Haematologica, Volume 105, Issue 9 by Haematologica

haematologica Journal of the Ferrata Storti Foundation

Editor-in-Chief Luca Malcovati (Pavia)

Deputy Editor Carlo Balduini (Pavia)

Managing Director Antonio Majocchi (Pavia)

Associate Editors Hélène Cavé (Paris), Monika Engelhardt (Freiburg), Steve Lane (Brisbane), PierMannuccio Mannucci (Milan), Simon Mendez-Ferrer (Cambridge), Pavan Reddy (Ann Arbor), Francesco Rodeghiero (Vicenza), Andreas Rosenwald (Wuerzburg), Davide Rossi (Bellinzona), Jacob Rowe (Haifa, Jerusalem), Wyndham Wilson (Bethesda), Swee Lay Thein (Bethesda)

Assistant Editors Anne Freckleton (English Editor), Britta Dorst (English Editor), Cristiana Pascutto (Statistical Consultant), Rachel Stenner (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 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 2019 are as following: Print edition

Institutional Euro 700

Personal Euro 170

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. Carlo Balduini; 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)

Associated with USPI, Unione Stampa Periodica Italiana. Premiato per l’alto valore culturale dal Ministero dei Beni Culturali ed Ambientali

haematologica Journal of the Ferrata Storti Foundation

Table of Contents Volume 105, Issue 9: Settembre 2020

About the cover 2187

100-year-old Haematologica images: the origin of megakaryocytes Carlo L. Balduini

Editorials 2188

Novel use for selective inhibitors of nuclear export in Î˛-thalassemia: block of HSP70 export from the nucleus via exportin Xpo1 improves ineffective erythropoiesis Susree Modepalli and Shilpa M. Hattangadi

2190

Combination therapy with interferon and ruxolitinib for polycythemia vera and myelofibrosis: are two drugs better than one? Richard T. Silver

2191

Mysteries of partial dihydroorotate dehydrogenase inhibition and leukemia terminal differentiation Yogen Saunthararajah

2194

Towards genomic-based prognostication and precision therapy for diffuse large B-cell lymphoma Marthe Minderman and Steven T. Pals

2196

Thrombin generation: a global coagulation procedure to investigate hypo- and hyper-coagulability Armando Tripodi

Persperctive Article 2200

BCR-ABL1-like acute lymphoblastic leukemia in childhood and targeted therapy Gunnar Cario et al.

Centenary Review Article 2205

Chronic lymphocytic leukemia: from molecular pathogenesis to novel therapeutic strategies Julio Delgado et al.

Review Articles 2218

Molecular heterogeneity of pyruvate kinase deficiency Paola Bianchi and Elisa Fermo

2229

The variable manifestations of disease in pyruvate kinase deficiency and their management Hanny Al-Samkari et al.

Articles Red Cell Biology & its Disorders

2240

XPO1 regulates erythroid differentiation and is a new target for the treatment of Î˛-thalassemia Flavia Guillem et al.

Granulocyte Biology & its Disorders

2250

Impaired microRNA processing in neutrophils from rheumatoid arthritis patients confers their pathogenic profile. Modulation by biological therapies Ivan Arias de la Rosa et al.

Haematologica 2020; vol. 105 no. 9 - September 2020 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

Myeloproliferative Neoplasms

2262

Ruxolitinib and interferon-α2 combination therapy for patients with polycythemia vera or myelofibrosis: a phase II study Anders Lindholm Sørensen et al.

Acute Myeloid Leukemia

2273

SETDB1 mediated histone H3 lysine 9 methylation suppresses MLL-fusion target expression and leukemic transformation James Ropa et al.

2286

ASLAN003, a potent dihydroorotate dehydrogenase inhibitor for differentiation of acute myeloid leukemia Jianbiao Zhou et al.

Non-Hodgkin Lymphoma

2298

Prognostic impact of somatic mutations in diffuse large B-cell lymphoma and relationship to cell-of-origin: data from the phase III GOYA study Christopher R. Bolen et al.

2308

DA-EPOCH-R combined with high-dose methotrexate in patients with newly diagnosed stage II-IV CD5-positive diffuse large B-cell lymphoma: a single-arm, open-label, phase II study Kana Miyazaki et al.

Plasma Cell Disorders

2316

RAL GTPases mediate multiple myeloma cell survival and are activated independently of oncogenic RAS Marcel Seibold et al.

Hemostasis

2327

Thrombin generation in cardiovascular disease and mortality – results from the Gutenberg Health Study Pauline C.S. van Paridon et al.

Coagulation & its Disorders

2335

A mutated factor X activatable by thrombin corrects bleedings in vivo in a rabbit model of antibody-induced hemophilia A Toufik Abache et al.

Letters to the Editor e440

Mortality in children with sickle cell disease in mainland France from 2000 to 2015 Emilie Desselas et al. http://www.haematologica.org/content/105/9/e440

e444

Circulating cell-free BRAF V600E during chemotherapy is associated with prognosis of children with Langerhans cell histiocytosis Lei Cui et al. http://www.haematologica.org/content/105/9/e444

e448

The long non-coding RNA Cancer Susceptibility 15 (CASC15) is induced by isocitrate dehydrogenase (IDH) mutations and maintains an immature phenotype in adult acute myeloid leukemia Sarah Grasedieck et al. http://www.haematologica.org/content/105/9/e448

e454

Transcription factor 4 (TCF4) expression predicts clinical outcome in RUNX1 mutated and translocated acute myeloid leukemia Florentien E. M. in 't Hout et al. http://www.haematologica.org/content/105/9/e454

e458

Whole exome sequencing identifies mutational signatures of vitreoretinal lymphoma Junwon Lee et al. http://www.haematologica.org/content/105/9/e458

e461

Safety and efficacy of brentuximab vedotin as a treatment for lymphoproliferative disorders in primary immunodeficiencies Thomas Pincez et al. http://www.haematologica.org/content/105/9/e461

Haematologica 2020; vol. 105 no. 9 - September 2020 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation e465

Expression and function of cathelicidin hCAP18/LL-37 in chronic lymphocytic leukemia Enrique Podaza et al. http://www.haematologica.org/content/105/9/e465

e470

Characterization of freshly isolated bone marrow mesenchymal stromal cells from healthy donors and patients with multiple myeloma: transcriptional modulation of the microenvironment Daniel Alameda et al. http://www.haematologica.org/content/105/9/e470

e474

Infection-related morbidity in a large study of transplant non-eligible newly diagnosed myeloma patients treated with UK standard of care Faouzi Djebbari et al. http://www.haematologica.org/content/105/9/e474

e480

Early relapse after autologous transplant for myeloma is associated with poor survival regardless of cytogenetic risk Jill Corre et al. http://www.haematologica.org/content/105/9/e480

Case Report e484

Chronic lymphocytic leukemia and prolymphocytic leukemia. Two coins or two sides of the same coin? Laura Magnano et al. http://www.haematologica.org/content/105/9/e484

Haematologica 2020; vol. 105 no. 9 - September 2020 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

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

Ancient Greek

Scientific Latin

Modern English

aÂma [haima] = blood a·matow [haimatos] = of blood lÒgow [logos]= reasoning haematologicus (adjective) = related to blood haematologica (adjective, plural and neuter, used as a noun) = hematological subjects The oldest hematology journal, publishing the newest research results. 2019 JCR impact factor = 7.116

ABOUT THE COVER 100-year-old Haematologica images: the origin of megakaryocytes Carlo L. Balduini Ferrata-Storti Foundation, Pavia, Italy E-mail: CARLO L. BALDUINI - carlo.balduini@unipv.it doi:10.3324/haematol.2020.262873

ome of the covers of recent issues of Haematologica were dedicated to the origin of platelets,1-3 one of the most hotly debated topics at the time the Journal was founded. Another subject of much discussion was the origin of megakaryocytes, the “giant bone marrow cells” discovered by Giulio Bizzozero in 1869.4 Most authors, including Adolfo Ferrata and Alexander A. Maximow, speculated that these cells derive from the common progenitor of all hematopoietic lineages (‘Emoistioblasto di Ferrata’ or stem cell) and that their multinuclearity is the consequence of nuclear divisions without subsequent cytodieresis. However, other authors thought differently. The image of the cover of this issue of the Journal has been taken from an article entitled "About the system of bone marrow giant cells" published in Haematologica by Giovanni di Guglielmo in 1925 (Figure 1).5 The author, at that time a young assistant to Ferrata at the Department of Medicine of the Pavia University Hospital, investigated the bone marrow of several animals both in the prenatal and postnatal periods and concluded that megakaryocytes derive from the Emoistioblasto di Ferrata, and that their multinuclearity is the consequence of the fusion of two or more of these elements. Di Guglielmo, who three years later would go on to publish his ground-breaking work

on acute erythremic myelosis (now known as pure erythroid leukemia),6,7 was wrong about the origin of the multinuclearity of megakaryocytes, but the beauty of the color plates that illustrate his article still amaze us. I think it is worth noting that Ferrata, who at the time was Editor in Chief of Haematologica, agreed to publish Di Guglielmo's article despite the fact that his pupil had arrived at conclusions that he did not share. Respect for the opinions of others, including those of his assistants, was an important characteristic of this old master of hematology.

References 1. Balduini CL. 100-Year Old Haematologica Images: The Quarrel about the Origin of Platelets (I). Haematologica. 2020;105(5):1169. 2. Balduini CL. 100-Year Old Haematologica Images: The Quarrel about the Origin of Platelets (II). Haematologica. 2020;105(6):1467. 3. Balduini CL. 100-Year Old Haematologica Images: The Quarrel about the Origin of Platelets (III). Haematologica. 2020;105(7):1751. 4. Bizzozero, G. [Sul midollo delle ossa]. Il Morgagni. 1869;11:617-646. 5. Di Guglielmo G. [Sul sistema delle cellule giganti midollari]. Haematologica. 1925;6:156-195. 6. Di Guglielmo G. [Le eritremie]. Haematologica. 1928;9:301-347. 7. Balduini CL. 100-Year Old Haematologica Images: Di Guglielmo Disease or Pure Erythroid Leukemia. Haematologica. 2020;105(3):105.

Figure 1. The origin of megakaryocytes. The images reproduce the hand-drawn color plates illustrating an article published in Haematologica by Giovanni Di Guglielmo in 1925, in which he suggested that megakaryocytes derive from common hematopoietic progenitors that merge and give rise to multinucleated giant cells.

haematologica | 2020; 105(9)

2187

EDITORIALS

Novel use for selective inhibitors of nuclear export in β-thalassemia: block of HSP70 export from the nucleus via exportin Xpo1 improves ineffective erythropoiesis

Susree Modepalli1 and Shilpa M. Hattangadi2 1

Georgetown University School of Medicine, Washington DC and 2National Institutes of Health, Bethesda, MD, USA

E-mail: SHILPA M. HATTANGADI - shilpam.hattangadi@nih.gov doi:10.3324/haematol.2020.254474

halassemia is the most prevalent monogenic recessive hereditary blood disorder, with around 300,000 children affected at birth worldwide.1 Imbalance between a- and non-a-globin chains results in premature death of differentiating erythroid precursors (erythroblasts), because as a-globin tetramers accumulate and precipitate, they form inclusion bodies that cause oxidative membrane damage and destruction by apoptosis. Premature death of differentiating erythroblasts that are dividing faster because of anemia is termed ineffective erythropoiesis, as increased precursors still results in fewer terminal erythrocytes. The hypoxic environment created by anemia leads to increases in factors such as erythropoietin (Epo) or

members of the transforming growth factor-beta (TGF-β) family and Activin receptor-II (ActR-II) trap ligands that stimulate erythropoiesis that continues to be ineffective.2 Continued apoptosis of abnormal precursors results in hemolytic anemia, aggravating other major complications of the disease like iron overload that stem from ineffective erythropoiesis. In the current issue of Haematologica, Guillem et al.3 present a novel potential therapeutic strategy for β-thalassemia, by mitigating the deleterious effect of excess a-globin chains through increased nuclear density of chaperone HSP70 by inhibiting its nuclear export. Multiple novel treatment modalities have been shown to ameliorate symptoms of thalassemia. These are directed either

Figure 1. KPT-251 inhibition of HSP70 export from the nucleus via exportin Xpo1 improves ineffective erythropoiesis in β-thalassemia. Under normal conditions, nuclear HSP70 protects GATA1 from caspase-3 cleavage. In thalassemia, excess free a-globin chains sequester HSP70 in the cytoplasm, prevent normal GATA1 target expression, and result in ineffective erythropoiesis. Treating erythroblasts with the Xpo1 inhibitor KPT-251 increases nuclear levels of HSP70, rescues GATA1 from caspase-3 cleavage, and improves terminal erythroid development.

2188

haematologica | 2020; 105(9)

Editorials at improving the balance between unbound a-globin and non-a-globin chains or correcting the ineffective erythropoiesis. Modified TFG-β family receptor antagonists like Sotatercept (ACE-011) and Luspatercept (ACE-536) block ligand binding to ActR-II receptors, and subsequent activation of the SMAD4 signaling pathway,4 improving erythroid maturation and red cell production. Correction of the aberrant ratio of unbound a-globin to non-a-globin chains has been achieved by successful gene therapy by CRISPR Therapeutics, backed by Boston's Vertex Pharmaceuticals. The somatic cell therapy, named CTX001, uses edited patient’s own hematopoietic stem cells (HSC) to stimulate fetal hemoglobin production.5 Targeting intracellular localization of HSP70 via XPO1 inhibition potentially merges these two treatment goals. Several lines of evidence suggest that erythroblasts use molecular chaperones to partition unstable excess a-globin chains during erythroid development,6-8 so it follows that targeting such chaperones could be useful in β-thalassemia when excess a-globin tetramers accumulate and create havoc. Numerous groups have noted that the molecular chaperone Hsp70 accumulates to high levels in erythroblasts9-11 and are important for streamlining erythroid maturation.11 Normal human erythroid maturation requires a transient activation of caspase-3 at the later stages of maturation in order to prevent excessive erythrocyte production. Activated caspases can cleave GATA-1 leading to maturation arrest and/or apoptosis.12 Ribeil et al. showed that Epo causes Hsp70 to translocate into the nucleus, bind GATA-1, and protect it from caspase-3 cleavage. Conversely, during Epo deprivation, Hsp70 is excluded from the nucleus and GATA-1 is cleaved by caspase-3, causing apoptotic death.13 Thus, alteration of the intracellular location of Hsp70 appears to play a critical role in erythroblast viability (Figure 1). The ineffective erythropoiesis observed in β-thalassemia is characterized by accelerated erythroid differentiation, maturation arrest and apoptosis at the polychromatophilic stage. During maturation of human β-thalassemia erythroblasts, HSP70 is sequestrated in the cytoplasm (Figure 1) directly by excess free a-globin chains.14 GATA-1 is no longer protected, resulting in end-stage maturation arrest and apoptosis. Transduction of a nuclear-targeted HSP70 mutant or a caspase-3-uncleavable GATA-1 mutant restores terminal maturation of β-thalassemia erythroblasts.14 In this issue of Haematologica, Guillem et al.3 follow up on this mechanism to show that exportin-1 (XPO1) regulates the nucleocytoplasmic location of HSP70 in erythroid progenitors under normal conditions. Guillem et al. confirm that treating erythroblasts with the Xpo1 inhibitor KPT-251 increased nuclear levels of HSP70, rescued GATA1 from caspase-3 cleavage, and improved terminal erythroid development (Figure 1). Although the use of selective inhibitors of nuclear export (SINE) for the treatment of lymphomas and multiple myeloma has been well reported,15-17 this is the first

haematologica | 2020; 105(9)

study showing the novel mechanism of SINE as a potential therapy for enhanced effective erythropoiesis in βthalassemia. The clinical diversity of thalassemia makes it hard to design a one-size-fits-all therapy, and current therapies are mainly aimed at improving one or more of the underlying pathologies, most importantly transfusion dependence and iron overload. In this regard, a targeted inhibitor aimed at Xpo1 promises a more specific and low-risk treatment. Targeting HSP70 nuclear translocation in erythroid precursors represents a novel and exciting therapeutic option to ameliorate the ineffective erythropoiesis of β-thalassemia.

