluation. Non-responders/nonevaluable had a PFS that ranged from 1.8 to 20.2 months, while responders tended to have a variable but longer PFS (4.2–34.0 months). As the vast majority of patients (79%) remain on single-agent ibrutinib, this analysis is limited in size and also to the patients who came off treatment fairly early (9 of the 16 patients who discontinued due to AEs did so in the first year). In total, 7 patients received subsequent therapy after ibrutinib at a median of 7.6 months following ibrutinib discontinuation (range, 1.2 to 20.8 months), including fludarabine-cyclophosphamiderituximab (n=3), bendamustine-rituximab (n=2), chlorambucil (n=1), and radiation (n=1). Six of these 7 patients (86%) remain alive with median follow up of 21 months (range, 9 to 25 months).

Discussion This extended analysis of RESONATE-2 with detailed clinical follow up demonstrates that ibrutinib continues to provide significant and sustained clinical benefits, improving the quality of responses, for the first-line treatment of older patients with CLL or SLL with a manageable safety profile over extended durations of treatment. Consistent with the initial report, ibrutinib demonstrates a significant 88% reduction in the risk of PFS events (progression or death) compared with chlorambucil (P<0.0001) with extended follow up. In addition, the OS benefit for ibrutinib compared with chlorambucil was maintained, despite crossover to treatment with ibrutinib for many patients in the chlorambucil arm (n=55). These data support the use

of ibrutinib in the first-line treatment of CLL as a chemotherapy-free option that can be taken continuously, achieving long-term disease control for the majority of patients including those with high risk features.10,13 Ibrutinib has a category 1 National Comprehensive Cancer Network® (NCCN®) recommendation as a singleagent first-line treatment for CLL without del(17p) in patients ≥65 years and for relapsed/refractory CLL without del(17p).18 The efficacy of ibrutinib in the first-line setting appears superior to that observed in relapsed or refractory patients.19 Nearly all patients randomly assigned to ibrutinib achieved rapid disease reduction, with an ORR of 92% translating to high rates of 24-month PFS and OS of 89% and 95%, respectively, with similar PFS and OS rates seen regardless of age. This observation suggests that ibrutinib may be most effective when used upfront before the acquisition of poor-risk molecular aberrations, which are selected for with chemotherapy.20,21 Additionally, sensitivity analyses to adjust for the effects of patients in RESONATE-2 who crossed over to ibrutinib found that treatment with ibrutinib was still associated with statically significant OS compared with chlorambucil.16 These results also demonstrate that depth of response substantially increases over time, with higher rates of CR during the extended follow up, indicating a persistent action of the drug rather than a simple maintenance effect. Similar findings were observed with long-term follow up of patients enrolled in the phase 2 trial of first-line ibrutinib.11 Within this previous study, 29% of patients achieved a CR, and 92% remained alive and progression free at 5 years.22 Given these data, the CR rate will likely continue to increase in the present study as long-term disease control and high tolerability with first-line use can be expected based on the earlier phase 2 results.12 In addition to the efficacy benefits overall, sustained

Figure 3. Response rates over time in ibrutinib-treated patients. CR: complete response; CRi: complete response with incomplete blood-count recovery; nPR: nodular partial response (defined according to the International Workshop on Chronic Lymphocytic Leukemia criteria for response16 as a complete response with lymphoid nodules in the bone marrow); PR: partial response; PR-L: partial response with lymphocytosis.

haematologica | 2018; 103(9)

1507

P.M. Barr et al.

robust outcomes were demonstrated in higher-risk groups. No difference in outcome was observed in patients with unmutated IGHV status, a traditional poor prognostic indicator for all chemoimmunotherapy regimens. Notably the rate of unmutated IGHV in this study of older patients was somewhat lower than other studies at 48% (vs. 58%-62% in CLL11), consistent with prior reports of higher frequency of unmutated IGHV in younger patients.4,23 For patients with del(11q), another traditionally high-risk subgroup, 100% of the 29 patients responded to treatment with ibrutinib, and there was a 99% reduction in the risk of progression or death, with only 1 del(11q) ibrutinib-treated patient experiencing disease progression after discontinuing therapy for an AE over the extended follow up. While this represents a relatively small patient subset (22%), ibrutinib demonstrates a particularly significant benefit in this population, which

historically experiences inferior outcomes with traditional chemotherapy or CD20-based regimens.24-26 Combined analysis of 3 randomized studies not only demonstrated superiority of ibrutinib over traditional chemotherapy and/or anti-CD20 comparators for patients with del(11q), but also equally positive PFS and OS outcomes irrespective of del(11q). These results suggest that current definitions of high-risk disease and the impact of prognostic biomarkers may need to be redefined with ibrutinib.27 The mechanism why del(11q) patients may have better outcomes when treated with ibrutinib is of high interest and is the subject of ongoing research. Safety of therapy administered to older patients over the long term is an area that requires close scrutiny. Firstline ibrutinib appears to be well tolerated with extended treatment as evidenced by over 80% of this older population being able to continue treatment for more than 2

Table 3. Characterization of select AEs of clinical interest in ibrutinib-treated patients observed at any time during follow up.a

AE Grade Diarrhea Visual disturbancesb Hypertensionc Arthralgia Atrial fibrillation Major hemorrhage

Any 61 (45) 30 (22) 27 (20) 27 (20) 14 (10) 9 (7)

Infections (grade â&#x2030;Ľ3) 31 (23)

Ibrutinib-treated patients n=135 n (%) 2 3 4 16 (12) 6 (4) 13 (10) 9 (7) 7 (5) 1 (<1) NA

5 (4) 0 7 (5) 3 (2) 6 (4) 7 (5)

0 0 0 0 0 1 (1)

Resolution, Median time to first event, n (%) days 5 Complete Partial Any 2 3 4 5 0 0 0 0 0 0

58 (95) 17 (57) 12 (44) 21 (78) 8 (57) 9 (100)

28 (21) 4 (3) 2 (1) 28 (90)

0 26 131 0 100 201 1 (4) 187 187 1 (4) 135 55 1 (7) 249.5 85 0 310 155 0

138

Median time from onset to resolution/improvement, days Any 2 3 4

219 NA 109.5 135 773.5 446

NA NA NA NA NA 254

NA NA NA NA NA NA

6 37.5 14 22 3 13.5

3 74.5 36 22 2 14.0

6.5 NA 9 15 7 11.0

NA NA NA NA NA 45.0

119

367.5

422

AE: adverse event; NA: not applicable. aFrom first dose of study treatment up to 30 days after last dose or initiation of subsequent anti-cancer therapy, whichever occurs earlier b Visual disturbances included the preferred terms blurred vision and reduced visual acuity. cHypertension (standardized MEDRA queries) group of preferred terms. .

(n=135) (n=123) (n=112)

1508

Figure 4. Safety and tolerability of ibrutinib over time. Rate of grade â&#x2030;Ľ3 AEs, discontinuations due to AEs, and dose reductions over different periods of time. AE, adverse events.

haematologica | 2018; 103(9)

Extended Phase 3 Results From RESONATE-2

years. This extended follow up allowed for new observations into the timing of when AEs occurred and time to resolution of AEs as well as the use of transfusions and growth factors. Diarrhea, while frequent, often occurred early during the first several months of treatment and was generally low grade and short-lived. Severe and treatmentlimiting AE rates decreased over time with extended ibrutinib treatment. A decrease in myelotoxicity and infectious complications over time was also observed. This contrasts with chemoimmunotherapy-associated AEs and was importantly associated with less medical resource utilization of neutrophil growth factors and slightly less transfusion need despite 4 times the treatment time with ibrutinib. While the rate of atrial fibrillation increased from 6% in the primary analysis13 to 10%, overall, ibrutinib dose reduction or discontinuation due to atrial fibrillation was uncommon and lessened with extended treatment in this population of older patients with CLL. Atrial fibrillation therefore appears manageable and does not frequently necessitate ibrutinib discontinuation. Additional information on the management and outcomes of atrial fibrillation along with associated anticoagulant therapy has been provided in a large pooled analysis of ibrutinib studies.28 Rates of major hemorrhage remained low despite half the patients receiving concomitant antiplatelet or anticoagulant medications. Previous work demonstrated that QOL is significantly compromised in patients with CLL, affecting physical fitness, cognitive function, levels of fatigue, and sleep. Worse scores were reported for patients receiving chemotherapy such as chlorambucil.29 Even with the addition of contemporary anti-CD20 agents (obinutuzumab), no significant benefit in QOL has been noted.4 However, this extended follow up provides the first analysis of QOL, as measured by FACIT Fatigue Scale, following ibrutinib treatment in previously untreated patients. Significantly greater improvements in QOL were observed with ibrutinib versus chlorambucil. In line with this and the favorable impact on QOL and tolerable safety profile, 79% of patients remained on first-line treatment with ibrutinib at the time of this later analysis with up to 3 years of therapy. Patients who discontinue treatment for CLL including ibrutinib may have varied outcomes dependent on the reason for discontinuation.30 In 1 study that included mostly patients with relapsed or refractory CLL, median OS following ibrutinib therapy was 33 months for those who discontinued because of AEs versus 16 months for those who discontinued because of disease progression. In our

References 1. Howlader N, Noone A, Krapcho M, et al. SEER Cancer Statistics Review 1975-2013. Bethesda, MD: National Cancer Institute; 2014. Available from: http://seer.cancer.gov/archive/csr/1975_20 13/, based on November 2015 SEER data submission, posted to SEER web site, April 2016. Accessed March 17, 2017. 2. Eichhorst BF, Busch R, Stilgenbauer S, et al. First-line therapy with fludarabine compared with chlorambucil does not result in a major benefit for elderly patients with advanced chronic lymphocytic leukemia.

haematologica | 2018; 103(9)

study, the 22 patients who discontinued therapy had a median follow up of 13 months after discontinuation. Of these 22 patients, 16 are still alive, while 2 of the 4 patients who progressed have died. Seven patients have received subsequent treatment, mostly chemoimmunotherapy (BR, FCR); 6 of those patients are still alive, with a median of 21 months of follow up. While retrospective analyses of real-world data have previously suggested that treatment with an alternate kinase inhibitor is more effective than chemoimmunotherapy following discontinuation of ibrutinib,31,32 our data suggests that patients who discontinue ibrutinib can respond to chemoimmunotherapy as second-line therapy. Continued follow up of patients in the RESONATE-2 trial who have discontinued ibrutinib will provide the needed further data as relatively few patients have progressed or stopped therapy to date. These data confirm that first-line treatment with ibrutinib results in long-term PFS in patients with CLL and that response quality continues to improve with ibrutinib over time, with substantial increase in patients achieving CR. In addition, rates of grade â&#x2030;Ľ3 AEs during treatment with ibrutinib decreased over time. The most common reasons for initiating first-line treatment in these patients, including marrow failure, disease burden, and disease symptoms, all improved to greater extents in patients treated with ibrutinib versus chemotherapy. Ongoing randomized studies, including ILLUMINATE (clinicaltrials.gov identifier 02264574), comparing ibrutinib-obinutuzumab with chlorambucil-obinutuzumab, and A041202 (clinicaltrials.gov identifier 01886872), comparing ibrutinib, ibrutinib-rituximab, and rituximab-bendamustine, will continue to define the role of ibrutinib for the first-line treatment of patients with CLL/SLL. Funding This study was supported by Pharmacyclics LLC, an AbbVie company, by grants (CA016672 and 5P01CA081534-14) from the National Institutes of Health, and by the MD Anderson Moon Shot Program in CLL. Pharmacyclics LLC, an AbbVie company, sponsored and designed the study. Study investigators and their research teams collected the data. The sponsor confirmed data accuracy and performed analysis of the data. Medical writing support was funded by the sponsor. Acknowledgments We thank all the patients who participated in this trial and their families and Jennifer Leslie, PhD, for medical writing supported by Pharmacyclics LLC, an AbbVie company.

Blood. 2009;114(16):3382-3391. 3. Eichhorst B, Fink AM, Bahlo J, et al. First-line chemoimmunotherapy with bendamustine and rituximab versus fludarabine, cyclophosphamide, and rituximab in patients with advanced chronic lymphocytic leukaemia (CLL10): an international, openlabel, randomised, phase 3, non-inferiority trial. Lancet Oncol. 2016;17(7):928-942. 4. Goede V, Fischer K, Busch R, et al. Obinutuzumab plus chlorambucil in patients with CLL and coexisting conditions. N Engl J Med. 2014;370(12):11011110. 5. de Gorter DJ, Beuling EA, Kersseboom R, et al. Bruton's tyrosine kinase and phos-

pholipase C 2 mediate chemokine-controlled B cell migration and homing. Immunity. 2007;26(1):93-104. 6. de Rooij MF, Kuil A, Geest CR, et al. The clinically active BTK inhibitor PCI-32765 targets B-cell receptor- and chemokinecontrolled adhesion and migration in chronic lymphocytic leukemia. Blood. 2012;119(11):2590-2594. 7. Herman SE, Gordon AL, Hertlein E, et al. Bruton tyrosine kinase represents a promising therapeutic target for treatment of chronic lymphocytic leukemia and is effectively targeted by PCI-32765. Blood. 2011;117(23):6287-6296. 8. Ponader S, Chen SS, Buggy JJ, et al. The

1509

P.M. Barr et al.

10.

11.

12.

13.

14. 15. 16.

17.

1510

Bruton tyrosine kinase inhibitor PCI-32765 thwarts chronic lymphocytic leukemia cell survival and tissue homing in vitro and in vivo. Blood. 2012;119(5):1182-1189. Spaargaren M, Beuling EA, Rurup ML, et al. The B cell antigen receptor controls integrin activity through Btk and PLCgamma2. J Exp Med. 2003;198(10):1539-1550. Byrd JC, Brown JR, O'Brien S, et al. Ibrutinib versus ofatumumab in previously treated chronic lymphoid leukemia. N Engl J Med. 2014;371(3):213-223. O'Brien S, Furman RR, Coutre SE, et al. Ibrutinib as initial therapy for elderly patients with chronic lymphocytic leukaemia or small lymphocytic lymphoma: an open-label, multicentre, phase 1b/2 trial. Lancet Oncol. 2014;15(1):48-58. Byrd JC, Furman RR, Coutre SE, et al. Three-year follow-up of treatment-naive and previously treated patients with CLL and SLL receiving single-agent ibrutinib. Blood. 2015;125(16):2497-2506. Burger JA, Tedeschi A, Barr PM, et al. Ibrutinib as initial therapy for patients with chronic lymphocytic leukemia. N Engl J Med. 2015;373(25):2425-2437. Ibrutinib [package insert]. Sunnyvale, CA; Pharmacyclics LLC; 2017. Ibrutinib [summary of product characteristics]. Beerse, Belgium; Janssen-Cilag International NV; 2017. Coutre S, Tedeschi A, Robak T, et al. Survival adjusting for crossover: phase 3 study of ibrutinib vs chlorambucil in older patients with untreated chronic lymphocytic leukemia/small lymphocytic lymphoma. Haematologica. 2017 Nov 23. [Epub ahead of print]. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on

18.

19.

20.

21.

22.

23.

24.

25.

Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008;111 (12):5446-5456. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Chronic lymphocytic leukemia/small lymphocytic leukemia. National Comprehensive Cancer Network; 2016. Available from: https://www.nccn.org/professionals/physician_gls/default.aspx#site. Accessed January 11, 2018. O'Brien PC, Byrd JC, Hillmen P, et al. Outcomes with ibrutinib by line of therapy in patient with CLL: Analyses from phase III data. J Clin Oncol. 2016;34(suppl). Abstract 7520. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. Cell. 2013;152(4):714-726. Nadeu F, Delgado J, Royo C, et al. Clinical impact of clonal and subclonal TP53, SF3B1, BIRC3, NOTCH1, and ATM mutations in chronic lymphocytic leukemia. Blood. 2016;127(17):2122-2130. O'Brien SM, Furman RR, Coutre SE, et al. Five-year experience with single-agent ibrutinib in patients with previously untreated and relapsed/refractory chronic lymphocytic leukemia/small lymphocytic leukemia. Blood. 2016;128(22):233-233. Parikh SA, Rabe KG, Kay NE, et al. Chronic lymphocytic leukemia in young (â&#x2030;¤ 55 years) patients: a comprehensive analysis of prognostic factors and outcomes. Haematologica. 2014; 99(1):140-147. Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343(26):1910-1916. Hallek M, Fischer K, Fingerle-Rowson G, et al. Addition of rituximab to fludarabine and

26. 27.

28.

29.

30.

31.

32.

cyclophosphamide in patients with chronic lymphocytic leukaemia: a randomised, open-label, phase 3 trial. Lancet. 2010;376 (9747):1164-1174. Kipps TJ, Stevenson FK, Wu CJ, et al. Chronic lymphocytic leukaemia. Nat Rev Dis Primers. 2017;3:16096. Kipps TJ, Fraser G, Coutre SE, et al. Outcomes of ibrutinib-treated patients with chronic lymphocytic leukemia/small lymphocytic leukemia (CLL/SLL) with high-risk prognostic factors in an integrated analysis of 3 randomized phase 3 studies. Abstract presented at: the XVII International Workshop on Chronic Lymphocytic Leukemia; May 12-15, 2017; New York, NY. Abstract 19. Brown JR, Moslehi J, O'Brien S, et al. Characterization of atrial fibrillation adverse events reported in ibrutinib randomized controlled registration trials. Haematologica. 2017;102(10):1796-1805. Holtzer-Goor KM, Schaafsma MR, Joosten P, et al. Quality of life of patients with chronic lymphocytic leukaemia in the Netherlands: results of a longitudinal multicentre study. Qual Life Res. 2015;24(12): 2895-2906. Jain P, Thompson PA, Keating M, et al. Longterm outcomes for patients with chronic lymphocytic leukemia who discontinue ibrutinib. Cancer. 2017;123(12):2268-2273. Mato AR, Nabhan C, Barr PM, et al. Outcomes of CLL patients treated with sequential kinase inhibitor therapy: a real world experience. Blood. 2016;128(18): 2199-2205. Mato AR, Hill BT, Lamanna N,et al. Optimal sequencing of ibrutinib, idelalisib, and venetoclax in chronic lymphocytic leukemia: results from a multicenter study of 683 patients. Ann Oncol. 2017; 28(5):1050-1056.

haematologica | 2018; 103(9)

ARTICLE

Chronic Lymphocytic Leukemia

Real-world outcomes and management strategies for venetoclax-treated chronic lymphocytic leukemia patients in the United States

Anthony R. Mato,1 Meghan Thompson,2 John N. Allan,3 Danielle M. Brander,4 John M. Pagel,5 Chaitra S. Ujjani,6 Brian T. Hill,7 Nicole Lamanna,8 Frederick Lansigan,9 Ryan Jacobs,10 Mazyar Shadman,11 Alan P. Skarbnik,12 Jeffrey J. Pu,13 Paul M. Barr,14 Alison R. Sehgal,15 Bruce D. Cheson,6 Clive S. Zent,14 Hande H. Tuncer,16 Stephen J. Schuster,2 Peter V. Pickens,17 Nirav N. Shah,18 Andre Goy,12 Allison M. Winter,7 Christine Garcia,15 Kaitlin Kennard,2 Krista Isaac,19 Colleen Dorsey,2 Lisa M. Gashonia,2 Arun K. Singavi,18 Lindsey E. Roeker,1 Andrew Zelenetz,1 Annalynn Williams,14 Christina Howlett,12 Hanna Weissbrot,8 Naveed Ali,17 Sirin Khajavian,11 Andrea Sitlinger,4 Eve Tranchito,7 Joanna Rhodes,2 Joshua Felsenfeld,3 Neil Bailey,5 Bhavisha Patel,20 Timothy F. Burns,9 Melissa Yacur,13 Mansi Malhotra,16 Jakub Svoboda,2 Richard R. Furman3 and Chadi Nabhan21

Ferrata Storti Foundation

Haematologica 2018 Volume 103(9):1511-1517

ARM and MT contributed equally to this work.

CLL Program, Leukemia Service, Division of Hematologic Oncology, Department of Internal Medicine, Memorial Sloan Kettering Cancer Center, New York, NY; 2Center for CLL, Division of Hematology and Oncology, University of Pennsylvania, Philadelphia, PA; 3 New York Presbyterian & Weill Cornell, NY; 4Division of Hematologic Malignancies and Cellular Therapy, Duke University, Durham, NC; 5Center for Blood Disorders and Stem Cell Transplantation, Swedish Cancer Institute, Seattle, WA; 6Georgetown University Hospital Lombardi Comprehensive Cancer Center, Washington, DC; 7Taussig Cancer Institute, Cleveland Clinic Foundation, OH; 8Columbia University Medical Center, New York, NY; 9Dartmouth-Hitchcock Medical Center, Lebanon, NH; 10Department of Hematologic Oncology and Blood Disorders, Levine Cancer Institute, Carolinas Healthcare System, Charlotte, NC; 11University of Washington/Fred Hutchinson Cancer Research Center, Seattle Cancer Care Alliance, WA; 12John Theurer Cancer Center, Hackensack Meridian Health, NJ; 13Penn State Health, Hershey, PA; 14Wilmot Cancer Institute Division of Hematology/Oncology, University of Rochester Medical Center, NY; 15 University of Pittsburgh Medical Center, PA; 16Tufts Medical Center, Boston, MA; 17 Abington Hem. Onc. Assoc., Inc., Willow Grove, PA; 18Division of Hematology & Oncology, Medical College of Wisconsin, Brookfield, WI; 19Internal Medicine, Lankenau Medical Center, Wynnewood, PA; 20Washington Hospital Center, DC and 21Cardinal Health, Dublin, OH, USA 1

ABSTRACT

enetoclax is a BCL2 inhibitor approved for 17p-deleted relapsed/refractory chronic lymphocytic leukemia with activity following kinase inhibitors. We conducted a multicenter retrospective cohort analysis of patients with chronic lymphocytic leukemia treated with venetoclax to describe outcomes, toxicities, and treatment selection following venetoclax discontinuation. A total of 141 chronic lymphocytic leukemia patients were included (98% relapsed/refractory). Median age at venetoclax initiation was 67 years (range 37-91), median prior therapies was 3 (0-11), 81% unmutated IGHV, 45% del(17p), and 26.8% complex karyotype (â&#x2030;Ľ 3 abnormalities). Prior to venetoclax initiation, 89% received a B-cell receptor antagonist. For tumor lysis syndrome prophylaxis, 93% received allopurinol, 92% normal saline, and 45% rasburicase. Dose escalation to the maximum recommended dose of 400 mg daily was achieved in 85% of patients. Adverse events of interest included neutropenia in 47.4%, thrombocytopenia in 36%, tumor lysis syndrome in 13.4%, neutropenic fever in 11.6%, and diarrhea in 7.3%. The overall response rate to venetoclax was 72% (19.4% complete remission). With a median follow up of 7 months, median progression free survival and overall survival for the entire cohort have not been reached. To date, 41 venetoclax treated patients have discontinued haematologica | 2018; 103(9)

Correspondence: matoa@mskcc.org

Received: March 17, 2018. Accepted: June 5, 2018. Pre-published: June 7, 2018. doi:10.3324/haematol.2018.193615 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1511 ÂŠ2018 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.

1511

A.R. Mato et al.

therapy and 24 have received a subsequent therapy, most commonly ibrutinib. In the largest clinical experience of venetoclax-treated chronic lymphocytic leukemia patients, the majority successfully completed and maintained a maximum recommended dose. Response rates and duration of response appear comparable to clinical trial data. Venetoclax was active in patients with mutations known to confer ibrutinib resistance. Optimal sequencing of newer chronic lymphocytic leukemia therapies requires further study.

Introduction Venetoclax is an oral second-generation BCL2 inhibitor with demonstrated activity and durable responses in relapsed/refractory (R/R) chronic lymphocytic leukemia (CLL), including those with chromosome 17p deletion (del(17p)), unmutated IGHV, fludarabine-resistance, bulky disease, and progression on or following ibrutinib or idelalisib.1-5 Overall response rate (ORR) to venetoclax monotherapy was 79% in the phase II trial in del(17p) R/R CLL.3 Moreover, ORRs were 65% and 67% in R/R CLL following ibrutinib or idelalisib, respectively.4,5 In addition to data with monotherapy, progression free survival (PFS) was superior for patients treated with venetoclax and rituximab compared to bendamustine and rituximab in R/R CLL, and more venetoclax-treated patients achieved undetectable minimal residual disease (MRD) (83.5% vs. 23.1%).6 Knowledge about venetoclax efficacy, dose-escalation, and toxicity in CLL patients has almost entirely been informed by experiences from clinical trials.3 As several real-world evidence series showed that toxicity profiles and outcomes for kinase inhibitor-treated patients may differ from those reported in the clinical trial setting, studying whether these differences apply to venetoclaxtreated patients is essential.7-11 Furthermore, there is one study to date regarding strategies for early identification of high-risk patients, particularly those previously treated with kinase inhibitor based therapy, for progression on venetoclax and how treatment is selected after its discontinuation is selected.2 We aimed to better understand disease characteristics and toxicities of CLL patients treated with venetoclax in clinical practice and contrast their outcomes to those reported in key clinical trials. We explored prognostic factors that predict early progression on venetoclax and studied treatment selection following discontinuation. To our knowledge, this analysis reports the largest series of CLL patients treated with venetoclax in a real-world setting with a focus on outcomes following venetoclax discontinuation.

Methods We conducted a multicenter, retrospective cohort study of all CLL patients treated with venetoclax across 19 United States academic and community cancer centers. The study was approved by the institutional review board at each US institution. Investigators conducted a detailed review of the institutional electronic medical records to identify patients with CLL treated with venetoclax. Collected data included demographics, clinical and genetic prognostic factors, venetoclax dose-escalation management, long-term dosing, toxicities, tumor lysis syndrome (TLS) prophylaxis strategies and outcomes, ORR, complete response (CR), survival out1512

comes, and reasons for discontinuation. Investigators were asked to follow the National Cancer Institute working group international workshop guidelines for CLL (iwCLL) published in 2008 to define rates of response and progression of disease.12 Disease burden as a predictor for TLS risk was categorized as low, medium, and high per the treating physician. Physicians were asked to use the venetoclax package insert to guide the categorization, which was developed based on United States approval of the drug in 2016.13 TLS events were defined as per Howard criteria, which specify criteria for laboratory and clinical TLS.14 Adverse events (AEs) were graded using the NCI Common Toxicity Terminology Criteria for Adverse Events v4.0 (CTCAE v4.0). Cytogenetics, FISH results and next generation sequencing (NGS) were reported for patients where available. The primary endpoint was PFS, defined as the time from venetoclax initiation until progression or death from any cause as per the Kaplan Meier method.15 Patients were censored at the time of last follow up or at the time of next therapy regardless of progression status. Outcomes were stratified by prognostic characteristics where available, including del(17p) status, complex karyotype (>3 abnormalities), and venetoclax monotherapy versus combinations. Secondary endpoints included overall survival (OS), venetoclax dosing and toxicities, TLS incidence, dose escalation schema, response rates, and reasons for discontinuation. Comparisons of survival outcomes data were made using the long rank (LR) test.16 Hazard ratios were estimated using Cox regression analyses.17 Other analyses were descriptive. Tests were two-sided at the 5% level. Statistical analyses were performed using STATA 10.1 (Stata Statistical Software: Release 10, 2007; StataCorp LP, College Station, TX).

Results Patient characteristics We identified 141 CLL patients treated with venetoclax. Males and Caucasians represented most patients at 66% and 87%, respectively. The median age at diagnosis was 59 years (range 30-88), and median age at venetoclax initiation was 67 years (range 37-91). The population consisted almost entirely of patients with R/R CLL, with only 2 (1.4% of 141) of patients being treatment-naĂŻve. Patients had received a median of 3 prior therapies (range 0-11). Venetoclax was administered in combinations in 18.4% (n=26 of 141) of patients. Ibrutinib (36%), obinutuzumab (32%), and rituximab (24%) were the most commonly used drugs with venetoclax. Almost 89% of patients were treated with a B-cell receptor signal transduction inhibitor prior to venetoclax; 82% (n=115/141) received ibrutinib. Patient characteristics are summarized in Table 1. Most patients in this cohort had at least one traditionally poor-risk feature: 45% (n=61/136 tested) patients had chromosome del(17p), 26% (n=34/131) had deletion of chromosome 11q (del(11q)), 44% (n=42/95) had p53 mutations, 26.8% (n=52/130) had a complex karyotype haematologica | 2018; 103(9)

Real world venetoclax experience

(>3 chromosomal abnormalities) and 26.8% (n=15/56) had a NOTCH1 mutation. Thirty-one patients (22%) had del(17p) and TP53 mutation, and 16 patients (11%) had del(17p), TP53 mutation, and complex karyotype.

Venetoclax dosing and adverse events All patients underwent 5-week dose escalation of venetoclax. During dose-escalation, 85% (n=120/141) of patients achieved a maximum dose of 400 mg daily, with 75% of the cohort (n=103/137) maintaining 400 mg daily as a long-term stable dose. Dose interruptions occurred in 30% of patients and 21% required dose reduction. Details are available in Online Supplementary Table S1. Reasons for interruptions were not available. Hematologic events were the most common AEs with neutropenia (defined as ANC < 1000 cells/microL) occurring in almost half of patients (47.4%, n=65/137) and thrombocytopenia (defined as platelets < 50,000 cells/microL) occurring in over one-third of patients (36.0%, n=49/136). Other AEs included TLS (13.4%, n=18/134), neutropenic fever (11.6%, n=16/138), and grade ≥ 2 diarrhea (7.3%, n=10/138). Opportunistic infections (OI) while on venetoclax were reported in 11 patients (7.8%) with the three most common being pneumocystis jirovecii pneumonia (PJP) (n=6), invasive fungal (n=2), and toxoplasmosis (n=2). Nine OI events occurred in patients with prior exposure to any kinase inhibitor, 8 OI events occurred in patients with prior exposure to ibrutinib, and 6 OI events occurred in patients with prior exposure to two prior kinase inhibitors. The median time from venetoclax start to OI event was 5 months (0.2 – 16 months).

TLS: prophylaxis and hospitalization practice patterns Of 134 patients with TLS data, 44.8% were low-risk (n=60), 35.8% were intermediate risk (n=48), and 19.4% were high risk (n=26). Eighty-nine of 131 patients (64.3%) had pre-venetoclax lymph node assessment by CT scan to inform TLS risk. To minimize the occurrence of TLS, allopurinol was used in almost all cases regardless of risk category (93.1% low risk, 87.5% intermediate risk, 100.0% for high risk), as was intravenous normal saline (82.1% low risk, 91.7% intermediate risk, 100.0% high risk). Rasburicase use as TLS prophylaxis varied by risk category: 17.2% of low risk, 31.3% of intermediate risk, and 46.2% of high risk patients. TLS risk stratification, prophylaxis patterns, and incidence are summarized in Table 2. Most patients had one or more planned hospitalizations during dose escalation regardless of risk category (Online Supplemental Table S2). Twenty-seven patients were not hospitalized at any point, including 20 of 58 (34.5%) low risk patients and 7 of 48 (14.6%) intermediate risk patients. All high-risk patients were hospitalized at least once for TLS monitoring and prophylaxis during the dose escalation phase. Among high-risk patients, 32.0% (n=8 of 25) were hospitalized for all five dose escalations. The mean number of days hospitalized during the 5-week dose escalation period for low risk, intermediate risk, and high-risk patients were 1.5, 1.7, and 3.1, respectively. Overall, the incidence of TLS events (laboratory and clinical) was 13.4% (n=18/134) with 5 events (3.7%) reported in low-risk, 4 (3.0%) in intermediate-risk, and 9 (6.7%) in high-risk patients. Of these events, 6 were recorded as clinical TLS events (2 low risk patients, 1 intermediate risk patient, and 3 high risk patients), and the haematologica | 2018; 103(9)

Table 1. Baseline characteristics of 141 patients treated with venetoclax.

Patient characteristics

Median (range)

Median age at diagnosis, years Median age at venetoclax start, years Median prior lines of therapy Follow up, months*

CLL characteristics

59 (30-88) 67 (37-91) 3 (0-11) 7 (0.1-38.4)

Frequency (n with characteristic/ total n with available data)

Relapsed/Refractory Treatment naive CLL genetics Del(17p) Del(11q) TP53 mutation NOTCH1 mutation Complex karyotype, ≥ 3 mutations Unmutated IGHV Prior ibrutinib exposure Ibrutinib resistance mutations BTK mutation PLCγ2 mutation Venetoclax administered in combination Venetoclax and ibrutinib Venetoclax and obinutuzumab Venetoclax and rituximab

ì98.6% (139/141) 1.4% (2/141) 44.9% (61/136) 26.0% (34/131) 44.2% (42/95) 26.8% (15/56) 26.8% (52/130) 83.3% (60/72) 81.6% (115/141) 35.3% (12/34) 12.5% (4/32) 18.4% (26/141) 36% (9/26) 32% (8/26) 24% (6/26)

*Median follow up calculated using overall survival.

remainder were laboratory events (n= 12). Of the clinical TLS patients, 4 of 6 achieved 400 mg venetoclax dosing. No TLS patient required hemodialysis. One TLS death was reported in a patient who was re-challenged with venetoclax after a delayed interruption without utilizing a dose escalation schedule or hospitalization for venetoclax re-escalation. We were unable to correlate TLS events with a threshold dose of venetoclax.

Outcomes The reported ORR and CR rate, stratified by selected risk factors, are summarized in Table 3. The ORR for the entire cohort was 72.1% and 19.4% of patients achieved a CR. The median time to best response was 2.1 months. Venetoclax had a similar ORR across several high-risk groups including patients with age ≥ 65 (ORR = 74.3%), del(17p) (71.4%), prior ibrutinib therapy (69.1%), BTK mutation (91.6%), and PLCγ2 mutation (75.0%). At a median follow up of 7 months, the median PFS and OS have not been reached for the entire cohort (Figure 1a and 1b). The projected PFS and OS for the entire cohort at 12 months were 68% and 88%, respectively. Patients with a TP53 interruption (del(17p) and/or TP53 mutation) had significantly shorter PFS than those with intact TP53 (Figure 1c), though OS for the two groups was not significantly different (Online Supplementary Figure S1). In univariate analyses, we identified TP53 interruption as a predictor of inferior PFS (HR 2.7, 95% CI 1.08-6.7, P=0.034) but not OS (HR 1.78, 95% CI .55-5.74, P=.332). 1513

A.R. Mato et al.

Complex karyotype (HR 1.36, 95% CI 0.66-2.84, P=0.4), prior ibrutinib therapy (1.74, 95% CI 0.61-5.0, P=0.3), and unmutated IGHV (HR 0.29, 95% CI 0.04-2.3, P=0.25) were not significantly associated with inferior PFS. TP53 interruption remained a significant predictor for inferior PFS in a multivariate analyses which included TP53 interruption, complex karyotype and prior ibrutinib therapy (HR 2.8, CI 1.22-6.4, P=0.03). The presence of del(11q) did not impact OS and had an observed protective effect on PFS (HR 0.31, 95% CI 0.11-0.90, P=0.03).

Venetoclax discontinuations and treatment selection following venetoclax Venetoclax was discontinued in 41 patients (29%). Progression of disease was the most common reason for discontinuation (53.8%, n=21) followed by toxicity (20.5%, n=9), two-thirds of which were hematologic. Other reasons for discontinuation included death not related to progressive disease (10.25%, n=4), second cancer (5.1%, n=2), physician/patient preference (2.5%, n=1), Richterâ&#x20AC;&#x2122;s transformation (2.5%, n=1), and planned alternate therapy including CD19 directed chimeric antigen receptor T cells (CAR-T, 2.5%, n=1) and transplantation (2.5%, n=1). Table 4 summarizes therapy selection and outcomes for individual cases following venetoclax. Notably, 17 of 34 patients (50%) who discontinued venetoclax and remain alive have not required a subsequent therapy. Reasons for discontinuation in the group of patients who have not yet been treated following venetoclax discontinuation

include toxicity (n=6), progression of CLL (n=4), death not secondary to toxicity or progression (n=4), secondary malignancy (n=2), and doctor or patient preference (n=1). Ibrutinib-based therapy was the most common choice after venetoclax; five of 24 (21%) patients receiving ibrutinib. Three of five of patients treated with ibrutinib had prior ibrutinib exposure. Of these five patients, 1 had a partial response, 2 had stable disease, and 2 had progressive disease. Other therapies selected included rituximab monotherapy (12.5%, n=3), anthracycline based regimens (12.5%, n=3), allogeneic stem cell transplant (12.5%, n=3), idelalisib-based therapy (8.3%, n=2), and CAR-T (8.3%, n = 2). Subsequent lines of therapies with their corresponding responses are detailed in Online Supplementary Table S3.

Discussion In the largest series of venetoclax-treated CLL patients treated in the U.S., response rates (ORR 72.1%) and survival data are comparable to those reported in published clinical trials.1,3-5 Toxicities were similar with hematologic toxicities being the most frequently observed. Rates of TLS were higher than prior reports. Collectively, these results suggest that the efficacy and safety profile of venetoclax demonstrated in the clinical trials setting are comparable to what has been observed in the real world. Consistent with previously published data, the ORR for the del(17p) population remained high at 71.4%. Whereas

Table 2. Tumor lysis syndrome prophylaxis and events.

TLS risk category n=134

Allopurinol

TLS prophylaxis Rasburicase

Normal saline

Total

TLS events Laboratory

Clinical

Low 44.8% (n=60) Intermediate 35.8% (n=48) High 19.4% (n=26)

93.1% (n=54/58) 87.5% (n=42/48) 100.0% (n=26/26)

17.2% (n=10/58) 31.3% (n=15/48) 46.2% (n=11/26)

82.1% (n=46/56) 91.7% (n=44/48) 100.0% (n=25/25)

HR 2.7, P=0.034

Figure 1. Survival analyses for patients following venetoclax initiation. (A) Progression free survival for the entire cohort. Median PFS has not been reached with median follow up of 7 months. Projected 12-month PFS is 68%. (B) Overall survival for the entire cohort. Median OS has not been reached with median follow up of 7 months. Projected 12-month OS is 88%. (C) Progression free survival by TP53 status. PFS is significantly superior for patients with intact TP53 compared to patients with TP53 interruption, either TP53 mutation or del(17p).

1514

haematologica | 2018; 103(9)

Real world venetoclax experience

TP53 interruption was significantly associated with inferior PFS, OS was not compromised. Complex karyotype was not associated with inferior PFS, despite being shown to be a risk factor for progression in patients receiving venetoclax in a recent study by Anderson et al.2 It is possible that our shorter follow up accounts for this discrepancy. Interestingly, del(11q) did not adversely affect PFS. Similar findings were reported by Kipps et al. in 620 patients treated with ibrutinib stratified by del(11q) status (HR 0.73, P=0.08).18 In subgroup analysis, the response rate for patients previously treated with ibrutinib was 69.1%, similar to the 65% demonstrated by Jones et al. in patients who were treated with venetoclax following ibrutinib.5 Finally, we did not observe a difference in PFS whether venetoclax was given alone or in combination. TLS has been suggested as the most critical toxicity with venetoclax, contributing to early treatment-related deaths in clinical trials, particularly before the current dose rampup schedule was implemented in trials to minimize TLS risk. In our study, TLS rates were higher than those reported in most recent trials. In this cohort, 18 patients (13.4%) had TLS; 12 (9.0%) cases were laboratory TLS events and 6 (4.5%) cases were clinical events. In the initial phase I venetoclax study, 18% of patients experienced TLS (12.5% laboratory, 5.6% clinical). However, once the dosing schedule was modified to minimize TLS risk, 1.7% of patients had laboratory TLS and none had clinical TLS.3 More recently reported clinical trials using the standard dose ramp-up protocol have shown laboratory TLS rates of 2.2%5 and 4.7%3 and clinical TLS rates of 0%.3-5 Despite the higher rates of TLS observed in our study, only 19.4% of patients were deemed high risk for TLS versus 25-49% of patients classified as high risk in the clinical trials setting.3-5 Overall, adherence to TLS prophylaxis recommendations was excellent. The majority (92%) received allopurinol, which is recommended for all patients regardless of risk category.13 Guidelines per the US venetoclax prescribing information document also recommend oral hydration and intravenous normal saline for high risk patients, which was followed for all patients in our cohort. Similarly, rasburicase is recommended for patients with elevated baseline uric acid. This was used in over onequarter of all cases and almost half of the high-risk cases. All high-risk patients had at least one planned hospitalization during dose escalation. Three of 25 high-risk patients forwent the second recommended hospitalization given lack of TLS development during the first hospitalization.