References 1. De Sanctis V, Kattamis C, Canatan D, et al. β-thalassemia distribution in the old world: an ancient disease seen from a historical standpoint. Mediterr J Hematol Infect Dis. 2017;9(1):e2017018. 2. Thein SL. Molecular basis of β thalassemia and potential therapeutic targets. Blood Cells Mos Dis. 2018;70:54-65. 3. Guillem F, Dussiot M, Colin E, et al. XPO1 regulates erythroid differentiation and is a new target for the treatment of β-thalassemia. Haematologica. 2020;105(9):2240-2249 4. Makis A, Hatzimichael E, Papassotiriou I, et al. 2017 Clinical trials update in new treatments of β‐thalassemia. Am J Hematol. 2016;91(11):1135-1145. 5. Lin MI, Paik E, Mishra B, et al. CRISPR/Cas9 genome editing to treat sickle cell disease and B-thalassemia: re-creating genetic variants to upregulate fetal hemoglobin appear well-tolerated, effective and durable. Blood. 2017;130(Supplement 1):284. 6. Shaeffer JR. Evidence for soluble alpha-chains as intermediates in hemoglobin synthesis in the rabbit reticulocyte. Biochem Biophys Res Comm. 1967;28(4):647-652. 7. Tavill AS, Grayzel AI, Vanderhoff GA, London IM. The control of hemoglobin synthesis. Trans Assoc Am Physicians. 1967;80:305313. 8. Olivieri NF, Weatherall DJ. Clinical aspects of beta thalassemia. In: Steinberg MH, Forget BG, Higgs DR, Nagel RL, eds. Disorders of Hemoglobin. Cambridge, United Kingdom: Cambridge University Press; 2001. pp. 277-341. 9. Morimoto R, Fodor E. Cell-specific expression of heat shock proteins in chicken reticulocytes and lymphocytes. J Cell Biol. 1984;99 (4):1316-1323. 10. Singh MK, Yu J. Accumulation of a heat shock-like protein during differentiation of human erythroid cell line K562. Nature. 1984;309(5969):631-633. 11. Banerji SS, Laing K, Morimoto RI. Erythroid lineage-specific expression and inducibility of the major heat shock protein HSP70 during avian embryogenesis. Genes Dev. 1987;1(9):946-953. 12. De Maria R, Zeuner A, Eramo A, et al. Negative regulation of erythropoiesis by caspase-mediated cleavage of GATA-1. Nature. 1999;401(6752):489-493. 13. Ribeil JA, Zermati Y, Vandekerckhove J, et al. Hsp70 regulates erythropoiesis by preventing caspase-3-mediated cleavage of GATA-1. Nature. 2007;445(7123):102-105. 14. Jean-Benoît A, Ribeil JA, Guillem F, et al. HSP70 sequestration by free a-globin promotes ineffective erythropoiesis in β-thalassaemia. Nature. 2014;514(7521):242-246. 15. Parikh K, Cang S, Sekhri A, Liu D. Selective inhibitors of nuclear export (SINE)–a novel class of anti-cancer agents. J Hematol Oncol. 2014;7:78. 16. Hing ZA, Fung HYJ, Ranganathan P, et al. Next-generation XPO1 inhibitor shows improved efficacy and in vivo tolerability in hematological malignancies. Leukemia. 2016;30(12):2364-2372. 17. Gandhi UH, Senapedis W, Baloglu E, et al. Clinical implications of targeting XPO1-mediated nuclear export in multiple myeloma. Clin Lymphoma Myeloma Leuk. 2018;18(5):335-345.

2189

Editorials

Combination therapy with interferon and ruxolitinib for polycythemia vera and myelofibrosis: are two drugs better than one? Richard T. Silver Richard T. Silver Myeloproliferative Neoplasms Center, Weill Cornell Medicine Division of Hematology-Oncology, New York, NY, USA E-mail: RICHARD T. SILVER - rtsilve@med.cornell.edu doi:10.3324/haematol.2020.256602

pproximately 35 years ago, recombinant interferon alpha-2 (rIFNa-2) was first reported to control myeloproliferation in essential thrombocythemia,1 polycythemia vera (PV),2 and the hyperproliferative phase of primary myelofibrosis (MF).3 In the ensuing years, these observations have been substantiated in thousands of patients with myeloproliferative neoplasms, indicating that rIFNa-2 is safe and effective for alleviating symptoms, diminishing organomegaly, reducing elevated platelet and white blood cell counts and, in PV, maintaining and controlling the hematocrit4 and decreasing the level of the JAK2 V617F allele burden.5 The effect of rIFNa on the molecular markers of essential thrombocythemia and MF have also been noted.6,7 The finding of a significant reduction of JAK2 V617F level in a subset of PV patients treated with a long-acting form of interferon owing to its pegylation (PEG- rIFNa-2a)8 led to more enthusiastic use of this treatment. Subsequently, normalization of marrow morphology and reduction of fibrosis were demonstrated.9 Interferon monotherapy normalized elevated blood cell counts within a few months, but the major molecular responses and marrow changes required 3-5 years of treatment, especially in MF.10 Discontinuation of rIFNa in PV could be sustained for a number of years.3,8 Patients with primary or secondary MF usually require continued treatment.10 As with all potent drugs used in treating hematologic cancers, rIFNa is associated with a significant number of side effects, but most commonly, constitutional symptoms such as fatigue, muscle aches, lethargy and, occasionally, fever.7 This is presumably due to cytokine effects. Discontinuation rates in rIFNa studies have ranged from 10-30% after 1-2 years,4,7,8 depending on the dose and frequency of administration of the drug, the severity of associated side effects, and the belief and enthusiasm of the physicians and patients, respectively, regarding its value. Recently, it has been suggested that considering side effects to be all related to dosing, per se, is too simplistic. Danish investigators indicated that chronic inflammation induced by interferon renders patients intolerant or refractory to the treatment.11 Adding an anti-inflammatory drug with an anti-JAK2 V617F effect, such as ruxolitinib, seemed logical (Figure 1).

The study by SĂ¸rensen et al., reported in this issue of Haematologica comprised 50 patients with PV or with primary or secondary MF, all of whom were resistant or refractory to rIFNaâ&#x2C6;&#x2019;2.12 After the addition of ruxolitinib to the treatment regimen, of 32 PV patients, ten (31%) achieved complete or partial remission; of 18 MF patients, eight (44%) achieved complete or partial remission. Combination treatment seemed to speed the time to remission, improve blood counts, reduce marrow cellularity and fibrosis and decrease the JAK2 allele burden, all with acceptable toxicity. The dropout rate at the end of 2 years was 6% for PV patients and 32% for MF patients. These results are most interesting and encouraging, but require confirmation, because it was a single-arm study. Other limiting features, as the authors point out, include the small number of cases, the lack of a dose-finding phase, and the duration of only 2 years of treatment. Detailed molecular reporting of the MF patients in particular would have been of interest, since it is known that initial molecular profile affects response to interferon treatment in early MF.10 Nevertheless, conceptually, the basis for the use of two potentially synergistic agents with different activities in untreated patients is logical. This then leads to the question of whether or not this combination should be used as initial therapy. The current tendency worldwide is to use hydroxyurea as firstline treatment when needed in PV and to adopt a watchand-wait attitude in treating primary MF. This approach seems illogical since it is not consistent with our concepts of cancer treatment in general, wherein treatment of early cancer yields results superior to those of treating metastatic disease. The reduction and/or elimination of symptoms of disease and avoidance of progression with agents that have a biological basis for use and that are tolerable seems far more rational. This requires testing in the immediate future.

References 1. Linkesch W, Gisslinger H, Ludwig H, Flener R, Sinzinger H. [Therapy with interferon (recombinant IFN-alpha-2C) in myeloproliferative diseases with severe thrombocytoses.] Acta Med Austriaca. 1985;12(5):123-127. Article in German.

Figure 1. Overcoming inflammatory-resistance to interferon with ruxolitinib.

2190

haematologica | 2020; 105(9)

Editorials 2. Silver RT. Recombinant interferon-alpha for treatment of polycythaemia vera. Lancet. 1988;2(8607):403. 3. Gilbert HS. Long term treatment of myeloproliferative disease with interferon-alpha-2b: feasibility and efficacy. Cancer. 1998;83(6): 1205-1213. 4. Silver RT. Long-term effects of the treatment of polycythemia vera with recombinant interferon-alpha. Cancer. 2006;107(3):451-458. 5. Jones AV, Silver RT, Waghorn K, et al. Minimal molecular response in polycythemia vera patients treated with imatinib or interferon alpha. Blood. 2006;107(8):3339-3341. 6. Barbui T, Tefferi A, Vannucchi AM, et al. Philadelphia chromosomenegative classical myeloproliferative neoplasms: revised management recommendations from European LeukemiaNet. Leukemia. 2018;32(5):1057-1069. 7. Silver RT, Kiladjian JJ, Hasselbalch HC. Interferon and the treatment of polycythemia vera, essential thrombocythemia and myelofibrosis. Expert Rev Hem. 2013;6(1):49-58.

8. Kiladjian JJ, Cassinat B, Chevret S, et al. Pegylated interferon-alfa-2a induces complete hematologic and molecular responses with low toxicity in polycythemia vera. Blood. 2008;112(8):3065-3072. 9. Pizzi M, Silver RT, Barel AC, Orazi A. Recombinant interferon-a in myelofibrosis reduces bone marrow fibrosis, improves its morphology and is associated with clinical response. Mod. Pathol. 2015;28(10):1315-1323. 10. Silver RT, Barel AC, Lascu E, et al. The effect of initial molecular profile on response to recombinant interferon-a (rIFNa) treatment in early myelofibrosis. Cancer. 2017;123(14):2680-2687. 11. Mikkelsen SU, Kjaer L, Bjorn ME, et al. Safety and efficacy of combination therapy of interferon a-2 and ruxolitinib in polycythemia vera and myelofibrosis. Cancer Med. 2018;7(8):3571-3581. 12. Sørensen AL, Mikkelsen SU, Knudsen TA, et al. Ruxolitinib and interferon a2 combination therapy for patients with polycythemia vera or myelofibrosis: a phase II study. Haematologica. 2020;105(9): 2262-2272.

Mysteries of partial dihydroorotate dehydrogenase inhibition and leukemia terminal differentiation Yogen Saunthararajah Department of Translational Hematology and Oncology Research, Taussig Cancer Institute, Cleveland Clinic, Cleveland, OH, USA E-mail: YOGEN SAUNTHARARAJAH - saunthy@ccf.org doi:10.3324/haematol.2020.254482

t is reasonable to wonder why inhibiting dihydroorotate dehydrogenase (DHODH), a protean and vital metabolic enzyme, would be expected to solve, not exacerbate, prevalent oncotherapy problems of toxicity and resistance. Yet, in addition to ASLAN003, described in this issue of Haematologica,1 at least four other DHODH inhibitors are being developed for oncotherapy.2 DHODH is the sole mitochondrial enzyme in the pathway of de novo pyrimidine synthesis, which makes pyrimidine nucleobases from glutamine and aspartate. Pyrimidines are not just building blocks for DNA and RNA, but are also key cofactors for glycoprotein, glycolipid and phospholipid synthesis. Moreover, the reaction that DHODH executes, reduction of dihydroorotate to orotate, is coupled to mitochondrial electron transport, to manufacture ATP independently of glucose and the Krebs cycle. Not surprisingly, therefore, DHODH is vital - its knock-out is lethal. Surprisingly, however, treatment of malignant cells with clinically tolerable concentrations of DHODH inhibitors induces not the cytotoxicity (apoptosis) expected from most anti-metabolite oncotherapeutics but terminal differentiation. Unbiased analyses illustrate this: of the thousands of genes most significantly up- and down-regulated by ASLAN003 treatment of acute myeloid leukemia (AML) cells, most are the same genes coordinately up- and downregulated during normal myeloid differentiation into granulocytes or monocytes (Figure 1A). Such ready recapitulation of normal lineage progression is rendered less astonishing upon recognition that malignant cells express very high levels of lineage differentiation-driving master transcription factors to begin with, e.g., SPI1, CEBPA, RUNX1 in AML cells.3,4 One function of these lineage master transcription factors is to activate lineage differentiation programs, but another is to cooperate with MYC for high-grade activation of proliferation – coupling of exponential proliferation and onward differentiation in this way is a feature of metazoan haematologica | 2020; 105(9)

biology sometimes called ‘transit amplification’. Oncogenic mutations decouple exponential proliferation from onward differentiation to create malignant self-replication.4 In short, partial DHODH inhibition reconnects circuitry already present to release malignant cells to complete lineage journeys already begun (Figure 1A). This modality for leukemia/cancer cytoreduction is worthy of investment for three fundamental reasons. First, cell cycle exiting by terminal differentiation does not require the p53 apoptosis machinery that mediates cytoreduction by antimetabolite chemotherapeutics in general, and thus offers activity even in chemorefractory disease with p53system mutations.4 Second, DHODH inhibitor-mediated induction of terminal differentiation is not restricted to rare AML or genetic subtypes of cancer, although differences in pyrimidine metabolism between histologically diverse cancers may influence this activity (discussed below). Last but not least, non-cytotoxic differentiation-based oncotherapy can spare normal dividing cells essential for health/normal lifespan, offering a good therapeutic index.4 To efficiently realize these fundamentals in the clinic, however, an obvious question needs an answer: how exactly does partial inhibition of DHODH, a protean metabolic enzyme, reconnect cancer cells to terminal lineage fates intended by their master transcription factor content? Given the contributions of DHODH and pyrimidines to so many fundamental cellular functions, it is difficult to know where to begin to answer this question. Fortunately, work spanning decades has provided excellent clues. One important observation is that the small molecule cyclopentenyl cytosine (CPEC), which inhibits the last step in de novo pyrimidine synthesis, uridine triphosphate (UTP) amination into cytidine triphosphate (CTP) by CTP synthase 2 (CTPS2), also releases AML and solid tumor cancer cells to terminal lineage fates.5,6 Moreover, exogenous cytidine that restored CTP but not UTP pools, and exogenous uridine 2191

Editorials

that restored both UTP and CTP pools, prevented induction of terminal differentiation by the DHODH inhibitor leflunomide, while neither deoxycytidine nor deoxythymidine could do this.7 CTP is the ribonucleotide present at limiting concentrations in our cells, having concentrations ~3fold lower than those of UTP or guanosine triphosphate (GTP), and ~30-fold lower than the concentration of adenosine triphosphate (ATP).8 CTP is also the ribonucleotide most upregulated in cancer versus normal cells, mediated at least in part by upregulation of CTPS2.8,9 It is of course no surprise that basic building blocks such as CTP are essential for cell proliferation, including malig-

nant proliferation. But how is CTP linked to lineage progression? Several mitochondrial products serve as cofactors for key epigenetic enzymes that remodel lineage-differentiation genes for activation or repression (Figure 1B).10 Alphaketoglutarate (AKG) is an essential cofactor for dioxygenases including TET DNA methylcytosine dioxygenases, and Jumonji domain-containing histone demethylases (KDM), which are components of multiprotein complexes (coactivators) that remodel chromatin to activate terminal differentiation genes. Acetyl-CoA is a cofactor for histone lysine acetyltransferases (HAT), which are also coactivator components. Flavin adenine dinucleotide (FAD) is a cofactor for

Figure 1. The dihydroorotate dehydrogenase inhibitor ASLAN003 recapitulates in acute myeloid leukemia cells the coordinated up- and down-regulation of thousands of genes that occurs with normal terminal granulocyte or monocyte differentiation. (A) The 1,000 genes most significantly up- or down-regulated upon addition of ASLAN003 to KG1 or MOLM14 acute myeloid leukemia (AML) cells were examined for their expression pattern in normal hematopoietic stem cells (HSC), multipotent progenitors (MPP), common myeloid progenitors (CMP), granulocyte-monocyte progenitors (GMP), granulocytes (gran) and monocytes (mono) (data from BloodPool15), and found to be genes normally up- or down-regulated with terminal granulocyte or monocyte differentiation. Experimental details of ASLAN003 treatment are described by Zhou et al.,1 gene expression by RNA-sequencing Geo Database GSE128950. (B) Proliferation and lineage differentiation are coupled through mitochondrial metabolism. AKG: alpha-ketoglutarate; Ac-CoA: acetyl-CoA; FAD: flavin adenine dinucleotide; NAD+: nicotinamide adenine dinucleotide; CTP: cytidine triphosphate. The putative epigenetic protein in eukaryotic cells for which CTP is a cofactor is unknown. (C) The dihydroorotate dehydrogenase (DHODH) inhibitor ASLAN003 acutely reconfigures pyrimidine metabolism in KG1 and MOLM14 AML cells (data from GSE1289501). Analysis using Broad Institute Morpheus software, P-value Marker Selection, 100 permutations.