Most low and intermediate-risk patients were hospitalized at least once, suggesting a conservative approach was utilized in this series during the dose ramp-up as outpatient dosing, with close monitoring. As per United States Food and Drug Administration (FDA) label guidelines, low and intermediate-risk patients can also be managed in the outpatient setting without hospitalization and we suspect future real-world series will demonstrate a higher proportion of low and intermediate-risk patients managed in the clinic during the dose ramp-up period. Potential reasons for increased TLS events may include difficulty in adhering to the exact dose-ramp up schedule or lack of physician/patient education surrounding importance of suggested prophylaxis, laboratory monitoring, and interventions for all patients. Patients may have had differences in comorbidities, such as impaired renal function, which would have made them ineligible for a venetoclax clinical trial and possibly at increased risk for TLS. Deviations in clinical practice initiation of venetoclax from that recommended in the FDA label, in particular the limited use of CT assessment (64.3%) to establish TLS risk prior to venetoclax initiation, could have led to risk misclassification. Additionally, while investigators were asked to use the Howard criteria to define TLS, it is possible that this mandate was not strictly followed when capturing data, leading to misclassification bias. To date, little is known regarding reasons for venetoclax discontinuation in clinical practice. In our study, 28% of all patients discontinued therapy; 53.8% of these patients discontinued due to progression of CLL excluding Richterâ&#x20AC;&#x2122;s transformation (RT) and 20.5% discontinued due to toxicity. As in clinical trials, we found CLL progression to be the most common reason for venetoclax discontinuation. In the phase I study of venetoclax for R/R CLL, the overall discontinuation rate of 56%, with 35% of discontinuations due to CLL progression (non RT) and 20% due to toxicity.1 In the phase II study of patients treated with prior B-cell receptor signal transduction inhibitors, the discontinuation rate was 49.5% at 14 months median follow up. CLL progression represented 49% of these discontinuations and AEs represented 11% of discontinuations.5 In this same series, RT was reported in 5% of patients who discontinued venetoclax. The median time to CLL progression was 8.4 months, and median time to RT was approximately one year.19 Anderson et al. reported that, in a group of heavily pretreated patients, 37% of patients progressed on venetoclax at a median follow up of 23 months, and that 8.2% of patients discontinued therapy

Table 3. Response rates.

Overall population Age n=129 >65 years n=82 ORR CR PR SD PD

72.1% 19.4% 52.7% 17.8% 10.1%

74.3% 18.3% 56.0% 17.1% 8.5%

Age <65 years n=47

Del 17p present n=56

Del 17p absent n=69

68% 21.2% 46.8% 19.1% 12.7%

71.4% 25.0% 46.4% 16.1% 12.5%

72% 16.0% 56.5% 18.8% 8.7%

Prior No prior ibrutinib ibrutinib therapy therapy n=107 n=22 69.1% 17.7% 51.4% 19.6% 11.2%

86.2% 27.2% 59.0% 9.0% 4.5%

BTK BTK PLCÎł2 mutation mutation mutation present absent present n=12 n=22 n=4 91.6% 8.3% 83.3% 8.3% 0%

72.6% 18.1% 54.5% 18.1% 9.0%

75.0% 0.0% 75.0% 25.0% 0%

PLCÎł2 mutation absent n=28 78.6% 14.3% 64.3% 14.3% 7.1%

CR: complete response; ORR: overall response rate; PD: progressive disease; PR: partial response; SD: stable disease. .

haematologica | 2018; 103(9)

1515

A.R. Mato et al. Table 4. First treatment following venetoclax discontinuation and treatment outcomes.

Treatment

Ibrutinib-based Idelalisib-based Rituximab monotherapy CAR-T Anthracycline-based (R-CHOP/R-EPOCH) Allogeneic SCT Other

Number treated with agent (Percentage of 24 patients who received subsequent line of therapy)

Patient level responses (n)

5 (20.8%) 2 (8.3%) 3 (12.5%) 2 (8.3%) 3 (12.5%) 3 (12.5%) 6 (25%)

PR (1), SD (2), PD (2) CR (1), No response assessment (1) PR (2), PD (1) No response assessment (2) PD (2), no response assessment (1) CR (2), no response assessment (1) PR (1), SD (1), PD (2), no response assessment (2)

due to other reasons with no patients discontinuing due to toxicity.2 We also note these results differ from recent real world BCR inhibitor series where AEs were the most common reason for drug discontinuation, followed by CLL progression.20 Even less is known about sequencing of therapies following venetoclax discontinuation. Anderson et al. report outcomes on 25 patients who progressed following venetoclax, 8 with progressive CLL and 17 with RT. In this series, 6 CLL patients with progression, all of whom were ibrutinib naĂŻve, were treated with a BTK inhibitor as the first therapy after discontinuation. Five of six (83%) initially achieved a partial response.2 Our study is unique in that we report patient level treatment data on 24 CLL patients progression following venetoclax, which represents the largest series reported to date. These patients are representative of the U.S. population currently treated with venetoclax in that they are treated in the R/R setting, and 89% had been exposed to a BCR inhibitor prior to venetoclax treatment, most commonly ibrutinib. We found that, following progression on venetoclax, ibrutinib was most commonly selected agent, accounting for 20.8% of the cases. However, idelalisibbased, rituximab monotherapy, CAR-T, anthracycline based therapy, and allogeneic stem cell transplant were also selected as next therapy in other cases. Interestingly, 3 patients underwent allogeneic stem cell transplant as first therapy after venetoclax. Two achieved a CR and 1 did not have an available response assessment. One patient received an allogeneic SCT as second therapy following venetoclax and achieved CR. Although interpretation of the SCT results are subject to selection bias, they suggest that there is a potential role for effective cellular therapies and should be explored. Our data demonstrate that no clear consensus exists for therapy selection following venetoclax failure and highlights the importance addressing sequencing strategies in future clinical trials. This will become increasingly important as more patients in practice are treated with venetoclax alone or in combination with antibodies and/or ibrutinib. Our study has several limitations. Data were collected retrospectively by multiple physicians and are subject to differences in clinical experience, practice style, and inconsistencies in chart review. Missing data varied with individual data points and were infrequent. To address this, we included absolute numbers and percentages to highlight any data that was not reported for individual data points. Additional data, including performance status, could offer additional insight but was not collected. Variables and out1516

comes, including TLS risk categorization, TLS events, and response were documented per physician assessment. Although we recommended the use of iwCLL response criteria, Howard criteria for TLS, and tumor burden classification per package insert, central review of outcomes was outside the scope of this study and, therefore, outcomes may have been subject to misclassification bias. While the case report form captured information on TLS prophylaxis and events, it was not designed to discern the detailed information that would be required to understand rate of TLS by management strategy. This was beyond the scope of this study but future research should consider examining this important information. Additionally, AE assessment was not comprehensive and included data on select AEs such as TLS, hematologic, infection or gastrointestinal toxicities. Applying CTCAE criteria retrospectively may result in underreporting of events and caution is emphasized in interpreting these findings. Indications for treatment were based on treating physiciansâ&#x20AC;&#x2122; discretion and were not specified. Detailed information regarding reason for discontinuation of line of therapy prior to venetoclax was not captured. As we know from prior studies, outcomes can differ significantly in patients who discontinue a kinase inhibitor due to toxicity as compared to progression.8 Future studies of venetoclax should consider stratifying patients by these subgroups. While we did include data from community practices, the fact that most patients were treated in academic centers could introduce a selection bias. Our median follow up of 7 months is short and does not capture progressions on venetoclax that occur later. Despite its inherent limitations, this series represents the largest real-world cohort of CLL patients treated with venetoclax. Neutropenia and thrombocytopenia were the most common toxicities, and progression was the leading cause of venetoclax discontinuation. Venetoclax was active in patients with mutations known to confer ibrutinib resistance as well as in patients with other poor risk features. However, TP53 interruption was associated with an inferior PFS. While we report the largest series of postvenetoclax outcomes, we demonstrate no clear sequencing pattern. Because the number of patients who discontinue venetoclax due to disease progression or toxicity within the first 2 years of initiating therapy is not trivial, understanding how these patients should be subsequently treated is a critical area of future research. Acknowledgments The authors would like to thank Joseph and Cindy Riggs for their support. haematologica | 2018; 103(9)

Real world venetoclax experience

References 1. Roberts AW, Davids MS, Pagel JM, Kahl BS, Puvvada SD, Gerecitano JF, et al. Targeting BCL2 with Venetoclax in Relapsed Chronic Lymphocytic Leukemia. N Engl J Med. 2016;374(4):311-22. 2. Anderson M, Tam C, Lew T, Juneja S, Juneja M, Westerman D, et al. Clinicopathological features and outcomes of progression of CLL on the BCL2 inhibitor venetoclax. Blood 2017; 129(25):3362-70. 3. Stilgenbauer S, Eichhorst B, Schetelig J, Coutre S, Seymour J, Munir T, et al. Venetoclax in relapsed or refractory chronic lymphocytic leukaemia with 17p deletion: a multicentre, open-label, phase 2 study. Lancet Oncol. 2016;17(6):768-78. 4. Coutre S, Choi M, Furman RR, Eradat H, Heffner L, Jones JA, et al. Venetoclax for patients with chronic lymphocytic leukemia who progressed during or after idelalisib therapy. Blood. 2018; 12;131(15):1704-11. 5. Jones J, Mato A, Wierda W, Davids M, Choi M, Cheson B, et al. Venetoclax for chronic lymphocytic leukaemia progressing after ibrutinib: an interim analysis of a multicentre, open-label, phase 2 trial. Lancet Oncol. 2018;19(1):65-75. 6. Seymour JF, Kipps TJ, Eichhorst B, Hillmen P, D'Rozario J, Assouline S, et al. Venetoclax-Rituximab in Relapsed or Refractory Chronic Lymphocytic Leukemia. N Engl J Med. 2018; 378(12):1107-20.

haematologica | 2018; 103(9)

7. Mato A, Nabhan C, Barr P, Ujjani C, Hill B, Lamanna N, et al. Outcomes of CLL patients treated with sequential kinase inhibitor therapy: a real world experience. Blood. 2016;128(18):2199-205. 8. Mato AR, Hill BT, Lamanna N, Barr PM, Ujjani CS, Brander DM, et al. Optimal sequencing of ibrutinib, idelalisib, and venetoclax in chronic lymphocytic leukemia: results from a multicenter study of 683 patients. Annals of oncology : official journal of the European Society for Medical Oncology. 2017;28(5):1050-6. 9. Forum UC. Ibrutinib for relapsed/refractory chronic lymphocytic leukemia: a UK and Ireland analysis of outcomes in 315 patients. Haematologica. 2016; 101(12): 1563-72. 10. Mato AR, Allan JN, Pagel JM, Brander DM, Hill BT, Cheson BD, et al. Front-Line Ibrutinib Therapy for Chronic Lymphocytic Leukemia (CLL) in the Real World: Responses, Toxicity, Outcomes and Subsequent Therapies. Blood 2017; 130(Supp 1):3011. 11. Barrientos JC, Kaur M, Mark A, Chung J, Driscoll N, Bender A, et al. Outcomes of Patients with Chronic Lymphocytic Leukemia (CLL) after Idelalisib Therapy Discontinuation. Blood. 2015;126(23): 4155. 12. Hallek M, Cheson B, Catovsky D, CaligarisCappio F, Dighiero G, Döhner H, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group

1996 guidelines. 2008;111(12):5445-56. 13. FDA. Highlights of Prescribing Information: Venclexta. 2018. 14. Howard S, Jones D, Pui C. The tumor lysis syndrome. N Engl J Med. 2011; 364(19):1844-54. 15. Bland J, Altman D. Survival probabilities (the Kaplan-Meier method). BMJ. 1998; 317(7172):1572. 16. Matthews D, Farewell V. 7 The Log-Rank or Mantel-Haenszel Test for the Comparison of Survival Curves In: Basel S, Karger A, eds. Using and Understanding Medical Statistics 2007:67-75. 17. Anderson P, Gill R. Cox’s regression model for counting processes: a large sample study. Ann Statist. 1982;10(4):1100-20. 18. Kipps TJ, Fraser G, Coutre SE, Brown JR, Barrientos JC, Barr PM, et al. Integrated analysis: outcomes of ibrutinib-treated patients with chronic lymphocytic leukemia/small lymphocytic leukemia (CLL?SLL) with high-risk Prognostic Factors. Hematological Oncology. 2017; 35(S2):109-11. 19. Mato A, Wierda W, Davids M, Cheson B, Coutre S, Choi M, et al. Analysis of PETCT to Identify Richter’s Transformation in 167 Patients with Disease Progression Following Kinase Inhibitor Therapy. ASH Abstract 817;2017. 20. Mato A, Nabhan C, Thompson M, Lamanna N, Brander D, Hill B, et al. Toxicities and outcomes of 616 ibrutinibtreated patients in the United States: a realworld analysis. Haematologica. 2018; 103(5):874-9.

1517

ARTICLE

Plasma Cell DIsorders

Ferrata Storti Foundation

)

Haematologica 2018 Volume 103(9):1518-1526

A phase I/II dose-escalation study investigating all-oral ixazomib-melphalanprednisone induction followed by single-agent ixazomib maintenance in transplant-ineligible newly diagnosed multiple myeloma Jesús F. San-Miguel,1 Maria-Asunción Echeveste Gutierrez,2 Ivan Špicka,3 María-Victoria Mateos,4 Kevin Song,5 Michael D. Craig,6 Joan Bladé,7 Roman Hájek,8 Christine Chen,9 Alessandra Di Bacco,10 Jose Estevam,10 Neeraj Gupta,10 Catriona Byrne,10 Vickie Lu,10 Helgi van de Velde10* and Sagar Lonial11

Clinica Universidad de Navarra, Centro Investigación Medica Aplicada (CIMA), El Instituto de Investigación Sanitaria de Navarra (IDISNA), Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), Pamplona, Spain; 2Hospital Universitario Donostia, San Sebastián, Spain; 31st Medical Department - Clinical Department of Haematology, First Faculty of Medicine and General Teaching Hospital, Charles University, Prague, Czech Republic; 4Hospital Universitario de Salamanca, Instituto Biosanitario de Salamanca (IBSAL), Spain; 5Division of Hematology, University of British Columbia, Vancouver, BC, Canada; 6Department of Medicine, West Virginia University, Morgantown, WV, USA; 7Department of Hematology, Hospital Clínic de Barcelona, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Spain; 8 Department of Haematooncology, University Hospital Ostrava, Faculty of Medicine, Ostrava University, Czech Republic; 9Cancer Clinical Research Unit, Princess Margaret Cancer Center, Toronto, ON, Canada; 10Millennium Pharmaceuticals, Inc., Cambridge, MA, USA, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited and 11 Winship Cancer Institute of Emory University, Atlanta, GA, USA 1

*Current affiliation: Sanofi, Cambridge, MA, USA

ABSTRACT

Correspondence: sanmiguel@unav.es

Received: February 2, 2018. Accepted: June 25, 2018. Pre-published: June 28, 2018. doi:10.3324/haematol.2017.185991 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1518 ©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.

1518

his phase I/II dose-escalation study investigated the all-oral ixazomib-melphalan-prednisone regimen, followed by single-agent ixazomib maintenance, in elderly, transplant-ineligible patients with newly diagnosed multiple myeloma. Primary phase I objectives were to determine the safety and recommended phase II dose of ixazomib-melphalan-prednisone. The primary phase II objective was to determine the complete plus very good partial response rate. In phase I, patients were enrolled to 4 arms investigating weekly or twice-weekly ixazomib (13 28-day cycles or nine 42-day cycles) plus melphalan-prednisone. In phase II, an expansion cohort was enrolled at the recommended phase II ixazomib dose. Of the 61 patients enrolled, 26 received the recommended phase II dose (ixazomib 4.0 mg [days 1, 8, 15] plus melphalan-prednisone 60 mg/m2 [days 1-4], 28-day cycles). Of the 61 enrolled patients, 36 (13 of 26 in the recommended phase II dose cohort) received single-agent ixazomib maintenance (days 1, 8, 15; 28-day cycles). In phase I, 10/38 patients reported dose-limiting toxicities in cycle 1, including grade 3 and/or 4 neutropenia (n=6) and thrombocytopenia (n=4). Complete plus very good partial response rate was 48% (48% at recommended phase II dose), including 28% (22%) complete response or better; responses deepened during maintenance in 34% (33%) of evaluable patients. After median follow up of 43.6 months, median progression-free survival was 22.1 months. Adverse events were mainly hematologic events, gastrointestinal events, and peripheral neuropathy. This study demonstrates the feasibility, tolerability, and activity of ixazomib-melphalan-prednisone induction and single-agent ixazomib maintenance in transplant-ineligible newly diagnosed multiple myeloma patients. clinicaltrials.gov identifier 01335685. haematologica | 2018; 103(9)

IMP phase I/II study in transplant-ineligible NDMM

Introduction

Methods

Although multiple myeloma (MM) remains, for most patients, an incurable hematologic malignancy, recent advances in treatment and diagnosis have led to substantial improvements in both progression-free survival (PFS) and overall survival (OS).1-4 As with many malignancies, younger, fitter patients usually achieve the best outcomes with initial treatment, while outcomes for elderly patients and those with comorbidities, who are unable to tolerate high-dose therapy (HDT) followed by autologous stem cell transplantation (ASCT), have traditionally lagged behind.3-8 For these elderly and frail patients, active, novel, frontline combination regimens are needed to achieve the best long-term outcomes. However, tolerability can be an issue for some patients,5,6 particularly in the case of long-term, continuous therapy, which is associated with improved outcomes.9-13 Following demonstration of favorable efficacy and tolerability in phase III trials,14-17 the combination of the proteasome inhibitor (PI) bortezomib plus melphalan and prednisone (VMP) is now, in many geographies, a standard-of-care regimen for the first-line treatment of elderly patients with newly diagnosed MM (NDMM) who are not eligible to receive HDT/ASCT because of age-related frailty and/or comorbidity.3,4 VMP represents an active, feasible frontline treatment option, including for patients with high-risk cytogenetic abnormalities (due to the activity of PI-based regimens in this population) and patients with renal impairment (as no starting dose adjustment is required), and offers a suitable option for patients in whom immunomodulatory drug-containing therapy is contraindicated.18-22 However, despite being a standard of care, the parenteral administration of bortezomib may create a burden for elderly patients, limiting its feasibility for long-term use. The combination of another parenterally administered PI, carfilzomib, and melphalan-prednisone (MP) was recently compared with VMP and demonstrated no statistically significant difference in PFS in transplant-ineligible NDMM; however, carfilzomib-MP was associated with a higher number of specific grade â&#x2030;Ľ3 adverse events (AEs), notably acute renal failure, cardiac failure, dyspnea, and hypertension, and fewer incidences of peripheral neuropathy (PN) than VMP.23 Therefore, there remains a need for a tolerable, efficacious, and convenient PI option for elderly patients with transplant-ineligible NDMM. Ixazomib is an oral PI with a safety profile amenable to extended dosing.24-26 Based on the results of the TOURMALINE-MM1 study, which led to its first approval in 2015, ixazomib has been approved in more than 50 countries worldwide, including the US, EU, and Japan, for use in combination with lenalidomide-dexamethasone (IRd) for the treatment of MM patients who have received at least one prior therapy.26-28 Recent phase I/II studies have demonstrated the activity and tolerability of ixazomib-based induction (IRd) and long-term ixazomib maintenance therapy in NDMM, demonstrating the feasibility of this approach.29,30 This phase I/II trial (clinicaltrials.gov identifier 01335685) was undertaken to evaluate the all-oral ixazomib-MP (IMP) induction regimen, followed by long-term maintenance with single-agent ixazomib, in predominantly elderly, transplant-ineligible patients with NDMM.

Study design

haematologica | 2018; 103(9)

This was a phase I/II, open-label, multicenter, dose-escalation study. The primary phase I objectives were to determine safety, maximum tolerated dose (MTD), and recommended phase II dose (RP2D) of ixazomib in combination with MP. A secondary objective was to characterize ixazomib pharmacokinetics. The primary phase II objective was to determine the complete response plus very good partial response (CR+VGPR) rate. Secondary objectives included overall response rate (ORR), time to response, duration of response, PFS, time to progression, OS, and safety (for details, see the Online Supplementary Material).

Patients Patients with previously untreated MM who were ineligible for HDT/ASCT due to age (â&#x2030;Ľ65 years) or comorbidity, and for whom standard MP treatment was indicated, were enrolled. Detailed eligibility criteria are presented in the Online Supplementary Material. The study complied with regulatory requirements, the Declaration of Helsinki, and Good Clinical Practice standards. Independent review boards/ethics committees approved the study. Patients gave written informed consent.

Treatment In phase I, patients were enrolled into one of four arms, as assigned by investigators under direction of the sponsor (Figure 1). In Arm A, patients received up to 9 42-day cycles of twice-weekly ixazomib (days 1, 4, 8, 11, 22, 25, 29, 32). In Arm B, patients received up to 13 28-day cycles of weekly ixazomib (days 1, 8, 15). In Arms C and D, patients received up to 9 42-day cycles of weekly ixazomib (days 1, 8, 15, 22, 29 for Arm C and days 1, 8, 22, 29 for Arm D). Patients also received melphalan 6 mg/m2 (Arm B) or 9 mg/m2 (Arms A, C, D) on days 1-4 and prednisone 60 mg/m2 (days 1-4) in each cycle. Ixazomib dose-escalation proceeded via a standard 3+3 design based on cycle 1 dose-limiting toxicities (DLTs; as defined in the Online Supplementary Material). The MTD required no more than 1 out of 6 DLT-evaluable patients to have a first-cycle DLT. Planned dose levels for ixazomib are shown in Online Supplementary Table S1. In phase II, an expansion cohort was enrolled at the RP2D, which was established by considering all available phase I toxicity (grade 3/4 AEs, serious AEs [SAEs], all-grade PN, and treatment discontinuation) and ORR over multiple cycles. After induction, patients with stable disease or better could receive single-agent ixazomib maintenance (at the dose tolerated for induction) on days 1, 8, 15 for up to 12 28-day cycles, or until disease progression or unacceptable toxicity (Figure 1).

Assessments Responses were assessed by investigators on day 1 of each cycle, at the end of induction, every 2 cycles during maintenance, and every 16 weeks during follow up until progression or start of subsequent antineoplastic therapy, according to International Myeloma Working Group criteria.31 AEs were monitored throughout and graded using the National Cancer InstituteCommon Terminology Criteria for AEs, version 4.03. Details of the pharmacokinetic, minimal residual disease (MRD), and safety assessments are provided in the Online Supplementary Material.

Analyses Analysis populations are defined in the Online Supplementary Material. Time-to-event endpoints were analyzed using survival 1519

J.F. San-Miguel et al.

Figure 1. Phase I study design. PD: progressive disease.

analysis techniques based on Kaplan-Meier estimates. All data were summarized using descriptive statistics.

Table 1. Patient demographics and disease characteristics at baseline (safety population).

Characteristic

Results Patient disposition and baseline characteristics A total of 61 patients were enrolled: 11, 34, 10, and 6 to Arms A, B, C, and D, respectively (Tables 1 and 2). All patients received ≥1 dose of any study drug and were included in the safety population; 26 of these patients received ixazomib at the RP2D (Tables 1 and 2). The baseline demographics and disease characteristics of the safety population are shown in Table 1. Seven patients had high-risk cytogenetics; all were enrolled in Arm B.

Dose-limiting toxicities and recommended phase II dose During phase I, 38 patients (9, 14, 9, and 6 in Arms A, B, C, and D, respectively) were evaluable for assessment of DLTs. Among these patients, 10 (26%) experienced a total of 16 DLTs in cycle 1. All DLTs were grade 3 or grade 4 in intensity. The RP2D was determined as weekly ixazomib 4.0 mg (days 1, 8, and 15 of 28-day cycles) based on the Arm B MTD, ORR (Online Supplementary Table S4), and observed rates of toxicity across multiple cycles. Baseline characteristics for this RP2D cohort were similar to those for the total study population (Table 1). Detailed descriptions of DLTs and determination of the RP2D can be found in the Online Supplementary Material.

Treatment exposure At final analysis, 4 patients remained on treatment. Primary reasons for discontinuation were progressive disease, patient withdrawal, and completion of protocol-specified treatment (Table 2). After a median fol1520

Total (N=61)

Median age, years (range) 74 (63-90) Male, n (%) 23 (38) Race, n (%) White 57 (93) Black / African American 1 (2) Asian 1 (2) Other 2 (3) ECOG performance status, n (%) 0 17 (28) 1 33 (54) 2 11 (18) ISS stage, n (%) I 13 (21) II 31 (51) III 17 (28)* Type of myeloma at initial diagnosis, n (%) IgG 36 (59) IgA myeloma 20 (33) Light-chain disease 5 (8) Extramedullary disease, n (%) 6 (10) High-risk cytogenetics, n (%)† 7 (12) Median β2M, mg/L (range) 4.3 (2.1-14.0) CrCl ≤60 mL/min, n (%) 34 (56)

RP2D 4.0 mg Arm B (N=26) 74 (67-84) 10 (38) 24 (92) 1 (4) 0 1 (4) 7 (27) 13 (50) 6 (23) 5 (19) 14 (54) 7 (27) 20 (77) 5 (19) 1 (4) 2 (8) 5 (21) 4.6 (2.4-14.0) 13 (50)

β2M: beta-2 microglobulin; CrCl: creatinine clearance; ECOG: Eastern Cooperative Oncology Group; Ig: immunoglobulin; ISS: International Staging System; RP2D: recommended phase II dose. The safety population was defined as all patients receiving ≥1 dose of any study drug. *Unknown for two patients; †High-risk cytogenetics includes del17p, t(4:14), and t(14:16) abnormalities.

haematologica | 2018; 103(9)

IMP phase I/II study in transplant-ineligible NDMM Table 2. Ixazomib dose received and primary reason for discontinuation by study arm (safety population).

Arm A

Arm B

Ixazomib dose received, mg

3.0

3.7

3.0

Patients, N Primary reason for discontinuation, n (%) PD AEs Patient withdrawal Completion of protocol-specified treatment Unsatisfactory therapeutic response Preference for immunomodulatory therapy given the PR Study terminated by sponsor

5 (71) 0 1 (14) 1 (14) 0 0 0

Arm C

Arm D

Total*

5.5

3.0

4.0

4.0 (RP2D) 26

2 (50) 0 2 (50) 0 0 0

2 (67) 0 0 1 (33) 0 0

11 (42) 8 (31) 2 (8) 2 (8) 1 (4) 1 (4)

3 (60) 2 (40) 0 0 0 0

3 (50) 0 0 0 1 (17) 0

1 (25) 1 (25) 1 (25) 1 (25) 0 0

2 (33) 2 (33) 0 1 (17) 0 0

29 (48) 13 (21) 6 (10) 6 (10) 2 (3) 1 (2)

1 (4)

2 (33)

1 (17)

4 (7)

AE: adverse event; PD: progressive disease; PR: partial response; RP2D: recommended phase II dose. *Discontinuation due to – PD: 8 during induction, 21 during maintenance; AEs: 11 during induction, 2 during maintenance; patient withdrawal: 3 during induction, 3 during maintenance; completion of protocol-specified treatment: all 6 during induction; unsatisfactory therapeutic response: both during induction; preference for immunomodulatory therapy: during induction; study terminated: all 4 during maintenance.

low up for OS of 43.6 months, patients had received a median of 16 cycles of ixazomib (12.5 cycles in the RP2D cohort; Online Supplementary Table S2). A total of 36 patients entered the maintenance phase (n=13 at the RP2D), and received a median number of maintenance cycles of 12, with a maximum duration of ixazomib treatment of 58 months (Online Supplementary Table S2). Thirteen patients (36%) remained on maintenance therapy for ≥13 cycles (≥1 year), and 5 patients (14%) remained on maintenance for ≥25 cycles (≥2 years). Mean relative dose intensity over the whole study for ixazomib was 82.8% (87.1% at the RP2D), and ≥90% for both melphalan and prednisone (Online Supplementary Table S2). Pharmacokinetic analyses are shown in Online Supplementary Table S3.

Efficacy Fifty-three patients were evaluable for response, including 23 at the RP2D. Among response-evaluable patients, the confirmed CR+VGPR rate at the end of study was 48%, including 28% ≥CR (Table 3). The confirmed ORR at end of study was 66%, with 86% of patients achieving a ≥50% reduction in serum M-protein. In both the total population and at the RP2D, 48% of patients demonstrated a 100% reduction in their serum M-protein (Table 3). Median time to ≥VGPR and CR was 3.7 and 11.6 months, respectively (Table 4). Of the 7 high-risk patients, 1 patient achieved a CR and 3 patients achieved a PR; 2 patients were not evaluable for response. Responses deepened during maintenance with single-agent ixazomib: in 11/32 (34%) response-evaluable patients overall (CR to sCR in 2 patients; VGPR to sCR in 5 patients; VGPR to CR in 3 patients; and PR to VGPR in 1 patient); and in 4/12 (33%) response-evaluable patients who received the RP2D (CR to sCR in 1 patient; VGPR to sCR in 2 patients; and PR to VGPR in 1 patient). The confirmed CR rate was 13% after induction, rising to 28% at the end of treatment (Table 3). Thirteen of 53 (24%) response-evaluable patients were assessed for minimal residual disease (MRD) by flow cytometry, 5 of whom were in the RP2D cohort. Of these 13 patients, 12 had a best confirmed response of ≥CR and one had a best confirmed response of VGPR. MRD results haematologica | 2018; 103(9)

in 9 of the 12 patients with ≥CR (75%) (3 at the RP2D) were found to be negative. Therefore, in the total study population, 9 of 53 response-evaluable patients (17%; 3 of 23 [13%] in the RP2D cohort) were MRD-negative. Evaluation of time-to-event data demonstrated the durability of responses (Table 4). Median time to best response (≥PR) was 4.6 months in the total study population and at the RP2D. Median duration of response was 22.6 months overall and 25.4 months in patients achieving ≥VGPR (Table 4). Median PFS was 22.1 months overall and 18.4 months at the RP2D after median follow up for PFS of 18.0 and 10.2 months, respectively (Figure 2A and Table 4). For patients who entered the maintenance phase, median PFS was 27.5 months (38.7 months at the RP2D) (Figure 2A and Table 4); median PFS for standard-risk patients who entered the maintenance phase was similar, at 28.8 months (38.7 months at the RP2D). Median OS was 54.4 months overall and not reached at the RP2D after median follow up of 43.6 months in the total population and 48.6 months in Arm B, respectively (Figure 2B and Table 4).

Safety Safety profiles during induction and maintenance are shown in Table 5, and the most common toxicities are shown in Table 6. The most common grade ≥3 AEs (≥10% incidence) were thrombocytopenia, neutropenia, lymphopenia, leukopenia, anemia, and diarrhea (Table 6). Hematologic toxicities were less common at the RP2D than in the total population. The most common SAE was pneumonia (n=6 [10%]; n=2 [8%] at the RP2D). The only AE to result in discontinuation of study treatment in more than one patient was thrombocytopenia (n=3, 5%). None of the 3 on-study deaths (all in the RP2D cohort; attributed to pneumonia, septic shock, and worsening of end-stage MM, respectively) were considered by investigators to be related to study treatment. There was a limited incidence of new-onset toxicities during single-agent ixazomib maintenance compared with IMP induction. Any-grade AEs with a ≥15% difference between patients who entered the maintenance period and those who did not were thrombocytopenia (64% for 1521

J.F. San-Miguel et al. Table 3. Response rates after induction and at end of study (response-evaluable population).

Table 4. Time-to-event outcomes with IMP induction and single-agent ixazomib maintenance.

n (%)

Outcome (in months unless otherwise stated)

Response after induction ORR (≥PR) CR (confirmed) sCR (confirmed) VGPR CR+VGPR (confirmed) ≥50% reduction in M-protein Response at end of study ORR CR (confirmed) sCR (confirmed) VGPR CR+VGPR (confirmed) ≥50% reduction in M-protein 100% reduction in M-protein

Total

RP2D 4.0 mg Arm B

N=53* 35 (66) 7 (13) 3 (6) 16 (30) 23 (43) 41 (82) N=53* 35 (66) 15 (28) 10 (19) 9 (17) 24 (48) 43 (86) 24 (48)

N=23 15 (65) 3 (13) 1 (4) 7 (30) 10 (43) 17 (77) n=23 15 (65) 5 (22) 4 (17) 6 (26) 11 (48) 18 (82) 11 (48)

CR: complete response; ORR: overall response rate; PR: partial response; R2PD: recommended phase II dose; sCR: stringent complete response; VGPR: very good partial response. Response-evaluable population was defined as patients receiving ≥5/8 (Arm A), ≥2/3 (Arm B), ≥4/5 (Arm C), or ≥3/4 (Arm D) doses of ixazomib during cycle 1 with measurable disease at baseline and 1 post-baseline response assessment. *Eight patients were not evaluable for response due to: no measurable disease (two patients); no post-baseline assessment (one patient); and incomplete dosing in cycle 1 (5 patients).

induction-only patients vs. 86% for maintenance patients), lymphopenia (28% vs. 44%), anemia (60% vs. 36%), constipation (52% vs. 33%), and rash (16% vs. 33%). Any-grade PN (classified by the high-level term peripheral neuropathies not elsewhere classified) considered to be study drug-related was reported in 24 patients (39%) (Table 6). PN was primarily low grade, with 12 patients (20%; 5 [19%] at the RP2D) and 19 patients (15%; 5 [19%] at the RP2D) reporting grade 1 and grade 2 PN, respectively. Three patients (5%) had grade 3 PN events. No patient had grade 4. Overall, 8 patients (13%) received dose reductions and 7 patients (11%) had study drug held due to PN. Twenty of the 24 patients (83%) who developed PN events during induction or maintenance had improved symptoms by the end of the study, with 17 (71%) having complete resolution of symptoms. Median time to resolution of PN events was 4.6 months (95% confidence interval: 1.6-14.3). Median time to resolution or improvement of PN events was 1.7 months (95% confidence interval: 1.1-6.4).

Discussion A PI, namely bortezomib, combined with MP has been shown to be an effective frontline treatment approach for NDMM patients unable to undergo HDT/ASCT due to advanced age and/or significant comorbidities, including those for whom immunomodulatory drugs are not an option.16,17,23 Most studies of the VMP regimen have utilized a fixed duration of treatment (often approximately 1 year) rather than extended or continuous therapy.17,19,32,33 Furthermore, in the real-world clinical 1522

Median time to first response (range)* Median time to first ≥VGPR (range)* Median time to first CR (range)* Median DOR (95% CI)*† Median DOR in patients achieving ≥VGPR (95% CI)* Median PFS (95% CI)†‡ Median PFS in patients who entered maintenance phase (95% CI)‡ Median time to progression (95% CI)†‡ Median OS (95% CI)†‡ Estimated 3-year OS rate, %†‡

Total (N=61)

RP2D 4.0 mg Arm B (n=26)

1.7 (1-7) 1.9 (1-7) 3.7 (1-13) 3.7 (1-13) 11.6 (1-23) 9.5 (5-22) 22.6 (15.9, 32.4) 25.2 (4.6, NR) 25.4 (15.0, 29.5) 29.5 (2.8, NR) 22.1 (18.0, 30.0) 18.4 (8.3, 38.7) 27.5 (18.7, 37.8) 38.7 (15.6, NR) 23.5 (18.0, 30.0) 22.1 (8.8, NR) 54.4 (39.7, NR) NR (35.0, NR) 73 68

CI: confidence interval; CR: complete response; DOR: duration of response; NR: not reached; OS: overall survival; PFS: progression-free survival; RP2D: recommended phase II dose; VGPR: very good partial response. *Response-evaluable population (total, N=53; RP2D 4.0 mg Arm B, n=23); †Median follow up: 18.4 (total) and 16.6 (RP2D) months for DOR; 18.0 (total) and 10.2 (RP2D) months for PFS and time to progression; and 43.6 (total) and 48.6 (RP2D) months for OS. ‡Safety population.

practice setting, early discontinuations due to toxicities are common.34 As long-term, continuous therapy is associated with improved outcomes,9-13 a tolerable, more convenient treatment regimen suitable for long-term use is needed, especially in elderly patients. The data from this study suggest that all-oral IMP induction followed by single-agent ixazomib maintenance is an active and well-tolerated frontline regimen in transplant-ineligible patients with NDMM. The regimen showed encouraging tolerability over a prolonged treatment period, with ≥50% of patients proceeding to maintenance, including at the RP2D, and duration of therapy of up to 4.8 years. The regimen also demonstrated high response rates, with an overall CR+VGPR rate of 48%, including 28% ≥CR. Additionally, lengthy outcomes were reported, with an overall median PFS of 22.1 months, and an overall median OS of 54.4 months. Differences in outcomes between the overall and RP2D populations should be interpreted with caution due to the relatively small numbers of patients involved and differences in patient characteristics (for example, 5 of the 7 patients with high-risk cytogenetic abnormalities were in the RP2D cohort). Additionally, the median follow up for PFS in the RP2D cohort was shorter than in the overall population, which also included patients who received more dose-intense regimens that may have resulted in improved short-term efficacy (ORR). While the efficacy data reported here were obtained in a relatively small number of patients within the context of a non-comparative early-phase trial, they appear similar to those seen with VMP or carfilzomib plus MP in phase III studies in transplant-ineligible patients with haematologica | 2018; 103(9)

IMP phase I/II study in transplant-ineligible NDMM

Figure 2. Kaplan-Meier analysis of PFS and OS. (A) PFS for the total patient population in the overall study (induction and maintenance phases) and in patients who went on to receive maintenance, and (B) OS in the overall study, for the total safety population and the subset of patients treated at the RP2D (4.0 mg) in Arm B. OS: overall survival; PFS: progression-free survival; RP2D: recommended phase II dose. Figure 2A, one patient in the RP2D group with PD entered maintenance, the patient was identified later following reassessment of the data.

NDMM.14,16,17,19,23,32,33 The CR+VGPR rates post-IMP (43%) and overall (48%) are comparable to rates reported for VMP (41–50% in the phase III VISTA, GIMEMA-MM-03-05, and ALCYONE trials17,19,33), and median PFS (22.1 months) also appeared similar to that reported with VMP and carfilzomib-MP (18.1–27.3 months).17,19,23,33 Similar CR+VGPR rates (47–49%) and median PFS (21–26 months) were seen with fixed-duration and continuous lenalidomide-dexamethasone (Rd) in the FIRST phase III trial in transplant-ineligible NDMM,35 and while responses and outcomes appeared better with daratumumab-VMP in ALCYONE (71% CR+VGPR, 18-month PFS 71.6%),33 and with IRd (58–63%, median PFS 29.4–35.4 months)29 and bortezomib-Rd (44%, median PFS 43 months) in NDMM patients,36 these differences should be considered in the context of the addition of daratumumab as a fourth induction agent and as maintenance therapy in ALCYONE, and the inclusion of a high proportion of transplant-eligible patients in the haematologica | 2018; 103(9)

IRd (35%)29 and VRd (69%) studies.36 The overall CR rate (28%) seen with IMP plus ixazomib maintenance was also comparable to those reported for VMP without maintenance (24–33%17,19,23,33). Although the post-IMP induction CR rate was lower than that reported for VMP in the VISTA study (13% vs. 31%), the CR rate increased to 28% during maintenance. This difference may simply reflect the longer median time to CR observed with IMP (11.6 months, compared with 4.2 months for VMP in the VISTA study).17 Indeed, the time to first response with IMP (1.7 months) was similar to that seen in the VISTA study (1.4 months). These findings together suggest that the response with IMP may mature over a longer period compared with VMP and similar response rates can be achieved with IMP followed by ixazomib maintenance and VMP without maintenance.17,19,23,33 It should also be noted that the weekly IMP regimen used at the RP2D was similar to the less-intense weekly VMP regimen used in the PETHEMA/GEM05 study, followed by bortezomibbased maintenance.14 In the overall study, the CR rate 1523

J.F. San-Miguel et al. Table 5. Safety profile with IMP induction and single-agent ixazomib maintenance (safety population).

Overall n (%)

Grade ≥3 AE Serious AE AE leading to discontinuation of any study drug AE leading to dose reduction of any study drug On-study death

Total (N=61)

RP2D 4.0 mg Arm B (N=26)

54 (89) 31 (51) 15 (25) 32 (52) 3 (5)

21 (81) 12 (46) 8 (31) 13 (50) 3 (12)

New-onset AE during induction Total RP2D 4.0 mg (N=61) Arm B (N=26)

New-onset AE during maintenance Total RP2D 4.0 mg (N=36) Arm B (N=13)

54 (89) 28 (46) 13 (21) 31 (51) 3 (5)*

18 (50) 8 (22) 2 (6) 3 (8) 0

21 (81) 11 (42) 6 (23) 12 (46) 3 (12)*

5 (38) 2 (15) 2 (15) 3 (23) 0

AE: adverse event; IMP: ixazomib-melphalan-prednisone; RP2D: recommended phase II dose. *On-study deaths: disease progression in 1 patient, and pneumonia and septic shock in 1 patient each (not considered drug related).

Table 6. Most common (≥30% incidence in either population) any-grade and grade ≥3 AEs (safety population).

Any-grade AE n (%)

Total (N=61)

RP2D 4.0 mg Arm B (n=26)

Total (N=61)

Thrombocytopenia Diarrhea Neutropenia Nausea Anemia Vomiting Constipation PN NEC* Lymphopenia Asthenia Decreased appetite Pyrexia Leukopenia Rashes, eruptions and exanthems NEC* Fatigue

47 (77) 38 (62) 38 (62) 33 (54) 31 (51) 28 (46) 25 (41) 24 (39) 23 (38) 22 (36) 22 (36) 20 (33) 19 (31) 18 (30) 17 (28)

16 (62) 17 (65) 12 (46) 11 (42) 9 (35) 11 (42) 6 (23) 11 (42) 8 (31) 9 (35) 8 (31) 6 (23) 7 (27) 6 (23) 9 (35)

30 (49) 7 (11) 27 (44) 0 9 (15) 2 (3) 1 (2) 3 (5) 18 (30) 4 (7) 0 1 (2) 12 (20) 4 (7) 2 (3)

Grade ≥3 AE

RP2D 4.0 mg Arm B (n=26) 7 (27) 4 (15) 6 (23) 0 4 (15) 0 0 1 (4) 5 (19) 1 (4) 0 0 3 (12) 1 (4) 2 (8)

AE: adverse event; NEC: not elsewhere classified; PN: peripheral neuropathy; RP2D: recommended phase II dose. *Higher-level terms including multiple preferred terms: PN NEC includes peripheral sensory neuropathy, neuropathy peripheral and polyneuropathy; rashes eruptions and exanthems NEC includes rash macular, rash maculo-papular, rash, rash papular, rash generalized.

post-VMP induction was 20%,14 and in a matched-pairs analysis comparing the PETHEMA/GEM05 and VISTA regimens, the CR rate in PETHEMA/GEM05 patients improved from 19% post-induction to 30% overall, following maintenance.32 As shown by the improved responses in >30% of patients during extended treatment with single-agent ixazomib maintenance, patients continued to derive clinical benefit from long-term single-agent ixazomib maintenance, an observation consistent with results from other early-phase ixazomib studies.29,30 The overall safety profile was as expected based on previous studies of ixazomib regimens and MP,24,26,37-41 with most AEs being hematologic and gastrointestinal, which are among the common AEs reported with melphalan and proteasome inhibitors, including ixazomib.42 Dose reductions or discontinuations of any study drug were required in approximately half and a quarter of patients, respectively. The incidence of PN was comparable to that reported in studies of IRd in NDMM,25,30 and appeared limited when compared with that reported with a VMP regimen incorporating twice-weekly intravenous bortezomib (13% grade ≥3).43 Importantly, >80% of reported PN 1524

events in the present study resolved or improved by study end. As suggested for IRd,26 the all-oral IMP regimen would be expected to be convenient for patients, and the number of planned visits to the clinic for administration of the regimen would be expected to be lower than with the VMP regimen, resulting in a lower patient burden. Our experience with regards to weekly and twice-weekly ixazomib dosing and dose level is in line with experience with IRd,44 which suggests that twice-weekly ixazomib may be associated with some additional toxicity, notably an increase in PN and rash.29,30 Importantly, the continued clinical benefit demonstrated with weekly single-agent ixazomib maintenance therapy was complemented by a favorable tolerability profile. The majority of AEs were observed during the induction period, only 6% of patients discontinued ixazomib maintenance because of AEs, and no on-study deaths occurred during maintenance. The number of patients who continued on long-term single-agent ixazomib maintenance therapy further emphasizes the tolerability of this regimen. Based on the efficacy and tolerability of single-agent maintenance seen in this and other trials, weekly ixazomib is under phase III haematologica | 2018; 103(9)

IMP phase I/II study in transplant-ineligible NDMM

investigation as MM maintenance therapy following ASCT (TOURMALINE-MM3; clinicaltrials.gov identifier 02181413). A second phase III study is also investigating weekly ixazomib as maintenance therapy after initial induction therapy without ASCT (TOURMALINE-MM4; clinicaltrials.gov identifier 02312258). Phase III investigation of IMP followed by ixazomib maintenance therapy is not currently planned. In conclusion, this study demonstrates the feasibility, tolerability, and antimyeloma activity of the all-oral IMP induction regimen followed by long-term maintenance with single-agent oral ixazomib in elderly, transplant-ineligible patients with NDMM. Oral dosing, coupled with a favorable safety profile at the RP2D, make ixazomib particularly suitable for long-term continuous therapy and may offer a more convenient, active, and well-tolerated alternative to a parenterally administered PI in this setting.

References 1. Genadieva Stavric S, Bonello F, Bringhen S, Boccadoro M, Larocca A. How is patient care for multiple myeloma advancing? Expert Rev Hematol. 2017;10(6):551-561. 2. Moreau P, de Wit E. Recent progress in relapsed multiple myeloma therapy: implications for treatment decisions. Br J Haematol. 2017;179(2):198-218. 3. 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. 4. Palumbo A, Rajkumar SV, San Miguel JF, et al. International Myeloma Working Group consensus statement for the management, treatment, and supportive care of patients with myeloma not eligible for standard autologous stem-cell transplantation. J Clin Oncol. 2014;32(6):587-600. 5. 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. 6. Pawlyn C, Gay F, Larocca A, Roy V, Ailawadhi S. Nuances in the Management of Older People With Multiple Myeloma. Curr Hematol Malig Rep. 2016;11(3):241251. 7. Wildes TM, Campagnaro E. Management of multiple myeloma in older adults: Gaining ground with geriatric assessment. J Geriatr Oncol. 2017;8(1):1-7. 8. Dimopoulos MA, Kastritis E, Delimpasi S, et al. Multiple myeloma in octogenarians: clinical features and outcome in the novel agent era. Eur J Haematol. 2012;89(1):1015. 9. Attal M, Lauwers-Cances V, Marit G, et al. Lenalidomide maintenance after stem-cell transplantation for multiple myeloma. N Engl J Med. 2012;366(19):1782-1791. 10. Brioli A, Tacchetti P, Zamagni E, Cavo M. Maintenance therapy in newly diagnosed multiple myeloma: current recommendations. Expert Rev Anticancer Ther. 2014;14(5):581-594. 11. Lipe B, Vukas R, Mikhael J. The role of maintenance therapy in multiple myeloma. Blood Cancer J. 2016;6(10):e485. 12. Mateos MV, Oriol A, Martinez-Lopez J, et

haematologica | 2018; 103(9)

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Acknowledgments The authors would like to thank all patients and their families, physicians, research nurses, study coordinators, and research staff participating in this study. The authors acknowledge Jane Saunders and Laura Webb of FireKite, an Ashfield company, part of UDG Healthcare plc, for writing assistance during the development of this manuscript, which was funded by Millennium Pharmaceuticals Inc., and complied with Good Publication Practice 3 ethical guidelines (Battisti WP, et al. Ann Intern Med 2015;163:461-4) and Renda Ferrari, PhD, of Millennium Pharmaceuticals, Inc., Cambridge, MA, USA, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited, for editorial support. Funding This study was sponsored by Millennium Pharmaceuticals, Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited.

al. Maintenance therapy with bortezomib plus thalidomide or bortezomib plus prednisone in elderly multiple myeloma patients included in the GEM2005MAS65 trial. Blood. 2012;120(13):2581-2588. Palumbo A, Hajek R, Delforge M, et al. Continuous lenalidomide treatment for newly diagnosed multiple myeloma. N Engl J Med. 2012;366(19):1759-1769. Mateos MV, Oriol A, Martinez-Lopez J, et al. GEM2005 trial update comparing VMP/VTP as induction in elderly multiple myeloma patients: do we still need alkylators? Blood. 2014;124(12):1887-1893. Niesvizky R, Flinn IW, Rifkin R, et al. Community-based phase IIIB trial of three UPFRONT bortezomib-based myeloma regimens. J Clin Oncol. 2015;33(33):39213929. San Miguel JF, Schlag R, Khuageva NK, et al. Persistent overall survival benefit and no increased risk of second malignancies with bortezomib-melphalan-prednisone versus melphalan-prednisone in patients with previously untreated multiple myeloma. J Clin Oncol. 2013;31(4):448-455. 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):906917. Zangari M, Guerrero J, Cavallo F, Prasad HK, Esseltine D, Fink L. Hemostatic effects of bortezomib treatment in patients with relapsed or refractory multiple myeloma. Haematologica. 2008;93(6):953-954. Palumbo A, Bringhen S, Rossi D, et al. Bortezomib-melphalan-prednisonethalidomide followed by maintenance with bortezomib-thalidomide compared with bortezomib-melphalan-prednisone for initial treatment of multiple myeloma: a randomized controlled trial. J Clin Oncol. 2010;28(34):5101-5109. Palumbo A, Rajkumar SV, Dimopoulos MA, et al. Prevention of thalidomide- and lenalidomide-associated thrombosis in myeloma. Leukemia. 2008;22(2):414-423. Sonneveld P, Avet-Loiseau H, Lonial S, et al. Treatment of multiple myeloma with highrisk cytogenetics: a consensus of the International Myeloma Working Group. Blood. 2016;127(24):2955-2962. Dimopoulos MA, Terpos E, Niesvizky R,

23.

24.

25.

26.

27.

28.

29.

30.

31.

Palumbo A. Clinical characteristics of patients with relapsed multiple myeloma. Cancer Treat Rev. 2015;41(10):827-835. Facon T, Lee JH, Moreau P, et al. Phase 3 study (CLARION) of carfilzomib, melphalan, prednisone (KMP) v bortezomib, melphalan, prednisone (VMP) in newly diagnosed multiple myeloma (NDMM). Clin Lymphoma Myeloma Leuk. 2017;17(suppl):e26-e27. Hou J, Jin J, Xu Y, et al. Randomized, double-blind, placebo-controlled phase III study of ixazomib plus lenalidomide-dexamethasone in patients with relapsed/refractory multiple myeloma: China Continuation study. J Hematol Oncol. 2017;10(1):137. Kumar SK, Berdeja JG, Niesvizky R, et al. Safety and tolerability of ixazomib, an oral proteasome inhibitor, in combination with lenalidomide and dexamethasone in patients with previously untreated multiple myeloma: an open-label phase 1/2 study. Lancet Oncol. 2014;15(13):1503-1512. Moreau P, Masszi T, Grzasko N, et al. Oral ixazomib, lenalidomide, and dexamethasone for multiple myeloma. N Engl J Med. 2016;374(17):1621-1634. Millennium Pharmaceuticals Inc. NINLAROÂŽ (ixazomib) capsules, for oral use. United States Prescribing Information. https://wwwninlarocom/downloads/prescribing-informationpdf, 2016. EMA. European Public Assessment Report: Ninlaro. http://wwwemaeuropaeu/docs/ e n _ G B / d o c u m e n t _ l i b r a r y / E PA R _ _Scientific_Conclusion/human/ 003844/ WC500217622pdf, 2016. Kumar SK, Berdeja J, Niesvizky R, et al. Deep and durable responses with weekly ixazomib, lenalidomide and dexamethasone in patients with newly diagnosed multiple myeloma: long-term follow-up of patients who did not undergo SCT. Haematologica. 2017;102(s2):142-143. Richardson P, Hofmeister C, Rosenbaum C, et al. Twice-weekly ixazomib plus lenalidomide-dexamethasone in patients with newly diagnosed multiple myeloma: longterm follow-up data for patients who did not undergo stem cell transplantation Haematologica. 2017;102(s2):317-318. Durie BG, Harousseau JL, Miguel JS, et al. International uniform response criteria for

1525

J.F. San-Miguel et al.

32.

33.

34.

35.

36.

1526

multiple myeloma. Leukemia. 2006;20(9):1467-1473. Mateos MV, Oriol A, Martinez-Lopez J, et al. Outcomes with two different schedules of bortezomib, melphalan, and prednisone (VMP) for previously untreated multiple myeloma: matched pair analysis using longterm follow-up data from the phase 3 VISTA and PETHEMA/GEM05 trials. Ann Hematol. 2016;95(12):2033-2041. Mateos MV, Dimopoulos MA, Cavo M, et al. Daratumumab plus bortezomib, melphalan, and prednisone for untreated myeloma. N Engl J Med. 2018;378(6):518528. Jagannath S, Roy A, Kish J, et al. Real-world treatment patterns and associated progression-free survival in relapsed/refractory multiple myeloma among US community oncology practices. Expert Rev Hematol. 2016;9(7):707-717. Facon T, Dimopoulos MA, Dispenzieri A, et al. Final analysis of survival outcomes in the phase 3 FIRST trial of up-front treatment for multiple myeloma. Blood. 2018; 131(3):301-310. Durie BG, Hoering A, Abidi MH, et al. Bortezomib with lenalidomide and dexamethasone versus lenalidomide and dexamethasone alone in patients with newly diag-

37.

38.

39.

40.

nosed myeloma without intent for immediate autologous stem-cell transplant (SWOG S0777): a randomised, open-label, phase 3 trial. Lancet. 2017;389(10068):519-527. Myeloma Trialists' Collaborative Group. Combination chemotherapy versus melphalan plus prednisone as treatment for multiple myeloma: an overview of 6,633 patients from 27 randomized trials. J Clin Oncol. 1998;16(12):3832-3842. Facon T, Mary JY, Hulin C, et al. Melphalan and prednisone plus thalidomide versus melphalan and prednisone alone or reduced-intensity autologous stem cell transplantation in elderly patients with multiple myeloma (IFM 99-06): a randomised trial. Lancet. 2007;370(9594):12091218. Hulin C, Facon T, Rodon P, et al. Efficacy of melphalan and prednisone plus thalidomide in patients older than 75 years with newly diagnosed multiple myeloma: IFM 01/01 trial. J Clin Oncol. 2009;27(22):36643670. Kumar SK, Bensinger WI, Zimmerman TM, et al. Phase 1 study of weekly dosing with the investigational oral proteasome inhibitor ixazomib in relapsed/refractory multiple myeloma. Blood. 2014; 124(7):1047-1055.

41. Richardson PG, Baz R, Wang M, et al. Phase 1 study of twice-weekly ixazomib, an oral proteasome inhibitor, in relapsed/refractory multiple myeloma patients. Blood. 2014; 124(7):1038-1046. 42. Kumar S, Moreau P, Hari P, et al. Management of adverse events associated with ixazomib plus lenalidomide/dexamethasone in relapsed/refractory multiple myeloma. Br J Haematol. 2017;178(4):571582. 43. Dimopoulos MA, Mateos MV, Richardson PG, et al. Risk factors for, and reversibility of, peripheral neuropathy associated with bortezomib-melphalan-prednisone in newly diagnosed patients with multiple myeloma: subanalysis of the phase 3 VISTA study. Eur J Haematol. 2011;86(1):23-31. 44. Gupta N, Yang H, Hanley MJ, et al. Dose and schedule selection of the oral proteasome inhibitor ixazomib in relapsed/refractory multiple myeloma: clinical and modelbased analyses. Target Oncol. 2017; 12(5):643-654. 45. Gupta N, Diderichsen PM, Hanley MJ, et al. Population pharmacokinetic analysis of ixazomib, an oral proteasome inhibitor, including data from the phase III TOURMALINEMM1 study to inform labelling. Clin Pharmacokinet. 2017;56(11):1355-1368.

haematologica | 2018; 103(9)

ARTICLE

Stem Cell Transplantation

Competing-risk outcomes after hematopoietic stem cell transplantation from the perspective of time-dependent effects

Ferrata Storti Foundation

Daniel Fuerst,1,2* Sandra Frank,3,4* Carlheinz Mueller,4,5 Dietrich W Beelen,4,6 Johannes Schetelig,7 Dietger Niederwieser,8 Jürgen Finke,9 Donald Bunjes,10 Nicolaus Kröger,11 Christine Neuchel,1,2 Chrysanthi Tsamadou,1,2 Hubert Schrezenmeier,1,2 Jan Beyersmann3 and Joannis Mytilineos1,2,4

Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Transfusion Service, Baden Wuerttemberg – Hessen and University Clinic Ulm; 2 Institute of Transfusion Medicine, University of Ulm; 3Institute of Statistics, University of Ulm; 4DRST – German Registry for Stem Cell Transplantation; 5Zentrales Knochenmarkspender-Register Deutschland (ZKRD - German Bone Marrow Donor Registry), Ulm; 6Department of Bone Marrow Transplantation, University Hospital, University of Duisburg-Essen, Essen; 7Department of Internal Medicine I, University Hospital Carl Gustav Carus, Dresden; 8Department of Hematology/Oncology, University of Leipzig; 9 Faculty of Medicine and Medical Center, University of Freiburg; 10Department of Internal Medicine III, University of Ulm and 11Department for Stem Cell Transplantation, University Cancer Center Hamburg, Germany 1

DF and SF contributed equally to this work.

Haematologica 2018 Volume 103(9):1527-1534

ABSTRACT

he success of hematopoietic stem cell transplantation is determined by multiple factors. Additional complexity is conferred by covariables showing time-dependent effects. We evaluated the effect of predictors on competing-risk outcomes after hematopoietic stem cell transplantation in a time-dependent manner. We analyzed 14951 outcomes of adult patients with hematologic malignancies who underwent a first allogeneic transplant. We extended the combined endpoints of disease-free and overall survival to competing-risk settings: disease-free survival was split into relapse and non-relapse mortality. Overall survival was divided into transplant-related mortality, death from other causes and death from unknown causes. For time-dependent effects we computed estimators before and after a covariable-specific cutpoint. Patients treated with reduced intensity conditioning had a constantly higher risk of relapse compared to patients treated with myeloablative conditioning. For non-relapse mortality, patients treated with reduced intensity conditioning had a reduced mortality risk but this effect was only seen in the first 4 months after transplantation (hazard ratio: 0.76, P<0.001) and not afterwards. Graft source exhibited a time-dependent effect on both transplant-related mortality (in first year: hazard ratio 0.70, P<0.001; after first year: hazard ratio 1.47, P=0.002) and non-relapse mortality (in first 8 months: hazard ratio 0.75, P<0.001; after first 8 months: hazard ratio 1.38, P<0.001). Patients with a poor Karnofsky performance score (<80) had a considerably higher risk of all endpoints in the first 4 months. The competing-risk analysis for overall survival and disease-free survival allows resolution of effects with different vectors early and later after stem cell transplantation, as shown for graft source. This information may be useful in risk assessment of treatment choices and for counseling patients on an individual basis.

haematologica | 2018; 103(9)

Correspondence: j.mytilineos@blutspende.de

Received: October 20, 2017. Accepted: May 30, 2018. Pre-published: June 7, 2018.

doi:10.3324/haematol.2017.183012 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1527 ©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.

1527

D. Fuerst et al.

Introduction Hematopoietic stem cell transplantation has been established as a curative treatment for various high-risk hematologic disorders.1 The success of hematopoietic stem cell transplantation is determined by multiple factors including disease-specific predictors, patient and donor characteristics, as well as treatment choices.2,3 Improvements in clinical care and identification of compatible donors have enhanced safety and efficacy leading to increasing numbers of patients being transplanted.4 Outcome is traditionally measured in terms of overall survival and disease-free survival. However, these endpoints represent a summary of events with different etiologies, but mainly events related to treatment complications and disease relapse. In order to characterize effects in clinical studies from the perspective of transplantation biology, subanalyses for event types are necessary.5 Competing-risk analysis is the standard approach to time-to-event analyses for endpoints which represent competing components of a composite outcome.6 Some clinical predictors are strongly associated with outcome and some of them are particularly involved in increased early mortality, leading to violation of the proportional hazards assumption in a standard Cox regression model.7,8 Such variables include poor Karnofsky performance score (KPS) and advanced disease stage at the time of transplantation as well as the pre-transplant toxicity of myeloablative conditioning. We have previously shown that these variables have a strong time-dependent effect on survival endpoints (overall and disease-free survival).9 In this analysis we aimed at investigating potential timedependent effects of different variables in an event-specific fashion. This procedure provides a deeper insight into the relation between covariables and outcome, extending the

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

Variable

Characteristics

Diagnosis

Acute myeloid leukemia Acute lymphoblastic leukemia Myelodysplastic syndromes NHL-indolent NHL-aggressive mean (SD) median (range) Bone marrow Peripheral blood stem cells Myeloablative Reduced intensity Early Intermediate Advanced 1976-2000 2001-2005 2006-2013 Good (80-100%) Poor (<80%) Missing Mismatched related Mismatched unrelated Matched related Matched unrelated

Patientsâ&#x20AC;&#x2122; age Graft source Conditioning Disease stage

Year of transplantation Karnofsky performance score Donor type

NHL: non-Hodgkin lymphoma; SD: standard deviation.

1528

N (%) 7133 (47.7%) 2696 (18.0%) 2380 (15.9%) 1545 (10.3%) 1197 (8.0%) 46.79 (14.1) 48 (18-78) 2303 (15.4%) 12648 (84.6%) 9684 (64.8%) 5267 (35.2%) 6238 (41.7%) 4511 (30.2%) 4202 (28.1%) 2727 (18.2%) 3713 (24.8%) 8511 (56.9%) 12244 (81.9%) 1018 (6.8%) 1689 (11.3%) 676 (4.5%) 2708 (18.1%) 5795 (38.8%) 5772 (38.6%)

scope of the previous analysis. The primary hypothesis was that poor KPS, conditioning toxicity, bone marrow as the graft source, and advanced disease stage are associated with significantly higher early mortality rates in analyses of mortality endpoints, i.e. transplant-related mortality and non-relapse mortality. In addition, we aimed to explore the time dependency in the effect of these covariables on relapse incidence.

Methods Patients We analyzed data from 14951 patients registered in the German Registry for Stem Cell Transplantation (DRST). Adult patients having received a first hematopoietic stem cell transplant for acute myeloid leukemia, acute lymphoblastic leukemia, myelodysplastic syndrome, and aggressive or indolent nonHodgkin lymphoma between 1976 and 2013 were included. Only transplants for which the graft source was bone marrow or peripheral blood were included in this study (Table 1).

Definitions The KPS at transplant was dichotomized into good (80-100%) and poor (<80%). Early disease stage was defined as transplantation in first complete remission for acute leukemia and as untreated or in first complete remission for myelodysplastic syndrome and non-Hodgkin lymphoma. Intermediate disease stage grouped together patients with acute leukemia transplanted in second complete remission, those with myelodysplastic syndrome transplanted in second complete or partial remission, and patients with lymphoma transplanted in second complete remission, partial remission or stable disease. Stages other than early or intermediate were classified as advanced disease stage.10 Conditioning regimen intensity was categorized into myeloablative and reduced intensity according to guidelines of the European Group for Blood and Marrow Transplantation (EBMT) Med-AB manual. Two competing risk models were considered: one with the endpoints transplant-related mortality, death from other causes and death from unknown causes, and a second with the endpoints relapse and non-relapse mortality (Table 2, Figure 1). Death from other causes comprises death due to secondary malignancies, relapse or progression of disease, and other causes (not transplant-related). Thus, death from other causes is one of three competing events, the other two being transplant-related mortality and death from unknown causes.

Table 2. Competing-risk characteristics.

Variable

Characteristics

CR setting: OS

Event indicator:

Censored 7073 TRM 3845 DOC 3673 Unknown death 360 Mean (SD): 919 days (1396) Median (range): 289 days (1-12059) Censored 6334 Relapse 4002 Death without prior relapse 4344 Missing 271 Mean (SD): 844 days (1374) Median (range): 218 days (1-12059)

Time: CR setting: DFS

Event indicator:

Time:

CR: competing risks; OS: overall survival; DFS: disease-free survival; TRM: transplant-related mortality, DOC: death of other cause; SD: standard deviation; censored: patients lost to follow up without having had an event or still event-free at data request.

haematologica | 2018; 103(9)

Time-dependent effects on competing-risk endpoints

Deaths from any cause without prior relapse of the original disease are events assigned to the non-relapse mortality endpoint. Its competing event is relapse. Transplant-related mortality represents deaths which were classified by the treating physician as directly related to the transplant. Non-relapse mortality comprises these events but, in addition, also includes all other deaths not related to transplantation and also not related to relapse or progression of the primary disease (other deaths). The patients’ consent to perform analysis of their clinical data was obtained upon their registration in the EBMT database. The study was approved by the ethical committee of Ulm University, Germany (n. 108/15).

competing-risk endpoints. Furthermore, a comparison of hazard ratios for selected covariables is given in Table 7.

Covariate: age Higher patients’ age increases risk following hematopoietic stem cell transplantation. This effect is constant over time (HR for transplant-related mortality: 1.015; for death from other causes: 1.009; for relapse: 1.003; and for non-relapse mortality: 1.019; all statistically significant estimates) (Tables 3-6). The estimates describe the increasing risk per life year for each endpoint.

Covariate: disease stage Statistical methods We used cause-specific Cox proportional hazards models to relate covariates to the competing survival outcomes. Covariates not satisfying the proportional hazard assumption were first identified using the test described by Therneau and Grambsch and subsequently modeled to have piecewise time-constant effects in order to facilitate interpretation.11 Preferential candidate breakpoints were as previously reported. They were confirmed using smoothed time-dependent regression coefficients resulting from the Therneau and Grambsch test and subsequently maximizing the maximal partial likelihood.12,13 Finally, the piecewise constant Cox model was fitted, thereby obtaining regression estimates before and after the individual cutpoints. Predictors evaluated in the models were age, disease stage, year of transplantation (scaled in the periods 1976-2000, 2001-2005, 2006-2013 in order to get conceivable effect estimates), graft source, conditioning treatment, and KPS (<80% poor versus 80-100% good). Missing data for the KPS were treated as ‘good’, choosing a conservative estimation in the sense of an underestimated effect of a poor KPS (see Table 1). Analyses were stratified for diagnosis, donor type and transplantation center, summarized in seven categories according to the total number of transplants. To detect only strong timedependent effects, the significance level for the proportional hazard assumption test was set at 0.01. For an effect-estimator a significance level of 0.05 was used.

Results The results of the standard Cox model with the proportional hazard test amended by the results of the piecewise constant Cox analysis are presented in Tables 3-6 for all

The Cox model showed a higher risk conferred by relapse for all endpoints. In particular, for patients transplanted in advanced disease stage, the risk of relapse in the first 8 months after transplantation was 2.92-fold higher than that of patients transplanted in early disease stage. This risk reduced afterwards markedly to a hazard ratio of 1.73. In line with this, patients in intermediate or advanced disease stage had a distinctly higher risk of death from other causes in the first 10 months. This trend remained subsequently, although it was moderated (Tables 3-7).

Covariate: year of transplantation Another covariate included in the Cox model was the year of transplantation. The effect of this covariate was time-dependent for relapse, non-relapse mortality, transplant-related mortality and death from other causes (1976-2000 versus 2006-2013). For all four endpoints, the event-specific cutpoints could be set at 8 months. Recent transplants showed lower risk estimates in comparison to earlier transplants (Tables 3-7).

Covariate: source of stem cells While for relapse and death from other causes the source of stem cells had no impact (Tables 4 and 5), we observed opposing effects for the different time phases in the analysis of non-relapse mortality and transplant-related mortality when considering the piecewise constant Cox model. In the first time period a peripheral blood stem cell graft showed a protective effect for both outcome endpoints, although this effect was subsequently inverted. The changing point of this effect was at 1 year for transplant-related

Figure 1. Schematic display of competing-risk settings. OS: overall survival; DFS: disease-free survival, CR: competing risks; TRM: transplant-related mortality; NRM: non-relapse mortality; sec.: secondary.

haematologica | 2018; 103(9)

1529

D. Fuerst et al.

mortality and 8 months for non-relapse mortality (nonrelapse mortality: HR for first 8 months: 0.75, HR for >8 months: 1.38; transplant-related mortality: HR for first year: 0.70, HR for >1 year: 1.47; all statistically significant estimates) (Tables 3 and 6). This relationship could not be revealed using the standard Cox model.

Covariate: conditioning treatment Reduced intensity conditioning protected against nonrelapse mortality only in the first 4 months and this benefi-

cial effect was no longer detectable afterwards. However, as far as its competing event, relapse, was concerned, reduced intensity conditioning seemed to increase the risk, which was constant over the whole time period (HR for relapse: 1.13, P<0.001; HR for non-relapse mortality within the first 4 months: 0.76, P<0.001, and after 4 months: 1.06, P=0.30) (Tables 5-7). In the other competing-risk setting, considering the events transplant-related mortality and death from other causes, reduced intensity conditioning showed a time-dependent effect with the optimal cutpoint at 4

Table 3. Competing risk model 1. Competing risk transplant-related mortality.

Variable Patient age Intermediate disease stage Advanced disease stage first 5 months after 5 months Year of transplantation 2001-2005 first 8 months after 8 months Year of transplantation 2006-2013 first 8 months after 8 months PBSC as graft source (vs. BM) first year after 1st year Reduced intensity (vs. MAC) first 4 months after 4 months KPS <80 (vs. 80-100) first 4 months after 4 months

HR 1.015 1.32 1.60

0.66 0.50

0.79

0.91

1.84

Cox model analysis P-value CI P-value PH-test <0.001 <0.001 <0.001

<0.001 <0.001

<0.001

0.018

<0.001

1.012-1.018 1.20-1.43 1.47-1.73

0.60-0.73 0.46-0.56

0.72-0.87

0.85-0.98

1.64-2.06

0.398 0.223 0.003

Piecewise constant Cox model analysis P-value CI P-value PH-test

HR 1.015 1.32

<0.001 <0.001

1.013-1.018 1.21-1.44

0.127 0.428

1.69 1.42

<0.001 <0.001

1.53-1.86 1.26-1.61

0.334 0.962

0.58 0.94

<0.001 0.487

0.51-0.65 0.78-1.13

0.792 0.973

0.44 0.78

<0.001 0.013

0.39-0.49 0.65-0.95

0.669 0.500

0.70 1.47

<0.001 0.002

0.64-0.78 1.16-1.86

0.231 0.269

0.78 1.10

<0.001 0.090

0.70-0.87 0.99-1.22

0.852 0.299

2.10 1.37

<0.001 0.004

1.84-2.40 1.10-1.70

0.428 0.587

<0.001 <0.001

<0.001

PBSC: peripheral blood stem cells; BM: bone marrow; MAC: myeloablative conditioning; KPS: Karnofsky performance score; HR: hazard ratio; CI confidence interval; PH-test: proportional hazard test.

Table 4. Competing risk model 1. Competing risk death of other cause.

Variable Patientsâ&#x20AC;&#x2122; age Intermediate disease stage first 10 months after 10 months Advanced disease stage first 10 months after 10 months Year of transplantation 2001-2005 Year of transplantation 2006-2013 first 8 months after 8 months PBSC as graft source (vs. BM) Reduced intensity (vs. MAC) first 4 months after 4 months KPS <80 (vs. 80-100) first 4 months after 4 months

HR 1.009 1.86 2.68

1.10 0.98

1.01 1.04

1.88

Cox model analysis P-value CI <0.001 <0.001 <0.001

0.078 0.714

0.784 0.317

<0.001

1.006-1.012 1.71-2.04 2.47-2.92

0.99-1.23 0.88-1.09

0.91-1.13 0.96-1.12

1.67-2.12

Piecewise constant Cox model analysis P-value CI P-value PH-test

P-value PH-test

0.752 <0.001

1.009

<0.001

1.006-1.012

0.459

2.35 1.45

<0.001 <0.001

2.08-2.66 1.27-1.66

0.936 0.261

3.39 2.07 1.10

<0.001 <0.001 0.082

3.02-3.81 1.83-2.34 0.99-1.22

0.881 0.541 0.259

0.88 1.13 1.01

0.041 0.073 0.836

0.77-0.99 0.99-1.29 0.91-1.12

0.811 0.034 0.256

0.89 1.11

0.130 0.024

0.77-1.03 1.02-1.21

0.971 0.168

2.75 1.34

<0.001 <0.001

2.23-3.25 1.13-1.59

0.856 0.396

<0.001

0.158 0.003

0.613 0.005

<0.001

PBSC: peripheral blood stem cells; BM: bone marrow; MAC: myeloablative conditioning; KPS: Karnofsky performance score; HR: hazard ratio; CI confidence interval; PH-test: proportional hazard test.

1530

haematologica | 2018; 103(9)

Time-dependent effects on competing-risk endpoints

months. The effect was significantly reduced for transplantrelated mortality in the first time period and increased for death from other causes in the second time period (HR for transplant-related mortality in first 4 months: 0.78, P<0.001 and after 4 months: 1.10, P=0.090; HR for death from other causes within the first 4 months: 0.89, P=0.130 and after 4 months: 1.11, P=0.024) (Tables 3, 4 and 7).

KPS the changing points of risk level could be fixed at 4 months for all outcome parameters. This means that patients with a poor KPS had a higher risk in the first 4 months after hematopoietic stem cell transplantation when compared to patients with a KPS of 80-100. In the first time period the risk for all outcome endpoints was increased at least 2-fold (Tables 3-7).

Covariate: Karnofsky performance score KPS exhibited a significant effect on all outcome events. While applying the proportional hazard test it was found that a covariate-specific cutpoint had to be defined for each of the competing risks examined in this study. For

Discussion To the best of our knowledge, our study is the first comprehensive analysis exploring time-dependent effects

Table 5. Competing risk model 2. Competing risk relapse.

Variable Patientsâ&#x20AC;&#x2122; age Intermediate disease stage first 10 months after 10 months Advanced disease stage first 8 months after 8 months Year of transplantation 2001-2005 first 8 months after 8 months Year of transplantation 2006-2013 first 8 months after 8 months PBSC as graft source (vs. BM) Reduced intensity (vs. MAC) KPS <80 (vs. 80-100) first 4 months after 4 months

HR 1.002 1.74 2.47

1.03 0.86

1.04 1.13 1.72

Cox model analysis P-value CI 0.074 <0.001 <0.001

0.577 0.030

0.423 <0.001 <0.001

1.000-1.005 1.60-1.89 2.28-2.68

0.93-1.14 0.80-0.99

0.94-1.15 1.05-1.21 1.53-1.94

Pvalue PH-test 0.031 <0.001

Piecewise constant Cox model analysis HR P-value CI P-value PH-test 1.003

0.059

0.999-1.005

0.015

1.87 1.52

<0.001 <0.001

1.69-2.07 1.36-1.77

0.226 0.907

2.92 1.73

<0.001 <0.001

2.65-3.21 1.50-1.99

0.418 0.323

0.95 1.21

0.391 0.025

0.84-1.07 1.02-1.42

0.616 0.181

0.82 1.10 1.04 1.13

<0.001 0.273 0.486 <0.001

0.72-0.92 0.93-1.30 0.94-1.15 1.05-1.22

0.757 0.122 0.165 0.472

2.10 1.26

<0.001 0.021

1.81-2.43 1.03-1.53

0.697 0.422

<0.001

0.002 <0.001

0.138 0.418 <0.001

PBSC: peripheral blood stem cells; BM: bone marrow; MAC: myeloablative conditioning; KPS: Karnofsky performance score; HR: hazard ratio; CI confidence interval; PH-test: proportional hazard test.

Table 6. Competing risk model 2. Competing risk non-relapse mortality.

Variable Patientsâ&#x20AC;&#x2122; age Intermediate disease stage Advanced disease stage Year of transplantation 2001-2005 first 8 months after 8 months Year of transplantation 2006-2013 first 8 months after 8 months PBSC as graft source (vs. BM) first 8 months after 8 months Reduced intensity (vs. MAC) first 4 months after 4 months KPS <80 (vs. 80-100) first 4 months after 4 months

HR 1.019 1.42 1.77 0.77 0.65

0.89

0.87

1.89

Cox model analysis P-value CI P-value PH-test <0.001 <0.001 <0.001 <0.001 <0.001

0.015

<0.001

1.016-1.022 1.31-1.54 1.64-1.91 0.69-0.85 0.59-0.71

0.81-0.98

0.81-0.94

1.70-2.10

0.473 0.269 0.053 <0.001

Piecewise constant Cox model analysis P-value CI P-value PH-test

1.019 1.42 1.75

<0.001 <0.001 <0.001

1.016-1.022 1.30-1.54 1.62-1.90

0.681 0.197 0.064

0.67 1.04

<0.001 0.654

0.59-0.75 0.87-1.25

0.863 0.324

0.56 0.99

<0.001 0.907

0.50-0.62 0.82-1.19

0.834 0.946

0.75 1.38

<0.001 <0.001

0.68-0.84 1.14-1.66

0.928 0.012

0.76 1.06

<0.001 0.299

0.69-0.83 0.95-1.18

0.813 0.396

2.29 1.26

<0.001 0.028

2.02-2.59 1.03-1.55

0.437 0.900

<0.001

PBSC: peripheral blood stem cells; BM: bone marrow; MAC: myeloablative conditioning; KPS: Karnofsky performance score; HR: hazard ratio; CI confidence interval; PH-test: proportional hazard test.

haematologica | 2018; 103(9)

1531

D. Fuerst et al.

on the outcome with hematopoietic stem cell transplantation in a setting with competing risks. We adjusted these variables in a piecewise-constant manner for the time-dependent effects, extending standard Cox-regression models in order to obtain more accurate risk estimates. For transplant-related mortality these were advanced disease stage, year of transplantation, source of stem cells, conditioning intensity and KPS. All these effects were highly significant (P<0.001). Year of transplantation before 2001, advanced disease stage, myeloablative conditioning and a poor KPS were associated with significantly increased early mortality after transplantation. This observation is related to effects induced by tissue damage caused by the acute toxicity of conditioning.14-16 After resolution of the acute toxicity these effects diminish.17 We estimate the time to resolution of acute toxicity to be around 4 months after transplantation according to the results obtained for the predictors conditioning intensity and KPS. The strikingly high transplantrelated mortality risk for patients with a poor KPS in the first 4 months (HR: 2.10) substantially decreased in patients surviving this critical phase (HR: 1.37). Our data further suggest that improvements in transplantation procedures and supportive care over time affect mainly mortality early after transplantation.18 Such a relationship may not be obvious in a Cox-regression approach without time-dependent effects. Peripheral blood stem cells as a graft source were associated with a lower risk of transplant-related mortality in the first year after transplantation but a higher risk thereafter. The protective effect of peripheral blood stem cell grafts is most likely due to the faster engraftment of these stem cells as compared to bone marrow, leading to a shorter period of aplasia and decreased vulnerability to infections.19 It has been reported that transplantation with peripheral blood stem cell grafts leads to an increased incidence of chronic graft-versus-host disease, which in turn may explain the higher

risk of transplant-related mortality more than 1 year after transplantation.20 Our analysis reveals here that the transplantation biology of bone marrow and peripheral blood stem cells is determined by two opposing effects, which would not have been detected in a more conventional analysis. Our approach enables a better understanding of the dynamic risk changes after hematopoietic stem cell transplantation. Death from other causes summarizes deaths due to relapse or progression of disease and deaths unrelated to hematopoietic stem cell transplantation. With regards to this endpoint, time-dependent effects for disease stage, year of transplantation (2006-2013), conditioning intensity and a poor KPS were detected.21,22 For disease stage, a considerably higher risk of death from other causes was observed in the first 10 months after transplantation, which reflects the higher incidence of relapse in patients in late disease stage.23 This is consistent with the results for relapse incidence within the composite endpoint of disease-free survival. This risk is considerably decreased in patients surviving more than 10 months after transplantation. Such information might allow physicians to reassure patients reaching this phase of follow-up. Patients transplanted in recent years had a slightly lower risk of transplant-related mortality and death from other causes in the first 8 months after transplantation, with no significant difference afterwards, which is most likely attributable to optimization in conditioning treatments and supportive care.24 Similarly, reduced intensity conditioning led to a lower early mortality in the first 4 months and a higher risk afterwards, highlighting the divergent effects on toxicity and relapse incidence.25 This cutpoint provides an estimate for the duration of conditioning toxicity after hematopoietic stem cell transplantation. A poor KPS is evoked by more severe disease burden and comorbidity. These problems can cause relapse and death from other causes. Consequently a poor KPS was associated

Table 7. Comparison of hazard ratios of selected covariables.