2192

haematologica | 2020; 105(9)

Editorials

amine oxidase domain-containing histone demethylases (KDM1A, KDM1B), which are components of corepressor complexes that repress terminal differentiation genes. Nicotinamide adenine dinucleotide (NAD+) is a cofactor for Sirtuin histone deacetylases (SIRT1, SIRT2) that participate in the regulation of tissue stem-cell genes. In short, sensing for these mitochondrial outputs is another way in which exponential proliferation and onward differentiation are coupled (Figure 1B), powerfully demonstrated by the natural experiment of neoplastic evolution: for example, recurrent mutations in isocitrate dehydrogenase genes (IDH1, IDH2) in cancer reduce or antagonize AKG to decouple exponential proliferation and onward differentiation.10 Since CTP is disproportionately elevated in malignancy, and interventions that decrease CTP reconnect to onward differentiation, it can be theorized that CTP is a cofactor for corepressor proteins that repress terminal-differentiation genes. In fact, chromosome partitioning ParB proteins in bacteria, which also serve as platforms for DNA-condensing proteins (condensins), require CTP as a cofactor.11 The identity of the putative corepressor component in eukaryotic cells for which CTP is a cofactor is, however, unknown. An alternative theory is that CTP is a negative regulator of coactivators that activate terminal differentiation genes, whose identity is also unknown. These observations are relevant to clinical translation. If a decrease in CTP and/or UTP mediates the induction of terminal differentiation by DHODH inhibitors, then the CTP/UTP content in cancer/leukemia cells is a candidate pharmacodynamic biomarker to guide the design of drug regimens. Moreover, the most likely mechanism for treatment failure can be anticipated: the pyrimidine metabolism network will respond automatically to preserve the amounts of CTP and/or UTP. For example, DHODH inhibitor-mediated depletion of CTP and UTP will relieve their allosteric inhibition of uridine cytidine kinase 2 (UCK2), which salvages cytidines and uridines from the extracellular environment, to automatically dampen any CTP/UTP decrease.12 CTP and/or UTP decreases can also be expected to trigger compensating shifts in the expression of key enzymes of pyrimidine metabolism: for example, ASLAN003 treatment of AML cells acutely upregulated expression of deoxycytidine kinase (DCK), which salvages deoxycytidines (Figure 1C). Treatment resistance that emerges automatically from the pyrimidine metabolism network in this way will be rapid and, importantly, will not be solved by simple escalation of DHODH-inhibitor dosages, since this would worsen the therapeutic index, which is a rationale for clinical development in the first place. Fortunately, compensatory metabolic responses can potentially be turned to advantage: DCK activates several oncotherapeutic prodrugs, e.g., decitabine that, like partial DHODH inhibition, can also operate in a non-cytotoxic, differentiation-based regime,13,14 and incorporation of DCKdependent prodrugs can potentially be timed to DHODH inhibitor-mediated DCK upregulation. Other pyrimidine metabolism responses, however, could be adverse to such combinations: the pyrimidine metabolism enzyme SAMand HD-containing deoxynucleoside triphosphate triphosphohydrolase 1 (SAMHD1) was also automatically upregulated by ASLAN003 (Figure 1C) and catabolizes the active nucleotide forms of decitabine and cytarabine which are haematologica | 2020; 105(9)

routinely used to treat AML. Ultimately, therefore, candidate solutions for resistance will require thorough experimental evaluation. Apoptosis/cytotoxicity is the standard pathway goal of oncotherapy, but is burdened by systemic toxicity and frequent futility, because of recurrent genetic attenuation of the p53-apoptosis axis in cancers and leukemias. DHODH inhibitors are appealing because of their potential to bolster a lagging inventory of p53-independent oncotherapeutics that cytoreduce malignancies by terminal differentiation instead. The pyrimidine metabolism network, however, will compensate automatically for reductions in CTP/UTP achieved by clinically viable doses of DHODH inhibitors. Our conditioned response to treatment failure is to escalate dosages toward more profound antimetabolite effects and cytotoxicity. Appreciation for, and fidelity to, terminal differentiation as the opportunity and pathway goal to be seized can help guard against these instincts â&#x20AC;&#x201C; been there, done that! â&#x20AC;&#x201C; and increase possibilities for clinical success. Funding National Heart, Lung and Blood Institute PO1 HL146372; National Cancer Institute P30 CA043703; National Cancer Institute RO1 CA204373

References 1. Zhou J, Quah JY, Ng Y, et al. ASLAN003, a potent dihydroorotate dehydrogenase inhibitor for differentiation of acute myeloid leukemia. Haematologica. 2020;105(9):2286-2297. 2. Madak JT, Bankhead A 3rd, Cuthbertson CR, Showalter HD, Neamati N. Revisiting the role of dihydroorotate dehydrogenase as a therapeutic target for cancer. Pharmacol Ther. 2019;195:111-131. 3. Gu X, Ebrahem Q, Mahfouz RZ, et al. Leukemogenic nucleophosmin mutation disrupts the transcription factor hub that regulates granulomonocytic fates. J Clin Invest. 2018;128(10):4260-4279. 4. Velcheti V, Schrump D, Saunthararajah Y. Ultimate precision: targeting cancer but not normal self-replication. Am Soc Clin Oncol Educ Book. 2018;38:950-963. 5. Huang M, Wang Y, Collins M, Graves LM. CPEC induces erythroid differentiation of human myeloid leukemia K562 cells through CTP depletion and p38 MAP kinase. Leukemia. 2004;18(11):1857-1863. 6. Bierau J, van Gennip AH, Helleman J, van Kuilenburg AB. The cytostatic- and differentiation-inducing effects of cyclopentenyl cytosine on neuroblastoma cell lines. Biochem Pharmacol. 2001;62(8):10991105. 7. Huang M, Wang Y, Collins M, Mitchell BS, Graves LM. A77 1726 induces differentiation of human myeloid leukemia K562 cells by depletion of intracellular CTP pools. Mol Pharmacol. 2002;62(3):463472. 8. Traut TW. Physiological concentrations of purines and pyrimidines. Mol Cell Biochem. 1994;140(1):1-22. 9. Williams JC, Kizaki H, Weber G, Morris HP. Increased CTP synthetase activity in cancer cells. Nature. 1978;271(5640):71-73. 10. Nieborak A, Schneider R. Metabolic intermediates - cellular messengers talking to chromatin modifiers. Mol Metab. 2018;14:39-52. 11. Soh YM, Davidson IF, Zamuner S, et al. Self-organization of parS centromeres by the ParB CTP hydrolase. Science. 2019;366(6469):1129-1133. 12. Huang M, Graves LM. De novo synthesis of pyrimidine nucleotides; emerging interfaces with signal transduction pathways. Cell Mol Life Sci. 2003;60(2):321-336. 13. Saunthararajah Y, Sekeres M, Advani A, et al. Evaluation of noncytotoxic DNMT1-depleting therapy in patients with myelodysplastic syndromes. J Clin Invest. 2015;125(3):1043-1055. 14. Ng KP, Ebrahem Q, Negrotto S, et al. p53 independent epigeneticdifferentiation treatment in xenotransplant models of acute myeloid leukemia. Leukemia. 2011;25(11):1739-1750. 15. Rapin N, Bagger FO, Jendholm J, et al. Comparing cancer vs normal gene expression profiles identifies new disease entities and common transcriptional programs in AML patients. Blood. 2014;123(6):894904.

2193

Editorials

Towards genomic-based prognostication and precision therapy for diffuse large B-cell lymphoma Marthe Minderman1,2 and Steven T. Pals1,2 1 Department of Pathology, Cancer Center Amsterdam, Amsterdam UMC, University of Amsterdam and 2Lymphoma and Myeloma Center Amsterdam – LYMMCARE, Amsterdam, the Netherlands.

E-mail: STEVEN T. PALS - s.t.pals@amsterdamumc.nl doi:10.3324/haematol.2020.255448

he development of precision medicine for diffuse large B-cell lymphoma (DLBCL) is complicated by its great clinical and molecular heterogeneity. Twenty years ago, two distinct cell-of-origin (COO) subtypes were identified based on distinct gene expression profiles that reflect different stages of B-cell development, i.e., germinal center B-cell-like (GCB) and activated B-celllike (ABC) DLBCL.1 The ongoing revolution in genomics has since shed light on the genetic landscape of these subtypes. Whereas ABC DLBCL is characterized by frequent mutations in nuclear factor-κB (NF-κB) pathway drivers and intermediates, including TNFAIP3, MYD88, CARD11 and CD79B, as well as loss of cell cycle regulators CDKN2A and CDKN2B, GCB DLBCL carry frequent mutations in epigenetic modifiers, such as CREBBP, KMT2D and EZH2.2,3 GCB DLBCL patients have a more favorable prognosis and a better clinical response to standard R-CHOP immunochemotherapy than those with ABC DLBCL. Nevertheless, the clinical outcome is heterogenous in both subtypes with an unfavorable outcome in a substantial proportion of patients, also in individuals with GCB DLBCL.1,4 Currently, the impact of somatic mutations and other genomic aberrations on the clinical outcome and therapy response is still not completely understood and there is a high need for targeted precision therapy. In this issue of Haematologica, Bolen et al. report the frequency and prognostic impact of genomic alterations in 499 untreated DLBCL patients enrolled in the GOYA study (clinicaltrials.gov identifier: NCT01287741).5 Using a well-validated targeted next-generation sequencing (NGS) approach, the authors demonstrate that only alterations of the BCL2 gene, translocations as well as single nucleotide variants (SNV), were significantly associated with reduced progression-free survival (PFS), independent of other molecular or clinical factors, including COO or the International Prognostic Index (IPI). In line with previous studies, BCL2 was the most frequently mutated gene in GCB DLBCL and there was a strong correlation between a BCL2 translocation and presence of BCL2 mutations, which presumably are a consequence of aberrant somatic hypermutation.3,6 While BCL2 translocations and SNV were highly enriched in the GCB subtype, BCL2 gene amplifications were more frequently detected in the ABC subtype. The findings suggest that a select subset of DLBCL patients may likely benefit from pharmacological inhibition of BCL2, as an addition to standard immunochemotherapy. The highly selective BCL2 inhibitor venetoclax strongly improved PFS in chronic lymphocytic leukemia (CLL) patients and is currently under investigation for DLBCL.7,8 Preliminary findings 2194

from the phase II CAVALLI trial indicate that addition of venetoclax to R-CHOP therapy can improve therapy outcome in patients with BCL2-positive lymphomas. The largest benefit was observed for BCL2-translocated and double-hit lymphomas, suggesting that the addition of venetoclax may be of particular value for these therapyresistant and aggressive subgroups. Although the other genetic aberrations identified by Bolen et al. did not offer prognostic value over well-established risk factors, such as COO and IPI, they nevertheless provide biological insights that are crucial for the design of precision therapies (Figure 1). For example, in line with previous studies, the authors demonstrate that a large fraction of GCB DLBCL (16%) carry gain-of-function mutations in the transcriptional repressor EZH2, implying that these patients could benefit from treatment with EZH2 inhibitors, such as tazemetostat. Indeed, preliminary results from a phase II clinical trial in 165 DLBCL and follicular lymphoma (FL) patients demonstrate that tazemetostat achieves favorable clinical responses in patients carrying activating EZH2 mutations.9 Likewise, the presence of inactivating mutations in the transcriptional activators KMT2D, CREBBP, EP300 and MEF2B implies that a subset of patients could benefit from treatment with histone deacetylase (HDAC) inhibitors. A phase II clinical trial evaluating the safety and therapeutic benefits of pan-HDAC inhibitor panobinostat showed durable responses in 11 out of 40 DLBCL patients, an effect that may be associated with mutations in transcriptional activator MEF2B.10 Activating mutations in MEF2B are present in approximately 10% of DLBCL and FL patients and contribute to lymphomagenesis by enhancing the transcription of the BCL6 oncogene.11 Two independent landmark studies published in 2018 by Schmitz et al.12 and Chapuy et al.13 have assessed the occurrence of somatic mutations, copy number alterations (CNA) and structural variants (SV) in a large cohort of DLBCL patients. Building on these observations, Wright et al. recently developed the ‘LymphGen algorithm’, which calculates the probability that a given tumor belongs to one of seven subtypes, based on its genetic features.14 The identified genetic subtypes are associated with differential responses to immunochemotherapy and, moreover, provide ample opportunity for the design of novel targeted (combination) treatments. In accordance with these and other previous studies, the study of Bolen et al.5 reports that B-cell receptor (BCR) complex component CD79B and Toll-like receptor (TLR) adaptor protein MYD88 are among the most frequently mutated genes in ABC DLBCL. These mutations define the ‘MCD’ cluster of Wright et al.,14 haematologica | 2020; 105(9)

Editorials

Figure 1. Prevalence of mutations in genes involved in various functional pathways in diffuse large B-cell lymphoma (DLBCL). Most frequently mutated genes as identified by Bolen et al.5 in a cohort of 482 DLBCL patients arranged according to the affected pathways. NF-κB: nuclear factor kappa B; TLR: Toll-like receptor; BCR: B-cell receptor.

comprising tumors characterized by a high prevalence and co-occurrence of MYD88 and CD79B mutations that are almost exclusively classified as ABC DLBCL and have an unfavorable prognosis.15 Intriguingly, this genetic subgroup is linked to primary extranodal lymphomas, including lymphomas arising in the central nervous system (CNS), ocular vitreo-retina and testis, all considered ‘immune-privileged’ sites as they tolerate allografts and permit only selective entrance of immune cells.16-19 Importantly, Wilson et al. established that ABC DLBCL harboring mutations in CD79B, particularly those with concurrent MYD88 mutations, were highly responsive to treatment with ibrutinib, a selective Bruton's tyrosine kinase (BTK) inhibitor.20 These observations suggest that (extranodal) lymphomas belonging to the MCD subtype could also be targeted by inhibition of BCR signaling. Indeed, a phase Ib study in a panel of 18 primary central nervous system lymhpomas (PCNSL) demonstrated that ibrutinib monotherapy reduced tumor mass in 94% of patients.21 Additionally, a second phase I clinical trial haematologica | 2020; 105(9)

showed clinical responses to ibrutinib in 10 out of 13 PCNSL patients, including five complete responses.22 Collectively, these suggest that patients with other primary extranodal lymphomas belonging to the MCD subtype, such as primary testicular or vitreoretinal lymphoma, might also benefit from treatment with BCR pathway inhibitors. In conclusion, the study by Bolen et al.5 fully confirms the previously described genetic heterogeneity and complexity of DLBCL. As a consequence of this complexity, well-established prognostic classifiers, such as COO and IPI, can only partially account for the differential responses to R-CHOP (and related) immunochemotherapy. The identification of alterations of the BCL2 gene as the only genetic abnormalities significantly associated with reduced PFS points towards targeting BCL2 as a rational addition to standard immunochemotherapy. Although not of (independent) prognostic value in the context of standard immunochemotherapy, genetic abnormalities defining potentially druggable targets/pathways were 2195

Editorials

identified in a large proportion of the tumors across the distinct COO subtypes (Figure 1). Applying the publicly available LymphGen algorithm on the GOYA dataset could help classify patients into well-defined molecularly and clinically distinct subgroups. These newly characterized subsets can identify patients with an unfavorable prognosis and may guide the development of new precision therapies for these aggressive lymphomas. Funding Funded by a grant from Lymph&Co.

References 1. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-511. 2. Pasqualucci L, Trifonov V, Fabbri G, et al. Analysis of the coding genome of diffuse large B-cell lymphoma. Nat Genet. 2011;43(9):830-837. 3. Lohr JG, Stojanov P, Lawrence MS, et al. Discovery and prioritization of somatic mutations in diffuse large B-cell lymphoma (DLBCL) by whole-exome sequencing. Proc Natl Acad Sci U S A. 2012;109(10):3879-3884. 4. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(25):1937-1947. 5. Bolen CR, Klanova M, Trneny M, et al. Prognostic impact of somatic mutations in diffuse large B-cell lymphoma and relationship to cellof-origin: data from the phase III GOYA study. Haematologica. 2020;105(9):2298-2307. 6. Schuetz JM, Johnson NA, Morin RD, et al. BCL2 mutations in diffuse large B-cell lymphoma. Leukemia. 2012;26(6):1383-1390. 7. Seymour JF, Kipps TJ, Eichhorst B, et al. Venetoclax-Rituximab in Relapsed or Refractory Chronic Lymphocytic Leukemia. N Engl J Med. 2018;378(12):1107-1120. 8. Morschhauser F, Feugier P, Flinn IW, et al. Venetoclax plus rituximab, cyclophosphamide, doxorubicin, vincristine and prednisolone (RCHOP) improves outcomes in BCL2-positive first-line diffuse large B-cell lymphoma (DLBCL): first safety, efficacy and biomarker analyses from the phase II CAVALLI study. Blood. 2018;132 (Supplement 1):782.

9. Morschhauser F, Salles G, McKay P, et al. Interim Report from a Phase 2 Multicenter Study of Tazemetostat, an EZH2 Inhibitor: Clinical Activity and Favorable Safety in Patients with Relapsed or Refractory B-Cell Non-Hodgkin Lymphoma. Clin Lymphoma Myeloma Leuk. 2017;17(Suppl 2):S380-381. 10. Assouline SE, Nielsen TH, Yu S, et al. Phase 2 study of panobinostat with or without rituximab in relapsed diffuse large B-cell lymphoma. Blood. 2016;128(2):185-194. 11. Ying CY, Dominguez-Sola D, Fabi M, et al. MEF2B mutations lead to deregulated expression of the oncogene BCL6 in diffuse large B cell lymphoma. Nat Immunol. 2013;14(10):1084-1092. 12. Schmitz R, Wright GW, Huang DW, et al. Genetics and Pathogenesis of Diffuse Large B-Cell Lymphoma. N Engl J Med. 2018;378 (15):1396-1407. 13. Chapuy B, Stewart C, Dunford AJ, et al. Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes. Nat Med. 2018;24(5):679-690. 14. Wright GW, Huang DW, Phelan JD, et al. A Probabilistic Classification Tool for Genetic Subtypes of Diffuse Large B Cell Lymphoma with Therapeutic Implications. Cancer Cell. 2020;37 (4):551-568.e14. 15. Vermaat JS, Somers SF, de Wreede LC, et al. MYD88 mutations identify a molecular subgroup of diffuse large B-cell lymphoma with an unfavorable prognosis. Haematologica. 2020;105(2):424-434. 16. Kraan W, Horlings HM, van Keimpema M, et al. High prevalence of oncogenic MYD88 and CD79B mutations in diffuse large B-cell lymphomas presenting at immune-privileged sites. Blood Cancer J. 2013;3(9):e139. 17. Yonese I, Takase H, Yoshimori M, et al. CD79B mutations in primary vitreoretinal lymphoma: Diagnostic and prognostic potential. Eur J Haematol. 2019;102(2):191-196. 18. Schrader AMR, Jansen PM, Willemze R, et al. High prevalence of MYD88 and CD79B mutations in intravascular large B-cell lymphoma. Blood. 2018;131(18):2086-2089. 19. Kraan W, van Keimpema M, Horlings HM, et al. High prevalence of oncogenic MYD88 and CD79B mutations in primary testicular diffuse large B-cell lymphoma. Leukemia. 2014;28(3):719-720. 20. Wilson WH, Young RM, Schmitz R, et al. Targeting B cell receptor signaling with ibrutinib in diffuse large B cell lymphoma. Nat Med. 2015;21(8):922-926. 21. Lionakis MS, Dunleavy K, Roschewski M, et al. Inhibition of B Cell Receptor Signaling by Ibrutinib in Primary CNS Lymphoma. Cancer Cell. 2017;31(6):833-843.e5. 22. Grommes C, Pastore A, Palaskas N, et al. Ibrutinib Unmasks Critical Role of Bruton Tyrosine Kinase in Primary CNS Lymphoma. Cancer Discov. 2017;7(9):1018-1029.