Covariable Intermediate disease stage first 10 months after 10 months Advanced disease stage first 5 |10| 8 months after 5 |10| 8 months Year of transplant 2001-2005 first 8 months after 8 months Year of transplant 2006-2013 first 8 months after 8 months RIC (vs. MAC) first 4 months after 4 months KPS <80 (vs. 80-100) first 4 months after 4 months

TRM

DOC

P-value

1.32

<0.001

P-value

Relapse P-value

2.35 1.45

<0.001 <0.001

1.87 1.52

<0.001 <0.001

3.39 2.07 1.10

<0.001 <0.001 0.082

2.92 1.73

<0.001 <0.001

0.95 1.21 0.82 1.10 1.13

1.69 1.42

<0.001 <0.001

0.58 0.94

<0.001 0.487

0.44 0.78

<0.001 0.013

0.88 1.13

0.041 0.073

0.78 1.10

<0.001 0.090

0.89 1.11

0.130 0.024

2.10 1.37

<0.001 0.004

2.75 1.34

<0.001 <0.001

2.10 1.26

NRM HR

P-value

1.42

<0.001

1.75

<0.001

0.391 0.025

0.67 1.04

<0.001 0.654

<0.001 0.273 <0.001

0.56 0.99

<0.001 0.907

0.76 1.06

<0.001 0.299

2.29 1.26

<0.001 0.028

<0.001 0.021

TRM: transplant-related mortality; DOC: death of other cause; NRM: non-relapse mortality; HR: hazard ratio; MAC: myeloablative conditioning; RIC: reduced intensity conditioning; KPS: Karnofsky performance score.

1532

haematologica | 2018; 103(9)

Time-dependent effects on competing-risk endpoints

with a high early mortality risk in the analysis of death from other causes, an effect probably related to higher relapse rates. The competing-risk endpoint relapse was derived from disease-free survival. Time-dependent variables that showed increased early relapse risk were intermediate and advanced disease stage, and a poor KPS.10 A poor KPS may be related to the primary disease or to any comorbidity/toxicity.26 It can be speculated that an increased relapse risk is mainly due to patients whose primary disease is the predominant reason for their poor KPS. We may, therefore, assume that, for this subset of patients, the risk may be even higher than our estimates. Interestingly, for the source of stem cells no significant differences could be detected between bone marrow and peripheral blood stem cells, which confirms the findings of a previous prospective trial.27 The analysis of non-relapse mortality (death without prior relapse) showed time-dependent effects for year of transplantation, source of stem cells, conditioning intensity and poor KPS. Of these variables myeloablative conditioning and a poor KPS were strongly associated with early mortality after transplantation.28 These effects are related to transplantation-associated morbidity, particularly toxicity in the case of conditioning intensity and comorbidity or disease burden in the case of a poor KPS. Again, the first 4 months after transplantation were confirmed to be a critical phase for conditioning toxicity and transplantation-related morbidity. The opposing effects of peripheral blood stem cell grafts on early and late mortality was prominent in this analysis, showing a highly significant protective effect in the first 8 months, probably due to a shorter period of aplasia, and highly significant adverse effects later on, most likely caused by an increase of chronic graft-versus-host disease.19,20 Improvements in conditioning therapy and supportive care are reflected in the reduction of early mortality in the first 8 months after transplantation in the more recent periods regarding year of transplantation.24 We considered two competing-risk settings, one with time until a specific cause of death and one with time until a first event (relapse or death) (Figure 1). The results of the analyses for non-relapse mortality and transplantrelated mortality are quite similar as there is a large, although not complete, overlap. Non-relapse mortality additionally includes deaths not related to transplantation or disease relapse/progression. In our study we needed both endpoints to construct different competing-risk settings (Figure 1). One difference between transplant-related mortality and non-relapse mortality is that the former requires the classification of deaths as transplant-related by a physician, which may be difficult in some situations, whereas the classification of non-relapse mortality does not require this assessment and is based instead on the plain and objective distinction between death in complete remission or not. Thus, the advantage of transplant-related mortality is that it is more specific for transplantationrelated adverse events while the advantage of non-relapse mortality is that it is easier and more objective to classify. Limitations of our analysis are that exact patterns of HLA-mismatches, regarding number and loci of mismatches, were not available for the majority of the transplants and so only a rough stratification according to donor type was feasible (i.e. matched related, mismatched related, matched unrelated, mismatched unrehaematologica | 2018; 103(9)

lated). A certain degree of heterogeneity is related to the large time span over which the transplants included in this study were performed, with changes in transplant procedures, graft source preferences, donor selection algorithms and supportive care over the years. This heterogeneity was only partly addressed by including the time period of transplantation as a covariate. In a separate analysis in which the only transplants included were those carried out in the periods 1998-2005 and 20062013, we found results comparable to those presented in this manuscript (i.e. inclusion of transplants from 1976 onwards). In order to be consistent with our previous publication, in which we used a similar approach to investigate the effect of various covariates on overall and disease-free survival in a time-dependent manner, we decided to carry out our analyses for this work based on the same data set as previously.9 Center-specific unobserved variables were included by using stratification according to frequency of transplants in the transplant center. Another highly predictive clinical variable is cytogenetic risk.2 This predictor could not be evaluated as information regarding cytogenetic risk is sparse in the DRST/EBMT database. Grathwohl et al. and Zwaan et al. highlighted the disease-related adjustment for time from diagnosis to transplantation.29,30 We chose not to include this variable in our models due to difficulties in its interpretation. Several factors influence time from diagnosis to transplantation. These are disease-inherent risk, pretreatment, clinical status of the patient, clinical urgency of transplantation, prompt availability of a suitable donor as well as the patientÂ´s choice with regard to alternative treatment options. The profile of these factors differs between distinct disease entities. Thus, the information provided by global estimates for this predictor is not meaningful. Several methods that allow modeling of time-dependent effects of covariables have been described.31 Here, we followed the approach of Fuerst et al.,9 and chose piecewise constant effects estimation for ease of interpretation.12,32 It is possible that the true underlying time-dependent effects are more complex, and a careful interpretation of our results would be that of averaged effects on the respective time intervals. In addition, we aimed to extend the results of Fuerst et al.9 to the more specific endpoints transplant-related mortality, death from other causes, relapse and non-relapse mortality. Consequently, preferential candidate time intervals were as described in Fuerst et al.,9 and were not cross-validated. In summary, this analysis of competing risks disentangles how the previously described net effect9 is achieved via the different outcomes summarized in the composite endpoint of overall survival (or disease-free survival). The description of time-dependent effects allows a better understanding of transplantation biology regarding competing-risk endpoints. Reasons for early mortality may be described and quantified more precisely. Predictors with ambivalent effects, such as graft source and conditioning toxicity, may be identified. These observations may support treatment choices, individual patient counseling and reassurance during follow-up. Acknowledgment The authors would like to thank the DRST Data administrators F. Hanke and H. Neidlinger for providing the clinical data for this analysis 1533

D. Fuerst et al.

References 1. Passweg JR, Baldomero H, Bader P, et al. Use of haploidentical stem cell transplantation continues to increase: the 2015 European Society for Blood and Marrow Transplant activity survey report. Bone Marrow Transplant. 2017;52(6):811-817. 2. Armand P, Gibson CJ, Cutler C, et al. A disease risk index for patients undergoing allogeneic stem cell transplantation. Blood. 2012;120(4):905-913. 3. Kollman C, Spellman SR, Zhang MJ, et al. The effect of donor characteristics on survival after unrelated donor transplantation for hematologic malignancy. Blood. 2016;127(2):260-267. 4. Gooley TA, Chien JW, Pergam SA, et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22):2091-2101. 5. Iacobelli S. Suggestions on the use of statistical methodologies in studies of the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant. 2013;48 (Suppl 1):S1-37. 6. Zhou B, Fine J, Latouche A, Labopin M. Competing risks regression for clustered data. Biostatistics. 2012;13(3):371-383. 7. Robin M, Porcher R, Ades L, et al. HLAmatched allogeneic stem cell transplantation improves outcome of higher risk myelodysplastic syndrome. A prospective study on behalf of SFGM-TC and GFM. Leukemia. 2015;29(7):1496-1501. 8. Wingard JR, Majhail NS, Brazauskas R, et al. Long-term survival and late deaths after allogeneic hematopoietic cell transplantation. J Clin Oncol. 2011;29(16):2230-2239. 9. Fuerst D, Mueller C, Beelen DW, et al. Time-dependent effects of clinical predictors in unrelated hematopoietic stem cell transplantation. Haematologica. 2016;101 (2):241-247. 10. Gratwohl A, Stern M, Brand R, et al. Risk score for outcome after allogeneic hematopoietic stem cell transplantation: a retrospective analysis. Cancer. 2009;115 (20):4715-4726. 11. Grambsch P, Therneau T. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81(3):515526. 12. Anderson JA, Senthilselvan A. A two-step regression model for hazard functions. J

1534

Applied Stat. 1982;31(1):44-51. 13. Zhang MH, Pasquini M, Ahn KW. Regression models in bone marrow transplantation - a case study. In: Klein J, Houwelingen H, Ibrahim J, Scheike TH, editors. Handbook of Survival Analysis. London, New York: CRC Press; 2014. 243262. 14. Carreras E, Diaz-Ricart M. The role of the endothelium in the short-term complications of hematopoietic SCT. Bone Marrow Transplant. 2011;46(12):1495-1502. 15. Paczesny S, Diaz-Ricart M, Carreras E, Cooke KR. Translational research efforts in biomarkers and biology of early transplantrelated complications. Biol Blood Marrow Transplant. 2011;17(1 Suppl):S101-S108. 16. Palomo M, Diaz-Ricart M, Carbo C, et al. Endothelial dysfunction after hematopoietic stem cell transplantation: role of the conditioning regimen and the type of transplantation. Biol Blood Marrow Transplant. 2010;16(7):985-993. 17. Gyurkocza B, Sandmaier BM. Conditioning regimens for hematopoietic cell transplantation: one size does not fit all. Blood. 2014;124(3):344-353. 18. Cheuk DK. Optimal stem cell source for allogeneic stem cell transplantation for hematological malignancies. World J Transplant. 2013;3(4):99-112. 19. Couban S, Simpson DR, Barnett MJ, et al. A randomized multicenter comparison of bone marrow and peripheral blood in recipients of matched sibling allogeneic transplants for myeloid malignancies. Blood. 2002;100(5):1525-1531. 20. Flowers ME, Parker PM, Johnston LJ, et al. Comparison of chronic graft-versus-host disease after transplantation of peripheral blood stem cells versus bone marrow in allogeneic recipients: long-term follow-up of a randomized trial. Blood. 2002;100(2): 415-419. 21. Finke J, Schmoor C, Bethge WA, et al. Long-term outcomes after standard graftversus-host disease prophylaxis with or without anti-human-T-lymphocyte immunoglobulin in haemopoietic cell transplantation from matched unrelated donors: final results of a randomised controlled trial. Lancet Haematol. 2017;4(6): e293-e301. 22. Lee SJ, Klein J, Haagenson M, et al. Highresolution donor-recipient HLA matching contributes to the success of unrelated

23.

24.

25.

26.

27.

28.

29. 30.

31.

32.

donor marrow transplantation. Blood. 2007;110(13):4576-4583. Craddock C, Nagra S, Peniket A, et al. Factors predicting long-term survival after T-cell depleted reduced intensity allogeneic stem cell transplantation for acute myeloid leukemia. Haematologica. 2010;95(6):989995. Remberger M, Ackefors M, Berglund S, et al. Improved survival after allogeneic hematopoietic stem cell transplantation in recent years. A single-center study. Biol Blood Marrow Transplant. 2011;17(11): 1688-1697. Abdul Wahid SF, Ismail NA, Mohd-Idris MR, et al. Comparison of reduced-intensity and myeloablative conditioning regimens for allogeneic hematopoietic stem cell transplantation in patients with acute myeloid leukemia and acute lymphoblastic leukemia: a meta-analysis. Stem Cells Dev. 2014;23(21):2535-2552. Sorror M, Storer B, Sandmaier BM, et al. Hematopoietic cell transplantation-comorbidity index and Karnofsky performance status are independent predictors of morbidity and mortality after allogeneic nonmyeloablative hematopoietic cell transplantation. Cancer. 2008;112(9):1992-2001. Anasetti C, Logan BR, Lee SJ, et al. Peripheral-blood stem cells versus bone marrow from unrelated donors. N Engl J Med. 2012;367(16):1487-1496. Gratwohl A, Brand R, Frassoni F, et al. Cause of death after allogeneic haematopoietic stem cell transplantation (HSCT) in early leukaemias: an EBMT analysis of lethal infectious complications and changes over calendar time. Bone Marrow Transplant. 2005;36(9):757-769. Gratwohl A. The EBMT risk score. Bone Marrow Transplant. 2012;47(6):749-756. Zwaan FE, Hermans J, Barrett AJ, Speck B. Bone marrow transplantation for acute nonlymphoblastic leukaemia: a survey of the European Group for Bone Marrow Transplantation (E.G.B.M.T.). Br J Haematol. 1984;56(4):645-653. Buchholz A, Sauerbrei W. Comparison of procedures to assess non-linear and timevarying effects in multivariable models for survival data. Biom J. 2011;53(2):308-331. Thomas L, Reyes EM. Tutorial: survival estimation for Cox regression models with time-varying coefficients using SAS and R. J Stat Software. 2014;61(1):1-16.

haematologica | 2018; 103(9)

ARTICLE

Quality of Life

Patient-reported outcomes and health status associated with chronic graft-versus-host disease

Ferrata Storti Foundation

Stephanie J. Lee,1 Lynn Onstad,1 Eric J. Chow,1 Bronwen E. Shaw,2 Heather S.L. Jim,3 Karen L. Syrjala,1 K. Scott Baker,1 Sarah Buckley1 and Mary E. Flowers1

1 Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA; 2Center for International Blood and Marrow Transplant Research, Department of Medicine, Medical College of Wisconsin, Milwaukee, WI and 3Department of Health Outcomes and Behavior, Moffitt Cancer Center, Tampa, FL, USA

Haematologica 2018 Volume 103(9):1535-1541

ABSTRACT

hronic graft-versus-host disease occurs in 20-50% of allogeneic hematopoietic cell transplant survivors. We surveyed patients about their quality of life, symptoms, health status, comorbid conditions and medications. Instruments included the Short-Form-36 (SF36), the Patient-Reported Outcomes Measurement Information System (PROMIS) Global and PROMIS-29 scales and the Lee Chronic Graft-versus-Host Disease Symptom Scale. Functional status was measured by self-reported Karnofsky performance status and work status. Of 3027 surveys sent to recipients surviving one or more years after transplantation, 1377 (45%) were returned. Among these, patients reported that their chronic graft-versus-host disease was mild (n=257, 18.7%), moderate (n=110, 8.0%) or severe (n=25, 1.8%). Another 377 (27.4%) had never had chronic graft-versus-host disease and 280 (20.3%) had had chronic graft-versus-host disease but it had resolved. We excluded 328 (23.8%) patients who did not answer the questions about chronic graft-versushost disease. Patients who reported moderate or severe chronic graft-versus-host disease reported worse quality of life, lower performance status, a higher symptom burden and were more likely to be taking prescription medications for pain, anxiety and depression compared to those with resolved chronic graft-versus-host disease. Self-reported measures were similar between patients with resolved chronic graft-versus-host disease and those who had never had it. Our data suggest that the PROMIS measures may be able to replace the SF-36 in the assessment of chronic graft-versus-host disease. Between 26.7-39.4% of people with active chronic graft-versus-host disease were unable to work due to health reasons, compared with 12.1% whose chronic graft-versus-host disease had resolved and 15.4% who had never had chronic graft-versus-host disease. Mouth, eye and nutritional symptoms persisted after resolution of chronic graft-versus-host disease. These results show that better prevention of and treatment for chronic graft-versus-host disease are needed to improve survivorship after allogeneic transplantation.

Introduction Approximately 20-50% of transplant survivors experience chronic graft-versus-host disease (GvHD), the most common late complication of allogeneic hematopoietic cell transplantation (HCT). Chronic GvHD is an iatrogenic complication of HCT that occurs when the donor’s immune system attacks the recipient’s tissues. The onset is typically 4-6 months after transplantation and three or more organs are involved in the majority of cases.1 Risk factors include older age of the recipient, use of peripheral haematologica | 2018; 103(9)

Correspondence: sjlee@fredhutch.org.

Received: March 7, 2018. Accepted: May 25, 2018. Pre-published: June 1, 2018.

doi:10.3324/haematol.2018.192930 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1535 ©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.

1535

S.J. Lee et al.

blood instead of bone marrow or cord blood, females donating to male patients, lack of in vivo or ex vivo T-cell depletion or post-transplant cyclosphosphamide, and HLA mismatching.2 Chronic GvHD may be mild, requiring only topical or local interventions or short-term systemic immunosuppression; it may also be moderate to severe and poorly controlled with available treatments, causing substantial morbidity or even death.3 Overall, chronic GvHD has been associated with worse quality of life (QOL) and functional status, and a higher symptom burden.4 Interventions that prevent chronic GvHD have been associated with a lower symptom burden and better functional status in randomized controlled trials.5,6 The impact of chronic GvHD on QOL as measured by the National Institutes of Health (NIH) Patient-Reported Outcomes Measurement Information System (PROMIS) instruments has not been reported7 and currently the PROMIS measures are not included in the list of recommended scales for chronic GvHD assessment in clinical trials. Interest in preventing and treating chronic GvHD has increased in the last 5-10 years with a growing recognition of the impact that this complication has on the long-term health of survivors. There are better options to prevent6,8-12 and treat chronic GvHD,13,14 and an increasing number of reports about biomarkers and insights from animal and human studies on its pathophysiology.15-17 In August 2017, ibrutinib became the first drug approved by the Food and Drug Administration for the management of chronic GvHD that persists or progresses after treatment with steroids.18 The goal of this study was to describe the QOL scores and health status of patients with active chronic GvHD of differing severity compared to those with resolved chronic GvHD or those who had never had chronic GvHD. We also aimed to investigate the PROMIS measures in chronic GvHD relative to established measures of QOL in longterm HCT survivors.

Methods Participants Questionnaires are sent annually to transplant survivors as part of the Fred Hutchinson Cancer Research Center (FHCRC) longterm follow-up research program. Non-responders to the initial survey are sent a reminder letter 1 month later. All patients who underwent allogeneic transplantation from any graft source at FHCRC and who consented to allow their information to be used for research purposes were included. Patients were aged 18 or greater at the time of the survey. The analyzed data were from surveys administered by mail or online, according to the participants’ choice, from July 1, 2015 to June 30, 2016. The dataset was sealed as of September 30, 2016. The Institutional Review Board of FHCRC approved the study, and all patients provided written informed consent at the time of their transplants.

Data collection instruments The core survey includes questions about QOL, symptoms, medical complications, medications, work status and chronic GvHD. In the chronic GvHD section, patients report having never had chronic GvHD, having had chronic GvHD but it resolved, or having current mild, moderate or severe involvement. Definitions were not provided for these categories. Patients also report their work/school status, whether they have had certain medical complications, and whether they are taking medications for specified conditions. An abbreviated comorbidity score consists of self1536

reported pulmonary disease, avascular necrosis, adrenal insufficiency and diabetes since these co-morbid conditions were previously associated with QOL.19 The core survey also includes the Medical Outcomes Study Short Form-36 (SF-36) health survey. The SF-36 is a 36-item multidimensional QOL instrument with population norms.20,21 Two summary scales, the physical component score (PCS) and mental component score (MCS), are normalized to a T score of 50 with a standard deviation of 10. Higher scores indicate better QOL. A clinically meaningful difference is 0.5 times the standard deviation, so scores <45 or >55 are considered to have clinical significance. There are also eight subscales: physical functioning, role physical functioning, emotional functioning, role emotional functioning, fatigue, social functioning, pain and general health. Each year, a different module is added to the long-term followup core survey and, for this analysis, included the PROMIS Global Health 10, PROMIS-29 v2.0, and the Lee chronic GvHD symptom scale (LSS). The PROMIS Global Health 10 measure comprises ten items with two summary scores for physical and mental functioning. The PROMIS-29 contains 29 scored items and seven subscales: for the physical and social functioning scales, higher scores indicate better functioning; for fatigue, pain, anxiety, depression, and sleep scales, higher scores indicate a greater symptom burden.22 Similar to the SF-36, scores are normalized to 50 with a standard deviation of 10 with higher scores indicating better functioning, and scores greater than 0.5 times the standard deviation (i.e., <45 or >55, compared to the general population) are considered clinically meaningful. The LSS is a 30-item measure with one summary score and seven domains: skin, mouth, eye, lung, psycho-emotional, vitality and nutrition.23,24 Higher scores indicate greater symptom burden. Half of the cases received the LSS with a 1-month recall period and half received a 1-week version; results were combined for analysis because there was no statistical difference between the aggregate scores for the two versions (data not shown). Scores may be calculated if more than half of the items in a subscale are answered. Scores range from 0-100 with higher scores indicating greater symptom burden. A difference of 6-7 points on the summary score is considered clinically meaningful but since this scale measures chronic GvHD symptoms, general population norms are not available.

Clinical and transplant variables Clinical information retrieved from the institutional database included patient’s age, sex, diagnosis, disease stage, conditioning regimen, donor type, graft source, GvHD prophylaxis, diagnosis of chronic GvHD, post-transplant relapse and vital status.

Statistical analysis Patient, transplant and chronic GvHD characteristics and known characteristics of responders and non-responders were compared using chi-square or Fisher exact tests and t-tests for categorical and continuous variables, respectively. Only patients who answered the chronic GvHD question on the survey were included in this analysis because this question was used to classify whether chronic GvHD was currently active and its severity. This information is not captured in the transplant database; only the onset date and maximum severity are routinely recorded. Patient-reported surveys were scored according to the directions of the developers, including methods for handling missing data. Multivariate linear regression models were used to examine the associations between the SF-36, PROMIS and LSS scores and potential explanatory variables including those from the medical record: patient’s age and sex, conditioning intensity, donor type and graft source, diagnosis of chronic GvHD, post-transplant relapse, haematologica | 2018; 103(9)

Chronic GvHD, PRO and health status

time since transplant, and variables from the patient-self-reported survey: severity of chronic GVHD, and presence of any of four comorbidities (pulmonary disease, avascular necrosis, adrenal insufficiency or diabetes) previously associated with QOL.19 Work or school status was categorized as: (i) work, school or homemaking full or part time, or retired by choice; (ii) not working, not in school or retired due to health problems, and unemployed. Because of multiple comparisons, P-values less than 0.01 were considered statistically significant.

Results The survey was sent to 3027 recipients surviving one or more years after allogeneic HCT and 1377 responded (overall response rate 45%). The median age of the respondents was 54 years [interquartile range (IQR) 42-63 years] and 55.2% were male. Non-respondents were younger (mean age 48.5 versus 56.4, P<0.001), more likely to have received bone marrow (55.9% versus 45.7%, P<0.001), myeloablative conditioning (86.6% versus 78.5%, P<0.001), or high dose total body irradiation (44.5% versus 34.7%, P<0.001) and to have survived longer since their transplant (mean 15.8 versus 13.8 years, P<0.001) than respondents. The median time since HCT was 11 years (IQR 4-20) for all respondents whose chronic GvHD status was known and 6 years (IQR 2-13) for those with current chronic GvHD. Table 1 shows the characteristics of the groups of patients divided according to their self-reported current chronic GvHD status, excluding the 328 who responded to the survey but did not answer the chronic GvHD questions. There were 377 who had never had chronic GvHD and 280 whose chronic GvHD had resolved by the time of the survey. Of those with current chronic GvHD, mild involvement was reported by 257, moderate involvement by 110, and severe involvement by 25. Overall, the groups were significantly different for all characteristics examined except for age. Patients who had never had chronic GvHD or whose chronic GvHD had resolved were more likely to have had matched related donors, myeloablative conditioning, high dose total body irradiation and to have survived longer since HCT. They were less likely to have received peripheral blood as the source of their graft.

Because of these cohort differences, all of these factors were included in the multivariate models. Starting in 1992, the database captures date last seen at FHCRC. Among the 1043 patients who had a date of last visit recorded, the median time since last being seen at FHCRC was 9.9 years (IQR 2.8-22.2) for patients with no chronic GvHD, 13.3 (IQR 7.0-20.9) for those whose GvHD had resolved, 1.2 (IQR 0.3-8.4) for mild, 0.8 (IQR 0.2-4.8) for moderate and 0.7 (IQR 0.2-9.5) for severe chronic GvHD. Table 2 shows the raw scores of the different instruments for the chronic GvHD subgroups, as well as the comparisons between never versus resolved chronic GvHD and, for those with active chronic GvHD, between selfidentified mild, moderate, and severe chronic GvHD. Patients whose chronic GvHD had resolved reported scores similar to those who reported never having had chronic GvHD on nearly all subscales. Exceptions were seen with the LSS in which patients with resolved chronic GvHD still reported more mouth, eye and nutritional symptoms and had higher summary scores than those never affected by chronic GvHD, although none of these differences was considered clinically meaningful (i.e., all <0.5 standard deviations). Patients with mild chronic GvHD had statistically significant worse scores than patients with resolved chronic GvHD in many domains (data not shown) but these differences were not considered clinically significant by virtue of the 0.5 standard deviation criterion. In contrast, among patients with current chronic GvHD, scores varied as expected between those with selfreported mild and those with moderate and severe chronic GvHD for all scales. Patients with moderate or severe chronic GvHD reported scores in multiple physical, mental and symptom domains that were both statistically and clinically worse than those of patients whose chronic GvHD had resolved, as shown by the bold values in Table 2. However, we did not detect any statistically significant differences between patients with moderate and severe chronic GvHD (data not shown). A similar pattern was seen for the physical functioning scales of the SF-36 and PROMIS Global Health (Figure 1A), the mental functioning scales of the SF-36 and PROMIS Global Health (Figure 1B) and the physical and social subscales of the PROMIS-29 (Figure 1C). Presence and severity of Chronic GvHD remained pre-

Table 1. Population characteristics.

Never (n=377)

Resolved (n=280)

Mild (n=257)

Moderate (n=110)

Severe (n=25)

P-value

135 (48.2) 58.8 (12.3) 136(48.6) 19 (6.8) 7 (2.5) 81 (28.9) 33 (11.8) 4 (1.4) 0 110 (39.3) 242 (86.4) 131 (46.8) 17.5 (8.4)

113 (44.0) 57.2 (12.6) 88 (34.2) 4 (1.6) 10 (3.9) 114 (44.4) 31 (12.1) 9 (3.5) 1 (0.4) 187 (72.8) 172 (66.9) 59 (23.0) 8.4 (7.5)

49 (44.5) 58.3 (10.8) 40 (36.4) 1 (0.9) 2 (8.1) 54 (49.1) 13 (11.8) 0 0 97 (88.2) 64 (58.2) 18 (16.4) 7.2 (5.8)

9 (36.0) 57.6 (15.0) 7 (28.0) 1 (4.0) 1 (4.0) 11 (44.0) 4 (16.0) 0 0 20 (80.0) 14 (56.0) 3 (12.0) 9.0 (8.1)

0.02 <0.001 <0.001

Female, n (%) 209 (55.4) Age, mean years (SD) 53.9 (13.5) Matched related 212 (56.2) Mismatched related 16 (4.2) Haplo-identical related 11 (2.9) Matched unrelated 95 (25.2) Mismatched unrelated 17 (4.5) Cord blood 19 (5.0) Syngeneic 6 (1.6) Peripheral blood, n (%) 150 (39.8) Myeloablative, n (%) 309 (82.0) High dose TBI, n (%) 133 (35.3) Years since HCT, mean, (SD) 14.9 (11.3)

<0.001 <0.001 <0.001 <0.001

TB: total body irradiation; HCT: hematopoietic cell transplant.

haematologica | 2018; 103(9)

1537

S.J. Lee et al.

dictors of QOL scores even after adjustment for sex, age, conditioning intensity, total body irradiation, prior relapse, donor and graft type and an abbreviated comorbidity score. Patients with any degree of chronic GvHD reported worse health and a greater symptom burden than patients with resolved chronic GvHD (Table 3). Table 4 shows health status according to chronic GvHD category. In general, the groups of patients differed significantly in self-reported Karnofsky performance status, organs still affected by chronic GvHD, diagnosis of osteopenia or osteoporosis and adrenal insufficiency, use of prednisone, other immunosuppressive medications and antibiotics to prevent infections, use of medications to prevent or treat osteoporosis, and prescription medication use for pain, anxiety and depression. Patients with active mild, moderate or severe chronic GvHD were less likely to work, be in school or be homemakers due to their health problems. No differences were detected in the proportions of patients with self-reported pulmonary disease, avascular necrosis or diabetes.

Discussion The goal of this study was to describe the QOL scores and health status among patients with active chronic

GvHD of differing severity and compare them to those of patients with resolved chronic GvHD or those who had never had chronic GvHD, with an emphasis on the PROMIS measures relative to the SF-36. Patients with resolved chronic GVHD reported QOL similar to patients who had never had chronic GvHD for most instruments, although they continued to have more mouth, eye, nutritional and overall symptoms. Among patients with current chronic GvHD, moderate or severe chronic GVHD was associated with worse patient-reported outcomes than mild disease for all QOL and symptom measures, confirming that PROMIS measures can detect these differences and are comparable to established measures of the impact of chronic GvHD on QOL. A previous 2006 study found similar QOL in patients without chronic GvHD and those whose chronic GvHD had resolved.25 Our findings confirm this observation and extend it by including a number of validated multidimensional QOL instruments and other measures of health status. These observations suggest that prevention and treatment of moderate-severe chronic GvHD rather than mild disease should be targeted in clinical trials. In support of this recommendation is our finding that patients with moderate or severe chronic GvHD were much more likely to still require immunosuppressive treatment and to be taking prescription medication for pain, anxiety and depression. Our

Table 2. Mean scores (standard deviation) for summary and subscale measures.

Instrument

Subscale

N PROMIS Global Health PROMIS 29

Short Form - 36

Lee symptom scale

Physical Mental Physical Social Anxiety Depression Fatigue Sleep Pain PCS MCS Physical Role physical Mental health Role emotional Social Vitality Pain General health Summary Skin Mouth Eye Vitality Lung Nutrition Psychological

Type of chronic graft-versus-host disease No active chronic GVHD Active chronic GVHD Never Resolved P-value: Mild Moderate Severe Pvalue: Clinically never vs. resolved Mild vs. meaningful moderate difference* vs. severe 375 282 257 110 25 52.6 (9.3) 52.9 (9.2) 50.4 (8.4) 55.3 (9.1) 47.1 (8.2) 46.4 (7.6) 46.5 (10.5) 47.1 (9.1) 48.0 (8.4) 49.6 (10.1) 52.9 (9.2) 49.7 (9.6) 49.0 (10.5) 54.1 (8.1) 50.2 (9.9) 50.5 (9.5) 53.4 (10.6) 52.5 (9.8) 49.8 (11.1) 8.5 (7.4) 5.4 (9.3) 2.0 (7.7) 15.5 (20.4) 18.0 (17.4) 3.2 (7.3) 1.3 (3.3) 14.5 (17.8)

52.9 (8.9) 53.0 (9.4) 51.1 (8.0) 55.7 (8.8) 46.5 (8.0) 45.3 (6.5) 46.6 (9.7) 46.9 (8.6) 47.9 (8.3) 49.3 (10.2) 53.4 (8.8) 49.4 (9.5) 49.2 (10.3) 53.6 (8.6) 50.6 (9.2) 50.4 (9.7) 53.8 (10.0) 51.8 (9.7) 49.8 (10.7) 11.1 (8.7) 6.1 (9.8) 4.5 (12.5) 28.3 (27.4) 19.0 (18.0) 3.6 (7.2) 2.4 (6.2) 14.2 (17.1)

0.750 0.913 0.378 0.593 0.459 0.074 0.858 0.761 0.876 0.655 0.516 0.760 0.830 0.442 0.616 0.893 0.605 0.342 0.958 <.001 0.387 0.004 <.001 0.468 0.454 <.001 0.296

48.4 (8.0) 49.4 (8.6) 47.1 (9.2) 51.1 (9.0) 48.7 (8.1) 47.0 (7.2) 49.8 (10.1) 48.5 (8.3) 49.7 (9.0) 44.5 (10.8) 50.7 (9.8) 46.0 (10.5) 43.6 (11.9) 51.5 (9.1) 48.0 (11.1) 46.3 (11.3) 49.9 (10.9) 49.7 (10.0) 43.8 (10.9) 16.2 (9.6) 11.1 (14.4) 13.7 (20.2) 36.5 (30.0) 24.9 (17.5) 4.9 (8.6) 3.7 (6.7) 18.0 (17.5)

42.3 (6.8) 44.7 (7.6) 41.7 (7.7) 46.3 (7.7) 52.9 (9.0) 52.2 (8.8) 55.1 (8.3) 51.0 (8.3) 53.9 (8.3) 38.1 (10.1) 47.8 (10.6) 40.5 (10.6) 38.7 (11.6) 47.9 (9.3) 44.9 (12.0) 41.9 (11.3) 44.8 (9.4) 43.8 (9.7) 37.3 (10.1) 26.2 (11.1) 20.4 (17.6) 26.0 (28.7) 56.5 (30.9) 37.4 (18.8) 8.3 (10.3) 8.0 (12.3) 26.0 (21.7)

42.0 (8.9) 42.2 (8.7) 40.9 (11.1) 46.2 (11.5) 54.6 (9.4) 51.4 (9.2) 55.5 (10.3) 52.8 (7.2) 51.7 (8.9) 36.2 (12.1) 42.9 (13.6) 36.1 (13.2) 34.0 (12.6) 44.8 (11.1) 40.2 (15.8) 37.9 (13.0) 39.7 (12.1) 46.5 (11.9) 35.1 (11.3) 26.9 (11.6) 22.3 (18.7) 12.5 (19.4) 57.5 (31.5) 43.9 (22.6) 13.0 (16.3) 6.5 (12.8) 32.5 (23.6)

<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.010 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.002 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

4.7 4.7 4.5 4.7 4.2 3.8 5.2 4.4 4.3 5.5 4.9 5.3 5.8 4.5 5.3 5.4 5.5 5.1 5.9 5.3 6.6 8.9 14.7 9.5 4.2 3.5 9.3

PCS: physical component score; MCS: mental component score. * 0.5 x standard deviation of the entire study population. Bold values are clinically significant, as defined by a greater than 0.5 x standard deviation difference compared to patients whose chronic GvHD had resolved.

1538

haematologica | 2018; 103(9)

Chronic GvHD, PRO and health status

data further suggest that current interventions are insufficient in controlling the impact of chronic GvHD on QOL. The fact that chronic GvHD remained a significant predictor of QOL despite inclusion of comorbidities suggests that chronic GvHD is associated with additional detrimental comorbidities or causal pathways beyond those included in the analysis. In addition to organ dysfunction and symptoms directly caused by chronic GvHD, patients with chronic GvHD require medications that often have side effects, need frequent appointments for medical monitoring, have functional deficits and environmental limitations, and suffer from frequent infections. These factors can all have a negative impact on QOL.

We found that the NIH PROMIS measures seemed to perform well in distinguishing patients who had never had chronic GvHD or whose chronic GvHD had resolved from those with mild disease and those with moderatesevere disease, suggesting that they could be used in place of the SF-36. The NIH PROMIS measures have a number of advantages over other commonly used instruments in HCT. First, they are available at no cost and feature valid translations in multiple languages. Second, the scales are available in various lengths and can be combined as needed allowing significant flexibility to investigators. Third, the scales can be administered with computer-assisted technology that uses item response theory to minimize

Table 3. Multivariate model of clinical characteristics predicting patient-reported outcomes.1

Factor Age, per decade Comorbidity burden, any vs. none2 Chronic GvHD Never vs. resolved Mild vs. resolved Moderate vs. resolved Severe vs. resolved

SF-36 PCS

SF-36 MCS

PROMIS-Global Physical: T score

PROMIS-Global Mental: T score

Lee symptom scale summary

-1.4 (-1.9, -0.9) <0.0001 -4.5 (-5.5, -3.4) <0.0001

0.8 (0.3, 1.3) 0.003 -1.5 (-2.6, -0.4) 0.005

-0.5 (-1.0, -0.02) 0.04 -3.6 (-4.6, -2.6) <0.0001

0.1 (-0.4, 0.6) 0.76 -2.0 (-3.1, -1.0) 0.0002

0.5 (0.05, 1.0) 0.03 2.1 (1.0, 3.1) < 0.0001

-0.5 (-2.1, 1.1) -4.2 (-6.0, -2.4) -9.6 (-11.9, -7.2) -13.0 (-17.4, -8.7) <0.0001

-0.4 (-1.9, 1.2) -3.0 (-4.7, -1.2) -5.8 (-8.0, -3.6) -9.0 (-13.1, -4.8) <0.0001

-0.8 (-2.2, 0.7) -4.3 (-6.0, -2.7) -9.6 (-11.7, -7.5) -10.8 (-14,9, -6.7) <0.0001

-0.4 (-1.9, 1.1) -3.6 (-5.3, -1.9) -7.7 (-10.0, -5.5) -9.7 (-14.1, -5.4) <0.0001

-2.3 (-3.8, -0.8) 5.3 (3.6, 7.0) 14.8 (12.6, 17.0) 14.9 (10.6, 19.2) < 0.0001

1 Regression coefficients (95% CI) and associated P-values adjusted for conditioning intensity, donor type, prior relapse, sex, graft type, and total body irradiation dose. 2includes pulmonary disease, avascular necrosis, adrenal insufficiency and diabetes, from patient self-reported data

C Figure 1. Box plots showing scores on the quality of life measures. (A) Short-Form 36 (SF-36) Physical Component Score (PCS) and Patient-Reported Outcomes Measurement Information System (PROMIS) Global Health Physical score (GH-Phys), according to whether a patient never (N) had chronic GvHD, had resolved chronic GvHD (R), or currently reported mild (Mi), moderate (Mo) or severe (S) chronic GvHD. (B) SF-36 Mental Component Score (MCS) and PROMIS Global Health Health Mental score (GH-Ment), and (C) PROMIS 29 subscales of physical and social functioning. Higher scores indicate better functioning, and the general population mean is 50 with a standard deviation of 10. The median and interquartile range are depicted by the box, and the range is represented by whiskers. N: never had chronic GvHD; R: resolved chronic GvHD; M: mild chronic GvHD; Mo: moderate chronic GvHD; S: severe chronic GvHD.

haematologica | 2018; 103(9)

1539

S.J. Lee et al.

respondent burden while delivering robust, reliable QOL scores. Finally, all scores are standardized for comparisons, and there are general population normative values for all the scales. Algorithms exist to convert between other commonly utilized and validated instruments using conversion factors.26 We recognize a number of caveats regarding our study. This is a single institution, cross-sectional study and

reflects the patient population and practices of one institution. The response rate was 45%, which is lower than ideal but adequate to conduct analyses given the sample size. This is a late survivorship population with a mean of 13.8 years since HCT, and only 25 participants reported severe chronic GvHD. We hypothesize that chronic GvHD may have already exacted a toll so that those with the highest morbidity died prior to the survey or did not

Table 4. Self-reported health status.1

Organs still affected by chronic GvHD

Resolved n %

Skin Mouth Eyes Intestine Esophagus Liver Lungs Joints Genitals

5 3 3 4 0 0 0 2 0

1.3 0.8 0.8 1.1 ---0.5 --

6 4 6 2 0 1 0 2 1

2.1 1.4 2.1 0.7 -0.4 -0.7 0.4

123 128 115 34 21 24 28 30 16

47.9 49.8 44.7 13.2 8.2 9.3 10.9 11.7 6.2

77 63 71 24 7 5 22 32 11

19 19 107

5.1 5.2 29.2

25 33 104

9.2 12.4 39.0

18 24 107

7.1 9.7 44.8

41 5

10.9 1.4

39 9

14.2 3.4

27 18

14 23

3.7 6.1

9 16

3.3 5.8

12.6

115 20

30.7 5.3

45 46 47

Never

Type of chronic GvHD Mild Moderate n % n %

P-value2

Severe n

70.0 57.3 64.5 21.8 6.4 4.5 20.0 29.1 10.0

14 10 15 3 5 3 8 6 2

56.0 40.0 60.0 12.0 20.0 12.0 32.0 24.0 8.0

<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

10 10 51

9.2 9.9 49.0

1 3 12

4.0 13.6 50.0

0.27 0.03 <0.0001

10.8 7.5

24 12

22.4 11.6

4 0

16.7 --

0.02 <0.0001

90 91

35.7 37.0

64 55

58.7 53.9

17 14

68.0 56.0

<0.0001 <0.0001

10.6

127

50.6

66.4

72.0

<0.0001

112 15

40.3 5.5

85 22

33.6 8.7

44 9

40.4 8.4

9 4

36.0 16.7

0.09 0.11

12.0 12.2 12.5

42 43 40

15.3 15.6 14.7

47 48 64

19.0 19.0 25.7

27 27 31

25.2 25.2 28.7

6 8 7

25.0 32.0 29.2

0.007 0.002 <0.0001

9.4

10.5

19.8

40.2

37.5

<0.0001

84.6

239

87.9

184

73.3

60.6

70.8

15.4

12.1

26.7

39.4

29.2

(90-100)

100

(90-100)

(80-100)

(70-90)

Comorbidities Pulmonary disease Avascular necrosis Osteopenia/ osteoporosis Diabetes Adrenal insufficiency

Current medications Prednisone Other immunosuppressive drugs Prophylactic antimicrobials Anti-hypertensives Treatment of arrhythmias Anxiolytics Antidepressants Prevention or treatment of osteoporosis Prescription pain medication

Work/school status Work, school or 308 homemaking full or part time, retired by choice Not working, not 56 in school, unemployed, or retired due to health KPS (median, IQR) 100

<0.0001

Based on endorsement of each item. Missing responses and “don’t know” were excluded. Karnofsky performance score (KPS) comparison based on the Wilcoxon rank sum test. All other factors were compared based on chi-squared or Fisher exact tests for patients across all five categories. 1

1540

haematologica | 2018; 103(9)

Chronic GvHD, PRO and health status

participate, thus potentially biasing our results and leading to an inability to differentiate moderate from severe chronic GvHD. This hypothesis is supported by a lack of difference in the frequency of well-known steroid-associated complications such as avascular necrosis and diabetes. Additional studies to show that the range of the PROMIS measures is sufficiently sensitive for patients with extreme levels of dysfunction are necessary. Nevertheless, we were able to show significantly compromised QOL and functional status for patients with moderate or severe chronic GvHD, and the PROMIS measures were able to detect differences between chronic GvHD categories. Another limitation is that the analysis is based on self-reported chronic GvHD severity since most patients did not have a recent clinical severity assessment by a transplant expert. Unfortunately we do not have concurrent clinician-assessed or NIH-classified severity information for this population. Finally, because this is a crosssectional study, patients were analyzed according to their self-reported category at the time of the survey. A longitudinal study is needed to confirm that patients who move from moderate-severe to resolved chronic GvHD have improved QOL, ideally, similar to those who had no or mild chronic GvHD. This information would be

References 1. Lee SJ, Flowers ME. Recognizing and managing chronic graft-versus-host disease. Hematology Am Soc Hematol Educ Program. 2008:134-141. 2. Flowers ME, Inamoto Y, Carpenter PA, et al. Comparative analysis of risk factors for acute graft-versus-host disease and for chronic graft-versus-host disease according to National Institutes of Health consensus criteria. Blood. 2011;117(11):3214-3219. 3. Lee SJ. Classification systems for chronic graft-versus-host disease. Blood. 2017;129 (1):30-37. 4. Pidala J, Kurland B, Chai X, et al. Patientreported quality of life is associated with severity of chronic graft-versus-host disease as measured by NIH criteria: report on baseline data from the Chronic GVHD Consortium. Blood. 2011;117(17):46514657. 5. Lee SJ, Logan B, Westervelt P, et al. Comparison of patient-reported outcomes in 5-year survivors who received bone marrow vs peripheral blood unrelated donor transplantation: long-term follow-up of a randomized clinical trial. JAMA Oncol. 2016;2(12):1583-1589. 6. Walker I, Panzarella T, Couban S, et al. Pretreatment with anti-thymocyte globulin versus no anti-thymocyte globulin in patients with haematological malignancies undergoing haemopoietic cell transplantation from unrelated donors: a randomised, controlled, open-label, phase 3, multicentre trial. Lancet Oncol. 2016;17(2):164-173. 7. Cella D, Riley W, Stone A, et al. The PatientReported Outcomes Measurement Information System (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005-2008. J Clin Epidemiol. 2010;63(11):1179-1194. 8. Bashey A, Zhang MJ, McCurdy SR, et al. Mobilized peripheral blood stem cells versus unstimulated bone marrow as a graft source

haematologica | 2018; 103(9)

10.