Thrombin generation: a global coagulation procedure to investigate hypo- and hyper-coagulability Armando Tripodi IRCCS Caâ&#x20AC;&#x2122; Granda Maggiore Hospital Foundation, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center and Fondazione Luigi Villa, Milano, Italy E-mail: ARMANDO TRIPODI - armando.tripodi@unimi.it doi:10.3324/haematol.2020.253047

he article by van Paridon et al.1 published in this issue of Haematologica on results of thrombin generation (TG) in cardiovascular disease and mortality, stemming from the Gutenberg Health Study, provides an opportunity to comment on TG as a global laboratory procedure to investigate hypo- and hyper-coagulability. TG as a laboratory test was developed in the early 1950s by McFarlane and Biggs2 and was based on the activation of coagulation in whole blood or plasma by triggers such as tissue factor or cephaline and calcium chlo2196

ride. The amount of thrombin generated over time was titrated by sampling the mixture at different time points into a fibrinogen solution and the resultant clotting times interpolated from a dose-response calibration curve to derive thrombin concentrations. Years later, Hemker et al. made substantial changes.3-5 The fibrinogen solution was replaced by a chromogenic substrate specific for thrombin, test plasma was defibrinated prior to testing and computer software was developed to derive the parameters stemming from the TG curve. These changes made haematologica | 2020; 105(9)

Editorials

Figure 1. Typical thrombin generation curve with related parameters that can be obtained after in vitro activation of coagulation in plasma. See text for more explanation. ETP: endogenous thrombin potential.

TG much easier to perform even in less specialized laboratories. In the early 1990s TG was further modified to include a slow-acting chromogenic substrate that allowed thrombin to be monitored continuously, avoiding the subsampling procedures. More recently, the chromogenic substrate has been replaced by a fluorogenic one, which obviates plasma defibrination and makes TG applicable to platelet-poor or platelet-rich plasma. Currently, TG is designed to operate on a microtiter plate-based fluorimeter, which, in combination with the computer software, displays TG curves and calculates the parameters illustrated in Figure 1: the lag-time, peak thrombin concentration, time to reach peak concentration and the endogenous thrombin potential (ETP). The lag-time is defined as the time needed for the first amounts of thrombin to be generated; it can be regarded as the conventional plasma clotting time and is expected to decrease or increase in conditions associated with hyper- or hypo-coagulability, respectively. The peak-thrombin defines the highest thrombin concentration that can be obtained under the experimental conditions and is expected to increase or decrease in conditions associated with hyper- or hypocoagulability, respectively. The time to reach the peak defines the velocity of TG and should be prolonged or shortened in conditions associated with hypo- or hypercoagulability, respectively. The area under the curve, known as the ETP, represents the net amount of thrombin that the test plasma can generate under the experimental conditions. ETP is considered to be the resultant of the two opposing drivers operating in plasma that favor TG (procoagulants) and thrombin decay (anticoagulants). Accordingly, when performed under standardized conditions, TG can be considered a laboratory tool suitable for investigating hypo- and hyper-coagulability much better than the conventional global tests of coagulation, such as the time-honored prothrombin time (PT) and activated partial thromboplastin time (APTT). PT and APTT are in fact static tests in which plasma clots soon after tiny amounts of thrombin (5%) are produced, leaving the haematologica | 2020; 105(9)

remaining part undetected.6 Furthermore, PT and APTT are responsive to procoagulant factors (which is why they are used to diagnose hemophilia and allied disorders), but they are much less responsive to naturally occurring anticoagulants (which is why they are normal rather than shortened in patients with congenital deficiency of protein C or antithrombin). It is therefore unsurprising that over the last two decades, TG attracted the attention of many researchers. Currently, there are 7,446 reports under the term â&#x20AC;&#x153;thrombin generationâ&#x20AC;? published in PubMed and most deal with the application of TG to investigate hypo- or hyper-coagulability. A scrutiny of the reported manuscripts reveals that TG has been applied to five categories: (i) to help elucidate the mechanisms of thrombogenesis in clinical conditions for which precise knowledge is still poor; (ii) laboratory diagnosis of disorders of hemostasis; (iii) monitoring treatment with pro-hemostatic agents; (iv) monitoring treatment with antithrombotic drugs; and (v) predicting the risk of recurrence of venous thromboembolism. The following paragraphs summarize the current state-of-the-art concerning the application of TG testing.

Elucidation of mechanisms of thrombogenesis This application has been the most productive in terms of results achieved. For example, the TG procedure when performed as such or modified by the addition of thrombomodulin has been instrumental in challenging the old dogma of liver cirrhosis as the prototype of acquired hemorrhagic coagulopathies.7,8 TG has also been useful for understanding the mechanisms of thrombogenesis in a variety of clinical conditions associated with increased risk of thrombosis (especially venous thrombosis) such as obesity,9 diabetes,10 and Cushing disease.11

Laboratory diagnosis of hemostatic disorders The TG procedure has been useful for understanding the contribution of the plasma levels of the individual coagulation factors in determining the bleeding pheno2197

Editorials

type in subjects with hemophilia12 or other rare congenital diseases.13

Monitoring treatment with pro-hemostatic agents In principle, every treatment aimed at preventing/treating hemorrhage accomplishes its goal by increasing thrombin generation. In this respect, the TG procedure is the most promising laboratory tool to monitor patients on treatment. As an example, TG has been used in hemophilia with inhibitors to factor VIII or IX to tailor treatment with conventional bypassing agents.14 These drugs achieve their hemostatic effect with no substantial modification of factors VIII/IX, thus making the measurement of these factors after infusion practically useless. However, the studies carried out so far, while providing evidence that TG is increased according to the dose of the drug infused to treat/prevent hemorrhage, failed to provide conclusive evidence that the laboratory endpoint is associated with clinical outcomes.15 Therefore, the procedure is not yet approved by the regulatory authorities for routine use in hemophilia patients being treated with bypassing agents.

Monitoring treatment with antithrombotic drugs There is no doubt that TG is decreased in a dose-dependent fashion in patients treated with any antithrombotic drug (e.g., heparins, parenteral direct thrombin inhibitors, warfarin or direct oral anticoagulants). However, there is still no conclusive evidence that the TG procedure is superior for treatment monitoring to the APTT (unfractionated heparin), anti-factor Xa assays (low molecular weight heparin), the PT-International Normalized Ratio (warfarin) or the specific measurement of the plasma concentration of direct oral anticoagulants.

atively long follow-up (up to 9.65 years). The authors found that some TG parameters were independently associated with overall mortality. In particular, the study showed that the ETP and lag-time were directly associated with overall mortality. Furthermore, these parameters were associated with some conditions (e.g., age, obesity, diabetes, dyslipidemia, use of oral contraceptives or hormonal replacement therapy) that are known risk factors for cardiovascular disease. The study by van Paridon et al. is the first to investigate the association of hypercoagulability (as measured by TG) and the risk of mortality in a large population.1 However, some issues emerge from the study that warrant consideration. While it is plausible that high ETP is directly associated with overall mortality, it is less plausible that the prolonged (not shortened) lagtime is associated with mortality. It should however be recognized that a prolonged lag-time has been hypothesized to be associated with increased levels of tissue factor pathway inhibitor, one of the naturally occurring anticoagulants. Unfortunately, the authors did not provide data on tissue factor pathway inhibitor in their cohort, nor did they report other parameters of TG such as the time to reach the peak concentration. One may wonder whether or not the latter parameter would be more representative than the lag-time to describe the real situation concerning the velocity of TG, which might help to understand the mechanism of thrombogenesis. Furthermore, the study did not investigate the association between TG and cardiovascular mortality. This would have been a more plausible association between hypercoagulability and clinical outcome. Notwithstanding these limitations, the study by van Paridon et al. should be considered an important step forward for the development of TG as a global procedure to assess the hemostatic system and its relationship with overall mortality.1

Predicting the risk of recurrent venous thromboembolism There is evidence stemming from clinical trials that the amount of thrombin generated after the in-vitro activation of coagulation is a good predictor of the risk of a recurrence of venous thromboembolism. In this respect, TG should be equivalent (as a risk predictor) to D-dimer,16,17 which, following the seminal observations of Palareti and co-workers,18 is used to make decisions on the optimal duration of anticoagulation to prevent recurrence of venous thromboembolism. Although these studies showed that TG and D-dimer are independently associated with the risk of recurrent venous thromboembolism, there is no conclusive evidence that performing the two measurements simultaneously can improve risk prediction in individual patients.18

New observations on thrombin generation The above observations are instrumental to supporting the old concept that hypercoagulability, when assessed by a global coagulation procedure, is one of the key mechanisms that explain the risk of thrombosis, the others being reduced blood flow and endothelial dysfunction (collectively known as Virchow triad). In this issue of the Journal van Paridon et al. add more evidence on the mechanisms of hypercoagulability and clinically relevant outcomes.1 They evaluated TG in a large population of subjects (n=5,000) enrolled in a prospective study with a rel2198

References 1. van Paridon PCS, Panova-Noeva M, van Oerle R, et al. Thrombin generation in cardiovascular disease and mortality - results from the Gutenberg Health Study. Haematologica. 2020;105(9):2327-2334. 2. Macfarlane RG, Biggs R. A thrombin generation test. J Clin Pathol. 1953;6(1):3-7. 3. Hemker HC, Willems GM, BeĚ guin S. A computer assisted method to obtain the prothrombin activation velocity in whole plasma independent of thrombin decay processes. Thromb Haemost. 1986;56(1):9-17. 4. Hemker HC, Wielders S, Kessels H, BeĚ guin S. Continuous registration of thrombin generation in plasma, its use for the determination of the thrombin potential. Thromb Haemost .1993;70(4):617-624. 5. Hemker HC, Giesen PL, Ramjee M, et al. The thrombogram: monitoring thrombin generation in platelet-rich plasma. Thromb Haemost. 2000;83(4):589-591. 6. Mann KG, Brummel K, Butenas S. What is all that thrombin for? J Thromb Haemost. 2003;1(7):1504-1514. 7. Tripodi A, Salerno F, Chantarangkul V, et al. Evidence of normal thrombin generation in cirrhosis despite abnormal conventional coagulation tests. Hepatology. 2005;41(3):553-558. 8. Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl J Med. 2011;365(2):147-156. 9. Tripodi A, Primignani M, Badiali S, et al. Body mass index reduction improves the baseline procoagulant imbalance of obese subjects. J Thromb Thrombolysis. 2019;48(1):52-60. 10. Tripodi A, Branchi A, Chantarangkul V, et al. Hypercoagulability in patients with type 2 diabetes mellitus detected by a thrombin generation assay. J Thromb Thrombolysis. 2011;31(2):165-172. 11. Tripodi A, Ammollo CT, Semeraro F, et al. Hypercoagulability in patients with Cushing disease detected by thrombin generation

haematologica | 2020; 105(9)

Editorials

12. 13.

14.

15.

assay is associated with increased levels of neutrophil extracellular trap-related factors. Endocrine. 2017;56(2):298-307. Santagostino E, Mancuso ME, Tripodi A, et al. Severe hemophilia with mild bleeding phenotype: molecular characterization and global coagulation profile. J Thromb Haemost. 2010;8(4):737-743. Zekavat OR, Haghpanah S, Dehghani J, et al. Comparison of thrombin generation assay with conventional coagulation tests in evaluation of bleeding risk in patients with rare bleeding disorders. Clin Appl Thromb Hemost. 2014;20(6):637-644. Tran HT, SĂ¸rensen B, BjĂ¸rnsen S, et al. Monitoring bypassing agent therapy - a prospective crossover study comparing thromboelastometry and thrombin generation assay. Haemophilia. 2015;21(2):275283. Tripodi A, Chantarangkul V, Novembrino C, Peyvandi F. Advances in the treatment of hemophilia: implications for laboratory testing.

haematologica | 2020; 105(9)

Clin Chem. 2019;65(2):254-262. 16. Tripodi A, Legnani C, Chantarangkul V, et al. High thrombin generation measured in the presence of thrombomodulin is associated with an increased risk of recurrent venous thromboembolism. J Thromb Haemost. 2008;6(8):1327-1333. 17. van Hylckama Vlieg A, Baglin CA, Luddington R, et al. The risk of a first and a recurrent venous thrombosis associated with an elevated D-dimer level and an elevated thrombin potential: results of the THE-VTE study. J Thromb Haemost. 2015;13(9):1642-1652. 18. Palareti G, Cosmi B, Legnani C, et al. D-dimer testing to determine the duration of anticoagulation therapy. N Engl J Med. 2006;355 (17):1780-1789. 19. Eichinger S, Hron G, Kollars M, Kyrle PA. Prediction of recurrent venous thromboembolism by endogenous thrombin potential and D-dimer. Clin Chem. 2008;54(12):2042-2048.

2199

PERSPECTIVE ARTICLE Ferrata Storti Foundation

BCR-ABL1-like acute lymphoblastic leukemia in childhood and targeted therapy Gunnar Cario,1* Veronica Leoni,2* Valentino Conter,2* André Baruchel,3# Martin Schrappe1# and Andrea Biondi2#

Pediatrics, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany; 2Clinica Pediatrica and Centro Ricerca Tettamanti, Università di Milano-Bicocca, Hospital S.Gerardo, Monza, Italy and 3Hôpital Universitaire Robert Debré (APHP) and Université de Paris, Paris, France 1

Haematologica 2020 Volume 105(9):2200-2204

GC, VL and VC contributed equally to this work as co-first authors.

AB, MS and AB contributed equally as co-senior authors.

Correspondence: ANDREA BIONDI abiondi.unimib@gmail.com Received: March 7, 2019. Accepted: May 8, 2020. Pre-published: May 15, 2020. doi:10.3324/haematol.2018.207019 ©2020 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.

2200

cute lymphoblastic leukemia (ALL) is a constellation of diseases driven by genetic alterations commonly derived from structural chromosome rearrangements, aneuploidy and co-operating mutations in genes that encode for transcription factors regulating lymphoid development, tumor suppressors, proteins regulating cell cycle progression, and epigenetic modifiers.1 Recent years have witnessed dramatic progress in ALL classification. Subtypes of ALL can be defined according to the nature of specific sentinel genetic aberrations that confer distinct biological and clinical characteristics. Some of them represent a therapeutic target for specific treatments, which may contribute to a further increase in cure rates, to reduce the intensity of conventional chemotherapy and/or the need for hematopoietic stem cell transplantation (HSCT). One of the first genetic aberrations identified was the Philadelphia chromosome (Ph), characterized by the t(9;22)(q34;q11) translocation that produces the BCR-ABL1 gene, and, in turn, a constitutively active tyrosine kinase. BCR-ABL1 fusion is present in 3-5% of pediatric ALL and in 25% of adult ALL patients. The evidence of this genetic aberration allowed the introduction of targeted therapy with tyrosine kinase inhibitors (TKI), which has dramatically improved the outcome of this subset of ALL.2-10 The pediatric COG AALL1131 and AALL0622 studies, and the contemporary EsPhALL2004 and subsequent EsPhALL2010 trials, in fact, showed a clear advantage in Ph positive (Ph+) ALL from early, continuous and protracted exposure to TKI combined with chemotherapy, challenging the indications for HSCT.5-10 Of note, however, the combination of chemotherapy and TKI may also be associated with increased toxicity, as shown in the EsPhALL2010 study.6,7 With advanced technologies, such as whole genome and transcriptome sequencing, novel genetic subtypes have recently been discovered. In 2009, among the so called “B-other”, a subgroup of B-cell precursor (BCP)-ALL lacking the known sentinel BCP-ALL genetic aberrations, a new category of ALL has been described by Mullighan11 and by den Boer12, and termed Philadelphia chromosome (Ph)-like and BCR-ABL1-like ALL, respectively. The second term is used in this paper. The two signatures are based on the prediction analysis of microarrays (PAM) classifier consisting of 257 gene probe sets trained on Ph+ ALL cases (Mullighan11) or on hierarchical clustering of 110 gene probe sets identified to predict the major pediatric ALL subtypes (den Boer), with only nine overlapping probe sets.12 BCR-ABL1-like ALL, defined by a gene expression profile greatly similar to that of Ph+ ALL, presents a high frequency of deletions of IKZF1, which encodes the lymphoid transcription factor IKAROS, and of other lymphoid transcription factor genes.11,13 BCR-ABL1–like ALL has been recognized as a provisional entity in the 2016 World Health Organization classification of myeloid neoplasms and acute leukemia;14 the prevalence varies with age from 12% in children to 21% in adolescents, 27% in young adults, and 20-24% in older adults with BCP-ALL. In addition to older age at diagnosis, BCR-ABL1–like ALL is associated with other high-risk clinical features, such as elevated leukocyte count at diagnosis and poor treatment response, i.e. high levels of end-induction minimal residual disease (MRD), increased risk of induction failure and of relapse.11,13,15-30 Importantly, BCR-ABL1-like ALL is not defined by a single unifying sentinel molecular aberration; but rather, it is characterized by a variety of genomic alterations that activate kinases and deregulate cytokine receptor signaling. Fusion genes involving at least 17 cytokine receptors or tyrosine kinases have been identified.23,29,31,32 These alterations can be grouped into several major subclasses: approximately 50% of BCRhaematologica | 2020; 105(9)