11.

12.

13. 14.

15.

16.

17.

encouraging to patients currently being treated for chronic GvHD. In conclusion, among patients with current chronic GvHD, moderate to severe chronic GvHD is associated with worse QOL scores whether measured by the SF-36, PROMIS Global Health, PROMIS-29 or LSS, and the SF-36 and PROMIS measures provide comparable score profiles. Prevention or better treatment of moderate to severe chronic GvHD is important in order to minimize adverse effects on QOL and health status for allogeneic transplant survivors. Patients whose chronic GvHD resolves appear similar to those who have never had chronic GvHD, except that they continue to have statistically but not clinically worse mouth, eye, nutritional and overall chronic GvHD symptoms. The freely available, concise and flexible PROMIS Global Health and PROMIS-29 measures seem to perform as well as the SF-36 in capturing the multi-dimensional QOL of these patients.26 Thus any of these multi-dimensional measures could be used along with the LSS for assessing QOL in HCT clinical trials. Acknowledgment This work was supported by the U.S. National Institutes of Health (grants CA018029, CA215134, and CA118953).

for T-cell-replete haploidentical donor transplantation using post-transplant cyclophosphamide. J Clin Oncol. 2017;35(26):30023009. Bleakley M, Heimfeld S, Loeb KR, et al. Outcomes of acute leukemia patients transplanted with naive T cell-depleted stem cell grafts. J Clin Invest. 2015;125(7):2677-2689. Finke J, Bethge WA, Schmoor C, et al. Standard graft-versus-host disease prophylaxis with or without anti-T-cell globulin in haematopoietic cell transplantation from matched unrelated donors: a randomised, open-label, multicentre phase 3 trial. Lancet Oncol. 2009;10(9):855-864. Kanakry CG, Bolanos-Meade J, Kasamon YL, et al. Low immunosuppressive burden after HLA-matched related or unrelated BMT using posttransplantation cyclophosphamide. Blood. 2017;129(10):1389-1393. Kroger N, Solano C, Wolschke C, et al. Antilymphocyte globulin for prevention of chronic graft-versus-host disease. N Engl J Med. 2016;374(1):43-53. Im A, Hakim FT, Pavletic SZ. Novel targets in the treatment of chronic graft-versus-host disease. Leukemia. 2017;31(3):543-554. MacDonald KPA, Betts BC, Couriel D. Emerging therapeutics for the control of chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2018;24(1):19-26. Cooke KR, Luznik L, Sarantopoulos S, et al. The biology of chronic graft-versus-host disease: a task force report from the National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2017;23(2):211-234. MacDonald KP, Hill GR, Blazar BR. Chronic graft-versus-host disease: biological insights from preclinical and clinical studies. Blood. 2017;129(1):13-21. Zeiser R, Blazar BR. Pathophysiology of chronic graft-versus-host disease and therapeutic targets. N Engl J Med. 2017;377(26): 2565-2579.

18. Miklos D, Cutler CS, Arora M, et al. Ibrutinib for chronic graft-versus-host disease after failure of prior therapy. Blood. 2017;130(21):2243-2250. 19. Khera N, Storer B, Flowers ME, et al. Nonmalignant late effects and compromised functional status in survivors of hematopoietic cell transplantation. J Clin Oncol. 2012;30(1):71-77. 20. Ware JE, Kosinski M, Keller SD. SF-36 physical and mental health summary scales: a user's manual. Boston: The Health Institute, New England Medical Center; 1994. 21. Ware JE, Snow KK, Kosinski M, Gandek B. SF-36 Health Survey: a Manual and Interpretation Guide. Boston: The Health Institute, New England Medical Center; 1993. 22. Jensen RE, Potosky AL, Moinpour CM, et al. United States Population-Based Estimates of Patient-Reported Outcomes Measurement Information System symptom and functional status reference values for individuals with cancer. J Clin Oncol. 2017;35(17):19131920. 23. Lee S, Cook EF, Soiffer R, Antin JH. Development and validation of a scale to measure symptoms of chronic graft-versushost disease. Biol Blood Marrow Transplant. 2002;8(8):444-452. 24. Merkel EC, Mitchell SA, Lee SJ. Content validity of the Lee Chronic Graft-versusHost Disease Symptom Scale as assessed by cognitive interviews. Biol Blood Marrow Transplant. 2016;22(4):752-758. 25. Fraser CJ, Bhatia S, Ness K, et al. Impact of chronic graft-versus-host disease on the health status of hematopoietic cell transplantation survivors: a report from the Bone Marrow Transplant Survivor Study. Blood. 2006;108(8):2867-2873. 26. Shaw BE, Syrjala KL, Onstad LE, et al. PROMIS measures can be used to assess symptoms and function in long-term hematopoietic cell transplantation survivors. Cancer. 2018;124(4):841-849.

1541

ARTICLE

Blood Transfusion

Ferrata Storti Foundation

Transfusion of packed red blood cells at the end of shelf life is associated with increased risk of mortality – a pooled patient data analysis of 16 observational trials

Monica S.Y. Ng,1,2 Michael David,3 Rutger A. Middelburg,4,5 Angela S.Y. Ng,1 Jacky Y. Suen,1 John-Paul Tung1,2 and John F. Fraser1

Critical Care Research Group, Faculty of Medicine, University of Queensland, Brisbane, Australia; 2Research and Development, Australian Red Cross Blood Service, Brisbane, Australia; 3School of Medicine and Population Health, The University of Newcastle, Callaghan, Australia; 4Centre for Clinical Transfusion Research, Sanquin Research, Leiden, the Netherlands and 5Department of Clinical Epidemiology, Leiden University Medical Center, the Netherlands

Haematologica 2018 Volume 103(9):1542-1548

ABSTRACT

Correspondence: monica.ng91@gmail.com

Received: February 22, 2018. Accepted: May 21, 2018. Pre-published: May 24, 2018.

bservational studies address packed red blood cell effects at the end of shelf life and have larger sample sizes compared to randomized control trials. Meta-analyses combining data from observational studies have been complicated by differences in aggregate transfused packed red blood cell age and outcome reporting. This study abrogated these issues by taking a pooled patient data approach. Observational studies reporting packed red blood cell age and clinical outcomes were identified and patient-level data sets were sought from investigators. Odds ratios and 95% confidence intervals for binary outcomes were calculated for each study, with mean packed red blood cell age or maximum packed red blood cell age acting as independent variables. The relationship between mean packed red blood cell age and hospital length of stay for each paper was analyzed using zero-inflated Poisson regression. Random effects models combined paper-level effect estimates. Extremes analyses were completed by comparing patients transfused with mean packed red blood cell aged less than ten days to those transfused with mean packed red blood cell aged at least 30 days. sixteen datasets were available for pooled patient data analysis. Mean packed red blood cell age of at least 30 days was associated with an increased risk of in-hospital mortality compared to mean packed red blood cell of less than ten days (odds ratio: 3.25, 95% confidence interval: 1.27-8.29). Packed red blood cell age was not correlated to increased risks of nosocomial infection or prolonged length of hospital stay.

doi:10.3324/haematol.2018.191932 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1542 ©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.

1542

Introduction Packed red blood cells (PRBCs) have a shelf life of 21-49 days depending on the additive solution used and jurisdiction.1 During in vitro storage, PRBCs accumulate cellular and biochemical changes collectively called the red blood cell (RBC) storage lesion. These changes occur secondary to RBC metabolism under artificial conditions such as refrigeration temperatures, limited nutrient supply and absent clearance mechanisms. RBC storage lesions clearly disturb physiological mechanisms and lead to harm in in vitro tests and animal models.2,3 Clinical data linking PRBC storage duration to adverse outcomes, such as mortality, nosocomial infection, prolonged hospital length of stay (HLOS), prolonged intensive care unit (ICU) length of stay, prolonged mechanical ventilation, acute renal failure and multiple organ dysfunction – have been consistently equivocal.2,4 The majority of studies demonstrating the damage caused by prolonged PRBC storage have been retrospective and observational in nature.5 Recent large randomized controlled trials (RCTs) have not linked PRBC storage duration to adverse outcomes.6-9 However, the two types of studies tended to ask different questions— the former, “Is prolonged PRBC storage harmful?” versus the latter, “Are fresh PRBCs better than standard practice?”. Furthermore, observational studies often had larger haematologica | 2018; 103(9)

Old PRBCs linked to increased mortality across 16 studies Table 1. Data synthesis method.

Demographic variables Age BMI Ejection fraction Outcome variables Mortality Surgical wound infection Any infection Duration of MV HLOS ILOS MODS

Calculated from date of birth and date of first transfusion Calculated from height and weight Classified as good >50%, fair 30-50%, poor <30% In-hospital mortality used whenever possible; when not reported, the mortality over the shortest duration used (e.g. 30-day mortality over 90-day mortality) General surgery: abdominal + perineal wound infections, intra-abdominal abscess Cardiothoracic surgery: sternal infection, anastomosis infection Any infectious outcome recorded including pneumonia, surgical wound infection, bloodstream infection, urinary tract infection Converted to days when reported as hours Calculated from hospital admission and discharge date Calculated from ICU admission and discharge date Peak SOFA score ≥6 provided that SOFA score<6 prior to intensive care unit admission

BMI: body mass index; HLOS; hospital length of stay; ICU: intensive care medicine; ILOS: intensive care length of stay; MODS: multiple organ dysfunction syndrome; MV: mechanical ventilation; SOFA: sepsis-related organ failure assessment score.

sample sizes with older mean ages of transfused PRBC, making them uniquely positioned to identify small adverse effects owing to PRBC units at the end of shelf life.10 One of the key issues with combining data from observational studies is the diverse methods of describing aggregate PRBC age, ranging from mean PRBC age to dichotomization at “x” days to maximum PRBC age transfused.5 Some paper level meta-analyses have used various adjustments with the aim of unifying aggregate PRBC age measurements and maximizing PRBC age differences between comparison groups.11 However, the temporal effect of storage-induced adverse outcomes has been relatively unmapped, and there is no evidence that different measures of aggregate PRBC age are interchangeable in their association with clinical outcomes.12 Moreover, the assumptions used to convert one aggregate measure to another (e.g., median to mean) could potentially lead to statistical inaccuracy. This issue could be abrogated via pooled patient data analysis allowing the use of one aggregate PRBC age measure. Pooled patient data analysis also touts improved subgroup analyses and consistency across studies compared to paper level analysis.13 This study analyzed pooled individual patient data (IPD) from 16 observational studies with the aim of quantifying the association of PRBC storage duration on mortality, nosocomial infection and HLOS. In representing 16 retrospective studies and over 17,000 patients – the study herein is one of the largest pooled patient data analysis completed for the investigation of storage-induced adverse PRBC transfusion outcomes to date.

Methods Study selection Institutional ethics approvals were sought from the University of Queensland and The Alfred Hospital prior to initiation of study. Observational studies reporting PRBC storage duration and clinical outcomes such as mortality, infection and HLOS were identified from PubMed and EMBASE using protocols haematologica | 2018; 103(9)

described previously.5 Corresponding investigators of each study were contacted to request the underlying patient-level dataset.

Data extraction and synthesis Demographic, intervention and outcome variables reported in more than three studies were combined into one Microsoft Excel 2016 spreadsheet by author Monica S.Y. Ng. and checked by author Angela S.Y. Ng. Patients who did not receive any PRBC units were excluded from the dataset. Variables not reported in the desired format were calculated from primary data where available. Table 1 demonstrates adjustments made to synthesize the aggregate datasheet.14 When individual PRBC unit ages were available, the aggregate ages of PRBC transfusions were expressed as mean age or maximum PRBC age. In so doing, aggregate PRBC ages were expressed in a time-independent manner for incorporation into logistic models.

Data analysis A two-stage meta-analysis using IPD was used to account for differences between study cohorts in general analyses. This approach has been shown to increase statistical power and avoid ecological bias when compared with the traditional approach which pools study estimates.15,16 In the first stage, the association between binary outcomes (such as in-hospital mortality or nosocomial infection) and PRBC age (expressed as mean PRBC age or maximum PRBC age) were calculated using binomial logistic regression for each study with age, sex and PRBC volume as continuous covariates. PRBC age (mean PRBC age, maximum PRBC age), recipient age and PRBC volume were incorporated as continuous variables, while sex was included as a binary variable. Each effect estimate was reported as an odds ratio (OR) with 95% confidence intervals (CI). Regarding HLOS, and due to an excessive number of zeroes, zero-inflated Poisson regression modelling was used to calculate the incidence rate ratio (IRR) for an additional day in hospital as a function of PRBC age (expressed as mean PRBC age or maximum PRBC age) for each paper. In the second stage, random effects models were used to combine the effect estimates for each paper. Funnel plots were generated for each analysis involving more than ten papers to assess for publication bias. This threshold was used as funnel 1543

M.S.Y. Ng et al.

plots with ten or less studies have insufficient power to identify heterogeneity.17 Sensitivity analyses were completed by calculating the adjusted OR for each geographical location (the Americas, Europe, other) and patient subgroup (cardiac surgery, intensive care unit (ICU), other). Time-lapse analyses involved adding studies sequentially to the random effects model in order of the recruitment period (i.e., from 1980-2011). In the extremes analyses for in-hospital mortality and nosocomial infection, logistic regression was used to calculate an aggregate OR and 95% CI comparing patients transfused with mean PRBC aged less than ten days to those transfused with mean PRBC aged at least 30 days old. The ten day threshold for fresh PRBC was selected to align with the Red Cell Storage Duration Study (RECESS). The 30 day threshold for stored PRBC was selected to maintain relevance for jurisdictions which store PRBC for 35 (e.g., China, The Netherlands, UK) and 42 days (e.g., Australia, USA, Canada). Furthermore, in vitro research suggests that PRBC storage lesions become clinically significant up to day 28.18 Subgroup analyses were completed to measure the modifying effects of leukoreduction status. Zero-inflated Poisson regression was used to calculate an aggregate IRR and 95% CI for HLOS extremes analysis. Age, sex and PRBC volume were included as covariates for each model. STATATM (StataCorp 2017, version 15.0) was utilized for all analyses.

Results Study characteristics Using the search strategy as previously specified, 3285 abstracts were retrieved from PubMed and EMBASE (Figure 1).5 After two sequential screens and a manual search, 64 clinical studies investigating clinical outcomes associated with PRBC storage duration were retained. Eight RCTs were removed, leaving 56 observational studies. Sixteen datasets were received from 14 investigators between January 2014 and January 2017. These studies covered 17,967 patients across burns, general surgery, ICU, oncology, acute medicine, trauma and cardiac surgery cohorts (Table 2). Overall, 77,962 units of PRBC were transfused across 16 studies with an average of 4.34 units transfused per patient. 8.1% of patients received more than ten units of PRBCs. The mean PRBC age transfused was less than ten days in 15.4% and at least 30 days in 18.9% of participants. The mean age of patients was 57.82, with 45.3% of patients being older than 65. On paper level comparisons, included studies had similar recruitment dates, mean sample size (n=904.56 versus n=1003.53) and rates of positive mortality findings (36.36% versus 33.33%) compared to studies for which

Figure 1. Outline of study selection. Fiftysix observational studies investigated the effects of PRBC storage duration on clinical outcomes, such as mortality, infection risk and hospital length of stay. Forty datasets were unavailable for various reasons: (1) no response from corresponding author after initial email, (2) lack of correspondence after initial contact, (3) institutional policy against data use by external investigators, (4) insufficient staff available to access data on site, (5) investigator retracted participation in study, and (6) other (e.g., data file corrupted). PRBC: packed red blood cells; yo: years old: RCT: randomized controlled trials.

1544

haematologica | 2018; 103(9)

Old PRBCs linked to increased mortality across 16 studies

datasets were unavailable (Online Supplementary Table S1). Received datasets included higher proportions of studies conducted in Europe (56.25% versus 25.00%) and ICU patients (31.25% versus 12.50%) compared to excluded studies. Unavailable datasets included higher proportions of studies conducted in the Americas (70.00% versus 25.00%) and trauma patients (32.50% versus 6.25%). Excluded studies also had increased rates of associations between PRBC age versus nosocomial infection (56.26% versus 12.50%) and HLOS (42.86% versus 25.00%). There were significant variations in demographic, treatment and outcome variables released by investigators (Online Supplementary Tables S2-S4). Age (n=15), sex (n=14) and PRBC volume (n=16) transfused were the most consistently reported demographic variables across studies. Mortality (n=16), any infection (n=10) and HLOS (n=12) were the most commonly reported outcome variables.

Mean PRBC age >30 days associated with an increased risk of in-hospital mortality Thirteen datasets contained the required covariate (age, sex, PRBC volume) and outcome data required for mortality analyses. These datasets covered 14,867 patients and 58,272 transfusions. There was no evidence of a significant association between mean PRBC age and post-transfusion in-hospital mortality (OR: 0.99, 95% CI: 0.98-1.00) (Figure 2). Similarly, there was no association between maximum transfused PRBC age and post-transfusion inhospital mortality (OR: 1.00, 95% CI: 0.98-1.01, Online Supplementary Figure S5). There was substantial heterogeneity in both analyses with I2 values of 63.6% and 59.0%, respectively. Sensitivity analyses demonstrated that effects estimates were similar across geographic location and patient subgroups (Online Supplementary Figure S6). Funnel plots for both analyses suggest that publication bias was not a significant concern. Time-lapse analyses demonstrated a trend of increasing OR in favor of fresh PRBCs from 1980-2011 (Online Supplementary Figure S7). Extremes analyses associated mean PRBC stored for at least 30 days with increased in-hospital mortality risk, as compared to mean PRBC stored for less than ten days (OR: 3.25, 95% CI: 1.27-8.29, Table 3). This association persisted in patients who received leukoreduced PRBCs (OR: 2.74, 95% CI: 1.39-5.36, Table 3).

No association between PRBC age and nosocomial infection Eight datasets were incorporated in the nosocomial infection analyses covering 2716 patients. Mean PRBC age (OR: 0.99, 95% CI: 0.97-1.02, Figure 3) and maximum

PRBC age (OR: 1.00, 95% CI: 0.96-1.04, Online Supplementary Figure S8) transfused were not correlated to nosocomial infection. There was moderate heterogeneity in both analyses with I2 values of 42.5% and 28.6%, respectively. Stratifying analyses by geographical location and patient subgroup did not significantly alter effect estimates (Online Supplementary Figure S9). Funnel plots were not constructed as less than ten datasets were analyzed. Nosocomial infection ORs remained consistent across recruitment periods on time lapse analyses (Online Supplementary Figure S10). Extremes analyses did not idenTable 2. Demographic features of patients in aggregate dataset.

Demographic Characteristic

PRBC dataset (n=17,967)

Age (years; mean)* <40 years old (count, %) 40-65 years old (count, %) >65 years old (count, %) Sex (male; count, %) Mean volume of PRBC transfused (units)* >10 units (count, %) Mean age of PRBC transfused (days)* <10 days (count, %) ≥ 30 days (count, %) In-hospital mortality (count, %) Nosocomial infection rate (count, %) Average HLOS <5 days (count, %) 5-14 days (count, %) >20 days (count, %) Clinical setting Cardiac surgery (count, %) General surgery (count, %) Acute medicine (count, %) Intensive care unit (count, %) Orthopedic surgery (count, %) Other (count, %) Geographic location The Americas (count, %) Europe (count, %) Middle East (count, %) ANZ (count, %)

57.82 ± 0.17 3094 (17.2) 6425 (35.8) 8146 (45.3) 9865 (54.9) 4.34 ± 0.04 1462 (8.1) 16.9 ± 0.12 2771 (15.4) 3399 (18.9) 5010 (27.4) 675 (19.2) 13.92 ± 0.15 3441 (18.8) 7418 (40.5) 2790 (15.2) 4403 (24.0) 412 (2.2) 9967 (54.4) 757 (4.1) 871 (4.8) 1904 (10.4) 1863 (12.5) 7258 (48.8) 1695 (11.4) 4051 (27.3)

*All means presented ± standard error of the mean. ANZ: Australia and New Zealand: PRBC: packed red blood cells; HLOS; hospital length of stay.

Table 3. Odds ratios from extremes analysis for in-hospital mortality and nosocomial infection risk as a function of mean PRBC age. Patient age, PRBC volume and sex are entered as covariates in the logistic model.

Independent variable Mean PRBC age Patient age PRBC volume Sex

Mortality Odds ratio 95% CI 3.25* 1.02* 1.03 1.14

1.27-8.29 1.01-1.03 0.98-1.07 0.90-1.45

Mortality (LR) Odds ratio 95% CI 2.74* 1.03* 1.02 1.16

1.39-5.37 1.02-1.03 0.97-1.07 0.92-1.47

Nosocomial infection Odds ratio 95% CI 1.57 1.00 0.97 1.99*

0.39-6.27 0.98-1.02 0.86-1.08 1.45-2.74

*=statistically significant (P<0.05). 95% CI: 95% confidence intervals; LR: leukoreduced patients only; PRBC: packed red blood cell.

haematologica | 2018; 103(9)

1545

M.S.Y. Ng et al.

tify any association between mean PRBC age and nosocomial infection risk (OR: 1.57, 95% CI: 0.39-6.27, Table 3).

No association between PRBC age and HLOS Ten datasets were analyzed in the HLOS analyses covering 14,063 patients. The incidence rates were found to decrease by 3.1% and 0.4% for each additional day of mean PRBC age and maximum PRBC age, respectively. However, neither rate was statistically significant after controlling for age, sex and PRBC volume. There was significant heterogeneity in the mean PRBC age analyses with an I2 value of 98.6%. Patient subgroup analyses identified ICU and other patients as major sources of heterogeneity compared to cardiac surgery patients (Online Supplementary Figure S11). However, all patient subgroups groups generated similar effect estimates. The effect estimate for studies originating from the USA was significantly different to studies from Europe and other countries, however, there was only one study in the USA group (Online Supplementary Figure S11).Time-lapse analyses found that the incidence rate ratio remained stable over time (Online Supplementary Figure S12). On extremes analysis, the incidence rates for days in hospital increased by 8.3% in patients with a mean PRBC of at least 30 days compared to less than ten days. Similar to the general analyses, this rate was not statistically significant after adjusting for age, sex and PRBC volume transfused.

Discussion The use of IPD enabled consistent treatment and outcome measures throughout analyses, leading to reduced variability and improved precision compared to paper level meta-analyses.5 The association between stored PRBCs and mortality aligned with the results of observational, but not RCT paper level meta-analyses.10,11,19 These RCT meta-analyses included small pilot trials and combined diverse populations. Furthermore, PRBCs also vary significantly within a blood bank due to donor, preparation and storage factors.20 The increased variability and confounding factors may have obscured associations between PRBC age and mortality.19 While large RCTs have not associated PRBC age with mortality, these RCTs did not test PRBCs at the end of shelf life.7-9,21 The extremes analyses addressed this issue by comparing patients transfused with a mean PRBC of at least 30 days to those transfused with a mean PRBC of less than ten days from observational studies. In this way, the pooled IPD analysis bridged the dissonance between clinical approaches (which tested stored PRBC aged 17 ± 13 days) and in vitro protocols (which sampled PRBCs over the course of shelf life).22 The association between PRBC >30 days old and mortality differed from secondary analyses of the Informing Fresh versus Old Red Cell Management (INFORM) study.23 In the secondary analyses, the use of maximum PRBC age to define stored PRBCs may have overestimated the aggregate PRBC age transfused.24 Patients transfused with predominantly fresh PRBC units, but with the addition of one PRBC unit >35 days old could inflate the number of “survivors” in the stored PRBC group. The finding that PRBC age was not associated with nosocomial infection corroborated the results from large RCTs.7,21 However, the datasets available for IPD analysis 1546

Figure 2. Forest plots and funnel plots for mortality analysis as a function of mean PRBC age. Mortality odds ratios were calculated for each study using logistic regression with mean PRBC age transfused as the independent variable. Age, sex and PRBC volume were covariates. (A) Odds ratios were then combined using random effects models. (B) Funnel plots were generated for each analyses to assess for publication bias. OR: odds ratio; CI: confidence interval.

had lower rates of associating PRBC age with nosocomial infections compared to unavailable datasets – potentially leading to an underestimation of the relationship between PRBC age and nosocomial infection risk.19 There was no association between PRBC age and HLOS. This may have occurred due to substantial HLOS variability as it is a composite indicator of disease severity, treatment efficacy and safety, that is heavily modulated by social factors. This pooled patient analysis was the first assessment of HLOS across multiple clinical PRBC age studies. Published studies describing PRBC storage effects have reported HLOS as stratified count data,25 median and interquartile range,26-28 median and range,29 mean and standard error,30 and Pearson’s correlation result31 – making paper level meta-analyses difficult. The IPD analysis abrogated this issue by using patient level HLOS data, allowing one consistent measure across all studies. The potential limitations of this study include long recruitment time, applicable to high-volume PRBC transfusion, potential confounding factors and selection bias. PRBC processing and transfusion guidelines may have haematologica | 2018; 103(9)

Old PRBCs linked to increased mortality across 16 studies

Figure 3. Forest plots for nosocomial infection analysis as a function of PRBC mean age. Nosocomial infection odds ratios were calculated for each study using logistic regression with mean PRBC age transfused as the independent variable. Age, sex and PRBC volume were entered into the model as covariates. Odds ratios were then combined using random effects models. OR: odds ratio; CI: confidence interval.

changed during the recruitment period (1980-2011), altering the effects of PRBC age on adverse events. Notably, the association between PRBC age and mortality persisted after leukoreduction â&#x20AC;&#x201C; one of the major changes in PRBC processing over the past 20 years. The mean volume of PRBC transfused in this study was approximately four units, signifying that the results likely have limited applicability to massive transfusion protocols where more than ten units can be transfused in one incident. This pooled patient data analysis may have been affected by confounding factors due to the use of observational studies, as patients were not randomly allocated to treatment groups. This pooled patient analysis incorporated three covariates â&#x20AC;&#x201C; age, sex and PRBC volume, which have been demonstrated to be the most influential confounders over multiple studies.32,33 The assessment of a limited subset of observational studies introduced the risk of selection bias. This effect was less likely to affect the observed association between PRBC age and mortality as it aligned with the results of existing meta-analyses.11 Furthermore, included studies were similar to excluded studies in terms of study size, recruitment dates and rates of positive mortality findings. Pooled IPD analysis found an association between PRBC >30 days and increased risks of in-hospital mortali-

References 1. Flegel WA, Natanson C, Klein HG. Does prolonged storage of red blood cells cause harm? Br J Haematol. 2014;165(1):3-16. 2. Antonelou MH, Seghatchian J. Insights into red blood cell storage lesion: toward a new appreciation. Transfus Apher Sci. 2016; 55(3):292-301. 3. Tung JP, Simonova G, Glenister K, Fraser JF, Fung YL. The sheep as a transfusion model:

haematologica | 2018; 103(9)

ty. This result was significant as 24.5% of all O negative PRBC units are transfused after 35-42 days of storage in Australia.34 These results align with the findings of in vitro studies and support the shortening of PRBC shelf life to 35 days, which has already occurred in some jurisdictions (The Netherlands, UK, Germany, China). Shortening PRBC shelf life should be approached with caution due to its implications for blood bank management, product wastage and PRBC access in remote locations. Ideally, these findings should be confirmed using extremes analyses of pooled patient data from recent large RCTs prior to implementing changes. Acknowledgments We would like to thank Professor Cartotto, Professor Edna, Professor Gajic, Professor Juffermans, Doctor Kadar, Professor Cooper, Professor Phelan, Doctor Sanders, Professor van de Watering, Doctor Voorhuis and Doctor Yap for providing access to their patient-level data. Funding Australian governments fund the Australian Red Cross Blood Service for the provision of blood, blood products and services to the Australian community. This project was partially funded by The Prince Charles Hospital Foundation.

comparison of the storage lesion of human and ovine red blood cell units. Transfus Med. 2013;23(s2):30-71. 4. Lelubre C, Vincent JL. Relationship between red cell storage duration and outcomes in adults receiving red cell transfusions: a systematic review. Crit Care. 2013; 17(2):R66. 5. Ng MS, Ng AS, Chan J, Tung JP, Fraser JF. Effects of packed red blood cell storage duration on post-transfusion clinical outcomes: a meta-analysis and systematic

review. Intensive Care Med. 2015;41(12): 2087-2097. 6. Lacroix J, Hebert P, Fergusson D, et al. The Age of Blood Evaluation (ABLE) randomized controlled trial: study design. Transfus Med Rev. 2011;25(3):197-205. 7. Cooper DJ, McQuilten ZK, Nichol A, et al. Age of red cells for transfusion and outcomes in critically ill adults. N Engl J Med. 2017;377(19):1858-1867. 8. Heddle NM, Cook RJ, Arnold DM, et al.

1547

M.S.Y. Ng et al.

10.

11.

12.

13.

14.

15.

16.

17.

1548

Effect of short-term vs. long-term blood storage on mortality after transfusion. N Engl J Med. 2016;375(20):1937-1945. Steiner ME, Ness PM, Assmann SF, et al. Effects of red-cell storage duration on patients undergoing cardiac surgery. N Engl J Med. 2015;372(15):1419-1429. Remy KE, Sun J, Wang D, et al. Transfusion of recently donated (fresh) red blood cells (RBCs) does not improve survival in comparison with current practice, while safety of the oldest stored units is yet to be established: a meta-analysis. Vox Sang. 2016;111(1):43-54. Wang D, Sun J, Solomon SB, Klein HG, Natanson C. Transfusion of older stored blood and risk of death: a meta-analysis. Transfusion. 2012;52(6):1184-1195. Zimring JC. Established and theoretical factors to consider in assessing the red cell storage lesion. Blood. 2015;125(14):21852190. Thomas D, Radji S, Benedetti A. Systematic review of methods for individual patient data meta- analysis with binary outcomes. BMC Med Res Methodol. 2014;14:79. Sakr Y, Lobo SM, Moreno RP, et al. Patterns and early evolution of organ failure in the intensive care unit and their relation to outcome. Crit Care. 2012;16(6):R222. Riley RD, Lambert PC, Abo-Zaid G. Metaanalysis of individual participant data: rationale, conduct, and reporting. BMJ. 2010;340:c221. Simmonds M, Stewart G, Stewart L. A decade of individual participant data metaanalyses: A review of current practice. Contemp Clin Trials. 2015;45(Pt A):76-83. Sterne JA, Sutton AJ, Ioannidis JP, et al.

18. 19.

20. 21.

22.

23.

24.

25.

26.

Recommendations for examining and interpreting funnel plot asymmetry in metaanalyses of randomised controlled trials. BMJ. 2011;343:d4002. Cohen B, Matot I. Aged erythrocytes: a fine wine or sour grapes? Br J Anaesth. 2013;111(Suppl 1):S62-70. Alexander PE, Barty R, Fei Y, et al. Transfusion of fresher vs older red blood cells in hospitalized patients: a systematic review and meta-analysis. Blood. 2016; 127(4):400-410. Sparrow RL. Red blood cell components: time to revisit the sources of variability. Blood Transfus. 2017;15(2):116-125. Lacroix J, Hebert PC, Fergusson DA, et al. Age of transfused blood in critically ill adults. N Engl J Med. 2015;372(15):14101418. Prudent M, Tissot JD, Lion N. In vitro assays and clinical trials in red blood cell aging: Lost in translation. Transfus Apher Sci. 2015;52(3):270-276. Cook RJ, Heddle NM, Lee KA, et al. Red blood cell storage and in-hospital mortality: a secondary analysis of the INFORM randomised controlled trial. Lancet Haematol. 2017;4(11):e544-e552. Bautista A, Wright TB, Meany J, et al. Red cell storage duration does not affect outcome after massive blood transfusion in trauma and nontrauma patients: a retrospective analysis of 305 patients. Biomed Res Int. 2017;2017:3718615. Keller ME, Jean R, LaMorte WW, Millham F, Hirsch E. Effects of age of transfused blood on length of stay in trauma patients: a preliminary report. J Trauma. 2002; 53(5):1023-1025. Sanders J, Patel S, Cooper J, et al. Red blood

27.

28.

29.

30.

31.

32.

33.

34.

cell storage is associated with length of stay and renal complications after cardiac surgery. Transfusion. 2011;51(11):2286-2294. Cywinski JB, You J, Argalious M, et al. Transfusion of older red blood cells is associated with decreased graft survival after orthotopic liver transplantation. Liver Transpl. 2013;19(11):1181-1188. Edgren G, Kamper-Jorgensen M, Eloranta S, et al. Duration of red blood cell storage and survival of transfused patients. Transfusion. 2010;50(6):1185-1195. Cartotto R, Yeo C, Camacho F, Callum J. Does the storage age of transfused blood affect outcome in burn patients? J Burn Care Res. 2013;24:186-197. Vamvakas EC, Carven JH. Length of storage of transfused red cells and postoperative morbidity in patients undergoing coronary artery bypass graft surgery. Transfusion. 2000;40(1):101-109. Min JJ, Bae JY, Kim TK, et al. Association between red blood cell storage duration and clinical outcome in patients undergoing off-pump coronary artery bypass surgery: a retrospective study. BMC Anesthesiol. 2014;14:95. Edgren G, Kamper-Jorgensen M, Eloranta S, et al. Duration of red blood cell storage and survival of transfused patients (CME). Transfusion. 2010;50(6):1185-1195. Middelburg RA, van de Watering LM, Briet E, van der Bom JG. Storage time of red blood cells and mortality of transfusion recipients. Transfus Med Rev. 2013; 27(1):36-43. Hirani R, Wong J, Diaz P, et al. A national review of the clinical use of group O D- red blood cell units. Transfusion. 2017; 57(5):1254-1261.

haematologica | 2018; 103(9)

ARTICLE

Coagulation & its Disorders

Frequency, risk factors, and impact on mortality of arterial thromboembolism in patients with cancer

Ferrata Storti Foundation

Ella Grilz,1 Oliver Königsbrügge,1 Florian Posch,1,2 Manuela Schmidinger,3 Robert Pirker,3 Irene M. Lang,4 Ingrid Pabinger1 and Cihan Ay1

Clinical Division of Hematology and Hemostaseology, Department of Medicine I, Medical University of Vienna; 2Division of Oncology, Department of Internal Medicine, Medical University of Graz; 3Clinical Division of Oncology, Department of Medicine I, Medical University of Vienna and 4Clinical Division of Cardiology, Department of Medicine II, Medical University of Vienna, Austria 1

Haematologica 2018 Volume 103(9):1549-1556

ABSTRACT

n contrast to venous thromboembolism, little is known about arterial thromboembolism in patients with cancer. The aim of this study was to quantify the risk and explore clinical risk factors of arterial thromboembolism in patients with cancer, and investigate its potential impact on mortality. Patients with newly-diagnosed cancer or progression of disease after remission were included in a prospective observational cohort study and followed for two years. Between October 2003 and October 2013, 1880 patients (54.3% male; median age 61 years) were included. During a median follow up of 723 days, 48 (2.6%) patients developed arterial thromboembolism [20 (41.7%) myocardial infarction, 16 (33.3%) stroke and 12 (25.0%) peripheral arterial events], 157 (8.4%) developed venous thromboembolism, and 754 (40.1%) patients died. The cumulative 3-, 6-, 12-, and 24-month risks of arterial thromboembolism were 0.9%, 1.1%, 1.7%, and 2.6%, respectively. Male sex (subdistribution hazard ratio=2.9, 95%CI: 1.5-5.6; P=0.002), age (subdistribution hazard ratio per 10 year increase=1.5, 1.2-1.7; P<0.001), hypertension (3.1, 1.7-5.5; P<0.001), smoking (2.0, 1.1-3.7; P=0.022), lung cancer (2.3, 1.2-4.2; P=0.009), and kidney cancer (3.8, 1.4-10.5; P=0.012) were associated with a higher arterial thromboembolism risk. Furthermore, the occurrence of arterial thromboembolism was associated with a 3.2-fold increased risk of all-cause mortality (hazard ratio=3.2, 95%CI: 2.2-4.8; P<0.001). Arterial thromboembolism is a less common complication in patients with cancer than venous thromboembolism. The risk of arterial thromboembolism is high in patients with lung and kidney cancer. Patients with cancer who develop arterial thromboembolism are at a 3-fold increased risk of mortality.

Introduction Cancer is associated with a hypercoagulable state which leads to an increased risk of venous thromboembolism (VTE).1,2 The risk of VTE varies significantly among different cancer types and is associated with an increased risk of mortality.3,4 In contrast to VTE, much less is known on the epidemiology of arterial thromboembolism (ATE) in patients with cancer.5–7 It is important to identify an association between ATE and cancer because cardiovascular diseases and cancer are becoming more prevalent in an aging population and may share common risk factors and pathobiology related to inflammation.8,9 Furthermore, recent advances in screening, diagnosis and therapy have improved the survival of many cancers, implying that the population of patients with cancer now lives longer and develops a risk for cardiovascular complications.10,11 The risk of ATE in cancer may also be associated with anti-neoplastic treatments, known for the risk of arterial complications.12 As data on the epidemiology, risk and burden of ATE in patients with cancer are haematologica | 2018; 103(9)

Correspondence: cihan.ay@meduniwien.ac.at

Received: March 1, 2018. Accepted: May 17, 2018. Pre-published: May 24, 2018. doi:10.3324/haematol.2018.192419 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1549 ©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.

1549

E. Grilz et al.

scarce, we conducted a study to describe the incidence of ATE, explore clinical risk factors, and investigate the impact of ATE on mortality in patients with cancer.

Methods Study design This study was performed within the framework of the Vienna Cancer and Thrombosis Study (CATS), which started in 2003 at the Medical University of Vienna. CATS is a single-center prospective observational cohort study, approved by the ethics committee (number: 126/2003, ethik-kom@meduniwien.ac.at), and conducted in accordance with the Declaration of Helsinki. Detailed information about the study design and procedures have been reported previously.13 Briefly, adult patients (≥18 years) with a newly diagnosed malignancy or progression of disease after complete or partial remission were eligible for inclusion. Patients were not included if they had received radiotherapy or surgery within the last two weeks or chemotherapy within the last three months before study inclusion. Furthermore, all patients with a thromboembolic event within the last three months or an overt bacterial or viral infection within the last six weeks before study inclusion were excluded. Patients with an indication for long-term prophylactic or therapeutic anticoagulation were excluded, but temporary treatment with low molecular heparin (e.g. for hospitalized patients) was allowed. Furthermore, patients on acetylsalicylic acid or other platelet inhibitors were not excluded.13,14 All patients gave their written informed consent and were prospectively followed for a maximum duration of two years, until the occurrence of VTE, loss of follow up, withdrawal of consent, or death. Until October 2013, 2004 patients were included in this study. After re-evaluation of the inclusion and exclusion criteria, 124 patients had to be excluded, because: 1) they did not fulfill inclusion (n=35) or exclusion criteria (n=60); 2) no follow up was available (n=20); 3) no material for laboratory analyses was available (n=7); or 4) patients withdrew consent (n=2). Thus, overall 1880 patients with active cancer between 17th October 2003 and 28th October 2013 were included in this analysis.

Outcome measurement Venous thromboembolism is the pre-defined primary outcome of CATS. Data on ATE were collected as a comorbid condition during the observation time. To verify diagnosis of ATE, we compiled data from: 1) the CATS database; 2) patients' follow-up letters; 3) telephone records with patients, their family members, treating oncologists/hemato-oncologists and general practitioners during follow up; 4) digital information systems of the General Hospital of Vienna and the Vienna Urban Health Care Providers; and 5) the Austrian death registry. The primary end point of this analysis was objectively confirmed symptomatic ATE, which was defined as a composite of acute myocardial infarction (ST-elevation myocardial infarction and non-ST-elevation myocardial infarction), peripheral arterial occlusion, if treated with an interventional procedure (i.e. a catheter-based or open surgical procedure to improve arterial blood flow in non-cardiac arteries, except the intracranial vessels), and ischemic stroke. Both minor (National Institute of Health Stroke Score ≤ 3) and major (National Institute of Health Stroke Score > 3) stroke were included.15,16 A panel of experts in cardiology, neurology, and vascular medicine adjudicated all events based on objective evidence. Objective evidence for diagnosis included: 1) computed tomography (CT); magnetic resonance imaging (MRI) and autopsy report for ischemic stroke; 2) Doppler-sonography, digital subtraction angiography, CT-angiography, and MR1550

angiography for peripheral arterial occlusion; 3) electrocardiography, echocardiography (e.g. hypokinetic/akinetic and hypotrophic myocardial section without any other existing reason), cardiac biomarkers, identification of an intracoronary thrombus by angiography, and autopsy evidence for myocardial infarction. No routine screening for ATE was carried out during the study. Asymptomatic arterial thrombosis (e.g. incidentally detected stroke on restaging CT scans) was considered an event if it was considered clinically significant by members of the adjudication committee. In cases where objective diagnostic evidence based on imaging or lab results was missing, the adjudication committee decided on the basis of documented clinical evidence. The adjudication committee also classified mortality into fatal ATE and death-from-any-cause-other-than-fatal-ATE, which was the secondary end point.

Statistical analysis Continuous variables were summarized as medians (25th-75th percentile), and count data as absolute frequencies (%). The reverse Kaplan-Meier method was used to estimate median follow-up time.17 The cumulative incidence of ATE was calculated using a competing risk estimator with 95% confidence intervals (95%CI).18 ATE incidences between groups were compared with Gray’s tests.19 We assume that cancer patients are not only at risk of ATE, but are also at risk of dying from cancer. If patients die from their underlying malignancy, the risk of ATE occurrence is instantly reduced to zero. Therefore, death represents a competing risk scenario.20,21 To address this issue, univariable and multivariable Fine & Gray competing risk regression models were used to analyze the subdistribution hazards of ATE.22 Overall survival was analyzed with Kaplan-Meier estimators, and hazards of death were modeled with uni- and multivariable Cox models. VTE and deathfrom-any-cause-other-than-fatal-ATE were considered competing events in all statistical analyses concerning the primary end point.20 Relative risks of ATE between patients with breast cancer and other tumor types could not be estimated because the risk of ATE was 0.0% in breast cancer patients. Thus, modeling for this comparison was performed with a generalized linear model from the Bernoulli family with an identity link.23 To quantify the impact of ATE occurring during follow up on mortality, we used Cox models treating ATE as a time-dependent variable (controlling for immortal time bias), and performed a landmark analysis. In the time-dependent Cox model, one day of survival time was added in one patient who developed ATE and died on the same day. Stata 14.0 (Stata Corp., Houston, TX, USA) and SPSS 24 (SPSS Inc., Chicago IL, USA) were used to perform all statistical analyses.

Results Patients’ and follow-up characteristics The study cohort included 1880 patients with a wide range of different cancer types (Table 1). The majority of patients (n=1385, 73.7%) had newly diagnosed cancer, while 495 (26.3%) patients had a progressive disease after complete or partial remission. The median follow-up time was 723 days [25th-75th percentile (Q1-Q3): 308-731, range: 1-731]. During the observation time 157 (8.4%) VTE events were observed and 754 (40.1%) patients died.

Risk of ATE in patients with cancer Forty-eight (2.6%) patients developed ATE during the observation time. Among those patients, 19 (39.6%) nonhaematologica | 2018; 103(9)

Arterial thromboembolism in patients with cancer

fatal myocardial infarctions, 16 (33.3%) ischemic strokes, 12 (25.0%) peripheral arterial events, and one (2.1%) fatal myocardial infarction were observed. Detailed information on cancer patients with ATE is listed in Table 2. The cumulative 3-, 6-, 12-, and 24-month ATE risks were 0.9% (95%CI: 0.6-1.4), 1.1% (0.7-1.7), 1.7% (1.2-2.4), and 2.6% (2.0-3.4), respectively (Figure 1). While the rate of VTE was highest during the first six months of follow up and strongly declined thereafter, ATE events did not have a peak incidence, but rather occurred at a relatively constant rate (Figure 2).

Table 1. Baseline characteristics of the total study population.

N. of patients Median age at study entry, years First to third quartile Median body mass index, kg/m2 First to third quartile Sex Female Male Site of cancer Lung Breast Lymphoma Brain Colorectal Prostate Pancreas Stomach Multiple myeloma Kidney Others Progression of tumor Localized Distant metastasis Not classifiablea Smoking status Smoker Ex-smoker (>1-year non-smoker) Non-smoker Hypertension at study entry Diabetes at study entry Known atherosclerotic cardiovascular disease at study entry Dyslipidemia at study entry History of venous thromboembolism* Platelet aggregation inhibitor use at study entry Lipid lowering agent use at study entry

61 52-68 25.1 22.4-28.3 860 1020

45.7 54.3

319 276 265 248 186 157 133 65 50 45 136

17.0 14.7 14.1 13.2 9.9 8.4 7.1 3.5 2.7 2.4 7.2

629 627 506

35.7 35.6 28.7

546 313 867 704 223 159

31.6 18.1 50.2 37.5 11.9 8.5

222 91 280

11.9 4.8 14.9

238

12.7

Continuous data are reported as medians with first and third quartiles. Categorical variables are given as absolute frequencies and percentages. aData on body mass index, progression of tumor, and smoking status are missing in 6, 118, and 154 patients, respectively. *Defined as venous thromboembolism (VTE) that had occurred more than three months before study inclusion. N: number.

haematologica | 2018; 103(9)

Risk factors for ATE in patients with cancer In univariable competing risk regression analysis, male sex [subdistribution hazard ratio (SHR)=2.9, 95%CI: 1.55.6; P=0.002], higher age (SHR per 10 year increase=1.5, 1.2-1.7; P<0.001), hypertension (SHR=3.1, 1.7-5.5; P<0.001), diabetes (SHR=2.2, 1.2-4.4; P=0.020), a positive

Table 2. Baseline characteristics of patients with ATE.

N. of patients Median age at study entry, years First to third quartile Median body mass index, kg/m2 First to third quartile Sex Female Male Type of arterial thromboembolic event Myocardial infarction Major stroke Minor stroke Peripheral arterial occlusion Site of cancer Lung Breast Lymphoma Brain Colorectal Prostate Pancreas Stomach Multiple myeloma Kidney Others Progression of tumor Localized Distant metastasis Not classifiablea Smoking status Smoker Ex-smoker (>1-year non-smoker) Non-smoker Hypertension at study entry Diabetes at study entry Known atherosclerotic cardiovascular disease at study entry Dyslipidemia at study entry History of venous thromboembolism* Platelet aggregation inhibitor use at study entry Lipid lowering agent use at study entry

66 60-69 26.3 23.2-29.0 11 37

22.9 77.1

20 13 3 12

41.7 27.1 6.3 25.0

15 0 5 7 3 8 2 2 0 4 2

45.8 0.0 10.4 14.6 6.3 16.7 4.2 4.2 0.0 8.4 4.2

19 14 10

44.2 32.6 23.3

19 12 16 31 11 12

40.4 25.5 34.0 64.6 22.9 25.0

8 2 22

16.7 4.2 45.8

29.2

Continuous data are reported as medians with first and third quartiles. Categorical variables are given as absolute frequencies and percentages. aData on progression of tumor are missing in 5 patients. Data on smoking status are missing in one patient. *Defined as venous thromboembolism (VTE) that had occurred more than three months before study inclusion. N: number.

1551

E. Grilz et al.

smoking history (SHR=2.0, 1.1-3.7; P=0.022), and a known arterial cardiovascular disease (e.g. history of stroke, peripheral arterial disease, coronary heart disease) at study entry (SHR=3.7, 1.9-7.2; P<0.001) were associated with a higher ATE risk. The body mass index (SHR=1.0, 1.0-1.1; P=0.115), and a prior history of VTE (SHR= 0.8, 0.2-3.5; P=0.814) were not associated with risk of ATE. Furthermore, dyslipidemia, which was defined as having at least one of the following diagnoses: 1) hypertriglyceridemia; 2) hyperlipidemia; or 3) hypercholesterolemia, was also not associated with the risk of ATE in patients with cancer (SHR=1.5, 0.7-3.2; P=0.302). Treatment with lipid lowering agents (SHR=2.9, 1.5-5.3; P<0.001) or platelet aggregation inhibitors (SHR=5.0, 2.8-8.8; P<0.001) at study entry was associated with ATE occurrence. Cancer stage was not associated with an increased risk of ATE occurrence (SHR=0.7, 0.4-1.5; P=0.398). Lung cancer (SHR=2.3, 1.24.2; P=0.009), and kidney cancer (SHR=3.8, 1.4-10.5; P=0.012) were associated with an increased risk of ATE. Cumulative incidences of ATE separated by cancer type are shown in Figure 3. In multivariable competing risk regression analyses, age [adjusted (adj.) SHR per 10 year increase=1.4, 95%CI: 1.2-1.7; P<0.001], male sex (adj. SHR=2.6, 1.3-5.2; P=0.006), and smoking (adj. SHR=2.1, 1.1-3.9; P=0.026) emerged as independently associated with the risk of ATE in patients with cancer when corrected for each other. After correction for age, male sex, and smoking, the association between hypertension (adj. SHR=2.4, 1.3-4.5; P=0.005), a known arterial cardiovascular disease (adj. SHR=2.6, 1.3-5.3; P=0.007), treatment with lipid lowering agents (adj. SHR=2.2, 1.2-4.4; P=0.013), and the use of platelet aggregation inhibitors (adj. SHR=3.7, 2.1-6.7; P<0.001) at study entry and ATE remained statistically significant. Diabetes did not reach statistical significance in this multivariable analysis (adj. SHR=1.8, 0.9-3.5; P=0.093). Furthermore, the association between ATE and kidney cancer prevailed (adj. SHR=3.7, 1.3-10.6; P=0.016), whereas it did not for lung cancer (adj.

SHR=1.6, 0.8-3.4; P=0.193). None of the 276 patients with breast cancer developed ATE during follow up (crude risk=0.0%). In contrast, 48 ATE events occurred in the 1604 patients with other tumor entities (crude risk=3.0%), for an absolute risk difference of 3.0% (95%CI: 1.0-5.0; P=0.004). However, breast cancer patients were significantly younger than patients with other tumors (median age 59 vs. 62 years; P<0.001). Nonetheless, breast cancer remained associated with a lower risk of ATE even after adjusting for age (adjusted absolute risk difference=2.8%, 95%CI: 0.7-4; P=0.007), and also after exclusion of male patients and adjustment for age (adjusted absolute risk difference=1.7%, 95%CI: 0.1-3.3; P=0.035). We also investigated the cumulative impact of cardiovascular risk factors on ATE risk in patients with cancer (Figure 4). We assigned one point for each of the following cardiovascular risk factors: hypertension, diabetes, known arterial cardiovascular disease, and dyslipidemia. Assuming a linear relationship between the number of risk factors and ATE risk, the SHR was 1.8 (1.4-2.2; P<0.001) per point increase. This association prevailed after adjustment for age and sex (SHR=1.6, 1.2-2.0; P=0.001). The 2-year cumulative incidence of ATE in patients with 0, 1, 2, 3, and 4 points was 1.4% (0.8-2.3), 2.7% (1.5-4.3), 5.8% (3.3-9.3), 5.6% (1.8-12.7), and 12.5% (2.1-32.8), respectively.

Association between ATE and survival of patients with cancer The 3-, 6-, 12-, and 24-month overall survival estimates of the study cohort were 94.0% (95%CI: 92.8-95.0), 87.5% (85.9-88.9), 74.8% (72.7-76.7), and 57.9% (55.560.2), respectively. In multistate modeling the occurrence of ATE was associated with a 3.2-fold relative increase in the risk of death from any cause [hazard ratio (HR)=3.2, 95%CI: 2.2-4.8; P<0.001). This association prevailed after adjusting for age (adj. HR=2.9, 2.0-4; P<0.001), as well as age and lung cancer as an indicator for a cancer type with poor prognosis (adj. HR=2.5, 1.7-3.7; P<0.001). In a land-

Figure 1. Cumulative incidence of arterial thromboembolism (ATE) in patients with cancer. For estimation we used a competing risk cumulative incidence estimator. Venous thromboembolism (VTE) and death-from-any-cause were considered as competing risk events. The dashed line represents 95% confidence bands.

1552

haematologica | 2018; 103(9)

Arterial thromboembolism in patients with cancer

mark analysis with the landmark set at three months after baseline, 2-year predicted overall survival was 62.3% in patients who did not develop ATE during the first three months of follow up, and 24.8% in patients who did develop ATE during the first three months (Mantel-Byar P<0.001) (Figure 5). The median survival time of patients with cancer after ATE was only 63 days (Q1-Q3: 36-233).

Discussion We analyzed the cumulative incidence of ATE, identified clinical risk factors, and evaluated the impact of ATE on mortality in patients with cancer in this prospective observational cohort study. During a follow-up period of up to two years, 2.6% of patients developed ATE, defined as the composite of myocardial infarction, stroke and peripheral artery disease, which was less common than VTE. The risk of ATE was increased in patients with higher age, male sex, hypertension, and a positive smoking history. In addition, the risk of ATE varied by cancer type with lung and kidney cancer having the highest, and breast cancer having the lowest risk. We also observed that the occurrence of ATE is associated with a 3-fold increased risk of mortality. The negative impact of ATE on cancer patientsâ&#x20AC;&#x2122; prognosis indicates an unmet need for better understanding of the burden of ATE in cancer, the identification of patients at risk of ATE, and improved strategies to prevent, treat and manage ATE in patients with cancer. An increased risk of stroke and coronary heart disease in patients with cancer has been reported in previous studies.24-30 In a recent retrospective analysis of a Surveillance, Epidemiology, and End Results (SEER) and Medicare linked dataset of patients with a primary cancer diagnosis between 2002 and 2011 in the USA, the 6month cumulative incidence of myocardial infarction and ischemic stroke was 4.7% compared to 2.2% in a

Figure 2. Time-dependent rates of arterial thromboembolism (ATE) and venous thromboembolism (VTE) over two years of follow up. While the rate of VTE was highest during the first six months of follow up and then declined, the risk of ATE did not have a peak and remained relatively constant during follow up. The curves were predicted with a flexible parametric regression model on the logcumulative-hazard scale (Stata routine stpm2).

haematologica | 2018; 103(9)

matched control group without cancer.31 The cumulative incidence of ATE in this study was higher than in our study (4.7% vs. 2.6%). However, increasing age is an important cardiovascular risk factor and patients in this retrospective study were older than in our study, where outcome data were prospectively collected (median age 74 vs. 61 years). It is likely that an increased rate of ATE may be attributable to some degree to differences in age and also to the presence of classical cardiovascular risk factors in the studied populations. In general, estimating the true risk of ATE may be challenging in cancer patients. Nonspecific ATE symptoms such as dyspnea and chest pain are highly prevalent in cancer patients and oncologists may attribute such symptoms often to the underlying cancer, which may lead to an under-diagnosis of ATE. On the other hand, cardiologists may be reluctant to schedule cancer patients for diagnostic coronary angiographies when cancer patients have a poor performance status or are deemed to have a poor prognosis, leading to a further underestimation of the ATE burden in cancer patients. Thus, oncologists and cardiologists should maintain a relatively high index of suspicion for ATE when dealing with cancer patients with ATE-related symptoms. The relative risk of ATE according to the cancer type in our study was similar to that of previous studies. Patients with lung and kidney cancer had the highest relative risk of ATE.29-31 The difference in ATE risk between cancer types may be driven by general risk factors that increase risk of both cancer and ATE, such as smoking, or by specific anti-cancer treatments, such as platinum-based combination chemotherapy, vascular endothelial growth factor (VEGF) / vascular endothelial growth factor receptor (VEGFR) inhibitors, which have been associated with increased risk of ATE.12,32-37 Prior studies also suggested an association between metastasis and the risk of ATE occurrence, which was not confirmed in our study.29-31 The risk of VTE in patients with cancer varies during the course of the cancer disease. It is highest during the first

Figure 3. Cumulative incidence of arterial thromboembolism (ATE) according to cancer type. A competing risk cumulative incidence estimator was used. For clarity, only selected tumor entities are compared. A Grayâ&#x20AC;&#x2122;s test was used to test for differences between cancer types (P<0.001), venous thromboembolism and death-from-any-cause were considered as competing risk event.

1553

E. Grilz et al. six months after diagnosis of cancer and then declines.38 Using flexible parametric modeling, we could demonstrate that in contrast to VTE, the rate of ATE does not appear to have a peak but remained relatively constant over the whole follow-up period. Hence, long-term strategies to effectively prevent ATE in cancer patients, and especially in cancer survivors, are needed. Further research has to be conducted to investigate the optimal management strategies. In this regard, a phase I trial is currently underway evaluating aspirin and statin in patients with cancer to prevent thrombosis (clinicaltrials.gov identifier: 02285738).39 We can speculate whether a primary antithrombotic prevention of ATE would have the greatest net-clinical-benefit in tumor entities with a high absolute ATE risk, such as in patients with lung and renal

cancer, or in patients with individual risk factors for ATE. The link between cancer and ATE could be explained in part by common risk factors, such as higher age and smoking. However, it is unlikely that the ATE burden in cancer patients is completely attributable to these and other general risk factors. In our study, only hypertension and a known arteriosclerotic cardiovascular disease were independently associated with the risk of ATE. We also examined the association of other cardiovascular risk factors and the occurrence of ATE in patients with cancer. Neither the body mass index nor diabetes, dyslipidemia, or prior VTE at study entry were independently associated with the risk of ATE. However, these analyses might be under-powered due to the low absolute number of patients with ATE. In contrast, the use of lipid lowering

Figure 4. Cumulative impact of cardiovascular risk factors on the risk of arterial thromboembolism (ATE). Competing risk analysis was used to analyze the incidence of ATE, considering venous thromboembolism (VTE) and death-from-anycause as competing risk events. SHR: subdistribution hazard ratio;CVRF: cardiovascular risk factors.

Figure 5. Landmark analysis of predicted overall survival according to arterial thromboembolism (ATE) status after three months of follow up. 1745 patients survived for at least three months (“landmark time”), of whom 13 had developed ATE within the first three months and had survived until the landmark time. Four patients had developed ATE within the first three months and died before the landmark time; 31 patients had developed ATE after the landmark time. “Predicted” instead of “observed” overall survival was chosen due to the small numbers of patients who had developed ATE within the first three months (n=13).

1554

haematologica | 2018; 103(9)

Arterial thromboembolism in patients with cancer

agents and platelet aggregation inhibitors were associated with ATE risk. However, previous studies show that a high pill burden, co-morbidities, and provision of care by multiple physicians compromise medication adherence;40,41 we assume that all of these are likely to be of relevance in patients with cancer. Therefore, we cannot exclude a bias in this analysis due to a possible lack of medication adherence or persistence. Another possible explanation for the association of use of lipid lowering drugs and platelet aggregation inhibitors could be that physicians had correctly identified those patients with a high cardiovascular risk, and consequently had prescribed this medication. Also cancer-specific risk factors such as anti-cancer treatments (e.g. radiotherapy, platinum-containing chemotherapy, treatment with monoclonal antibodies, tyrosine kinase inhibitors) are known to increase the risk of ATE.24,34-36,42 The major limitation of our study is that we were not able to specifically model the time-dependent contribution of selected cancer therapies, such as platinum and VEGF-targeted agents, to risk of ATE, because not all information on treatment regimens was available. Furthermore, patients receiving therapeutic or prophylactic anticoagulation were excluded from the study, which might mean that this study population is less representative. Another main finding of our study was that the occurrence of ATE during follow up is associated with an increased risk of mortality. Patients who developed ATE had a 3-fold higher risk of mortality. This is consistent with a recent study that reported a 4-fold increased risk of mortality in patients with cancer and ATE.31 Collectively, these results demonstrate that incidental ATE is a major contributor to death in patients with cancer, and that special medical attention is needed for patients with cancer and arterial thromboembolic complications to improve their prognosis. With regard to fatal ATE, we have to mention that the rate of fatal ATE in our study cohort is lower

References 1. Ay C, Pabinger I. Predictive potential of haemostatic biomarkers for venous thromboembolism in cancer patients. Thromb Res. 2012;129 Suppl 1:S6-9. 2. Riedl J, Pabinger I, Ay C. Platelets in cancer and thrombosis. Hamostaseologie. 2014;34(1):54-62. 3. Pabinger I, Thaler J, Ay C. Biomarkers for prediction of venous thromboembolism in cancer. Blood. 2013;122(12):2011-2018. 4. Königsbrügge O, Pabinger I, Ay C. Risk factors for venous thromboembolism in cancer: novel findings from the Vienna Cancer and Thrombosis Study (CATS). Thromb Res. 2014;133 Suppl 2:S39-43. 5. Di Nisio M, Ferrante N, Feragalli B, et al. Arterial thrombosis in ambulatory cancer patients treated with chemotherapy. Thromb Res. 2011;127(4):382-383. 6. El Sakka K, Gambhir RPS, Halawa M, Chong P, Rashid H. Association of malignant disease with critical leg ischaemia. Br J Surg. 2005;92(12):1498-1501. 7. Javid M, Magee TR, Galland RB. Arterial

haematologica | 2018; 103(9)

9. 10.

11.

12.

13.

than in non-cancer patients.43,44 Therefore, we cannot exclude a possible reporting bias because no routine autopsies were performed in patients included in CATS and it is likely that physicians generally tend to attribute death to the underlying malignancy. Although previous studies had a larger sample size, these data are taken from retrospective analyses, selected cancer entities or from post-hoc analyses of studies that were not designed to investigate ATE in cancer.45-48 To our knowledge, this is the first study in which data on ATE in cancer patients were collected prospectively. To minimize any potential bias, we retrospectively evaluated different sources for ascertainment of the ATE outcome data and a committee adjudicated all events. In conclusion, this study used subject-level data of a well-characterized cancer cohort to define the epidemiology of ATE in patients with malignant diseases. ATE in patients with cancer is a less frequent complication than VTE. In certain cancer types, such as lung and kidney cancer, the risk of ATE is still as high as the risk of VTE. Furthermore, occurrence of ATE in patients with cancer is associated with an increased risk of mortality, calling for future clinical research efforts to better characterize patients with cancer at risk of ATE and reduce the burden of arterial thromboembolic complications. Acknowledgments We thank the members of the adjudication committee: Thomas Gremmel (Clinical Division of Angiology, Department of Medicine II, Medical University of Vienna) and Fritz Leutmezer (Department of Neurology, Medical University of Vienna). Funding The study was funded by the Anniversary Fund of the Austrian National Bank (“Jubiläumsfonds der Österreichischen Nationalbank (OeNB)” project numbers #10935, #12739 and #14744) and the Austrian Science Fund (FWF), Special Research Program (SFB) 54.

Thrombosis Associated with Malignant Disease. Eur J Vasc Endovasc Surg. 2008; 35(1):84-87. Prisco D, D’Elios MM, Cenci C, Ciucciarelli L, Tamburini C. Cardiovascular oncology: a new discipline inside internal medicine? Intern Emerg Med. 2014;9(4):359-364. Abe J, Martin JF, Yeh ETH. The Future of Onco-Cardiology. Circ Res. 2016; 119(8):896-899. Vasu S, Hundley WG. Understanding cardiovascular injury after treatment for cancer: an overview of current uses and future directions of cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2013; 15:66. DeSantis CE, Lin CC, Mariotto AB, et al. Cancer treatment and survivorship statistics, 2014. CA Cancer J Clin. 2014; 64(4):252-271. Choueiri TK, Schutz FAB, Je Y, Rosenberg JE, Bellmunt J. Risk of Arterial Thromboembolic Events With Sunitinib and Sorafenib: A Systematic Review and Meta-Analysis of Clinical Trials. J Clin Oncol. 2010;28(13):2280-2285. Ay C, Simanek R, Vormittag R, et al. High plasma levels of soluble P-selectin are pre-

14.

15. 16. 17. 18. 19.

20.

dictive of venous thromboembolism in cancer patients: results from the Vienna Cancer and Thrombosis Study (CATS). Blood. 2008;112(7):2703-2708. Ay C, Vormittag R, Dunkler D, et al. Ddimer and prothrombin fragment 1 + 2 predict venous thromboembolism in patients with cancer: results from the Vienna Cancer and Thrombosis Study. J Clin Oncol. 2009;27(25):4124-4129. Thygesen K, Alpert JS, Jaffe AS, et al. Third universal definition of myocardial infarction. Eur Heart J. 2012;33(20):2551-2567. Fischer U, Baumgartner A, Arnold M, et al. What Is a Minor Stroke? Stroke. 2010;41(4):661-666. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials. 1996;17(4):343-346. Coviello V, Boggess M. Cumulative incidence estimation in the presence of competing risks. Stata J. 2004;4(2):103-112. Gray RJ. A Class of K-Sample Tests for Comparing the Cumulative Incidence of a Competing Risk. Ann Stat. 1988; 16(3):1141-1154. Ay C, Posch F, Kaider A, Zielinski C, Pabinger I. Estimating risk of venous

1555

E. Grilz et al.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

1556

thromboembolism in patients with cancer in the presence of competing mortality. J Thromb Haemost. 2015;13(3):390-397. Delgado J, Pereira A, Villamor N, LopezGuillermo A, Rozman C. Survival analysis in hematologic malignancies: recommendations for clinicians. Haematologica. 2014;99(9):1410-1420. Fine JP, Gray RJ. A Proportional Hazards Model for the Subdistribution of a Competing Risk. J Am Stat Assoc. 1999; 94(446):496-509. Terbuch A, Posch F, Annerer LM, et al. Long-term cardiovascular complications in stage I seminoma patients. Clin Transl Oncol. 2017;19(11):1400-1408. McGale P, Darby SC, Hall P, et al. Incidence of heart disease in 35,000 women treated with radiotherapy for breast cancer in Denmark and Sweden. Radiother Oncol. 2011;100(2):167-175. Chen PC, Muo CH, Lee YT, Yu YH, Sung FC. Lung cancer and incidence of stroke: a population-based cohort study. Stroke. 2011;42(11):3034-3039. Kuan AS, Teng CJ, Wu HH, et al. Risk of ischemic stroke in patients with ovarian cancer: a nationwide population-based study. BMC Med. 2014;12(1):53. Kuan AS, Chen SC, Yeh CM, et al. Risk of Ischemic Stroke in Patients With Gastric Cancer. Medicine (Baltimore). 2015; 94(37):e1336. Navi BB, Reiner AS, Kamel H, et al. Association between incident cancer and subsequent stroke. Ann Neurol 2015; 77(2):291-300. Zöller B, Ji J, Sundquist J, Sundquist K. Risk of coronary heart disease in patients with cancer: A nationwide follow-up study from Sweden. Eur J Cancer. 2012;48(1):121-128. Zöller B, Ji J, Sundquist J, Sundquist K. Risk of haemorrhagic and ischaemic stroke in patients with cancer: A nationwide followup study from Sweden. Eur J Cancer. 2012;48(12):1875-1883.

31. Navi BB, Reiner AS, Kamel H, et al. Risk of Arterial Thromboembolism in Patients With Cancer. J Am Coll Cardiol. 2017;70(8):926-938. 32. Ranpura V, Hapani S, Chuang J, Wu S. Risk of cardiac ischemia and arterial thromboembolic events with the angiogenesis inhibitor bevacizumab in cancer patients: A meta-analysis of randomized controlled trials. Acta Oncol. 2010;49(3):287-297. 33. Pantaleo MA, Mandrioli A, Saponara M, et al. Development of coronary artery stenosis in a patient with metastatic renal cell carcinoma treated with sorafenib. BMC Cancer. 2012;12(1):231. 34. Sonpavde G, Bellmunt J, Schutz F, Choueiri TK. The Double Edged Sword of Bleeding and Clotting from VEGF Inhibition in Renal Cancer Patients. Curr Oncol Rep. 2012; 14(4):295-306. 35. Qi W-X, Shen Z, Tang L-N, Yao Y. Risk of arterial thromboembolic events with vascular endothelial growth factor receptor tyrosine kinase inhibitors: An up-to-date meta-analysis. Crit Rev Oncol Hematol. 2014;92(2):71-82. 36. Proverbs-Singh T, Chiu SK, Liu Z, et al. Arterial thromboembolism in cancer patients treated with cisplatin: a systematic review and meta-analysis. J Natl Cancer Inst. 2012;104(23):1837-1840. 37. Schmidinger M, Zielinski CC, Vogl UM, et al. Cardiac toxicity of sunitinib and sorafenib in patients with metastatic renal cell carcinoma. J Clin Oncol. 2008; 26(32):5204-5212. 38. Khorana AA, Carrier M, Garcia DA, Lee AYY. Guidance for the prevention and treatment of cancer-associated venous thromboembolism. J Thromb Thrombolysis. 2016;41(1):81-91. 39. Anti-Platelet and Statin Therapy to Prevent Cancer-Associated Thrombosis. https://clinicaltrials.gov/ct2/show/NCT022 85738 [Last accessed 9 Oct 2017]. 40. Lindhardsen J, Ahlehoff O, Gislason GH, et

41. 42.

43.

44.

45.

46.

47.

48.

al. Initiation and adherence to secondary prevention pharmacotherapy after myocardial infarction in patients with rheumatoid arthritis: a nationwide cohort study. Ann Rheum Dis. 2012;71(9):1496-1501. Brown MT, Bussell JK. Medication Adherence: WHO Cares? Mayo Clin Proc. 2011;86(4):304-314. Scappaticci FA, Skillings JR, Holden SN, et al. Arterial thromboembolic events in patients with metastatic carcinoma treated with chemotherapy and bevacizumab. J Natl Cancer Inst. 2007;99(16):1232-1239. Roe MT, Messenger JC, Weintraub WS, et al. Treatments, Trends, and Outcomes of Acute Myocardial Infarction and Percutaneous Coronary Intervention. J Am Coll Cardiol. 2010;56(4):254-263. Grysiewicz RA, Thomas K, Pandey DK. Epidemiology of ischemic and hemorrhagic stroke: incidence, prevalence, mortality, and risk factors. Neurol Clin. 2008;26(4):871-895, vii. Mellema WW, van der Hoek D, Postmus PE, Smit EF. Retrospective evaluation of thromboembolic events in patients with non-small cell lung cancer treated with platinum-based chemotherapy. Lung Cancer. 2014;86(1):73-77. Reiner AS, Navi BB, DeAngelis LM, Panageas KS. Increased risk of arterial thromboembolism in older men with breast cancer. Breast Cancer Res Treat. 2017;166(3):903-910. Haguet H, Douxfils J, Mullier F, Chatelain C, Graux C, Dogné JM. Risk of arterial and venous occlusive events in chronic myeloid leukemia patients treated with new generation BCR-ABL tyrosine kinase inhibitors: a systematic review and meta-analysis. Expert Opin Drug Saf. 2017;16(1):5-12. Hultcrantz M, Pfeiffer RM, Björkholm M, et al. Elevated risk of venous but not arterial thrombosis in Waldenström macroglobulinemia/lymphoplasmacytic lymphoma. J Thromb Haemost. 2014;12(11):1816-1821.

haematologica | 2018; 103(9)

ARTICLE

Platelet Biology & its Disorders

Impaired mitochondrial activity explains platelet dysfunction in thrombocytopenic cancer patients undergoing chemotherapy

Ferrata Storti Foundation

Constance C. F. M. J. Baaten,1 Floor C. J. I. Moenen,2* Yvonne M. C. Henskens,3* Frauke Swieringa,1,4 Rick J. H. Wetzels,3 René van Oerle,3,5 Harry F. G. Heijnen,6 Hugo ten Cate,1,5 Graham P. Holloway,7 Erik A. M. Beckers,2 Johan W. M. Heemskerk1 and Paola E. J. van der Meijden1

Department of Biochemistry, Cardiovascular Research Institute Maastricht, Maastricht University Medical Centre, the Netherlands; 2Department of Hematology, Maastricht University Medical Centre, the Netherlands; 3Central Diagnostic Laboratory, Maastricht University Medical Centre, the Netherlands; 4Department of Protein Dynamics, Leibniz Institute for Analytical Sciences - ISAS-e.V., Dortmund, Germany; 5Laboratory for Clinical Thrombosis and Hemostasis, Department of Internal Medicine, Cardiovascular Research Institute Maastricht, Maastricht University Medical Center, the Netherlands; 6 Department of Cell Biology and Department of Clinical Chemistry and Haematology, University Medical Center Utrecht, the Netherlands and 7Department of Human Health and Nutritional Sciences, University of Guelph, Ontario, Canada 1

Haematologica 2018 Volume 103(9):1557-1567

*FCJIM and YMCH contributed equally to this work.

ABSTRACT

evere thrombocytopenia (≤50x109 platelets/L) due to hematological malignancy and intensive chemotherapy is associated with an increased risk of clinically significant bleeding. Since the bleeding risk is not linked to the platelet count only, other hemostatic factors must be involved. We studied platelet function in 77 patients with acute leukemia, multiple myeloma or malignant lymphoma, who experienced chemotherapy-induced thrombocytopenia. Platelets from all patients independent of disease or treatment type - were to a variable extent compromised in Ca2+ flux, integrin aIIbβ3 activation and P-selectin expression when stimulated with a panel of agonists. The patients' platelets were also impaired in spreading on fibrinogen. Whereas the Ca2+ store content was unaffected, the patients’ platelets showed ongoing phosphatidylserine exposure, which was not due to apoptotic caspase activity. Interestingly, mitochondrial function was markedly reduced in platelets from a representative subset of patients, as evidenced by a low mitochondrial membrane potential (P<0.001) and low oxygen consumption (P<0.05), while the mitochondrial content was normal. Moreover, the mitochondrial impairments coincided with elevated levels of reactive oxygen species (Spearman’s rho=-0.459, P=0.012). Markedly, the impairment of platelet function only appeared after two days of chemotherapy, suggesting origination in the megakaryocytes. In patients with bone marrow recovery, platelet function improved. In conclusion, our findings disclose defective receptor signaling related to impaired mitochondrial bioenergetics, independent of apoptosis, in platelets from cancer patients treated with chemotherapy, explaining the low hemostatic potential of these patients. Introduction Platelets are indispensable for maintaining vascular integrity and accomplishing hemostatic plug formation.1 A sufficient platelet count as well as an adequate platelet function is required for prevention of bleeding. Patients with hematological malignancies, such as leukemia, multiple myeloma or malignant lymphoma, are commonly treated with combination chemotherapy, frequently followed by bone marrow transplantation. This treatment impairs the proliferation of megakaryocytes and the production of proplatelets. As a consequence, severe thrombocyhaematologica | 2018; 103(9)

Correspondence: p.vandermeijden@maastrichtuniversity.nl

Received: November 22, 2017. Accepted: June 5, 2018. Pre-published: June 7, 2018. doi:10.3324/haematol.2017.185165 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1557 ©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.

1557

C.C.F.M.J. Baaten et al. topenia, i.e., a platelet count of ≤50x109/L, develops in virtually all treated patients.2 These patients are at high risk of bleeding, with up to 43% experiencing clinically significant bleeding (World Health Organization [WHO] grade 2 or higher), and 1% experiencing life-threatening bleeding.3 Prophylactic transfusion with platelet concentrates for preventing bleeding is given as standard care once the count drops below 10x109/L, or in case of active bleeding.2,4 Randomized clinical trials have indicated that the bleeding risk in this patient group is reduced by platelet transfusion, although it does not completely eliminate hemorrhagic events.3,5 Since bleeding is relatively infrequent in non-malignant thrombocytopenia,6,7 it can be considered that a low platelet count is not the sole risk factor for bleeding in chemotherapy-treated patients. Earlier studies on patients with acute myeloid leukemia, of whom none received chemotherapy, have provided indications for impaired platelet function due to disease, as apparent from low platelet aggregation, reduced granule secretion and weak thromboxane B2 production.8-10 It was proposed that low expression of the a-granule glycoprotein, P-selectin, can be used as a prognostic marker for hemorrhage.11 However, bleeding in combination with thrombocytopenia is more frequently observed in cancer patients treated with chemotherapy.12 The literature thus far only indicates that the anthracycline daunorubicin inhibits integrin aIIbβ3 activation, aggregation and secretion of platelets upon agonist stimulation.13,14 Daunorubicin and its analogue idarubicin were found to induce integrin activation and secretion in resting platelets.15 However, to what extent and by which mechanism myelosuppressive chemotherapy in general affects platelet function has remained largely unclear. In this study, we evaluated the platelet activation processes and coagulant activity in 77 patients with hematological malignancies treated with chemotherapy. Our results point to multiple functional defects in the patients' platelets which are related to impaired mitochondrial activity, independent of classical apoptosis. In the majority of patients, low platelet activity could be improved by platelet transfusion.