Targeted therapy for pediatric BCR-ABL1-like ALL

ABL1-like cases harbor rearrangements of the cytokine receptor like factor 2 (CRLF2) resulting in upregulation of CRLF2 expression, in the vast majority as a consequence of either a translocation resulting in IGH-CRLF2 juxtaposition or a deletion of the PAR region of the X chromosome leading to the P2RY8-CRLF2 fusion. Frequent concomitant activating gene mutations occur in Janus kinases or other regulators of JAK-STAT signaling, with about 50% of CRLF2 rearranged cases presenting JAK1 or JAK2 point mutations.15,16,33,34 However, although the P2RY8-CRLF2 rearrangement is associated with an intermediate to poor outcome, its role with respect to relapse disposition is not fully clear, as the P2RY8-CRLF2 rearrangement has been reported in some cases to be lost at relapse, particularly when it has been identified initially in a sub-clone only.35-38 About one-third of BCR-ABL1-like non-CRLF2 rearranged ALL cases present chromosomal rearrangements that result in constitutive deregulation of a cytokine receptor or the formation of kinase fusion genes: a major subgroup includes ABL-class alterations involving ABL1, ABL2, CSF1R, LYN, PDGFRA and PDFGRB. A second major group regards rearrangements that activate JAK family kinases, including JAK2, EPOR, TYK2 and IL2RB. A third group constitutes a variety of other kinases or cytokine receptor alterations such as NTRK3, FLT3, FGFR1 and BLNK, and the RAS signaling pathway.11,13,23, 29,32-39 The limited data available confirm that BCR-ABL1-like ALL is associated with high-risk features also in pediatric patients. A single institution reported that the outcome in BCR-ABL1–like ALL patients, although inferior to that of other patients, was favorable with MRD-driven therapy and with the majority of patients treated in the higher risk arms and 15% undergoing HSCT.23,24 Subsequently, the COG found that, within standard risk ALL patients defined by National Cancer Institute (NCI) criteria, Ph-like ALL patients had a still good, but significantly lower, event-freesurvival and no significant difference in survival when compared to non-Ph-like NCI standard risk ALL.40 In keeping with these data, Boer reported an increased cumulative incidence of relapse in BCR-ABL1–like ALL compared to nonBCR-ABL1–like B-other ALL.26 Finally, the AIEOP-BFM study group has recently reported the outcome of ABL-class fusion positive BCP-ALL in a retrospective study, which, although limited by its retrospective nature, and especially by a potential selection bias towards cases with a poor treatment response, indicates that these patients have an overall poor prognosis.41 The role of CRLF2 abnormalities on BCR-ABL1-like ALL outcome is still controversial. The COG showed that, while high CRLF2-expression predicted a dismal outcome in high-risk patients, the two specific genomic CRLF2-lesions did not confer independent prognostic significance.42 Similarly, CRLF2-rearrangements had no independent prognostic value in the Medical Research Council ALL97 trial,43 while the AIEOP-BFM study group reported that P2RY8CRLF2 positive patients allocated in the non-HR group had a poorer prognosis.35,36 However, it should be remembered that data on CRLF2-rearranged BCP-ALL are not exclusively restricted to cases with BCR-ABL1-like gene expression signature. Outcome data on BCR-ABL1-like ALL are summarized in Table 1. Overall, these data confirm that there is a clinical need for innovative targeted therapies which may be effective in this ALL subtype, as suggested by pre-clinical studies. In vitro studies have, in fact, demonstrated constitutive haematologica | 2020; 105(9)

activation of kinase signaling networks in subsets of BCRABL1–like ALL harboring JAK pathway aberrations,42,44,45 and in vivo studies have demonstrated anti-leukemic activity of the type I JAK2 inhibitor ruxolitinib and of the dual PI3K/mTOR inhibitor gedatolisib given as a monotherapy in patient-derived xenograft models of JAK pathway– mutant BCR-ABL1-like ALL.44,46-50 Other studies have reported superior anti-leukemic efficacy with the type II JAK inhibitor CHZ868, which synergizes with dexamethasone to induce apoptosis, suggesting that type II JAK2 inhibition may be more effective to target CRLF2-rearranged BCPALL. This may be because type II inhibitors stabilize JAK2 in the inactive conformation, and overcome the JAK2 hyperphosphorylation observed with type I JAK inhibitors which target the ATP binding pocket and stabilize JAK2 in the active conformation.51 Likewise, pre-clinical experimental studies have shown that cell lines and human cells expressing ABL-class fusions, as well as patient-derived xenograft models, have marked sensitivity to the TKI such as imatinib and dasatinib, similarly to BCR-ABL1 cells.52 Clinical studies are still very limited. A COG phase I trial (ADVL1011; clinicaltrials.gov identifier: 01164163) demonstrated the safety of JAK2 inhibitor ruxolitinib, given as monotherapy in children with relapsed or refractory cancers,53 while anecdotal reports have provided evidence of efficacy of TKI (imatinib and dasatinib) to induce remission and clear MRD in patients with ABL-class fusions with poor response to previous chemotherapy.54-56 Optimal clinical management of pediatric BCR-ABL1-like ALL, thus, remains to be defined. The heterogeneous genomic landscape and the diverse array of targetable kinase-activating lesions of BCR-ABL1–like ALL require precise diagnostic strategies. Initially, the DCOG group used a validated Affymetrix gene expression array which included 110 probe sets, while the COG and SJCRH used an Affymetrix gene expression array with 255 probe sets to screen patients for BCR-ABL1-like ALL signature. Subsequently, COG first utilized a quantitative reverse transcriptase polymerase chain reaction (RT-PCR)-based low density array (LDA) platform to identify patients with BCR-ABL1-like ALL enrolled in their ALL COG front-line AALL1131 trial. As a second step, a series of multiplex RTPCR assays, fluorescence in situ hybridization (FISH), and DNA sequencing to identify the underlying genomic aberration were applied.57,58 The COG is now using Archer targeted RNA sequencing instead of multiplex RT-PCR assays. Alternatively, combined FISH or targeted RNA-next generation sequencing (NGS) strategies with probes capturing the recurrently fused genes can be successfully applied.59 In the future, NGS-based whole transcriptome sequencing should allow the detection of all relevant gene fusions and mutations in one step, as recently demonstrated by Gu et al.60 and Li et al.61 This approach will facilitate the timed diagnosis and the early implementation of specific treatments. Of note, despite the large number of individual kinase alterations identified, the majority converge on a limited number of pathways that can be targeted. The best therapeutic strategy for this subgroup of patients remains a matter of investigation. Several ongoing studies are assessing the role of the addition of TKI or ruxolitinib on top of chemotherapy in pediatric BCP-ALL harboring ABL-class fusions or CRLF2/JAK pathway alterations. In the current COG AALL1131 and AALL1521 (clinicaltrials.gov identifier: 02883049 and 02723994, respectively) and SJCRH Total Therapy XVII trials (clinicaltrials.gov identi2201

G. Cario et al. Table 1. Outcome of BCR-ABL1-like acute lymphoblastic leukemia among different study groups.

Study group

Protocol

Roberts et al.23

Total therapy XV, Total therapy XVI, P9906 AALL0232, E2993, C19802, C10102 and C10403

Roberts et al.24

N. BCR-ABL1(Ph) like ALL patients/total BCP-ALL

Outcome (CIR, EFS, OS)

264/1725

5-years pEFS 58.2±5.3%, 41.0±7.4%, and 24.1±10.5% for children with high-risk ALL, adolescents, and young adults; 5-years pOS 72.8±4.8%, 65.8±7.1%, and 25.8±9.9% for children with high-risk ALL, adolescents, and young adults. Across all age groups OS rates were inferior to those among patients with non–Ph-like ALL (P<0.001 for both comparisons)

Total therapy XV

40/344

Roberts et al.40

COG AALL0331

206/1023 Standard-Risk ALL

Boer et al.26

DCOG ALL-8, ALL-9, ALL10 COALL 06-97 and COALL 07-03 AIEOP BFM ALL 2000 and AIEOP BFM ALL 2009

5 -years pEFS 90.0% ± 4.7% vs. 88.4% ± 0.9%, P=0.41 in BCRABL1–like ALL vs. non-BCR-ABL1–like ALL; 5-years pOS 92.5% ± 4.2% vs. 95.1% ± 1.3%, P=0.41 in BCR-ABL1–like ALL vs. non-BCR-ABL1–like ALL 7-years pEFS 82.4 ± 3.6% vs. 90.7 ± 1.0%, P=0.0022, Ph-like ALL vs. non–Ph-like ALL; 7-years pOS 93.2 ± 2.4% vs. 95.8 ± 0.7%, P=0.14, Ph-like ALL vs. non–Ph-like ALL 8-years pCIR 35% vs. 17%, P=0.07, BCR-ABL1–like ALL vs. non BCR-ABL1–like B-other ALL 5-years pEFS was 49.1±8.9% , 5-years pOS 69.6±7.8% and 5-years CIR was 25.6±8.2%

Cario et al.41

77/574 46 ABL-class fusion positive ALL

ALL: acute lymphoblastic leukemia; BCP-ALL: B-cell precursor acute lymphoblastic leukemia; CIR: cumulative incidence of relapse; EFS: event-free survival; OS: overall survival.

fier: 03117751), ALL patients with NCI high-risk characteristics or poor early MRD response are screened for ABLclass fusions and JAK pathway mutations. In patients positive for these alterations, dasatinib and ruxolitinib, respectively, are given in combination with conventional frontline chemotherapy from the consolidation phase until the end of maintenance therapy.57,62 Patients with NCI standard risk characteristics and early good MRD response are not included because available data on their outcome are very limited.48,52 Other phase I/II trials conducted at the MD Anderson Cancer Center (clinicaltrials.gov identifier: 02420717) are testing dasatinib or low doses of ruxolitinib in combination with hyper-CVAD (cyclophosphamide, vincristine, doxorubicin, dexamethasone) in adolescents and adults with relapsed/refractory ALL and ABL-class fusions or CRLF2/JAK mutations, respectively; interim data analysis demonstrates the safety of these combinations with limited efficacy.63 A recent phase I trial (clinicaltrials.gov identifier: 03571321) at the University of Chicago and other institutions is studying ruxolitinib in combination with the pediatric-inspired CALBG 10403 chemotherapy regimen in adolescents with newly diagnosed Ph-like ALL harboring CRLF2/JAK alterations, with a planned phase II expansion study if safety is demonstrated.64,65 In Europe, the AIEOP-BFM ALL and ALLTogether study groups are also investigating the addition of innovative or targeted therapy on top of chemotherapy in BCR-ABL1-like ALL. In the AIEOP-BFM ALL 2017 trial (clinicaltrials.gov identifier: 03643276), patients are screened at diagnosis for IKZF1 deletions, which are frequently found in BCR-ABL1like ALL, and for additional deletions of genes relevant for B-cell development. Those cases defined as IKZF1 plus positive66 with any MRD positivity after induction treatment are randomized to receive the proteasome inhibitor bortezomib in addition to chemotherapy during consolidation and to receive the bispecific T-cell engager (BiTE) antibody blinatumomab during post-consolidation treatment. 2202

Especially the approach to apply immunotherapy instead of extremely intensive high-risk blocks may be of advantage for ABL-class-fusion positive cases, bearing in mind the high rate of severe treatment-related complications in Ph+ ALL patients treated with high-risk chemotherapy plus TKI. In the ALLTogether study, patients are screened for ABL-class fusions at diagnosis and those positive receive TKI on top of chemotherapy from day 15 of induction onward. In both AIEOP-BFM and ALLTogether studies, these patients have an indication for HSCT in case of poor MRD response. Likewise, the French CALL-F01 protocol (clinicaltrials.gov identifier: 02716233) has been amended in 2018 to bring to RNA sequencing all B-other ALL in case of induction failure or end of induction MRD above or equal to 10-3: these patients are to receive imatinib on top of chemotherapy in the high-risk group. Then, according to subsequent MRD and effective exposure to imatinib, they either continue TKI plus chemotherapy or go to HSCT. A similar approach in the early introduction of a TKI in addition to chemotherapy in ABL-class positive BCP-ALL is planned within the EsPhALL2017/COGAALL1631 protocol (clinicaltrials.gov identifier: 03007147), the first intercontinental collaborative trial for the treatment of pediatric Ph+ ALL involving COG and EsPhALL study groups. In this trial, an amendment is ongoing to extend the eligibility to patients with ABL-class fusion positive BCP-ALL and, thus, treat them with imatinib given early after diagnosis and continuously on top of high-risk chemotherapy. Actually, in pediatric patients there is no clear evidence for superiority of a specific type of TKI. In the COG AALL0622 study, dasatinib (60 mg/m2) was substituted for imatinib (340 mg/m2) on top of the same chemotherapy backbone of the AALL0031 study with no benefit. The same dose of dasatinib was used also in a joint COG/EsPhALL study (BMS CA180372) on top of the EsPhALL therapeutic strategy with preliminary results which appear similar to the contemporary EsPhALL study haematologica | 2020; 105(9)

Targeted therapy for pediatric BCR-ABL1-like ALL

which used imatinib (300 mg/m2). A very recent study, where dasatinib was used at a higher dose (80 mg/m2) and randomized versus imatinib (300 mg/m2), showed a superiority of dasatinib; however, follow up of this study was relatively short, and results in the cohort treated with imatinib were inferior to those obtained by the EsPhALL and COG groups with the use of imatinib, thus, challenging the evidence of superiority itself. Other TKI such as nilotinib, bosutinib and ponatinib are still being investigated as phase I and II trials in pediatric cancers. At this moment, the choice of both imatinib or dasatinib appears to be reasonable as TKI in frontline ALL protocols for children and adolescents.7,9,10,67,68 In summary, there are still some challenges to implanting targeted therapy into frontline ALL treatment. There is a need for an early identification of BCP-ALL harboring ABL-class and JAK-pathway alterations to allow prompt intervention with targeted therapy to reduce intensity of chemotherapy, and refine HSCT indications, as already shown for Ph+ ALL.5-10 Diagnostic technologies such as RNA sequencing and similar strategies should be implemented in a timely fashion for all “B-other ALL”.

References 1. Hunger SP, Mullighan CG. Redefining ALL classification: toward detecting high-risk ALL and implementing precision medicine. Blood. 2015;125(26):3977-3987. 2. Bernt KM, Hunger SP. Current concepts in pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia. Front Oncol. 2014;4:54. 3. Fielding AK. Treatment of Philadelphia chromosome-positive acute lymphoblastic leukemia in adults: a broader range of options, improved outcomes, and more therapeutic dilemmas. Am Soc Clin Oncol Educ Book. 2015;2e352-359. 4. Aricò M, Schrappe M, Hunger SP, et al. Clinical outcome of children with newly diagnosed Philadelphia chromosome-positive acute lymphoblastic leukemia treated between 1995 and 2005. J Clin Oncol. 2010;28(31):4755-4761. 5. Biondi A, Schrappe M, De Lorenzo P, et al. Imatinib after induction for treatment of children and adolescents with Philadelphiachromosome positive acute lymphoblastic leukaemia (EsPhALL): a randomised, openlabel, intergroup study. Lancet Oncol. 2012;13(9):936-945. 6. Biondi A, Cario G, De Lorenzo P, et al. Longterm follow up of pediatric Philadelphia positive acute lymphoblastic leukemia treated with the EsPhALL2004 study: high white blood cell count at diagnosis is the strongest prognostic factor. Haematologica. 2019;104 (1):13-16. 7 Biondi A, Gandemer V, De Lorenzo P, et al. Imatinib treatment of paediatric Philadelphia chromosome-positive acute lymphoblastic leukaemia (EsPhALL2010): a prospective, intergroup, open-label, singlearm clinical trial. Lancet Haematol. 2018;5(12):641-652. 8 Schultz KR, Bowman WP, Aledo A, et al. Improved early event-free survival with imatinib in Philadelphia chromosome-posi-

haematologica | 2020; 105(9)

11.

12.

13.

14.

15.

16.

Although ABL-class and JAK-pathway alterations account for most BCR-ABL1-like ALL cases, there are also several alterations involving kinases that are not inhibited by either TKI or JAK inhibitors. Future studies are required to assess the potential of targeted inhibitors of these kinases in model systems and human leukemic cells. In the meantime, for this subgroup of BCR-ABL1-like cases without known targetable lesions, the optimal treatment should be based on MRD response, and might include innovative therapies such as immunotherapy. Moreover, all ALL patients treated with targeted approaches should be registered and closely followed up on the molecular level as recently discussed by Elitzur and Izraeli in order to understand response and resistance to targeted treatment.69 Due to the rarity of these clinical entities, collaborative international efforts are strongly needed to conduct successful studies. Acknowledgments The authors would like to thank AIRC 2017 20564; CRUK/AIRC/FC AECC 22791 and AIRC Molecular Clinical Oncology 5 per mille 21147.

tive acute lymphoblastic leukemia: a children’s oncology group study. J Clin Oncol. 2009;27(31):5175-5181. Schultz KR, Carroll A, Heerema NA, et al. Long-term follow-up of imatinib in pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia: Children’s Oncology Group Study AALL0031. Leukemia. 2014;28(7):1467-1471. Slayton WB, Schultz KR, Kairalla JA, et al. Dasatinib plus intensive chemotherapy in children, adolescents, and young adults with philadelphia chromosome-positive acute lymphoblastic leukemia: results of Children’s Oncology Group Trial AALL0622. J Clin Oncol. 2018;36(22):23062314. Mullighan CG, Su X, Zhang J, et al; Children’s Oncology Group. Deletion of IKZF1 and prognosis in acute lymphoblastic leukemia. N Engl J Med. 2009;360(5):470480. Boer JM, Marchante JR, Evans WE, et al. BCR-ABL1-like cases in pediatric acute lymphoblastic leukemia: a comparison between DCOG/Erasmus MC and COG/St. Jude signatures. Haematologica. 2015;100(9):e354357. Den Boer ML, van Slegtenhorst M, De Menezes RX, et al. A subtype of childhood acute lymphoblastic leukaemia with poor treatment outcome: a genome-wide classification study. Lancet Oncol. 2009;10(2):125134. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391-2405. Harvey RC, Mullighan CG, Chen IM, et al. Rearrangement of CRLF2 is associated with mutation of JAK kinases, alteration of IKZF1, Hispanic/Latino ethnicity, and a poor outcome in pediatric B-progenitor acute lymphoblastic leukemia. Blood. 2010;115(26):5312-5321. Russell LJ, Capasso M, Vater I, et al.

17.

18.

19.

20.

21.

22.

23.

24.