Methods Materials and methods See Online Supplementary Material.

platelet count: 357x109/L). Patient exclusion criteria were: sepsis, splenomegaly, signs of active bleeding at the time of blood withdrawal, previous platelet transfusion within three days (excluding the presence of donor platelets), and/or use of antithrombotic medication during the previous 14 days. For clinical care, blood samples were collected before and during chemotherapeutic treatment at multiple time points: 1) before the start of chemotherapy, 2) before myelosuppression, 3) during myelosuppression (platelet count ≤50x109/L), 4) during myelosuppression: before (platelet count ≤10x109/L) and one hour after platelet transfusion, and 5) during bone marrow recovery (platelet count ≤50x109/L). Patient blood samples were obtained via a central venous catheter, rinsed with 100 mL saline to remove residual traces of heparin (verified by measurement of thrombin time). Blood samples from healthy control subjects were obtained via venipuncture of the antecubital vein using a Vacutainer 21-gauge needle (Becton-Dickinson Bioscience, NJ, USA). Blood collection was always into 3.2% (w/v) trisodium citrate (Greiner Bio-One Vacuette, Alphen a/d Rijn, The Netherlands). For clinical care (hematological parameters), separate samples from patients were drawn into vacuette tubes containing K2-ethylenediaminetetraacetic acid (EDTA; Becton-Dickinson Bioscience, NJ, USA).

Experimental setup Within the limitations of medical ethical permission, a total of 52 blood samples from patients (platelet count ≤50x109/L) could be obtained during myelosuppression (study A). In all these samples, platelet responsiveness was assessed using flow cytometry. Due to the limited blood volume and the low platelet counts, a restricted number of additional analyses was carried out per sample. When there remained sufficient sample volume, platelet function was further characterized by measuring the following platelet responses: platelet spreading, intracellular calcium signaling and phosphatidylserine (PS) exposure. To gain a deeper understanding of the underlying mechanisms of platelet dysfunction, subsequent blood samples could be obtained from 25 additional patients (platelet count ≤50x109/L) during the myelosuppression phase (study B). The samples were used to investigate apoptotic signaling (caspase activity; western blotting for caspase-mediated protein cleavage), mitochondrial respiration and structure (high-resolution respirometry, citrate synthase activity, transmission electron microscopy) or reactive oxygen species (ROS). The maximum of care was taken that for all measurements patients from the major treatment classes were represented (see Figure Legends). For 36 of the patients in study A, blood samples could also be obtained at one hour after transfusion with platelet concentrate. Again, platelet responsiveness was determined by flow cytometry.

Patients and control subjects

Statistical analysis

The study was approved by the local ethics committee (METC11-4-097). All participating patients and healthy volunteers gave written informed consent according to the Helsinki declaration. Patients, reporting at the hospital, fulfilling the inclusion criteria and providing informed consent, were consecutively included in the period between November 2014 and April 2018. Eligible patients were ≥18 years of age, received chemotherapy for treatment of a confirmed hematologic malignancy (acute myeloid leukemia, acute lymphocytic leukemia, multiple myeloma or malignant lymphoma), and had, or were expected to have, thrombocytopenia (platelet count ≤50x109/L). Morning platelet counts were monitored daily as part of routine clinical care. According to standard practice, when the morning platelet count was <10x109/L, patients received prophylactic transfusion with one batch of platelet concentrate (leukocyte-depleted pooled buffy coat from five donors, median storage time: six days, median

Data are represented as medians with interquartile ranges. Paired data were compared using the Wilcoxon signed-rank test, otherwise the Mann-Whitney U test was used. When comparing more than two groups, the Kruskal Wallis H test was used. P-values <0.05 were considered significant. Graphs were made using GraphPad Prism v6 (San Diego, CA, USA). Statistical analysis was performed using the SPSS Statistics 23 package (IBM, Armonk, NY, USA).

1558

Results Variable impairment of platelet activation in cancer patients with thrombocytopenia after chemotherapy Blood samples were obtained from a total of 77 patients, who were diagnosed with acute myeloid leukemia or haematologica | 2018; 103(9)

Platelet dysfunction in chemotherapy

acute lymphocytic leukemia (AML/ALL, n=37), multiple myeloma (n=21), malignant lymphoma (n=15) or other hematologic malignancies (n=4). All patients experienced severe thrombocytopenia due to chemotherapy, which was stopped at a median of eight days before blood sample analysis (Table 1). The median age of the patient group was 60 years, and 41% was female (Table 1). Leukocyte and platelet counts were below normal, as was the hemoglobin level. Standard coagulation parameters were determined in plasmas from 43 patients following their chemotherapy treatment. For 70% of the patients, values of activated partial thromboplastin time (aPTT), prothrombin time and thrombin time were within reference ranges (Online Supplementary Table S1). Fibrinogen and von Willebrand factor (VWF) levels were slightly elevated, while D-dimer levels were substantially increased in patient plasmas. On the other hand, factor VII activity levels were decreased. Treatment regimens in accordance with national guidelines varied with disease type.16-18 Since these regimens consisted of multiple chemotherapeutic compounds, the distribution of the drugs was evaluated among patients with different diagnoses. Therefore, the various drugs were assigned to one of five pharmacological classes: A, antitumor antibiotics & topoisomerase II inhibitors; B, antimetabolites; C, alkylating agents; D, mitotic inhibitors; E, other (Online Supplementary Table S2).19 Most patients were treated with anti-tumor antibiotics/topo-isomerase inhibitors, antimetabolites and/or alkylating agents (Online Suplementary Table S3). The patients diagnosed with AML/ALL and lymphoma usually received drugs from one or more of these three classes, while the patients diagnosed with multiple myeloma only received alkylating agents. Of all 77 patients, 50 had undergone hematopoietic stem cell transplantation before inclusion, of which 39 patients received an autologous transplant and 11 an allogenic transplant (Table 1). Blood samples were obtained at eight days (median) after the last administration of chemotherapy or at eight days (median) after stem cell transplantation. Responsiveness of washed platelets was determined by flow cytometry, using a platelet count of 10x109/L, for 52 patients and 27 healthy control subjects. In the absence of agonists, surface activation markers were low for both patient and control platelets. After stimulation with adenosine diphosphate ([ADP]; P2Y1/12 agonist), collagen-related peptide (CRP-XL; Glycoprotein VI (GPVI) agonist) or thrombin (PAR1/4 agonist) at maximal doses, integrin aIIbÎ˛3 activation (Figure 1A) and P-selectin expression (Figure 1B) of the patients' platelets were reduced to a variable extent, when compared to the controls, irrespective of the agonist used. Detailed analysis indicated that the overall platelet responsiveness (median=36.8% interquartile range [IQR]=29.7- 46.7%), defined as the average fraction of platelets positive for integrin activation and P-selectin expression for the three agonists: (i) was not different between diagnoses, i.e., AML/ALL, multiple myeloma, lymphoma and other hematological malignancies (Kruskal Wallis H test, P=0.192); (ii) was not affected by stem cell transplantation, i.e., no transplant, autologous or allogenic stem cell transplantation (Kruskal Wallis H test, P=0.640); (iii) was similar for the four major treatment classes, i.e., A+B, A+B+C, B+C, C (Kruskal Wallis H test, P=0.512; Online Supplementary Figure S1); and (iv) did not correlate with the whole blood platelet count (Spearmanâ&#x20AC;&#x2122;s haematologica | 2018; 103(9)

Table 1. Characteristics and hematological parameters of patients during myelosuppression.

Patients characteristics

Number / Value

Age (years) Female/male (n) Diagnosis (n) AML/ALL Multiple myeloma Lymphoma Other Stem cell transplantation (n) Autologous Allogeneic Time since chemotherapy (days) Time since stem cell transplantation (days)

Blood parameters 9

Leukocyte count (x 10 /L) Hemoglobin (mM) Platelet count (x 109/L) Absolute immature platelet number (x 109/L) Immature platelet fraction (%)

60 (60) 32/45 (20/32) 37 (25) 21 (12) 15 (13) 4 (2) 39 (26) 11 (8) 8 (9) 8 (8)

Value

Reference range

0.15 (0.22) 5.7 (5.7) 8 (7) 0.31 (0.26)

3.5 - 11.0 7.5 - 11.0 150 - 400

3.9 (3.6)

1.1 -6.147

Data are for total number of patients (n=77). Patient information for study A (n=52) is indicated between brackets. Median values are given. AML/ALL: acute myeloid leukemia or acute lymphocytic leukemia.

rho=0.175, P=0.239). Together, this suggested that the variability in platelet responsiveness among patients was not directly linked to the disorder, treatment type or number of (residual) circulating platelets. Additional functional analyses were performed with platelets, invariably from patients in the major treatment classes. For 36 of the patients, blood samples could be obtained before and one hour after platelet transfusion. As expected, platelet count increased after transfusion (Online Supplementary Table S4). The clinical efficacy of transfusion was evaluated from the corrected count increment (CCI: [platelet count increment x body surface area]/[number of transfused platelets x 1011]).20 This was adequate for 96% of the patients, as indicated by a CCI value of >7.5 (median: 14.8, IQR: 11.3-18.0). Flow cytometric analysis of integrin activation and P-selectin expression demonstrated that at one hour after transfusion, platelet responsiveness was improved for most patients (Online Supplementary Figure S2). Whenever possible, platelets were also isolated from the remainder of the transfusion concentrates. It appeared that the activity of the circulating platelets after transfusion approached that of the platelets of the concentrates when triggered with thrombin or CRP-XL. However, the responsiveness to ADP of the circulating platelets after transfusion was higher than in concentrates (integrin aIIbÎ˛3 activation, P=0.002). The improved platelet responses after transfusion underlined the low responsiveness of the autologous platelets after chemotherapy.

Impaired platelet responsiveness during myelosuppression To determine whether the reduced platelet responsiveness was linked to the treatment phase, flow cytometric 1559

C.C.F.M.J. Baaten et al.

analysis of platelet responsiveness was performed during the decreasing period of platelet count (50-11 x109/L and ≤10 x109/L), and the recovery of platelet count (11-50 x109/L). The latter was defined as a sustained increase in the platelet count (observed for patient care), independent of platelet transfusion. Of the eight patients included in this category, three had received an autologous transplant and one patient an allogeneic stem cell transplant, prior to recovery. In the decreasing period, integrin activation and P-selectin expression following stimulation with thrombin or CRP-XL were comparable in patients with platelet counts in the range of 50-11 x109/L and ≤10 x109/L (Figure 2). In contrast, platelet responsiveness to thrombin and CRP-XL significantly improved in the case of count recovery (P<0.001). For stimulation with ADP, these differences were less pronounced, with only P-selectin expression increased during count recovery. These results indicated that platelet count alone is not a good marker of platelet activity. For five patients (one AML, three multiple myeloma, one lymphoma), blood samples could also be analyzed at an earlier time point, i.e., after the stop of chemotherapy, but before severe thrombocytopenia occurred. Remarkably, in all these samples, platelet function was within the normal range for the three agonists (integrin activation 69-86%, P-selectin expression 49-85%). Furthermore, in vitro treatment of control blood with clinically relevant concentrations of cytarabine and/or melphalan did not affect platelet reactivity (Online Supplementary Figure S3A,B). These results argue against a direct effect of chemotherapeutics on the platelet activation properties.

Impaired platelet spreading and Ca2+ signaling of platelets after chemotherapy treatment To further characterize the patient platelets, they were allowed to adhere and spread for ten minutes on a fibrinogen surface, interacting with platelet integrin aIIbβ3. The

observed morphology of the cells was divided into three stages: 1) formation of filopodia, 2) formation of lamellipodia, and 3) full spreading. Most of the platelets from control subjects were in stages 2-3, while the patient platelets predominantly stayed in stage 1 (forming filopodia only), with few platelets being fully spread (Figure 3A). The patients’ platelets displayed a slightly decreased expression of glycoprotein (GP)Iba and GPVI, but not in integrin β3 expression (data not shown). This suggested a diminished integrin activity and outside-in signaling in the patient platelets. We further examined agonist-induced Ca2+ signaling after loading the platelets with Fluo-4. Stimulation with thrombin or CRP-XL induced only a small rise in [Ca2+]i in patient platelets when compared to control platelets (Figure 3B,C). On the other hand, the [Ca2+]i rise induced by thapsigargin (an inhibitor of endoplasmic reticulum Ca2+-ATPases), as a measure of Ca2+ store content,21 was similar for patient and control platelets. Together, this pointed to a defective agonist-induced Ca2+ signaling machinery, independently of receptor type (i.e., PAR1/4 or GPVI receptors).

Impaired mitochondrial bioenergetics but no apoptosis in platelets after chemotherapy Given the cytotoxicity of chemotherapeutic compounds, we evaluated if patient platelets showed characteristics of apoptosis, since this process is known to lead to dysfunctional signaling.22 As a marker of apoptosis, PS exposure was determined by fluorescein isothiocyanate (FITC)-annexin A5 binding. In contrast to control platelets, the patient platelets were prone to expose PS upon shortterm storage without external stimuli (Figure 4A). Upon stimulation with the BH3 mimetic ABT-737, triggering the intrinsic pathway of apoptosis,22 PS exposure was initially accelerated in the patient platelets, when compared to control platelets (Figure 4B). As expected, preincubation with the pan caspase inhibitor quinoline-val-asp-difluo-

1560

Figure 1. Variable impairment of integrin aIIbβ3 activation and P-selectin expression in stimulated platelets from cancer patients with thrombocytopenia after chemotherapy. Washed platelets (10x109/L) from healthy control subjects (healthy ctrl) and thrombocytopenic patients receiving chemotherapy were activated with thrombin (4 nM), CRP-XL (10 μg/mL) or 2MeS-ADP (1 μM) in the presence of 2 mM CaCl2. After 15 min activation, integrin aIIbβ3 activation (A) and P-selectin expression (B) were measured by flow cytometry using PAC-1 and anti-P-selectin antibody, respectively. Medians with IQR; data from 52 patients (25 AML/ALL, 12 multiple myeloma, 13 lymphoma, two other), 27 healthy controls, ***P<0.001. CRP: collagen-related peptide; ADP: adenosine diphosphate.

haematologica | 2018; 103(9)

Platelet dysfunction in chemotherapy

rophenoxymethyl ketone (Q-VD-OPh) fully inhibited the PS exposure triggered by ABT-737. However, Q-VD-OPh failed to affect the storage-dependent PS exposure (Figure 4C). Furthermore, whereas ABT-737 stimulation resulted in high caspase-3 activity, no such activity could be detected during storage (Figure 4D). Additional confirmation for the absence of apoptotic signaling was obtained by assessing the caspase-dependent cleavage of the integrin-binding protein, kindlin-3.23 Western blot analysis indicated that, in platelets from control subjects, ABT-737 treatment induced full cleavage of kindlin-3, which was prevented by Q-VD-OPh (Figure 4E). In the patient platelets (with

confirmed functional impairment of integrin activation and P-selectin expression), however, no kindlin-3 cleavage could be detected in the absence of ABT-737. Platelet activation is known to rely on mitochondrial activity for sufficient ATP production.24 Given that mitochondrial impairment can lead to PS exposure,25,26 we assessed the activity of mitochondria in several ways. As part of the initial characterization of the patient platelets, the mitochondrial membrane potential was assessed by staining with TMRE. Whereas control platelets displayed high TMRE fluorescence, the patient platelets showed much less fluorescence intensity (Figure 5A). This suggest-

Figure 2. Impaired platelet responsiveness in relation to phase of treatment and/or recovery. Platelet integrin aIIbÎ˛3 activation and P-selectin expression were measured (see Figure 1). Patients (n=52) were divided into two categories: (i) decreasing platelet count 50-11 x109/L (n=15) and (ii) decreasing platelet count â&#x2030;¤10 x109/L (n=37). Furthermore, from a subset of patients a sample could be collected when the platelet count increased independently of platelet transfusion (iii): 1150 x109/L (n=8). Data are expressed as % of platelets positive for PAC-1 or anti-P-selectin staining in the absence of stimulation (A), or after stimulation with thrombin (B), CRP-XL (C) or 2MeS-ADP (D). Medians with IQR for patients and healthy controls (n=27); **P<0.01 and ***P<0.001. plt: platelet.

haematologica | 2018; 103(9)

1561

C.C.F.M.J. Baaten et al.

ed a depolarization of the platelet mitochondria, which was independent of diagnosis or treatment class (KruskalWallis H test, P=0.656 and P=0.126, respectively). The low TMRE fluorescence correlated well with the reduced platelet responsiveness (Spearman’s rho=0.569, P=0.001). However, cyclosporin A-induced inhibition of mitochondrial permeability pore formation did not affect PS exposure (data not shown). We subsequently assessed platelet mitochondrial activity

by measuring mitochondrial respiration via high-resolution respirometry.27 With saturating amounts of complex I-II substrates of the oxidative phosphorylation (OXPHOS) chain, i.e., pyruvate, malate, ADP, glutamate and succinate, the maximal ADP-supported respiration of mitochondria was significantly lower in platelets from patients than from controls (Figure 5B). To exclude that the mitochondrial content was altered, we measured the citrate synthase activity.28 However, this was unchanged in the patients’

Figure 3. Impaired platelet spreading and Ca2+ signaling of platelets from patients. (A) Platelets from patients or healthy controls were allowed to spread on a fibrinogen surface for 10 min, after which microscopic images were captured. Spreading state per platelet was classified in three stages based on morphology: (i) filopodia, (ii) lamellipodia, or (iii) fully spread. Percentages of platelets per category are shown. Medians (with IQR) for nine patients, seven control subjects. (B, C) Fluo-4loaded platelets from patients (n=7) and controls (n=5) were stimulated with thrombin (4 nM), CRP-XL (10 μg/mL) or thapsigargin (0.5 μM) in the presence of 2 mM CaCl2. Changes in Fluo-4 fluorescence were measured in time by flow cytometry. (B) Representative Fluo-4 traces in time. (C) Relative increases in cytosolic Ca2+. Medians with IQR, **P<0.01. Overall platelet responsiveness of the patients was 31.5 – 57.9% (IQR). CRP: collagen-related peptide.

1562

haematologica | 2018; 103(9)

Platelet dysfunction in chemotherapy

platelets (Figure 5C). Transmission electron microscopic images were also recorded, and these did not reveal structural abnormalities of the mitochondria (data not shown). Chemotherapeutics like anthracycline analogues can cause (cardio)myopathy and neuropathy by inducing mitochondrial damage, a process mediated by oxidative stress.29,30 To determine whether a similar process is operative in the platelet lineage, activation markers, mitochondrial function (TMRE) and reactive oxygen species (ROS) levels were measured in platelets from seven patients prior to chemotherapy, and two days after said chemotherapy. Additional blood samples were analyzed when severe thrombocytopenia occurred (median ten days after last treatment; median count 11x109/L). Before the start of chemotherapy, platelet reactivity in these patients was comparable to that of healthy controls (Figure 6A,B). After two days of therapy, the platelet count was slighter lowered (median decrease: 15x109/L, IQR: 12.5-24.5), but platelet reactivity was not significantly changed. In contrast, reactivity in response to all agonists decreased markedly when the patients became thrombocytopenic. Similarly, TMRE fluorescence only decreased in the latter case (Figure 6C), which only then was accompanied by a higher ROS production (Figure 6D). The reduction in TMRE fluorescence correlated with the production of ROS (Spearman’s rho=-0.459, P=0.012). Treatment of con-

trol platelets in vitro with chemotherapeutics affected neither the mitochondrial membrane potential nor the production of ROS (Online Supplementary Figure S3C,D). Together, these results strongly suggest that mitochondrial dysfunction is not caused by a direct effect of chemotherapeutics on platelets, but by affecting the platelet precursor cells, the megakaryocytes.

Discussion In this paper, we provide novel evidence that the platelets from thrombocytopenic patients suffering from hematological malignancies and treated with myeloablative chemotherapy are dysfunctional in multiple aspects. We found that key agonist-induced responses of the patients' platelets, such as integrin activation, secretion and Ca2+ fluxes are impaired, at a remarkably variable extent. Furthermore, the platelets from almost all patients showed agonist-independent exposure of PS upon storage, which was not linked to apoptotic caspase activity, in contrast to the platelets from healthy subjects which did not display PS exposure. In the patients' platelets, the defective activation could be linked to an impaired mitochondrial membrane potential and a decreased mitochondrial respiratory activity.

Figure 4. Increased PS exposure in platelets from patients receiving chemotherapy in the absence of apoptosis. Isolated platelets from patients and controls were incubated at 37°C for 90 min, and stained for PS exposure with FITC-annexin A5. (A) Percentages of PS-exposing platelets, (patients n=15, controls n=12). (B) PS exposure measured after indicated times with vehicle or 5 µM ABT-737 to induce apoptosis, (n=6-9). Platelets (10x109/L) from patients or controls were pretreated with caspase inhibitor Q-VD-OPh (10 μM), as indicated, and then stimulated with ABT-737 (5 μM) or vehicle. (C) Fractions of platelets with PS exposure, measured with FITC-annexin A5, (n=8). (D) Caspase-3 activity determined with a fluorometric assay, (n=4). (E) Absence of caspase-dependent kindlin-3 cleavage in western blots from patient platelets. Control platelets were stimulated with ABT-737 with(out) Q-VD-OPh pretreatment; patient platelets were analyzed during the decreasing and recovery phases of platelet count, (n=7). Overall platelet responsiveness of the patients was 30.5 – 48.4% (IQR). Medians with IQR, *P<0.05, **P<0.01 and ***P<0.001. plt: platelets.

haematologica | 2018; 103(9)

1563

C.C.F.M.J. Baaten et al.

The impaired platelet responsiveness after myeloablative chemotherapy (median of eight days) was only weakly correlated to the whole blood platelet count, thus indicating that the extent of thrombocytopenia was not a main factor in the dysfunction. In agreement with this conclusion, in patients with a recovering platelet count after transplantation, the functionality of the platelets was enhanced. Detailed analysis indicated that neither disease type nor chemotherapy regimen could explain the interpatient variation in platelet responsiveness. This points to other factors determining the severity of dysfunction, such as a different sensitivity of megakaryocytes in the bone marrow to the previous chemotherapy treatment.

As the sensitivity of megakaryocytic precursor cells to chemotherapeutics is known to vary,31 the extent of platelet dysfunction might be a combined result of the sensitivity of the precise drugs administered and their dosage. The dysfunction of platelets identified in this patient group differs markedly from the so-called â&#x20AC;&#x2DC;exhaustedâ&#x20AC;&#x2122; platelets, which have been described for patients with solid tumors.32 Exhausted platelets were characterized by a high integrin activation and P-selectin expression in the absence of stimulating agents, and a reduced increase in the parameters after agonist stimulation. These changes might point to platelet activation in vivo, resulting in a sec-

P=0.071 P=0.054

C Figure 5. Impaired mitochondrial bioenergetics in patient platelets. A) Initial screening of TMRE staining of washed platelets from patients. To assess the mitochondrial membrane potential, platelets were stained with TMRE and subsequently analyzed by flow cytometry. Shown are mean fluorescence intensities of TMRE ((n=39: treatment classes: A+B: n=10; A+B+C: n=5; B+C: n=7; C: n=13) and healthy controls (n=27)). B) High resolution respirometry to measure mitochondrial respiration in washed platelets from additionally included patients (n=7) and controls (n=9). Depicted is oxygen consumption due to sequential addition of saturating amounts of pyruvate (P), malate (M), ADP, glutamate (G), succinate (S) and cytochrome C (Cyto C). C) Citrate synthase activity in washed platelets from patients (n=6) and controls (n=7) to assess mitochondrial content. Medians with IQR, *P<0.05, ***P<0.001. Overall platelet responsiveness of the patients was 28.7â&#x20AC;&#x201C; 46.9% (IQR). ADP: adenosine diphosphate.

1564

haematologica | 2018; 103(9)

Platelet dysfunction in chemotherapy

ondary loss of function.33 Given that in the present patient group P-selectin expression and integrin activation were low without stimulation, there is no evidence for in vivo platelet activation linked to chemotherapy treatment. On the other hand, the patients’ platelets showed a tendency to expose PS, which is compatible with an apoptotic process, as apoptotic platelets are known to be defective in aggregation and secretion.22 However, ongoing apoptotic signaling could be excluded, since: (i) treatment with the pancaspase inhibitor Q-VD-OPh did not prevent PS exposure, (ii) measurable caspase-3 activity was absent, and (iii) caspase-dependent cleavage of kindlin-3 could not be detected. Platelets rely on mitochondrial ATP production, in particular upon activation when their energy demand increases.24 While the mitochondrial content and ultrastructure appeared normal in the patients’ platelets, we noticed a marked reduction of the platelet mitochondrial membrane potential and the mitochondrial oxidative phosphorylation. Other authors have shown that anti-tumor antibiotics (anthracyclines), an important class of chemotherapeutic agents used to treat hematological malignancies, induce cardiotoxicity and muscle weakness due to the impairment of mitochondrial function via an increased production of ROS.29,34,35 In cardiac cells, the accumulation of iron inside the mitochondria may contribute to the pro-

duction of ROS.36 Furthermore, the mitochondrial activity in myocardial and hepatic cells is known to be impaired by the chemotherapeutics cyclophosphamide and carmustine (BCNU).37-39 Our results suggest that a similar mechanism of ROS-linked mitochondrial dysfunction is operative in the platelet precursor cells, as deduced from the strong correlation (at >2 days after treatment) between mitochondrial dysfunction and elevated ROS levels. The fact that platelet activation induced by strong agonists (CRP-XL, thrombin) was more affected than platelet activation by ADP suggests a relatively larger role of mitochondrial ATP production upon stimulation with stronger agonists.40 The slight decrease in GPVI (and GPIba) receptor levels might contribute to the lower responsiveness of platelets, although this can also be the consequence of receptor shedding induced by ROS and mitochondrial stress.41 Taken together, our findings suggest that ROS-induced dysfunction in the mitochondria (before the production of platelets) impairs platelet activity and induces PS exposure, thus leading to a shortened platelet lifetime. This conclusion is supported by a recent study in mice, developing thrombocytopenia after 5-fluororacil treatment. In these animals, low-level laser therapy was found to increase the mitochondrial activity of megakaryocytes, resulting in a normalization of hemostasis.42 Another pos-

Figure 6. Decreased responsiveness of patient platelets is accompanied by mitochondrial membrane depolarization and ROS production. Platelets (10x109/L) were isolated from healthy controls (healthy ctrl) and from patients at three time points; namely 1: directly before the start of chemotherapy, 2: at two days of chemotherapy and, 3: upon severe thrombocytopenia (count ≤50x109/L). Washed platelets were activated with thrombin (4 nM), CRP-XL (10 μg/mL) or 2MeS-ADP (1 μM) in the presence of 2 mM CaCl2. After 15 min activation, integrin aIIbβ3 activation (A) and P-selectin expression (B) were measured by flow cytometry using labeled PAC-1 and anti-P-selectin antibody, respectively. Depicted is mean platelet responsiveness to thrombin, CRP-XL and ADP. Platelet samples were loaded with TMRE (C) to assess mitochondrial membrane potential, indicative of mitochondrial function, or with H2DCFDA (D) to measure ROS levels. Platelets from healthy controls were treated with CCCP as a positive control (h.c. CCCP). Medians with interquartile ranges (IQR); n=7-10 (patients) and n=7 (healthy controls), *P<0.05,**P<0.01.

haematologica | 2018; 103(9)

1565

C.C.F.M.J. Baaten et al.

sible strategy to improve mitochondrial function after chemotherapy are the administration of antioxidants to reduce ROS, considering that patients with hematological malignancies have low levels of vitamin C.30,43 Alternatively, treatment with metformin to improve the mitochondrial energy metabolism could be beneficial.30 Due to ethical limitations, we could not assess whether the platelet dysfunction after chemotherapeutic treatment was linked to abnormal (pro)platelet formation from megakaryocytes in the bone marrow. The available literature suggests that the progenitor cells are more vulnerable towards chemotherapy than matured megakaryocytes.31 In patients who received chemotherapy and had not yet developed thrombocytopenia, we observed a normal platelet activity comparable to that before treatment had started. Furthermore, in vitro treatment of whole blood from healthy controls with cytarabine and/or melphalan affected neither platelet reactivity nor mitochondrial function. This agrees with an indirect drug effect via the megakaryocytes or precursor cells, rather than a direct effect on the circulating platelets. With regard to the coagulant state, the reduced level of factor VII found in combination with high circulating Ddimers in the patients' plasmas is suggestive for a mild ongoing state of tissue factor-triggered coagulation.44 However, the data do not provide evidence for appreciable consumption of other coagulation factors. Given that factor VII has a short half-life in blood,45 it will be the first coagulation factor to decline upon ongoing coagulation. Chemotherapy can induce endothelial cell activation and upregulate tissue factor levels,46 which also can explain the elevated VWF levels in patients. The increased bleeding tendency is most likely the result of the impaired platelet function, without compensation by a higher coagulant

References 1. Versteeg HH, Heemskerk JW, Levi M, Reitsma PH. New fundamentals in hemostasis. Physiol Rev. 2013;93(1):327-358. 2. Apelseth TO, Hervig T, Bruserud Ă&#x2DC;. Current practice and future directions for optimization of platelet transfusions in patients with severe therapy-induced cytopenia. Blood Rev. 2011;25(3):113-122. 3. Stanworth SJ, Estcourt LJ, Powter G, et al. A no-prophylaxis platelet-transfusion strategy for hematologic cancers. N Engl J Med. 2013;368(19):1771-1780. 4. Estcourt LJ, Stanworth SJ, Doree C, et al. Prophylactic platelet transfusion for prevention of bleeding in patients with haematological disorders after chemotherapy and stem cell transplantation. Cochrane Database Syst Rev. 2012;16(5): CD004269. 5. Wandt H, Schaefer-Eckart K, Wendeling K, et al. Therapeutic platelet transfusion versus routine prophylactic transfusion in patients with haematological malignancies: an openlabel, multicentre, randomised study. Lancet. 2012;380(9850):1309-1316. 6. Friedmann AM, Sengul H, Lehmann H, Schwartz C, Goodman S. Do basic laboratory tests or clinical observations predict bleeding in thrombocytopenic oncology patients? A reevaluation of prophylactic

1566

8. 9.

10.

11.

12.

13.

14.

activity. Moreover, although the relative number of PS positive platelets is high, given their fast clearance from circulation it is unlikely that this platelet population would significantly compensate for primary hemostasis. The study herein has several limitations. Given that the number of isolated platelets was limited due to severe thrombocytopenia, only a restricted subset of measurements could be performed per patient blood sample, with the consequence that different patient samples needed to be used for some of the measurements. Furthermore, platelet samples were analyzed from patients with different disease types (AML/ALL, multiple myeloma and malignant lymphoma) after receiving chemotherapy in distinct treatment regimens. Herein, we wish to stress the fact that a reduced platelet function was detected in all patient groups and all therapeutic regimens. Current guidelines for prophylactic transfusion during myelosuppression are based on platelet count only. Our novel findings indicate that, along with the platelet count, the activity of circulating platelets also needs to be considered for an optimal control of hemostasis. Hence, this work encourages an inclusion of platelet function assays for the prediction of bleeding in this patient group. Acknowledgments We thank all medical students involved for assisting in patient inclusion. Funding Funding for this project was provided by the EHA-ISTH Research Fellowship granted by the European Hematology Association and the International Society of Thrombosis and Haemostasis to PvdM. FS is supported by the Alexander von Humboldt Foundation.

platelet transfusions. Transfus Med Rev. 2002;16(1):34-45. Slichter SJ. Relationship between platelet count and bleeding risk in thrombocytopenic patients. Transfus Med Rev. 2004;18(3):153-167. Cowan DH, Graham RC, Baunach D. The platelet defect in leukemia. J Clin Invest. 1975;56(1):188-200. Woodcock BE, Cooper PC, Brown PR, et al. The platelet defect in acute myeloid leukaemia. J Clin Pathol. 1984;37(12):13391342. Leinoe EB, Hoffmann MH, Kjaersgaard E, Johnsen HE. Multiple platelet defects identified by flow cytometry at diagnosis in acute myeloid leukaemia. Br J Haematol. 2004;127(1):76-84. Leinoe EB, Hoffmann MH, Kjaersgaard E, et al. Prediction of haemorrhage in the early stage of acute myeloid leukemia by flow cytometric analysis of platelet function Br J Haematol. 2005;128(4):526-532. Kuter DJ. Managing thrombocytopenia associated with cancer chemotherapy. Oncology (Williston Park). 2015;29(4):282294. Pogliani EM, Fantasia R, LambertenghiDeliliers G, Cofrancesco E. Daunorubicin and platelet function. Thromb Haemost. 1981;45(1):38-42. Lanzi C, Banfi P, Ravagnani F, Gambetta RA.

15.

16.

17.

18.

19.

Diversity of effects of two antitumor anthracycline analogs on the pathway of activation of PKC in intact human platelets. Biochem Pharmacol. 1988;37(18):34973504. Foss B, Ulvestad E, Hervig T, Bruserud Ă&#x2DC;. Effects of cytarabine and various anthracyclins on platelet activation: characterization of in vitro effects and their possible clinical relevance in acute myelogenous leukemia. Int J Cancer. 2002;97(1):106-114. 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):iv5261. HOVON - the Haemato Oncology Foundation for Adults in the Netherlands. Trials (by type) 2017. Available from: http://www.hovon.nl/trials/trials-bytype/all.html. Last accessed: November 2017. HOVON - the Haemato Oncology Foundation for Adults in the Netherlands. Trials NHL 2017. Available from: http://www.hovon.nl/trials/trials-bytype/nhl.html. Last accessed: November 2017. American Cancer Society. How chemotherapy drugs work 2016 [updated February 11, 2016]. Available from: https://www.cancer.org/treatment/treat-

haematologica | 2018; 103(9)

Platelet dysfunction in chemotherapy

20. 21.

22.

23.

24.

25.

26.

27.

28.

ments-and-side-effects/treatment-types/ chemotherapy/how-chemotherapy-drugswork.html. Last accessed November 2017. CBO. Blood transfusion guidelines- the Netherlands. 2011. Smeets EF, Heemskerk JW, Comfurius P, Bevers EM, Zwaal RF. Thapsigargin amplifies the platelet procoagulant response caused by thrombin. Thromb Haemost. 1993;70(6):1024-1029. Vogler M, Hamali HA, Sun XM, et al. BCL2/BCL-XL inhibition induces apoptosis, disrupts cellular calcium homeostasis, and prevents platelets activation. Blood. 2011;117(26):7145-7154. Solari FA, Mattheij NJ, Burkhart JM, et al. Combined quantification of the global proteome, phosphoproteome, and proteolytic cleavage to characterize altered platelet functions in the human Scott syndrome. Mol Cell Proteomics. 2016;15(10):31543169. Kramer PA, Ravi S, Chacko B, Johnson MS, Darley-Usmar VM. A review of the mitochondrial and glycolytic metabolism in human platelets and leukocytes: Implications for their use as bioenergetic biomarkers. Redox Biol. 2014;2:206-210. Mattheij NJ, Gilio K, van Kruchten R, et al. Dual mechanism of integrin aiibb3 closure in procoagulant platelets. J Biol Chem. 2013;288(19):13325-13336. van Kruchten R, Mattheij NJA, Saunders C, et al. Both TMEM16F-dependent and TMEM16F-independent pathways contribute to phosphatidylserine exposure in platelet apoptosis and platelet activation. Blood. 2013;121(10):1850-1857. Lanza IR, Nair KS. Mitochondrial metabolic function assessed in vivo and in vitro. Curr Opin Clin Nutr Metab Care. 2010;13(5):511517. Larsen S, Nielsen J, Hansen CN, et al. Biomarkers of mitochondrial content in skeletal muscle of healthy young human subjects. J Physiol. 2012;590(14):3349-3360.

haematologica | 2018; 103(9)

29. Gilliam LAA, Fisher-Wellman KH, Lin CT, et al. The anticancer agent doxorubicin disrupts mitochondrial energy metabolism and redox balance in skeletal muscle. Free Radic Biol Med. 2013;65:988-996. 30. Ma J, Kavelaars A, Dougherty PM, Heijnen CJ. Beyond symptomatic relief for chemotherapy-induced peripheral neuropathy: Targeting the source. Cancer. 2018;Epub ahead of print. 31. Zeuner A, Signore M, Martinetti D, et al. Chemotherapy-induced thrombocytopenia derives from the selective death of megakaryocyte progenitors and can be rescued by stem cell factor. Cancer Res. 2007; 67(10):4767-4773. 32. Riedl J, Kaider A, Marosi C, et al. Decreased platelet reactivity in patients with cancer is associated with high risk of venous thromboembolism and poor prognosis. Thromb Haemost. 2017;117(1):90-98. 33. Baaten CC, Ten Cate H, van der Meijden PE, Heemskerk JW. Platelet populations and priming in hematological diseases. Blood Rev. 2017;31(6):389-399. 34. Sorensen JC, Cheregi BD, Timpani CA, et al. Mitochondria: inadvertent targets in chemotherapy-induced skeletal muscle toxicity and wasting? ' Cancer Chemother Pharmacol. 2016;78(4):673-683. 35. Gouspillou G, Scheede-Bergdahl C, Spendiff S, et al. Anthracycline-containing chemotherapy causes long-term impairment of mitochondrial respiration and increased reactive oxygen species release in skeletal muscle. Sci Rep. 2015;5:8717. 36. Ichikawa Y, Ghanefar M, Bayeva M, et al. Cardiotoxicity of doxorubicin is mediated through mitochondrial iron accumulation. J Clin Invest. 2014;124(2):617-630. 37. al-Nasser IA. In vivo prevention of cyclophosphamide-induced Ca2+ dependent damage of rat heart and liver mitochondria by cyclosporin A. Comp Biochem Physiol A Mol Integr Physiol. 1998; 121(3):209-214.

38. Prasad SB, Rosangkima G, Nicol BM. Cyclophosphamide and ascorbic acid-mediated ultrastructural and biochemical changes in Dalton's lymphoma cells in vivo. Eur J Pharmacol. 2010;645(1-3):47-54. 39. Kang PT, Chen CL, Ren P, Guarini G, Chen YR. BCNU-induced gR2 defect mediates Sglutathionylation of complex I and respiratory uncoupling in myocardium. Biochem Pharmacol. 2014;89(4):490-502. 40. Corona de la Peña N, Gutiérrez-Aguilar M, Hernández-Reséndiz I, Marín-Hernández Á, Rodríguez-Enríguez S. Glycoprotein Ib activation by thrombin stimulates the energy metabolism in human platelets. PLoS One. 2017;12(8):e0182374. 41. Bergmeier W, Piffath CL, Cheng G, et al. Tumor necrosis factor-alpha-converting enzyme (ADAM17) mediates GPIbalpha shedding from platelets in vitro and in vivo. Circ Res. 2004;95(7):677-683. 42. Zhang Q, Dong T, Li P, Wu MX. Noninvasive low-level laser therapy for thrombocytopenia. Sci Transl Med. 2016; 8(349):349ra101. 43. Huijskens MJ, Wodzig WK, Walczak M, Germeraad WT, Bos GM. Ascorbic acid serum levels are reduced in patients with hematological malignancies. Results Immunol. 2016;6:8-10. 44. Mackman N, Tilley RE, Key NS. Role of the extrinsic pathway of blood coagulation in hemostasis and thrombosis. Arterioscler Thromb Vasc Biol. 2007;27(8):1687-1693. 45. Hoffbrand AV, Moss PAH. Essential Haematology. 6th ed: Wiley-Blackwell; 2011. 46. Giordano P, Molinari AC, Del Vecchio GC, et al. Prospective study of hemostatic alterations in children with acute lymphoblastic leukemia. Am J Hematol. 2010;85(5):325330. 47. Briggs C, Kunka S, Hart D, Oguni S, Machin SJ. Assessment of an immature platelet fraction (IPF) in peripheral thrombocytopenia. Br J Haematol. 2004; 126(1):93-99.