Deregulated expression of cytokine receptor gene, CRLF2, is involved in lymphoid transformation in B-cell precursor acute lymphoblastic leukemia. Blood. 2009;114(13): 2688-2698. Herold T, Schneider S, Metzeler K, et al. Adults with Philadelphia chromosome-like acute lymphoblastic leukemia frequently have IGH-CRLF2 and JAK2 mutations, persistence of minimal residual disease and poor prognosis. Haematologica. 2017; 102(1):130-138. Boer JM, Koenders JE, van der Holt B, et al. Expression profiling of adult acute lymphoblastic leukemia identifies a BCR-ABL1like subgroup characterized by high nonresponse and relapse rates. Haematologica. 2015;100(7):261-264. Roberts KG, Mullighan CG. Genomics in acute lymphoblastic leukaemia: insights and treatment implications. Nat Rev Clin Oncol. 2015;12(6):344-357. Harvey RC, Mullighan CG, Wang X, et al. Identification of novel cluster groups in pediatric high-risk B-precursor acute lymphoblastic leukemia with gene expression profiling: correlation with genome-wide DNA copy number alterations, clinical characteristics, and outcome. Blood. 2010;116(23):4874-4884. Loh ML, Zhang J, Harvey RC, et al. Tyrosine kinome sequencing of pediatric acute lymphoblastic leukemia: a report from the Children’s Oncology Group TARGET Project. Blood. 2013;121(3):485-488. van der Veer A, Waanders E, Pieters R, et al. Independent prognostic value of BCR-ABL1like signature and IKZF1 deletion, but not high CRLF2 expression, in children with Bcell precursor ALL. Blood. 2013;122(15): 2622-2629. Roberts KG, Li Y, Payne-Turner D, et al. Targetable kinase-activating lesions in Phlike acute lymphoblastic leukemia. N Engl J Med. 2014;371(11):1005-1015. Roberts KG, Pei D, Campana D, et al. Outcomes of children with BCR-ABL1–like

2203

G. Cario et al.

25.

26.

27.

28.

29.

30. 31.

32.

33.

34.

35.

36.

37.

38.

39.

2204

acute lymphoblastic leukemia treated with risk-directed therapy based on the levels of minimal residual disease. J Clin Oncol. 2014;32(27):3012-3020. Imamura T, Kiyokawa N, Kato M, et al. Characterization of pediatric Philadelphia negative B-cell precursor acute lymphoblastic leukemia with kinase fusions in Japan. Blood Cancer J. 2016;6:e419. Boer JM, Steeghs EM, Marchante JR, et al. Tyrosine kinase fusion genes in pediatric BCRABL1-like acute lymphoblastic leukemia. Oncotarget. 2017;8(3):4618-4628. Herold T, Schneider S, Metzeler KH, et al. Adults with Philadelphia chromosome-like acute lymphoblastic leukemia frequently have IGHCRLF2 and JAK2 mutations, persistence of minimal residual disease and poor prognosis. Haematologica. 2017;102 (1):130-138. Jain N, Roberts KG, Jabbour E, et al. Ph-like acute lymphoblastic leukemia: a high-risk subtype in adults. Blood. 2017;129(5):572581. Reshmi SC, Harvey RC, Roberts KG, et al. Targetable kinase gene fusions in high-risk B-ALL: a study from the Children’s Oncology Group. Blood. 2017;129(25): 3352-3361. Tran TH, Loh ML. Ph-like acute lymphoblastic leukemia. Hematology Am Soc Hematol Educ Program. 2016;2016(1):561-566. Roberts KG, Gu Z, Payne-Turner D, et al. High frequency and poor outcome of philadelphia chromosome-like acute lymphoblastic leukemia in adults. J Clin Oncol. 2017;35(4):394-401. Roberts KG, Morin RD, Zhang J, et al. Genetic alterations activating kinase and cytokine receptor signaling in high-risk acute lymphoblastic leukemia. Cancer Cell. 2012;22(2):153-166. Mullighan CG, Collins-Underwood JR, Phillips LA, et al. Rearrangement of CRLF2 in B-progenitor- and Down syndrome-associated acute lymphoblastic leukemia. Nat Genet. 2009;41(11):1243-1246. Yoda A, Yoda Y, Chiaretti S, et al. Functional screening identifies CRLF2 in precursor Bcell acute lymphoblastic leukemia. Proc Natl Acad Sci USA. 2010;107(1):252-257. Cario G, Zimmermann M, Romey R, et al. Presence of the P2RY8-CRLF2 rearrangement is associated with a poor prognosis in non-high-risk precursor B-cell acute lymphoblastic leukemia in children treated according to the ALL-BFM 2000 protocol. Blood. 2010;115(26):5393-5397. Palmi C, Vendramini E, Silvestri D, et al. Poor prognosis for P2RY8-CRLF2 fusion but not for CRLF2 over-expression in children with intermediate risk B-cell precursor acute lymphoblastic leukemia. Leukemia. 2012;26(10):2245-2253. Morak M, Attarbaschi A, Fischer S, et al. Small sizes and indolent evolutionary dynamics challenge the potential role of P2RY8-CRLF2-harboring clones as main relapse-driving force in childhood ALL. Blood. 2012;120(26):5134-5142. Attarbaschi A, Morak M, Cario G, et al. Treatment outcome of CRLF2-rearranged childhood acute lymphoblastic leukaemia: a comparative analysis of the AIEOP-BFM and UK NCRI-CCLG study groups. Br J Haematol. 2012;158(6):772-777. Roberts KG, Yang Y, Turner DP, et al. Oncogenic role and therapeutic targeting of

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

ABL-class and JAK-STAT activating kinase alterations in Ph-like ALL. Blood Adv. 2017;1(20):1657-1671. Roberts KG, Reshmi SC, Harvey RC, et al. Genomic and outcome analyses of Ph-like ALL in NCI standard-risk patients: a report from the Children's Oncology Group. Blood. 2018;132(8):815-824. Cario G, Leoni V, Conter V, et al. Relapses and treatment-related events contributed equally to poor prognosis in children with ABL-class fusion positive B-cell acute lymphoblastic leukemia treated according to AIEOP-BFM protocols. Haematologica. 2019 Oct 10. [Epub ahead of print]. Chen IM, Harvey RC, Mullighan CG, et al. Outcome modeling with CRLF2, IKZF1, JAK, and minimal residual disease in pediatric acute lymphoblastic leukemia: a Children’s Oncology Group study. Blood. 2012;119(15):3512-3522. Ensor HM, Schwab C, Russell LJ, et al. Demographic, clinical, and outcome features of children with acute lymphoblastic leukemia and CRLF2 deregulation: results from the MRC ALL97 clinical trial. Blood. 2011;117(7):2129-2136. Maude SL, Tasian SK, Vincent T, et al. Targeting JAK1/2 and mTOR in murine xenograft models of Ph-like acute lymphoblastic leukemia. Blood. 2012;120(17): 3510-3518. Tasian SK, Doral MY, Borowitz MJ, et al. Aberrant STAT5 and PI3K/mTOR pathway signaling occurs in human CRLF2rearranged B-precursor acute lymphoblastic leukemia. Blood. 2012;120(4):833-842. Suryani S, Bracken LS, Harvey RC, et al. Evaluation of the in vitro and in vivo efficacy of the JAK inhibitor AZD1480 against JAKmutated acute lymphoblastic leukemia. Mol Cancer Ther. 2015;14(2):364-374. Tasian SK, Teachey DT, Li Y, et al. Potent efficacy of combined PI3K/mTOR and JAK or ABL inhibition in murine xenograft models of Ph-like acute lymphoblastic leukemia. Blood. 2017;129(2):177-187. van Bodegom D, Zhong J, Kopp N, et al. Differences in signaling through the B-cell leukemia oncoprotein CRLF2 in response to TSLP and through mutant JAK2. Blood. 2012;120(14):2853-2863. Weigert O, Lane AA, Bird L, et al. Genetic resistance to JAK2 enzymatic inhibitors is overcome by HSP90 inhibition. J Exp Med. 2012;209(2):259-273. Wu SC, Li LS, Kopp N, et al. Activity of the type II JAK2 inhibitor CHZ868 in B cell acute lymphoblastic leukemia. Cancer Cell. 2015; 28(1):29-41. Koppikar P1, Bhagwat N, Kilpivaara O, et al. Heterodimeric JAK-STAT activation as a mechanism of persistence to JAK2 inhibitor therapy. Nature. 2012;489(7414):155-159. Roberts KG, Morin RD, Zhang J, et al. Genetic alterations activating kinase and cytokine receptor signaling in high-risk acute lymphoblastic leukemia. Cancer Cell. 2012;22(2):153-166. Loh ML, Tasian SK, Rabin KR, et al. A phase 1 dosing study of ruxolitinib in children with relapsed or refractory solid tumors, leukemias, or myeloproliferative neoplasms: A Children's Oncology Group phase 1 consortium study (ADVL1011). Pediatr Blood Cancer. 2015;62(10):1717-1724. Lengline E, Beldjord K, Dombret H, et al. Successful tyrosine kinase inhibitor therapy

55.

56.

57. 58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

in a refractory B-cell precursor acute lymphoblastic leukemia with EBF1PDGFRB fusion. Haematologica. 2013;98(11):146148. Weston BW, Hayden MA, Roberts KG, et al. Tyrosine kinase inhibitor therapy induces remission in a patient with refractory EBF1PDGFRB-positive acute lymphoblastic leukemia. J Clin Oncol. 2013;31 (25):413416. Tanasi I, Ba I, Sirvent N, et al. Efficacy of tyrosine kinase Inhibitors in Ph-like acute lymphoblastic leukemia harboring ABLclass rearrangements. Blood. 2019;134(16): 1351-1355. Tasian SK, Loh ML, Hunger SP. Philadelphia chromosome-like acute lymphoblastic leukemia. Blood. 2017;130(19): 2064-2072. Schwab C, Ryan SL, Chilton L, et al. EBF1PDGFRB fusion in pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL): genetic profile and clinical implications. Blood. 2016;127(18):2214-2218. Grioni, A, Fazio G, Rigamonti S, et al. A simple RNA target capture NGS strategy for fusion genes assessment in the diagnostics of pediatric B-cell acute lymphoblastic leukemia. HemaSphere. 2019;3(3):e250. Gu Z, Churchman ML, Roberts KG, et al. PAX5-driven subtypes of B-progenitor acute lymphoblastic leukemia. Nat Genet. 2019;51(2):296-307. Li JF, Dai YT, Lilljebjörn H, et al. Transcriptional landscape of B cell precursor acute lymphoblastic leukemia based on an international study of 1,223 cases. Proc Natl Acad Sci U S A. 2018;115(50):11711-11720. Maese L, Tasian SK, Raetz EA. How is the Ph-like signature being incorporated into therapy? Best Pract Res Clin Haematol. 2017;30(3):222-228. Jain N, Jabbour EJ, McKay PZ, et al. Ruxolitinib or dasatinib in combination with chemotherapy for patients with relapsed/refractory Philadelphia (Ph)-like acute lymphoblastic leukemia: a phase I-II trial. Blood. 2017;130(suppl 1):1322. Curran E, Stock W. How I treat acute lymphoblastic leukemia in older adolescents and young adults. Blood. 2015;125(24):37023710. Stock W, Luger SM, Advani AS, et al. A pediatric regimen for older adolescents and young adults with acute lymphoblastic leukemia: results of CALGB 10403. Blood. 2019;133(14):1548-1559. Stanulla M, Dagdan E, Zaliova M, et al. IKZF1plus defines a new minimal residual disease-dependent very-poor prognostic profile in pediatric B-cell precursor acute lymphoblastic leukemia. J Clin Oncol. 2018;36(12):1240-1249. Hunger SP, Saha V, Devidas M, et al. CA180372: An international collaborative phase 2 trial of dasatinib and chemotherapy in pediatric patients with newly diagnosed Philadelphia chromosome positive acute lymphoblastic leukemia (Ph+ ALL) Blood. 2017;130 (Supplement 1):98. Shen S, Chen X, Cai J, et al. Effect of dasatinib vs imatinib in the treatment of pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia: a randomized clinical trial. JAMA Oncol. 2020 Jan 16. [Epub ahead of print]. Elitzur S, Izraeli S. Genomic precision medicine: on the TRK. Blood. 2018;132(8):773774.

haematologica | 2020; 105(9)

CENTENARY REVIEW ARTICLE

Chronic lymphocytic leukemia: from molecular pathogenesis to novel therapeutic strategies

Ferrata Storti Foundation

Julio Delgado,1,2,3 Ferran Nadeu,2,3 Dolors Colomer,2,3,4 and Elias Campo2,3,4

1 Department of Hematology, Hospital Clínic, University of Barcelona, Barcelona; 2Centro de Investigación Biomédica en Red en Oncologia (CIBERONC), Madrid; 3Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona and 4 Hematopathology Section, Hospital Clínic, University of Barcelona, Barcelona, Spain

ABSTRACT

hronic lymphocytic leukemia is a well-defined lymphoid neoplasm with very heterogeneous biological and clinical behavior. The last decade has been remarkably fruitful in novel findings, elucidating multiple aspects of the pathogenesis of the disease including mechanisms of genetic susceptibility, insights into the relevance of immunogenetic factors driving the disease, profiling of genomic alterations, epigenetic subtypes, global epigenomic tumor cell reprogramming, modulation of tumor cell and microenvironment interactions, and dynamics of clonal evolution from early steps in monoclonal B-cell lymphocytosis to progression and transformation into diffuse large B-cell lymphoma. All this knowledge has offered new perspectives that are being exploited therapeutically with novel, targeted agents and management strategies. In this review we provide an overview of these novel advances and highlight questions and perspectives that need further progress to translate this biological knowledge into the clinic and improve patients’ outcome.

Haematologica 2020 Volume 105(9):2205-2217

History Chronic lymphocytic leukemia (CLL) is a lymphoid malignancy characterized by the proliferation and accumulation of mature CD5+ B cells in the blood, bone marrow and lymphoid tissues. The diagnosis of CLL requires the presence of ≥5 x109/L monoclonal B cells of typical phenotype in the blood. Patients with <5 x109/L circulating CLL-type cells may be diagnosed with small lymphocytic lymphoma if they also present with either lymphadenopathy, organomegaly or extramedullary disease; or with monoclonal B-cell lymphocytosis (MBL) if they do not.1 CLL is the most prevalent type of leukemia in adults in Western countries, with an age-adjusted incidence rate of 4.9 cases per 100,000 inhabitants per year. There is a stark difference between the incidence in men (6.8 cases per 100,000/year) and women (3.5 cases per 100,000/year) and also between Caucasians (7.3 and 3.8 cases per 100,000/year for men and women, respectively), African Americans (4.9 and 2.4 cases per 100,000/year for men and women, respectively) and Asian Americans (1.5 and 0.7 cases per 100,000/year for men and women, respectively).2 The disease may have a stable course but also become aggressive, with frequent relapses, or even transform into an aggressive lymphoma, typically diffuse large B-cell lymphoma (DLBCL) (Richter transformation). In the last decade, genomic and epigenomic studies have expanded our knowledge of the pathogenesis of CLL remarkably, unraveling a large number of novel alterations that might drive the evolution of the disease.3–7 Moreover, understanding the crosstalk between tumor cells and their microenvironment has been fundamental in the development of new, targeted agents, which are transforming the way we manage the disease. In this review we provide an overview of these novel advances and how they relate to our understanding of the pathogenesis and current management of CLL.

Pathogenesis Genetic predisposition

Correspondence: ELIAS CAMPO ecampo@clinic.cat Received: April 29, 2020. Accepted: June 18, 2020. Pre-published: July 2, 2020. doi:10.3324/haematol.2019.236000 ©2020 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.

Family studies have consistently shown that first-degree relatives of patients with CLL have a 2- to 8-fold increased risk of developing the disease.8 Genomehaematologica | 2020; 105(9)

2205

J. Delgado et al.

wide association studies have identified up to 45 susceptibility loci, mostly mapping to non-coding regions of the genome.8 The mechanisms linking these susceptibility variants and the development of the disease are being elucidated thanks to integrated genome-wide association/ transcriptome/epigenome studies. These analyses recently revealed that 93% of the susceptibility loci are located in active promoters or enhancers and modify the binding sites of a number of transcription factors (e.g., FOX, NFAT and TCF/LEF) that, in turn, alter the expression of more than 30 genes involved in immune response, cell survival, or Wnt signaling (Figure 1).9 Despite these advances, molecular analysis for predisposition to CLL remains investigational.

Cell of origin Hematopoietic stem cells derived from patients with CLL seem epigenetically primed to clonal expansions of CLL-like cells when implanted in mice. Interestingly, these clonal expansions do not always carry the same genomic aberrations as the original disease.10 Moreover, hematopoietic stem cells derived from patients with CLL express higher levels of transcription factors, such as TCF3, IKZF1 or IRF8, than those from healthy donors, which is intriguing if we consider that some susceptibility loci increase TCF3 binding or IRF8 expression.9 Mutations in driver genes such as NOTCH1 or SF3B1 may be acquired by hematopoietic stem cells, but also at more

advanced stages of B-cell differentiation, explaining why these genomic aberrations are frequently subclonal.11–13 These alterations observed in early steps of B-cell development are also consistent with the identification of shared mutations in CLL and myeloid cells and the detection of oligo- and multi-clonality in patients with MBL/CLL.14–16 The B-cell receptor (BCR) is crucial for CLL pathogenesis and is composed of immunoglobulin (IG) molecules plus CD79a/b subunits. From an immunogenetic point of view, two major molecular subgroups have been identified: those harboring unmutated IG heavy-chain variable region (IGHV) genes (U-CLL, ≥98% identity with the germline) and those with mutated IGHV genes (MCLL).17,18 U-CLL originates from B cells that have not experienced the germinal center, whereas M-CLL originates from post-germinal center B cells.19 In addition, around 30% of patients have highly homologous amino acid sequences derived from almost identical IG rearrangements, known as stereotypes.20 Several hundred stereotypes have been identified, of which 19 are considered major due to their frequency. The prognostic importance of several stereotypes has been prospectively validated.21 The presence of stereotypes and the remarkable bias in the use of certain IGHV genes highlight the relevance of antigen selection in CLL clonal expansion. Interestingly, the IG portion of the BCR may also recognize homotypic epitopes that trigger downstream signaling.22,23 In this sense, the acquisition of the mutation at position 110

Figure 1. Genetic susceptibility mechanisms. Most susceptibility loci map to non-coding regions of the genome, are mainly located in active promoters or enhancers, and modify the binding sites of a number of transcription factors. As a consequence, the binding sites for SPI1 and NF-κB are disrupted, whereas there is an increased affinity for members of the FOX, NFAT and TCF/LEF families. This, in turn, alters the expression of genes involved in the immune response (SP140, IRF8), cell survival (BCL2, BMF, CASP8, BCL2L11) or Wnt signaling (UBR5, LEF1).9

2206

haematologica | 2020; 105(9)

CLL pathogenesis and management

(G>C, glycine-arginine) of IGLV3-21*01 mediated by somatic hypermutation confers autonomous BCR signaling.24 This change is present in 7-18% of CLL and seems responsible for the adverse outcome associated with the use of IGLV3-21 independently of the mutational status of the IGHV.24,25 Epigenetic studies have shown that, although both CLL subtypes are antigen-experienced, M-CLL keeps a methylation signature of germinal center-experienced cells (memory-like B cells), whereas U-CLL has a pre-germinal center, naïve-like methylation signature.5,26 Of note, these epigenetic studies also identified a third subtype with an intermediate profile made of cases with moderate IGHV mutation levels. All three epigenetic subsets have different usage of IGHV genes, stereotypes, genomic aberrations and clinical outcome (Table 1).27 Their prognostic relevance has been validated in retrospective cohorts and clinical trials.26–28 The intermediate epigenetic subtype may be more heterogeneous than initially thought since it includes most stereotype subset 2 cases with aggressive behavior whereas other cases may behave more indolently. The understanding of the biological significance of this subtype requires further analysis.

The microenvironment in chronic lymphocytic leukemia CLL cells are highly dependent on signals coming from the microenvironment for proliferation and survival. Tumor cells proliferate primarily in lymph nodes, and to less extent in bone marrow,29 where they are in intimate contact with extracellular matrix, T cells, nurse-like cells, follicular dendritic cells and other stromal cells (Figure 2). The interactions between CLL cells and this complex microenvironment are mediated by a network of adhesion molecules, cell surface ligands, chemokines, cytokines, and their respective receptors. CLL cells organize their supportive inflammatory milieu and promote an immunosuppressive microenvironment through different mechanisms, such as secretion of soluble factors, cell-tocell contact, and release of extracellular vesicles (Figure 2).29,30 Environmental or self-antigens and homotypic interactions trigger BCR and Toll-like receptor (TLR) signaling, amplifying the response of CLL cells to other signals from the microenvironment and increasing the activation of anti-apoptotic and proliferation pathways.31,32 Genomic studies have identified recurrent mutations in genes regulating tumor cell-microenvironment interactions, which are already required for tumor cell growth. Thus, NOTCH1 mutations are dependent on the presence of Notch ligands in the microenvironment and activate processes such as cell migration, invasion and angiogenesis.33,34 BCR and NOTCH1 pathways are functionally linked, mutually enhancing their activation.35 MYD88 mutations activate the NF-κB pathway in response to TLR ligands, increasing the cytokine release involved in recruiting stromal and T cells.36 Tumor cells also reconfigure the function of T- and myeloid-derived cells towards a leukemia-supportive and immunosuppressive microenvironment.30,37 Thus, tumor cells reduce T-cell motility and the effector function of CD4+ cells while inducing CD8+-cell exhaustion38–41 and monocyte differentiation towards macrophages with protumoral functions (M2like) and nurse-like cells.37 Many studies have confirmed the fundamental role of BCR activation for CLL pathogenesis.42 Several proteins, haematologica | 2020; 105(9)

including phosphatidylinositol 3 kinase (PI3K), Bruton tyrosine kinase (BTK) and spleen tyrosine kinase (SYK) are essential for BCR signal transduction.43 The effect of BCR-mediated signaling varies according to IGHV mutation status: M-CLL cells are generally driven towards anergy, whereas U-CLL cells are more directed towards cell growth and proliferation.44 Moreover, anergic cells normally retain a higher susceptibility to apoptosis unless anti-apoptotic proteins such as BCL2 are overexpressed, as is the case for CLL cells.45 Indeed, most major therapeutic advances occurring in the last decade are related to the inhibition of BCR and BCL2-mediated signaling.

Structural genomic aberrations Initial chromosome banding analysis revealed that deletions or trisomies were relatively common but only observed in fewer than half of the patients.46 With the advent of fluorescent in situ hybridization (FISH), genomic aberrations were identified in more than 80% of patients, the more relevant being trisomy 12, 13q deletion [del(13q)], 11q deletion [del(11q)] and 17p deletion [del(17p)];47 and FISH became the gold standard for genomic evaluation in CLL. Later, the introduction of more effective mitogens expanded the use of chromosome banding analysis in CLL and revealed other aberrations that could not be detected by FISH, including chromosome translocations in 20-35% of the cases.48 These translocations may occur in the context of complex karyotypes. The most common rearrangements involve 13q14, with multiple partners, and the IGH locus. The genes most commonly rearranged with IGH are BCL2 [t(14;18)(q32;q21)] (2% of cases, usually M-CLL);3,49 and BCL3 [t(14;19)(q32;q13)] or BCL11A [t(2;14)(p16;q32)] (<1% of cases, usually U-CLL with atypical features).50,51 Chromosomal microarray analysis has identified novel copy number alterations and also copy number neutral loss of heterozygosity.3,52 This latter is observed in 5% of patients, typically affects deleted regions such as 11q, 17p and 13q, and is associated with mutations of the target gene, particularly TP53.52,53 A novel use of both chromo-

Table 1. Clinical and biological characteristics of the three epigenetic subtypes in chronic lymphocytic leukemia.20,27,137

Methylation cell signature

Naïve-like

Intermediate

Memory-like

Typical IG genes

IGHV1, -5, -7 IGHD6-19 IGHJ4 IGKV1-39 1 Unmutated

IGHV3-21 IGHJ6 IGLV3-21

IGHV4-34 IGHD5-18 IGHJ6 IGKV2-30 4 Mutated

Typical stereotype subset IGHV mutations

Mutated drivers

Clinical outcome27

NOTCH1 NFKBIE TP53 Aggressive; TTFT at 10 years = 97%

2 Mutated or unmutated (around the 98% cutoff) SF3B1 del(11q) rarely TP53 Intermediate; TTFT at 10 years = 38%

del(13q) Indolent; TTFT at 10 years = 24%

IG: immunoglobulin; IGHV: immunoglobulin heavy-chain variable region; TTFT: time to first treatment; ND: not determined.

2207

J. Delgado et al.

Figure 2. The chronic lymphocytic leukemia microenvironment. Communication between chronic lymphocytic leukemia (CLL) cells and stromal cells, T cells and nurse-like cells (NLC) is established and maintained by direct contacts, chemokine/cytokine receptors, adhesion molecules and ligand-receptor interactions. CLL cells migrate to tissues attracted by the chemokines CXCL12 secreted by NLC and stromal cells, CXCL13 by follicular dendritic cells (FDC), and CCL19/CCL21 by highendothelial venules, which interact with the CLL receptors CXCR4, CXCR5 and CCR7, respectively. Adhesion molecules (e.g., a4β1 integrin, LFA-1) and their ligands (VCAM1, ICAM, among others) facilitate tumor cell migration and homing. Environmental or auto-/self-antigens and homotypic IG interactions trigger B-cell receptor (BCR) activation capable of driving CLL proliferation.22,138 Interactions between CD40 and CD40 ligand (CD40L) on activated CD4+ T cells are critical in the context of antigen presentation and induction of normal B-cell responses. Activated CLL cells secrete chemokines (CCL2, CCL3, and CCL4) and angiogenic factors that attract T cells and different stromal cells.30,139 Suppressive factors (IL-10)140 and immune inhibitory molecules (PD-L1 among others)141 facilitate tumor cells to evade immuneresponse and maintain tolerance. Anti-tumor CD8+ T cells become exhausted by constant exposure to tumor-derived antigens leading to cell exhaustion.40 Regulatory T cells (Treg) exert an inhibitory effect on CD4+ and CD8+ cells through secretion of suppressive cytokines.142 Tumor-released extracellular vesicles carrying noncoding RNA and proteins induce an inflammatory phenotype in T cells, monocytes, and stromal cells.143

some banding analysis and chromosomal microarray analysis is the identification of complex karyotypes, observed in up to 20% of patients with CLL. A complex karyotype appears not only prognostic but also predictive in the context of treatment with both conventional and novel agents.54,55 Intriguingly, a subset of patients with complex karyotypes carrying trisomy 12, trisomy 19, and additional trisomies seem to correspond to a particular genetic subgroup with favorable outcome.54–56 In contrast to patients with other hematologic malignancies, patients with five or more aberrations have a worse prognosis compared to those with “less complex” karyotypes (3 or 4 aberrations).55

Mutational landscape On average, CLL tumors accumulate around 2,500 somatic mutations with a clear difference between MCLL and U-CLL (3,000 vs. 2,000 somatic mutations on average, respectively). This increased mutation burden observed in M-CLL has limited transformation potential as patients with M-CLL have fewer mutated drivers and better clinical outcomes than patients with U-CLL.3 The mutational landscape of the disease is remarkably heterogeneous, with only a handful of genes mutated in more 2208

than 5% of patients at diagnosis (NOTCH1, SF3B1, TP53, ATM) followed by a long tail of genes mutated at lower frequencies.3,4 Despite this diversity, most mutated drivers cluster in a number of cell pathways (Figure 3), such as NOTCH1 signaling (NOTCH1, FBXW7); BCR and TLR signaling (EGR2, BCOR, MYD88, TLR2, IKZF3); the MAPK-ERK pathway (KRAS, NRAS); NF-κB signaling (BIRC3, NFKB2, NFKBIE, TRAF2, TRAF3); chromatin modifiers (CHD2, SETD2, KMT2D, ASXL1); cell cycle (ATM, TP53, CCND2, CDKN1B, CDKN2A); DNA damage response (ATM, TP53, POT1); and RNA splicing and metabolism (SF3B1, U1, XPO1, DDX3X, RPS15). This clustering suggests that mutations belonging to the same cell process may have similar functional and clinical impacts,57,58 but this hypothesis requires further confirmation. A detailed descriptions of these cell processes can be found elsewhere.51 This accumulation of low-frequency driver alterations highlights the striking interpatient heterogeneity of CLL, which is partly determined by three major factors: (i) the cell of origin: M-CLL have fewer driver mutations than UCLL and some mutated genes are almost exclusively or predominantly seen in one of the two subtypes (e.g MYD88 and PAX5 in M-CLL and U1, NOTCH1, and haematologica | 2020; 105(9)

CLL pathogenesis and management

Figure 3. Recurrently mutated genes and pathways in chronic lymphocytic leukemia. The main molecular pathways affected by mutations in chronic lymphocytic leukemia are depicted. Genes mutated at higher frequencies (>5%) in newly diagnosed patients are highlighted in bold.

POT1 in U-CLL), whereas others are seen in both subtypes; (ii) age of the patients: MYD88 mutations seem to be more frequent in younger patients; (iii) disease evolution: some mutations (SF3B1, POT1, ATM) are more frequent in patients requiring therapy compared to those with stable disease, and some others (TP53, BIRC3, MAP2K1, NOTCH1) are more frequent in patients with progressive disease after chemoimmunotherapy (CIT).3,4,59 The co-occurrence of many of these driver alterations within the same tumor complicates the analysis of their relative clinical relevance. For instance, mutations in SF3B1, POT1 or XPO1 are generally associated with poor prognosis, but they rarely appear on their own.13 This has led some investigators to propose a multi-hit model in which the accumulation of driver mutations, regardless of the individual genes targeted by each of these mutations, gradually impairs patientsâ&#x20AC;&#x2122; outcome.3,13 Indeed, the survival of patients in whom no driver aberrations are identified is comparable to that of individuals in the general population, further reinforcing this concept.3 Deep, targeted next-generation sequencing has revealed that subclonal mutations (i.e., those present in only a fraction of tumor cells) can be detected for all driver genes and are associated with rapid disease progression haematologica | 2020; 105(9)

and poor outcome.11â&#x20AC;&#x201C;13 This is particularly relevant for TP53 mutations given the fact that, as explained below, CLL therapy is based on the presence or absence of these mutations. The current consensus is that, apart from clonal mutations, subclonal mutations with a variant allelic frequency ranging from 5 to 10% (and therefore below the threshold of detection by conventional molecular techniques) could also be reported, whereas those with a variant allelic frequency lower than 5% should not, but there is much controversy around these issues and this recommendation may well change in the future.60,61 Furthermore, the analysis of clonal and subclonal aberrations has also allowed the reconstruction of each tumorâ&#x20AC;&#x2122;s phylogeny. Thus, clonal aberrations, which are mostly structural abnormalities [e.g. trisomy 12, del(13q)] generally correspond to earlier driving events, while subclonal mutations in driver genes (e.g., SF3B1, POT1, NOTCH1) are acquired later over the course of the disease.13 Moreover, some genes appear to be specifically selected at relapse. For instance, small clones harboring TP53 mutations typically expand and dominate the disease after CIT, which explains the poor prognosis associated with these subclonal mutations.12,62 Apart from TP53, mutations in IKZF3 and SAMHD1 have also been recur2209

J. Delgado et al.

rently selected in small cohorts of patients after CIT.63,64 Clonal evolution plays an important role not only in resistance to CIT, but also to novel agents. Indeed, different point mutations have been identified in the BTK and PLCG2 genes in patients previously treated with the BTK inhibitor ibrutinib,65 and in the BCL2 gene in patients relapsing after treatment with the BCL2 antagonist venetoclax.66 Resistance to these agents has been associated with these mutations in around 70% of cases, although they are usually subclonal and their specific role causing resistance needs to be proven.67 Other resistance mechanisms involve upregulation of BCL-XL and MEK/ERK, and cell reprogramming and transdifferentiation to cell subtypes that do not require BCR signaling.65,67–70

Epigenomic landscape The genome of CLL features widespread hypomethylation, and a large fraction of the differences between UCLL and M-CLL can be attributed to their different cell of origin in germinal center-independent or -experienced B cells, respectively.5 Major hypomethylation changes occur at transcription factor binding sites such as TCF3, PU.1/SPIB, NFAT and EGR, and enhancers that modulate genes relevant for CLL pathogenesis involved in B-cell function, BCR signaling, and NF-κB activation among others. This methylation profile is already acquired at the MBL stage3 and remains relatively stable over time. However, some CLL have intratumor variability in certain regions, which may alter the expression of several genes and facilitate tumor evolution.71 Of note, this variability is greater in U-CLL than in M-CLL and is associated with increasing number of subclones.7,71 Several groups have evaluated the full reference epigenome of CLL, providing a genome-wide map of histone marks and three-dimensional chromatin architecture.6,72–74 Surprisingly, there was a significant variability in active regulatory regions among individual patients. This variability, and also the total number of active sites, was larger in U-CLL than in M-CLL. Around 80% of these active sites were also present in normal naïve, germinal center, memory, or plasma cells.6 Some of these active regions are seen in all CLL cases but in none of the normal B-cell subtypes, and may therefore be crucial for CLL pathogenesis. Most of these de novo active regions target regulatory loci and super-enhancers enriched in transcription binding motifs of NFAT, FOX, TCL/LEF, and PAX5, which have been shown to play a role in CLL pathogenesis and could potentially be targeted pharmacologically.26,72 Somatic mutations in chromatin remodeler genes could modify the epigenomic landscape of CLL, but they are uncommon in this malignancy compared to other lymphoid neoplasms. CHD2 is mutated in 5% of CLL and 7% of MBL.75 The histone methyltransferase SETD2 and ARID1A are also mutated in a small proportion of patients. Of note, MYD88 mutations and trisomy 12 are associated with specific remodeling of chromatin activation and accessibility regions. More specifically, the epigenomic profile induced by MYD88 mutations targets regulatory regions related to NF-κB signaling,6 whereas the epigenetic configuration of trisomy 12 CLL is characterized by a subtype-specific hypomethylation signature associated with increased H3K27 acetylation, which leads to the overexpression of 25 target genes including RUNX3.76 2210

Pathogenic mechanisms in the evolution of the disease CLL is always preceded by an often unnoticed premalignant state known as high-count MBL.77 Low-count MBL may persist for a long time but the risk of progression is negligible.78 Yearly, 1% of cases of high-count MBL evolve into CLL requiring therapy,79 and 2-10% of patients with symptomatic CLL eventually develop Richter transformation.80 At the other end of the spectrum, around 30% of patients with CLL never require any CLL-specific therapy and die of other causes, and 1-2% of them even experience spontaneous regression of their disease.81 It is therefore evident that the rate and pattern of growth (or even decline) of the disease can vary greatly among patients. Patients with high-count MBL carry mutations in driver genes which may be detected at a median of 41 months prior to progression to CLL.3,82 The mutation rates for the most common drivers (e.g., SF3B1, DDX3X, BIRC3, ATM) are comparable between MBL and CLL, with only a few genes being more commonly mutated in CLL (NOTCH1, TP53, XPO1).82,83 Patients with MBL with mutated drivers have a shorter time to first treatment compared to cases without mutations. Once CLL is established, the growth dynamics of tumor cells is heterogeneous. Some patients exhibit a logistic-like behavior in which the clone stabilizes over time, whereas some others show an exponential-like growth pattern.84 This exponential growth, clinically defined as “short lymphocyte doubling time” is still considered an adverse prognostic parameter in CLL.85,86 As expected, the median number of driver mutations (both clonal and subclonal) is higher in patients with exponential growth, and this patient population also displays unmutated IGHV genes more frequently. In addition, the rate of clonal evolution after therapy (i.e., with a significant shift in at least one subclone) is also higher in patients with exponential growth (Figure 4). Transformation of CLL into an aggressive lymphoma occurs in 2-10% of patients and in most of them (>90%) corresponds to a DLBCL, but Hodgkin lymphoma may also occur.87 The DLBCL usually emerges as a linear evolution of the same CLL clone with only rare cases deriving from a branching divergent subclone.88 CLL carrying stereotyped subset 8 (IGHV4-39), NOTCH1 or TP53 mutations and complex karyotypes are at higher risk of transformation after CIT. Transformed DLBCL frequently add CDKN2A deletions and MYC translocations or amplifications on top of the genomic alterations already present in the original CLL, but lack the common mutations observed in primary DLBCL indicating that they may correspond to a different biological category.80 Richter transformation also occurs in patients treated with BTK inhibitors. These tumors do not usually acquire BTK or PLCG2 mutations but, if these were present in the original CLL, subclones may emerge with additional independent mutations.89,90 In the rare instances in which the disease regresses spontaneously, the patients uniformly have mutated IGHV genes, no stereotypes, low proliferative activity and poor migration to proliferation centers as exemplified by a high CXCR4 expression. Interestingly, patients with spontaneous regression show reduced T-cell exhaustion and increased T-cell proliferation, confirming the important role of the immune system in CLL progression.91 Moreover, driver mutations are also present in these patients but they always remain stable without subclonal shifts.91 haematologica | 2020; 105(9)

CLL pathogenesis and management

Figure 4. Evolutionary steps and growth dynamics of chronic lymphocytic leukemia. (Left) The progression of monoclonal B-cell lymphocytosis (MBL) to chronic lymphocytic leukemia (CLL) is a linear process discriminated by the total number of lymphocytes. The presence of driver alterations is associated with rapid progression. Although a few alterations are enriched in CLL compared to MBL, both phases share a similar driver composition. (Middle) The main growth dynamics during the pre-treatment phase of the CLL are shown, including spontaneous regression, logistic growth and exponential growth. The main characteristics associated with each pattern are specified. WBC; peripheral white blood cell count. (Right) Richter transformation to diffuse large B-cell lymphoma (DLBCL) is associated with subset #8, NOTCH1 or TP53 mutations and complex karyotype. It follows a linear evolution from the CLL clone through the recurrent acquisition of CDKN2A and MYC alterations.