1567

ARTICLE

Platelet Biology & its Disorders

Ferrata Storti Foundation

NLRP3 regulates platelet integrin aIIbβ3 outside-in signaling, hemostasis and arterial thrombosis

Jianlin Qiao,1,2,3# Xiaoqing Wu,1# Qi Luo,1# Guangyu Wei,1 Mengdi Xu,1,2,3 Yulu Wu,1 Yun Liu,1 Xiaoqian Li,2 Jie Zi,2 Wen Ju,1,2,3 Lin Fu,1,2,3 Chong Chen,1,2,3 Qingyun Wu,1,2,3 Shengyun Zhu,1,2,3 Kunming Qi,2 Depeng Li,2 Zhenyu Li,1,2,3 Robert K. Andrews,4 Lingyu Zeng,1,3* Elizabeth E. Gardiner5* and Kailin Xu1,2,3*

Haematologica 2018 Volume 103(9):1568-1576

1 Blood Diseases Institute, Xuzhou Medical University, China; 2Department of Hematology, the Affiliated Hospital of Xuzhou Medical University, China; 3Key Laboratory of Bone Marrow Stem Cell, Jiangsu Province, Xuzhou, China; 4Australian Centre for Blood Diseases, Monash University, Melbourne, Australia and 5ACRF Department of Cancer Biology and Therapeutics, John Curtin School of Medical Research, Australian National University, Canberra, Australia

#JQ, XW and QL, contributed equally to this study.

*LZ, EEG and KX contributed equally to this study.

ABSTRACT

Correspondence: lihmd@163.com

Received: February 20, 2018. Accepted: May 17, 2018. Pre-published: 24 May, 2018.

doi:10.3324/haematol.2018.191700 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/9/1568 ©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.

1568

n addition to their hemostatic function, platelets play an important role in regulating the inflammatory response. The platelet NLRP3 inflammasome not only promotes interleukin-1β secretion, but was also found to be upregulated during platelet activation and thrombus formation in vitro. However, the role of NLRP3 in platelet function and thrombus formation in vivo remains unclear. In this study, we aimed to investigate the role of NLRP3 in platelet integrin aIIbβ3 signaling transduction. Using NLRP3-/- mice, we showed that NLRP3-deficient platelets do not have significant differences in expression of the platelet-specific adhesive receptors aIIbβ3 integrin, GPIba or GPVI; however, NLRP3-/platelets transfused into wild-type mice resulted in prolonged tail-bleeding time and delayed arterial thrombus formation, as well as exhibiting impaired spreading on immobilized fibrinogen and defective clot retraction, concomitant with decreased phosphorylation of c-Src, Syk and PLCγ2 in response to thrombin stimulation. Interestingly, addition of exogenous recombinant interleukin-1β reversed the defect in NLRP3-/platelet spreading and clot retraction, and restored thrombin-induced phosphorylation of c-Src/Syk/PLCγ2, whereas an anti-interleukin-1β antibody blocked spreading and clot retraction mediated by wild-type platelets. Using the direct NLRP3 inhibitor, CY-09, we demonstrated significantly reduced human platelet aggregation in response to threshold concentrations of collagen and ADP, as well as impaired clot retraction in CY-09-treated human platelets, supporting a role for NLRP3 also in regulating human platelet aIIbβ3 outside-in signaling. This study identifies a novel role for NLRP3 and interleukin-1β in platelet function, and provides a new potential link between thrombosis and inflammation, suggesting that therapies targeting NLRP3 or interleukin-1β might be beneficial for treating inflammation-associated thrombosis.

Introduction The primary platelet-specific receptors glycoprotein (GP)VI, which binds collagen and fibrin, and GPIba, which binds von Willebrand factor, initiate platelet aggregation (thrombus formation) by recognition of exposed von Willebrand factor/collagen in the damaged blood vessel wall.1,2 Engagement of platelet receptors initiates intra-platelet signaling pathways, which shift platelet integrin aIIbβ3 from haematologica | 2018; 103(9)

NLRP3 regulates platelet function and thrombosis

a low- to a high-affinity state (inside-out signaling) and enable platelet aggregation and thrombus formation through binding of soluble fibrinogen and other aIIbβ3 ligands.3 Ligand binding to aIIbβ3 also triggers aIIbβ3 outside-in signaling,4 leading to tyrosine phosphorylation of signaling proteins,5-7 including c-Src, spleen tyrosine kinase (Syk) and phospholipase Cγ2 (PLCγ2), and initiates downstream platelet responses, such as granule secretion, platelet spreading and clot retraction.8,9 Platelets also have roles in the inflammatory response and in inflammatory pathology associated with atherosclerosis, malarial or dengue infection and rheumatoid arthritis.10-13 Inflammasomes are multiprotein complexes that mediate responses to various inflammatory stimuli by controlling secretion of the pro-inflammatory cytokine, interleukin 1β (IL-1β).14 Upon stimulation, NLRP3 undergoes oligomerization, leading to conversion of pro-caspase-1 to active caspase-1, which then cleaves pro-IL-1β to mature IL-1β.14 For example, in dengue virus infection, platelet NLPR3 is activated, triggering IL-1β secretion.13 NLPR3 also contributes to platelet activation, aggregation and thrombus formation in vitro, as shown by caspase activity measurements and pharmacological inhibition or genetic ablation of the NLPR3-associated adaptor protein, Bruton tyrosine kinase (BTK).15 In this study using NLRP3deficient platelets, we demonstrated a specific contribution of NLRP3 to aIIbβ3 outside-in signaling, and hemostasis and arterial thrombosis in vivo.

Quantitative real-time polymerase chain reaction The mRNA expression of GPIba, GPVI and IL-1β was measured by quantitative real-time polymerase chain reaction (PCR) as described previously.19,20 In brief, RNA was reversely transcribed into cDNA using oligo(dT) and M-MLV Reverse Transcriptase (Thermo Fisher Scientific) and PCR amplification was performed in triplicate on a LightCycler® R480 II (Roche Life Science) with a total volume of 20 μL, consisting of 10 μL SYBR Green qPCR Super Mix, 0.5 μL forward primer (10 μM), 0.5 μL reverse primer (10 μM), 5 μL cDNA and 4 μL sterile water. The primers for GPIba, GPVI and IL-1β were designed as follows: GPIba forward primer: 5’-AGTTCATACTACCCACTGGAGCC-3’, reverse primer: 5’-GTGGGTTTATGAGTTGGAGGC-3’; GPVI forward primer: 5’-AGGAGACCTTCCATCTTACCCA-3’, reverse primer: 5’-GAGCAAAACCAAATGGAGGG-3’; IL-1β forward primer: 5’CCTGAACTCAACTGTGAAATGC-3’, reverse primer: 5’-GATGTGCTGCTGCGAGATT-3’. The relative mRNA expression of GPIba, GPVI and IL-1β was calculated using the 2-DDCt method and normalized to an internal control (β-actin). Detailed methods on RNA extraction are provided in the Online Supplement.

Tail bleeding time Tail bleeding assays were performed as previously described.18 In brief, a 10-mm segment of tail tip was cut off and the tail was then immersed in pre-warmed sterile saline solution (37°C). Tail bleeding time was calculated as the time taken for bleeding to stop.

FeCl3-induced thrombosis in vivo Methods Animals NLRP3-/- C57BL/6 mice16 were purchased from Jackson Laboratories. All experimental procedures were approved by the Ethics Committee of Xuzhou Medical University.

Platelet preparation Procedures involving collection of mouse and human blood were approved by the Medical Ethics Committee of Xuzhou Medical University. For mouse platelet studies, blood was collected from the retro-orbital plexus using ACD (85 mM trisodium citrate, 83 mM dextrose, and 21 mM citric acid) as anticoagulant and diluted in modified Tyrode buffer (12 mM NaHCO3, 138 mM NaCl, 5.5 mM glucose, 2.9 mM KCl, 2 mM MgCl2, 0.42 mM NaH2PO4, 10 mM HEPES, pH 7.4). Platelets were then pelleted by centrifugation at 180 g in the presence of PGE1 (0.1 μg/mL) and apyrase (1 U/mL) (Sigma-Aldrich), washed twice with CGS buffer (120 mM sodium chloride, 12.9 mM trisodium citrate, 30 mM Dglucose, pH 6.5) and re-suspended in modified Tyrode buffer. Isolated platelets were allowed to rest for 1 h at room temperature before use. For human platelet studies, venous human blood was collected and then the platelets were isolated as described previously.18

Platelet analyses in vitro Platelet receptor expression, activation, aggregation and immunoblotting were studied as previously described. 17,18 Antibodies against c-Src (anti-Tyr-416, Cell Signaling Technology; pan-c Src, Proteintech), Syk (anti-Tyr-525 and pan-Syk, Bioworld Technology) and PLCγ2 (anti-Tyr-1217 and pan-PLCγ2; Bioworld Technology), IL-1β (Cell Signaling Technology) and Caspase-1 (BioVision) were used. Detailed methods of the electron microscopy of platelet spreading, and clot retraction are provided in the Online Supplement. haematologica | 2018; 103(9)

Platelets isolated from wild-type and NLRP3-/- mice were labeled with calcein and infused into wild-type mice via tail vein injection. Injury to mesenteric arterioles was induced by 0.62 M FeCl3 and thrombus formation was monitored by fluorescence microscopy (Olympus BX53).

Statistical analysis Data are represented as mean ± standard deviation (SD) or standard error (SE) where indicated and analyzed by the Student t-test, one-way or two-way ANOVA.

Results NLRP3 deficiency in platelets impairs in vivo hemostasis and thrombosis To evaluate whether NLRP3 deficiency affects platelet production or clearance, we measured platelet count, mean platelet volume, platelet distribution width and plateletcrit and found similar values in wild-type and NLRP3-/- mice (P>0.05) (Figure 1A). Platelet receptors GPIba, GPVI and integrin aIIbβ3 are critical for platelet function.21,22 Evaluation of these receptors by flow cytometry and reverse transcriptase PCR revealed equivalent mRNA and protein levels in wild-type and NLRP3-/- mice (P>0.05) (Figure 1B). Electron microscopy analysis indicated that NLRP3 deficiency did not affect platelet ultrastructural organization, or the number and size of a- and dense granules (Figure 1C). Together, these data suggest that NLRP3 does not affect platelet production, expression of platelet receptors or granules. In order to investigate whether NLRP3 influences platelet function in vivo, we performed tail bleeding assays and monitored FeCl3-induced mesenteric arteriole thrombus formation. As NLRP3-/- mice demonstrated decreased tail bleeding times (mean ± SD) of 33.33 ± 14.43 s (n = 6), 1569

J. Qiao et al.

D(i)

D(ii)

E(i)

E(ii)

E(iii)

Figure 1. Platelet parameters, adhesion receptor expression, ultrastructure analysis, tail bleeding and arterial occlusion time in wild-type or NLRP3-/- mice. (A) Platelet count, mean platelet volume (MPV), platelet distribution width (PDW) and plateletcrit (PCT) determined by an automatic blood analyzer (mean ± SE). (B) Platelet aIIbβ3 surface expression was determined by flow cytometry using FITC-conjugated anti-mouse aIIβ monoclonal antibody; meanwhile total RNA was isolated from washed platelets in order to assess the expression of GPIba and GPVI by quantitative real-time PCR. Data are represented as a ratio relative to an internal control (β-actin) (mean ± SE, n = 5-7) (Student t-test). (C) Analysis of platelet ultrastructure (a-granules and dense granules) by electron microscopy. Scale bar: 2 μm for upper panel (x 15,000 magnification) and 1 μm for lower panel (x 30,000 magnification). Black arrow: a-granule; white arrow: dense granule. The numbers of agranules and dense granules were counted in 60 wild-type (WT) and 60 NLRP3 knockout platelets (mean ± SE) (Student t-test). (D)-i Tail bleeding time analysis of WT and thrombocytopenic (TP) mice after injection of anti-aIIb antibody, followed by infusion of donor platelets (1 x 108) from WT mice at 9 h after antibody administration (mean ± SD, n= 6-7) (Student t-test). (D)-ii To investigate hemostasis in vivo, washed platelets (1 x 108) were isolated from WT or NLRP3-/- mice and then infused into mice made thrombocytopenic by intraperitoneal injection of 0.1 mg/kg (body weight) rat anti-mouse aIIb antibody (MWReg 30) for 9 h, followed by analysis of tail bleeding time (mean ± SD, n = 5-6) (Student t-test). (E) For analysis of thrombosis in vivo, FeCl3-induced arterial thrombus formation was initiated after washed platelets from WT or NLRP3-/- mice had been infused into WT mice and the time to vessel occlusion was recorded (mean ± SD, n = 9) (Student t-test) (i). Representative image of thrombus formation (ii) and the relative fluorescence (mean ± SD, n = 5) (one-way ANOVA) (iii) at different time points are shown. Movies showing real-time platelet adhesion and thrombus formation in FeCl3-induced injured mesenteric arterioles were provided in the supporting information. *P<0.05;**P<0.01; ***P<0.001; ****P<0.0001. ns: not significant.

1570

haematologica | 2018; 103(9)

NLRP3 regulates platelet function and thrombosis

Figure 2. Platelet aggregation and activation. (A) Platelet aggregation was induced by addition of collagen, CRP, thrombin or ADP to platelets from wild-type (WT) or NLRP3-/- mice. Representative aggregation traces using platelets from WT (green) or NLRP3-/- (red) mice are shown together with combined data (mean ± SE) for three mice (Student t-test). (B) Platelet P-selectin expression and (C) activated aIIbβ3 (JON/A binding) were assessed in platelets before and after treatment with the indicated concentrations of collagen, CRP, ADP or thrombin for 10 min by flow cytometry using phycoerythrin-conjugated anti-P-selectin antibody (Ebioscience) or JON/A antibody (emfret ANALYTICS) (mean ± SE, n = 3-6). *P<0.05.

and reduced prothrombin times of 8.54 ± 0.30 s (n = 8) compared with the wild-type control values of 63.43 ± 26.94 s (n = 7) and 9.07 ± 0.37 s (n = 10), respectively, as well as elevated levels of coagulation factors VIII and IX (data not shown), platelets from wild-type or NLRP3-/- mice were injected into thrombocytopenic or wild-type mice to eliminate non-platelet hemostatic or thrombotic factor variations. The platelet numbers in mice receiving infusion of wild-type or NLRP3-/- platelets either before, after injection of anti-aIIb antibody or after infusion were comparable (Online Supplementary Figure S1). In addition, injection of anti-aIIb antibody did not affect the function of donor platelets as the tail bleeding time of thrombocytopenic mice receiving donor platelets (9 h after antibody) was similar to that of wild-type mice (Figure 1D-i). A significantly prolonged tail bleeding time (P<0.001) (Figure 1D-ii), delayed arterial thrombus formation (Figure 1E-i) as well as reduced platelet accumulation (relative fluorescence at the site of vascular injury) (P<0.0001) (Figure 1E-ii and iii) were detected in mice with NLRP3-/- platelets, suggesting that platelet NLRP3 deficiency significantly impairs hemostasis and arterial thrombus formation in vivo.

NLRP3-/- platelets display mildly reduced platelet aggregation and normal degranulation and aIIbβ3 activation

As platelet aggregation plays an important role in platelet function, we further investigated the effect of NLRP3 on platelet aggregation and activation. As shown in Figure 2A, platelet aggregation in response to a low dose of collagen (0.5 μg/mL), thrombin (0.05 U/mL) and ADP (10 μM) was haematologica | 2018; 103(9)

significantly reduced in NLRP3-/- platelets compared with that of wide-type platelets (P<0.05), which was consistent with the findings of a previous study showing that NLRP3 deficiency affects platelet aggregation and activation.15 However, no differences of platelet aggregation were found in response to a relatively high dose of collagen (1 μg/mL), thrombin (0.1 U/mL), ADP (20 μM) as well as collagen-related peptide (CRP: 0.5 and 1 μg/mL) (Figure 2A). As platelet granule secretion induced by agonist stimulation plays critical roles in the amplification of platelet signaling and subsequent activation and aggregation, we also measured platelet degranulation, represented by P-selectin expression and aIIbβ3 activation by flow cytometry. Interestingly, P-selectin upregulation (Figure 2B) and aIIbβ3 activation (Figure 2C) were normal in NLRP3-/- platelets after stimulation by collagen, CRP, thrombin or ADP. This is in contrast to the findings of Murthy and colleagues15 and could be related to subtle differences in platelet preparations and experimental conditions. A direct link between NLRP3 and a granule release has not been described, and differences in rate and extent of P-selectin exposure are consistent with studies that have demonstrated agonist-related differential kinetics of platelet degranulation and secretion.23-25

NLRP3 deficiency in platelets affects integrin aIIbβ3 outside-in signaling

Platelet spreading and clot retraction, two processes regulated by early- and late-aIIbβ3 outside-in signaling, respectively,26 were evaluated. Spreading of NLRP3-deficient platelets (Figure 3A) but not platelet adhesion (Figure 1571

J. Qiao et al.

3B) was significantly inhibited on immobilized fibrinogen, together with reduced phosphorylation of c-Src and PLCγ2 (Figure 3C), which mediate platelet spreading.5,7 Addition of thrombin (2 U/mL) did not reverse the defective spreading of NLRP3-/- platelets (Figure 3A), indicating that defective spreading did not result from insufficient activation. Interestingly, impaired platelet spreading was also observed in apyrase (1 U/mL)-treated platelets from wild-type mice, similar to the defective spreading of NLRP3-/- platelets. Defective spreading of NLRP3-/-platelets

E(i)

was rescued by the addition of ADP (Online Supplementary Figure S2), suggesting that ADP secretion might be impaired in NLRP3-/- platelets. NLRP3 regulates inflammation through processing pro-IL-1β to IL-1β,27 and platelets excise introns from IL-1β pre-mRNA, yielding a mature message and translated protein.28 Wild-type and NLRP3deficient platelets expressed equivalent levels of IL-1β mRNA (Figure 3D-i) in RNA extracts that contained minimal RNA from contaminating leukocytes (Figure 3D-ii); however, thrombin stimulation triggered release of IL-1β

D(ii)

D(i)

E(ii)

Figure 3. Platelet spreading and interleukin-1β secretion. (A) Platelet spreading on immobilized fibrinogen in the presence or absence of 10 ng/mL IL-1β, 0.5 μg/mL anti-IL-1β or 2 U/mL thrombin. Scale bar = 20 μm. Covered area was quantified by Image J software and analyzed by one-way ANOVA for comparison. Images (X100) are representative of three independent experiments (mean ± SD, n = 3). Compared with wild-type (WT), **P<0.01; ns: not significant. (B) Platelet adhesion on fibrinogen for 90 min. The number of platelets that adhered to fibrinogen-coated glass coverslips was calculated using Image J software (mean ± SE, n = 3) (Student t-test). (C) Phosphorylation of c-Src and PLCγ2 in platelets after spreading on fibrinogen for 90 min. Data were quantified using Image J software and are represented as a ratio relative to total level (mean ± SD, n = 3) (Student t-test). *P<0.05, **P<0.01. (D) Total RNA was isolated from 5 x 108/mL platelets in order to measure IL-1β mRNA expression by RT-PCR (mean ± SD, n = 3) (Student t-test) (i); Washed platelet preparations (5 x 108/mL) were found to contain approximately 2 x 104/mL leukocytes. RNA was isolated from 5 x 108/mL platelets or 2 x 104/mL leukocytes followed by reverse transcription into cDNA which was used to measure IL-1β mRNA expression by PCR. PCR products were evaluated on 1.5% agarose gel (ii). (E) Washed platelets were stimulated with 0.5 U/mL thrombin before the level of IL-1β in supernatants was measured by enzyme-linked immunosorbent assay (mean ± SE, n = 3-5) (Student t-test) (i) and western blot (mean ± SD, n = 3) (ii). (F) Active caspase-1 expression after thrombin stimulation (mean ± SD, n = 3) (Student t-test). The western blot analysis of IL-1β or active caspase-1 expression was quantified as fold change to the level without treatment. *P<0.05;**P<0.01; ***P<0.001. ns: not significant.

1572

haematologica | 2018; 103(9)

NLRP3 regulates platelet function and thrombosis

from wild-type but not NLRP3-/- platelets (Figure 3E-i). Consistently, western blot analysis showed significantly reduced IL-1β expression after thrombin stimulation (Figure 3E-ii). A preliminary analysis suggested that the expression of active caspase-1, which is responsible for processing IL-1β, was significantly lower in NLRP3-/platelets than in wild-type platelets after thrombin stimulation (Figure 3F). Supporting a role for IL-1β in platelet spreading, anti-IL-1β antibody treatment significantly impaired wild-type platelet spreading (Figure 3A) and when NLRP3-/- platelets were pre-treated with recombinant mouse IL-1β, platelet spreading was restored (Figure 3A). Since IL-1β also regulates platelet activation,29 we hypothesized that defective spreading of NLRP3-/- platelets might be due to decreased secretion of IL-1β. To test this, we measured IL-1β release from platelets after spreading for 90 min and found a small but significant decrease in IL1β secretion from NLRP3-/- platelets (59.0 ± 2.4 pg/mL) compared to that from wild-type platelets (52.5 ± 2.5 pg/mL) (mean ± SD; P=0.03;), suggesting that efficient spreading requires mature IL-1β. To then determine whether IL-1β exerted its effect on platelet spreading through IL-1 receptor (IL-1R) engagement, we pre-treated platelets with recombinant IL-1R antagonist (IL-1RA) and found defective spreading of wild-type platelets (Online Supplementary Figure S3). Furthermore, IL-1RA abolished IL-1β-mediated rescue of spreading of NLRP3-/- platelets

(Online Supplementary Figure S3), suggesting that IL-1β affects platelet spreading via IL-1R. Consistent with the ablated platelet spreading, NLRP3-/platelets showed a significant impairment of clot retraction (Figure 4A) which was recovered by IL-1β addition, suggesting that NLRP3 regulates clot retraction via an IL-1βdependent mechanism. Impaired clot retraction was also evident in wild-type platelets treated with anti-IL-1β antibody (Figure 4A). Furthermore, IL-1RA treatment impaired clot retraction of normal platelets and abolished the effect of IL-1β on the recovery of clot retraction in NLRP3-/- platelets (Online Supplementary Figure S4). As activation of aIIbβ3 outside-in signaling leads to phosphorylation of c-Src, Syk, and PLCγ2, and clot retraction,5,6 we measured the phosphorylation status of these signaling proteins. Consistent with the defective platelet spreading and clot retraction, NLRP3-/- platelets exhibited significantly reduced phosphorylation of c-Src (Figure 4B), Syk (Figure 4C), and PLCγ2 (Figure 4D) following thrombin stimulation compared with that of wild-type platelets. However, treatment with recombinant IL-1β reversed the decreased phosphorylation of signaling proteins (Figure 4B-D). Interestingly, robust phosphorylation of c-Src, Syk and PLCγ2 in response to CRP/GPVI engagement, which does not require aIIbβ3 signaling, was achieved in NLRP3-/- platelets (Online Supplementary Figure S5). Together, these findings support a specific role for

Figure 4. Impaired clot retraction and phosphorylation of c-Src, Syk and PLCγ2 in platelets from NLRP3-/- mice after thrombin stimulation. (A) Clot retraction was studied using washed platelets treated with 1 U/mL thrombin in the presence/absence of 10 ng/mL recombinant mouse IL-1β or 0.5 μg/mL anti-IL-1β antibody at 37°C. Representative images at 15, 30, 45, 60, 75, 90, 105 and 120 min from three independent experiments are shown. Data were quantified as the clot volume (%) and are presented as mean values (two-way ANOVA). Western blots of total and phosphorylated (B) c-Src (Tyr-416), (C) Syk (Tyr-525) and (D) PLCγ2 (Tyr-1217) in platelets treated with 1 U/mL thrombin in the presence of 2 mM Ca2+ and 0.5 mg/mL fibrinogen with or without 10 ng/mL recombinant mouse IL-1β pre-treatment at different time points were quantified (as a ratio of phosphorylated to total protein level) using Image J software and analyzed by two-way ANOVA for comparison (mean ± SD, n = 3). Images are representative of three independent western blot experiments. Compared with wild-type (WT): *P<0.05; **P<0.01; ***P<0.001.

haematologica | 2018; 103(9)

1573

J. Qiao et al. NLRP3/IL-1β in the regulation of platelet integrin aIIbβ3 outside-in signaling.

Inhibition of NLRP3 impairs clot retraction in human platelets Using a selective and direct NLRP3 inhibitor (CY-09),30 we evaluated the role of NLRP3 in integrin aIIbβ3 signaling

transduction in human platelets. We found significantly reduced human platelet aggregation in response to low doses but not high doses of collagen and ADP following treatment with CY-09 (Figure 5A). Interestingly, CY-09 treatment did not affect platelet aggregation in response to CRP stimulation (Figure 5A). Using flow cytometry we found no differences in platelet degranulation (P-selectin

C Figure 5. Effect of NLRP3 inhibition on human platelet function. (A) Human platelet-rich plasma was incubated with CY-20 (20 μΜ) for 30 min at 37°C and then platelet aggregation in response to collagen (1 and 2 μg/mL), ADP (2.5 and 5 μΜ) or CRP (0.5 μg/mL) was measured in a light transmittance aggregometry (Helena Aggram, Helena Laboratories, Beaumont, USA). Maximum platelet aggregation (%) was recorded (mean ± SE, n = 3) (two-way ANOVA). (B) Platelet P-selectin expression and (C) aIIbβ3 activation in response to collagen, ADP or CRP stimulation was measured by flow cytometry using phycoerythrin-conjugated antiP-selectin antibody and FITC-conjugated PAC-1 antibody, respectively, and represented as mean ± SE (n = 3) (one-way ANOVA). (D) Clot retraction was initiated using CY-09-treated washed human platelets stimulated with 1 U/mL thrombin in the presence/absence of recombinant human IL-1β (10 ng/mL). Representative images at 30, 60, 90, and 120 min from three independent experiments are shown and data were quantified as the clot volume (%) (mean ± SD, n = 3) (two-way ANOVA). *P<0.05. Compared with vehicle, **P<0.01; ***P<0.001. Compared with CY-09, #P<0.05. Figure 6. Role of NLRP3 in the regulation of platelet integrin aIIbβ3 outside-in signaling. Engagement of G protein coupled receptors (GPCR) by thrombin induces platelet intracellular reactive oxygen species (ROS) production (1), which activates NLRP3, leading to assembly of the NLRP3 inflammasome and subsequent activation of caspase-1, which processes immature pro-IL-1β into mature IL-1β. Once released, IL-1β binds to IL-1 receptor (IL-1R) and initiates IL-1R intracellular signaling transduction, resulting in phosphorylation of c-Src and Syk, which regulates platelet spreading and clot retraction. Meanwhile, ligation of GPCR also induces ATP release (2), which can activate NLRP3 through binding to P2XR. LRR: Leucine-rich repeat; NACHT: NACHT, NAIP, CIITA, HET-E and TP1; PYD: Pyrin domain; ASC: Apoptosis-associated speck-like protein containing a CARD.

1574

haematologica | 2018; 103(9)

NLRP3 regulates platelet function and thrombosis expression) and aIIbβ3 activation (PAC-1 binding) in collagen, ADP or CRP-stimulated human platelets after CY-09 treatment (Figure 5B), consistent with minimal effects of NLRP3 depletion on other platelet activation pathways. Thrombin-initiated clot retraction was significantly impaired in CY-09-treated human platelets (Figure 5C). Addition of recombinant human IL-1β rescued the impaired clot retraction of CY-09-treated platelets, suggesting that NLRP3 might also be involved in the regulation of human platelet integrin aIIbβ3 outside-in signaling. Collectively, our data support a role for NLRP3 in regulating platelet aIIbβ3 outside-in signaling in human and murine platelets.

Discussion Systemic inflammation has been demonstrated to be a potent prothrombotic stimulus, with mechanisms including upregulation of procoagulant factors, inhibition of natural anticoagulants and fibrinolytic activity as well as increased platelet reactivity.31,32 However, the relationship between inflammation and platelet function remains poorly understood. In this study, we identified a specific contribution of the NLRP3 inflammasome to aIIbβ3 outside-in signaling, and hemostasis and arterial thrombosis in vivo. Inflammasomes are multiprotein complexes that respond to various inflammatory stimuli by controlling secretion of the pro-inflammatory cytokine, IL-1β.14 The NLRP3 inflammasome is one of the largest and most studied cytosolic inflammasomes. It undergoes oligomerization upon stimulation, leading to activation of caspase1, which mediates the maturation of IL-1β.14 Along with immune cells, platelets express NLRP3 inflammasomes. Platelet NLRP3 has been shown to be involved in the regulation of endothelial permeability after dengue infection by processing pro-IL-1β and releasing mature IL-1β.13 Additionally, NLRP3 contributes to platelet activation, aggregation and thrombus formation in vitro.15 Consistent with this, we showed reduced platelet aggregation in response to threshold concentrations of collagen, thrombin and ADP in NLRP3-deficient platelets. However, degranulation (P-selectin expression) and aIIbβ3 activation were normal in NLRP3-/- platelets. As platelet NLRP3 deficiency significantly impaired hemostasis and arterial thrombosis in vivo, we suggest that NLRP3 is a novel molecular link between inflammation and thrombosis. Platelet spreading and clot retraction, regulated by aIIbβ3 outside-in signaling, play important roles in stabilizing thrombus formation.26 To investigate the role of NLRP3 in aIIbβ3 outside-in signaling, we measured platelet spreading and clot retraction and showed that NLRP3 deficiency significantly decreased platelet spreading on immobilized fibrinogen and impaired clot retraction. As IL-1β has been demonstrated to increase platelet adhesion to collagen and fibrinogen,29 we hypothesized that the effect of NLRP3 on aIIbβ3 signaling might be through IL-1β. To test

References 1. Qiao JL, Shen Y, Gardiner EE, Andrews RK. Proteolysis of platelet receptors in humans and other species. Biol Chem. 2010;391 (8):893-900. 2. Qiao J, Arthur JF, Gardiner EE, Andrews RK,

haematologica | 2018; 103(9)

that, we measured IL-1β secretion from thrombin-stimulated platelets and found significantly reduced IL-1β release from NLRP3-deficient platelets, indicating that NLRP3 is responsible for processing IL-1β in platelets. Moreover, addition of IL-1β rescued impaired platelet spreading and clot retraction, supporting the role of NLRP3/IL-1β in platelet integrin aIIbβ3 signaling. Through the use of a selective and direct NLRP3 inhibitor, CY-09,30 we also evaluated whether NLRP3 exerts the same functional effects on human platelets and showed reduced platelet aggregation in response to threshold doses of collagen and ADP without differences in degranulation and aIIbβ3 activation. However, inhibition of NLRP3 by CY-09 significantly ablated clot retraction in human platelets. Importantly, addition of recombinant human IL-1β rescued impaired clot retraction of CY-09-treated human platelets, indicating that NLRP3 might also participate in the regulation of aIIbβ3 outside-in signaling in human platelets, which is consistent with findings in mouse platelets. A potential limitation of our findings is that we cannot exclude the interesting possibility that NLRP3 depletion in mice affects platelet hemostatic/thrombotic function through as yet unknown indirect pathways. Further studies are required to explore the exact mechanisms of NLRP3/IL1β signaling and the specific role of this inflammasome complex in other thrombotic models and human disease. In conclusion, our results demonstrate that NLRP3 regulates platelet spreading and clot retraction by a mechanism involving IL-1β. Moreover, impaired hemostasis and arterial thrombosis were observed in vivo in mice with NLRP3-/platelets. Furthermore, inhibition of NLRP3 impairs clot retraction in human platelets. These data identify a unique role for NLRP3 in the regulation of platelet function and thrombus formation (Figure 6), and provide a novel molecular link between thrombosis and inflammation. Acknowledgments This research was supported by the National Natural Science Foundation of China (grant n. 81400082, 81370602, 81570096, 81671584, 81641151 and 81700178), the Natural Science Foundation of Jiangsu Province (grant n. BK20141138 BK20140219), funding for the Distinguished Professorship Program of Jiangsu Province, the Shuangchuang Project of Jiangsu Province, the National Health and Medical Research Council of Australia, the Six Talent Peaks Project of Jiangsu Province (WSN-133), the 333 projects of Jiangsu Province (BRA2017542), the Key University Science Research Project of Jiangsu Province (17KJA320008), Jiangsu Province’s Key Provincial Talents Program (ZDRCA2016054), the Colleges’ Science Foundation of Jiangsu Province (16KJB320013), Postgraduate Research Innovation Project of Jiangsu Province (KYCX18_2186), and Key University Science Research Project of Jiangsu Province (18KJA320010). We thank Prof. Rongbin Zhou (University of Science and Technology of China, Hefei, China) for kindly providing the NLRP3 inhibitor CY-09.

Zeng L, Xu K. Regulation of platelet activation and thrombus formation by reactive oxygen species. Redox Biol. 2017;14:126130. 3. Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell. 1992;69(1):11-25.

4. Li Z, Delaney MK, O'Brien KA, Du X. Signaling during platelet adhesion and activation. Arterioscler Thromb Vasc Biol. 2010;30(12):2341-2349. 5. Clark EA, Shattil SJ, Ginsberg MH, Bolen J, Brugge JS. Regulation of the protein tyrosine kinase pp72syk by platelet agonists and the

1575

J. Qiao et al.

9. 10.

11. 12. 13.

14. 15.

1576

integrin alpha IIb beta 3. J Biol Chem. 1994;269(46):28859-28864. Suzuki-Inoue K, Hughes CE, Inoue O, et al. Involvement of Src kinases and PLCgamma2 in clot retraction. Thromb Res. 2007;120(2):251-258. Wonerow P, Pearce AC, Vaux DJ, Watson SP. A critical role for phospholipase Cgamma2 in alphaIIbbeta3-mediated platelet spreading. J Biol Chem. 2003;278 (39):3752037529. Phillips DR, Jennings LK, Edwards HH. Identification of membrane proteins mediating the interaction of human platelets. J Cell Biol. 1980;86(1):77-86. Shattil SJ, Kashiwagi H, Pampori N. Integrin signaling: the platelet paradigm. Blood. 1998;91(8):2645-2657. Morrell CN, Aggrey AA, Chapman LM, Modjeski KL. Emerging roles for platelets as immune and inflammatory cells. Blood. 2014;123(18):2759-2767. Kapur R, Semple JW. The nonhemostatic immune functions of platelets. Semin Hematol. 2016;53 (Suppl 1):S2-6. Duerschmied D, Bode C, Ahrens I. Immune functions of platelets. Thromb Haemost. 2014;112(4):678-691. Hottz ED, Lopes JF, Freitas C, et al. Platelets mediate increased endothelium permeability in dengue through NLRP3-inflammasome activation. Blood. 2013;122(20):34053414. Haneklaus M, O'Neill LA. NLRP3 at the interface of metabolism and inflammation. Immunol Rev. 2015;265(1):53-62. Murthy P, Durco F, Miller-Ocuin JL, et al. The NLRP3 inflammasome and Bruton's

16.

17.

18.

19.

20.

21.

22.

23. 24.

tyrosine kinase in platelets co-regulate platelet activation, aggregation, and in vitro thrombus formation. Biochem Biophys Res Commun. 2017;483(1):230-236. Kovarova M, Hesker PR, Jania L, et al. NLRP1-dependent pyroptosis leads to acute lung injury and morbidity in mice. J Immunol. 2012;189(4):2006-2016. Qiao J, Wu Y, Liu Y, et al. Busulfan triggers intrinsic mitochondrial-dependent platelet apoptosis independent of platelet activation. Biol Blood Marrow Transplant. 2016;22(9): 1565-1572. Qiao J, Wu Y, Wu X, et al. An absence of platelet activation following thalidomide treatment in vitro or in vivo. Oncotarget. 2017;8(22):35776-35782. Qiao J, Liu Y, Li D, et al. Imbalanced expression of Bcl-xL and Bax in platelets treated with plasma from immune thrombocytopenia. Immunol Res. 2016;64(2):604-609. Qiao J, Huang Y, Xia Y, et al. Busulfan and cyclosphamide induce liver inflammation through NLRP3 activation in mice after hematopoietic stem cell transplantation. Sci Rep. 2015;5:17828. Rivera J, Lozano ML, Navarro-Nunez L, Vicente V. Platelet receptors and signaling in the dynamics of thrombus formation. Haematologica. 2009;94(5):700-711. Swieringa F, Kuijpers MJ, Heemskerk JW, van der Meijden PE. Targeting platelet receptor function in thrombus formation: the risk of bleeding. Blood Rev. 2014;28(1):9-21. Pagel O, Walter E, Jurk K, Zahedi RP. Taking the stock of granule cargo: platelet releasate proteomics. Platelets. 2017;28(2):119-128. Jonnalagadda D, Izu LT, Whiteheart SW.

25.

26.

27.

28.

29.

30.

31.

32.

Platelet secretion is kinetically heterogeneous in an agonist-responsive manner. Blood. 2012;120(26):5209-5216. Italiano JE, Richardson JL, Patel-Hett S, et al. Angiogenesis is regulated by a novel mechanism: pro- and antiangiogenic proteins are organized into separate platelet alpha granules and differentially released. Blood. 2008;111(3):1227-1233. Flevaris P, Stojanovic A, Gong H, Chishti A, Welch E, Du X. A molecular switch that controls cell spreading and retraction. J Cell Biol. 2007;179(3):553-565. Davis BK, Wen H, Ting JP. The inflammasome NLRs in immunity, inflammation, and associated diseases. Annu Rev Immunol. 2011;29:707-735. Denis MM, Tolley ND, Bunting M, et al. Escaping the nuclear confines: signal-dependent pre-mRNA splicing in anucleate platelets. Cell. 2005;122(3):379-391. Beaulieu LM, Lin E, Mick E, et al. Interleukin 1 receptor 1 and interleukin 1beta regulate megakaryocyte maturation, platelet activation, and transcript profile during inflammation in mice and humans. Arterioscler Thromb Vasc Biol. 2014;34(3): 552-564. Jiang H, He H, Chen Y, et al. Identification of a selective and direct NLRP3 inhibitor to treat inflammatory disorders. J Exp Med. 2017;214(11):3219-3238. Aksu K, Donmez A, Keser G. Inflammationinduced thrombosis: mechanisms, disease associations and management. Curr Pharm Des. 2012;18(11):1478-1493. Esmon CT. Inflammation and thrombosis. J Thromb Haemost. 2003;1(7):1343-1348.

haematologica | 2018; 103(9)

ERRATA CORRIGE

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

Farid Boulad,1,2 Tsiporah Shore,3 Koen van Besien,3 Caterina Minniti,4 Mihaela Barbu-Stevanovic,5 Sylvie Wiener Fedus,6 Fabiana Perna,2 June Greenberg,7 Danielle Guarneri,7 Vijay Nandi,5 Audrey Mauguen,8 Karina Yazdanbakhsh,5 Michel Sadelain2 and Patricia A. Shi4,5

Department of Pediatrics, BMT Service, Memorial Sloan Kettering Cancer Center, New York; 2Center for Cell Engineering, Memorial Sloan Kettering Cancer Center, New York; 3Bone Marrow and Hematopoietic Stem Cell Transplant Program, Weill Cornell Medicine/ New York Presbyterian Hospital, New York; 4Sickle Cell Program, Division of Hematology, Albert Einstein College of Medicine, Bronx; 5 Lindsley F. Kimball Research Institute, New York Blood Center, NY; 6Laboratory Medicine, Memorial Sloan Kettering Cancer Center, New York; 7Division of Hematology and Oncology, Weill Cornell Medicine /New York Presbyterian Hospital, NY and 8Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, USA 1

doi:10.3324/haematol.2018.199414

An incorrect version of sentence appared on May 2018 Issue, pages 773. Our target goal of mobilizing at least 30 CD34+ cells/μL was, however, reached in only 50% of patients given the plerixafor dose of 80 μg/kg, 33% of patients given 160 μg/kg, and 33% of patients given 240 μg/kg.

The corrected version of sentence is published below. Our target goal of mobilizing at least 30 CD34+ cells/μL was, however, reached in only 50% of patients given the plerixafor dose of 80 μg/kg, 67% of patients given 160 μg/kg, and 67% of patients given 240 μg/kg.

An incorrect version of sentence appared on May 2018 Issue, pages 777. Eight of 15 patients (53%) with SCD treated with plerixafor reached the peripheral blood CD34 cell target count of at least 30 CD34+ cells/μL, including three of six patients treated at a dose of 240 μg/kg.

The corrected version of sentence is published below. Nine of 15 patients (60%) with SCD treated with plerixafor reached the peripheral blood CD34 cell target count of at least 30 CD34+ cells/μL, including four of six patients treated at a dose of 240 μg/kg.

haematologica | 2018; 103(9)

1577