Clinical and prognostic implications of novel discoveries The clinical course of CLL is rather heterogeneous, ranging from a fairly asymptomatic disease that may even regress spontaneously to a progressive disease that eventually leads to the patient’s death, so there has always been remarkable interest in determining the prognosis of individual patients. Even though many prognostic markers have been identified over the past decades, only a few prevail. For many years, the prognosis of patients with CLL was defined using purely clinical parameters, such as those included in the Rai and Binet staging systems,92,93 the IGHV mutational status,17,18 and numerical aberrations as determined by FISH.47 With the advent of next-generation sequencing, novel drivers were discovered (NOTCH1, SF3B1, BIRC3) and incorporated into these prognostic systems, but none of these attempts succeeded in becoming standard of care.94–96 Indeed, the International Workshop on CLL (iwCLL) guidelines only recommend evaluating the IGHV status and presence/absence of TP53 aberrations in routine practice.86 The recent CLL International Prognostic Index (CLL-IPI) incorporates both clinical and cytogenetic/genomic data (age, clinical staging, β2microglobulin serum concentration, IGHV mutation status and TP53 aberrations) into one prognostic score.97 The CLL-IPI was developed in cohorts of patients treated with CIT and has been validated in retrospective series.98–100 Among the five items, both TP53 and IGHV have the strongest impact on a patient’s outcome, and it is therefore not surprising that simplified versions of the CLL-IPI incorporating only these two markers have been proposed.101 A recent study has determined that a score based on the presence of unmutated IGHV, absolute lymphocyte count >15 x109/L, and palpable lymph nodes predicts for a shorter time to first treatment in patients with early, asymptomatic disease.102 On the other hand, several groups are advocating for the incorporation of novel markers, such as a complex karyotype55 or epigenetic subsets,27,28 into clinical practice. All these novel prognostic haematologica | 2020; 105(9)

and/or predictive models will need to be validated in cohorts of patients treated with targeted agents.

Treatment Treatment for CLL has changed remarkably in the last decade (Figure 5). The mainstay of therapy used to be CIT - a combination of conventional chemotherapeutic agents plus a monoclonal antibody, such as rituximab or obinutuzumab, - although this is no longer the case, at least for most patients. Novel, targeted agents are now the preferred option, and among them, the drugs currently approved by both the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) are the BTK inhibitor ibrutinib, the BCL2 inhibitor venetoclax and the PI3K inhibitor idelalisib, while the second-generation BTK inhibitor acalabrutinib and PI3K inhibitor duvelisib have already been approved by the FDA and are under evaluation by the EMA.

Frontline therapy Not all patients with CLL require therapy. Despite all recent advances, the iwCLL still recommends watchful observation for patients with asymptomatic disease.86 This recommendation is based on at least two randomized trials comparing observation to either chlorambucil monotherapy or fludarabine, cyclophosphamide and rituximab (FCR).103,104 Both trials concluded that early therapy in asymptomatic patients was not associated with a prolonged overall survival. Very recently, preliminary results from a third trial comparing ibrutinib versus observation were presented.105 Patients receiving ibrutinib had a longer event-free survival, but no overall survival advantage, although the results were still immature. Moreover, although severe adverse events rates were comparable between groups, patients receiving ibrutinib had a higher incidence of some specific adverse events such as bleeding, hypertension and atrial fibrillation. For patients with symptomatic disease requiring therapy, ibrutinib is often recommended based on four phase III 2211

J. Delgado et al.

Figure 5. Recommended therapy for patients with symptomatic chronic lymphocytic leukemia. M-CLL, mutated IGHV; U-CLL, unmutated IGHV; FCR: fludarabine + cyclophosphamide + rituximab; BR: bendamustine + rituximab; V: venetoclax; VR: venetoclax + rituximab; VO: venetoclax + obinutuzumab; I: ibrutinib; IO: ibrutinib + obinutuzumab; A: acalabrutinib; ClbO: chlorambucil + obinutuzumab; R: rituximab; D: duvelisib; AlloHCT: allogeneic hematopoietic cell transplantation; R-CHOP, rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone; IGHV, immunoglobulin heavy-chain variable region. *Acalabrutinib (A) is approved for both treatment-naïve and relapsed disease by the FDA but not the EMA. #Venetoclax (V) monotherapy is approved for patients with TP53 aberrations who are refractory or intolerant to ibrutinib or patients without TP53 aberrations who are refractory or intolerant to both chemoimmunotherapy and ibrutinib. Venetoclax plus rituximab (VR) is approved for any patient with relapsed disease. †Duvelisib (D) monotherapy is approved for relapsed disease (minimum of two prior therapies) by the FDA but not the EMA. ‡AlloHCT is recommended for appropriate patients with high-risk disease, defined by TP53 aberrations and/or complex karyotype in whom ibrutinib and/or venetoclax has failed. Allogeneic HCT is also recommended for appropriate patients with transformed disease who have responded to salvage chemotherapy (e.g., R-CHOP).

randomized clinical trials comparing ibrutinib with chlorambucil monotherapy106 and other commonly used CIT combinations, namely FCR, bendamustine plus rituximab and chlorambucil plus obinutuzumab (ClbO).107–109 Ibrutinib was superior to chlorambucil and all CIT combinations in terms of response rate and progression-free survival, and even conferred a longer overall survival compared to that provided by chlorambucil monotherapy and FCR.106,107 In these trials, ibrutinib was sometimes combined with a monoclonal antibody, either rituximab or obinutuzumab, and sometimes given as monotherapy, but the true added value of the monoclonal antibody in this context is unknown.108,110 In terms of toxicity, ibrutinib was less toxic than CIT combinations when severe adverse events or toxic deaths were considered.107–109 Apart from ibrutinib, patients with M-CLL, devoid of 2212

TP53 aberrations and fit enough to tolerate FCR therapy, may still be good candidates for the latter, with the benefit being that this treatment can be completed in 6 months while ibrutinib must be taken indefinitely. This option would be particularly valuable for non-compliant patients or those in whom ibrutinib is contraindicated. If FCR is the treatment of choice, caution must be taken in patients with NOTCH1 mutations, in whom rituximab appears to have little added value.59 Other genomic subgroups, such as patients with BIRC3 mutations appear to derive little benefit from CIT,111,112 but these results should be further validated. Unfit patients also have the alternative of venetoclax plus obinutuzumab (VO) as frontline therapy. This is based on a phase III trial that compared VO with ClbO in elderly/unfit patients.113 VO was superior in terms of haematologica | 2020; 105(9)

CLL pathogenesis and management

response rate and progression-free survival, and had a comparable safety profile. In this trial VO was administered for a definite period of time (2 years), which is quite appealing for older/unfit patients. Moreover, many well established adverse prognostic markers, including U-CLL, ATM aberrations or NOTCH1/BIRC3 mutations, lost their negative effect in patients treated with VO. The only factor that remained predictive of a shorter progression-free survival in this cohort of patients was TP53 aberrations.112 Finally, the alternative BTK inhibitor acalabrutinib was recently approved by the FDA (not by the EMA yet) as frontline therapy in view of the results of a phase III trial comparing acalabrutinib versus ClbO.114

Relapsed/refractory disease Treatment for relapsed/refractory disease must be decided depending on prior therapy and also the reason why the original treatment was no longer appropriate (e.g., refractoriness vs. intolerance). Ibrutinib is the current gold standard therapy for patients with relapsed/refractory disease, based on the results of several phase I-III trials,115â&#x20AC;&#x201C;119 but this is also changing for two main reasons: (i) an increasing proportion of patients currently receive ibrutinib as frontline therapy; and (ii) a few serious contenders have appeared in the last year. Venetoclax is one of the best alternatives in this situation, including patients with high-risk genomic aberrations. The drug was already proven effective and safe in several phase I-II trials, in patients who had previously received either CIT or BTK/PI3K inhibitors.120â&#x20AC;&#x201C;123 The formal confirmation of this promising activity came with a phase III trial in which venetoclax combined with rituximab was superior to bendamustine plus rituximab in terms of response rate, progression-free survival and overall survival, leading to its full approval for patients with relapsed/refractory CLL.124 Other possibilities are PI3K inhibitors and alternative BTK inhibitors. Idelalisib, in combination with rituximab, was the first PI3K inhibitor approved for the treatment of relapsed/refractory CLL based on the results of a phase III trial,125,126 and yet it is infrequently used because of its less favorable adverseevent profile. It may have a role in patients with complex karyotypes,127 who have a higher risk of progression and/or transformation when treated with ibrutinib or venetoclax,90,128 or in older patients who also tend not to tolerate ibrutinib well,129 but there are no randomized data to substantiate this potential superiority. Duvelisib was the second PI3K inhibitor approved by the FDA, also based on a phase III randomized trial.130 The efficacy and safety profile of the drug appear comparable with those of idelalisib, if not slightly advantageous. Regarding alternative BTK inhibitors, there are several products in development, but only acalabrutinib is approved by the FDA for the treatment of relapsed/refractory CLL. This is based on a phase III trial in which acalabrutinib was superior to either bendamustine plus rituximab or idelalisib plus rituximab.131 In this trial, prior ibrutinib therapy was not allowed, but a separate trial has shown that 85% of patients who were intolerant to ibrutinib were subsequently able to take acalabrutinib, with a 76% response rate.132 Despite all recent therapeutic advances, a proportion of patients will still fail to respond and should be considered for curative therapy. Currently, only allogeneic hematopoietic cell transplantation can be considered potentially curative, but it is also associated with considhaematologica | 2020; 105(9)

erable morbidity and mortality. Over the past decades, the number of patients referred for allogeneic hematopoietic cell transplantation has dropped significantly,133 but the procedure should be recommended to young/fit patients in whom BCR/BCL2 inhibitor treatment fails, particularly in those with TP53 aberrations, or in the case of Richter transformation.134,135 Moreover, although chimeric antigen receptor T cells could also be appropriate in this situation and the results are promising,136 none of the commercially available products is yet approved for this indication.

Disease transformation Richter transformation remains an ominous event for patients with CLL, particularly when it is clonally related to the original CLL, because none of the recently approved novel agents is truly effective. Indeed, disease transformation is a relatively common cause of failure to benefit from these drugs.90,128,129 Histological confirmation is always recommended since it can guide prognosis (i.e., Hodgkin lymphoma and clonally unrelated tumors have more favorable prognosis). Patients with transformed disease should be offered conventional CIT (e.g., RCHOP: rituximab plus cyclophosphamide, doxorubicin, vincristine, prednisone) followed by allogeneic hematopoietic cell transplantation in the case of response. Autologous hematopoietic cell transplantation remains an option if allogeneic transplantation is considered inappropriate.134 Chimeric antigen receptor T cells may also be effective but, unfortunately, none of the approved products is current available for patients with Richter transformation.

Conclusions and perspectives Recent molecular studies have provided many insights into the processes that govern the development and progression of CLL, including many novel mutated genes clustered in different functional pathways. The CLL epigenome is reprogrammed through the modulation of regulatory regions that appear de novo in the disease, whereas other regions maintain functions already present in different stages of B-cell differentiation. Analysis of the CLL microenvironment has provided clues to understand the survival of tumor cells and resistance to therapy. All this knowledge has offered new perspectives that are being exploited therapeutically with novel agents and strategies. However, these studies are also raising new questions. The relationship between the remarkable molecular heterogeneity of the disease and the clinical diversity is not well understood. The disease is always preceded by a premalignant state (MBL) which shares most molecular drivers with overt CLL. In many cases, these molecular drivers remain constant over time. However, clonal evolution is also possible and is usually associated with exponential tumor growth, progressive disease and, in some cases, disease transformation. Most studies have been performed in pretreated patients and it is not fully understood how the genome and epigenomic alterations and microenvironmental interactions influence the evolution of the disease. Translating new knowledge into clinical practice will require an effort to obtain an integrated view of all these factors in order to understand the disease better and design effective treatments and management strategies. 2213

J. Delgado et al.

Acknowledgments The authors would like to thank Silvia Beà, Jose Ignacio Martín-Subero, and Armando Lopez-Guillermo (Hospital Clinic of Barcelona and IDIBAPS) for their helpful comments on the manuscript, and Neus Giménez (IDIBAPS) for her assistance with the figure design. JD is supported by a grant from Generalitat de Catalunya (PERIS IPFE SLT006/17/301). FN is supported by a predoctoral fellowship from the Ministerio de Ciencia e Innovación (MCI) (BES2016-076372). DC is supported by MCI (RTI2018–094584B-I00), Generalitat de Catalunya Suport Grups de Recerca (AGAUR, grant 2017-SGR-1009) and CIBERONC (CB16/12/00334). EC is supported by grants from “La

References 1. Campo E, Ghia P, Montserrat E, MüllerHermelink HK, Stein H, Swerdlow SH. Chronic lymphocytic leukaemia/small lymphocytic lymphoma. In: WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Rev 4th Ed. Lyon: WHO Press; 2017:216-221. 2. Howlader N, Noone A, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2016, National Cancer Institute. Bethesda, MD, USA. (Last accessed: January 15, 2020). 3. Puente XS, Beà S, Valdés-Mas R, et al. Noncoding recurrent mutations in chronic lymphocytic leukaemia. Nature. 2015;526 (7574):519-524. 4. Landau DA, Tausch E, Taylor-Weiner AN, et al. Mutations driving CLL and their evolution in progression and relapse. Nature. 2015;526(7574):525-530. 5. Kulis M, Heath S, Bibikova M, et al. Epigenomic analysis detects widespread gene-body DNA hypomethylation in chronic lymphocytic leukemia. Nat Genet. 2012;44(11):1236-1242. 6. Beekman R, Chapaprieta V, Russiñol N, et al. The reference epigenome and regulatory chromatin landscape of chronic lymphocytic leukemia. Nat Med. 2018;24(6):868-880. 7. Oakes CC, Claus R, Gu L, et al. Evolution of DNA methylation is linked to genetic aberrations in chronic lymphocytic leukemia. Cancer Discov. 2014;4(3):348-361. 8. Berndt SI, Camp NJ, Skibola CF, et al. Metaanalysis of genome-wide association studies discovers multiple loci for chronic lymphocytic leukemia. Nat Commun. 2016;7(1): 10933. 9. Speedy HE, Beekman R, Chapaprieta V, et al. Insight into genetic predisposition to chronic lymphocytic leukemia from integrative epigenomics. Nat Commun. 2019;10(1): 3615. 10. Kikushige Y, Ishikawa F, Miyamoto T, et al. Self-renewing hematopoietic stem cell is the primary target in pathogenesis of human chronic lymphocytic leukemia. Cancer Cell. 2011;20(2):246-259. 11. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. Cell. 2013;152(4):714-726. 12. 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.

2214

Caixa” Foundation (CLLEvolution-LCF/PR/HR17/ 52150017), the Instituto de Salud Carlos III and the European Regional Development Fund (FEDER – “Una Manera de Hacer Europa”) (PMP15/00007), MCI (grants RTI2018094274-B-I00 and SAF2016-81860-REDT), CIBERONC (CB16/12/00225) and AGAUR (2017-SGR-1142). EC is an Academia Researcher of the Institució Catalana de Recerca i Estudis Avançats (ICREA) of the Generalitat de Catalunya. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (BCLLatlas - 810287). We apologize to authors whose work was not included due to space limitations.

2016;127(17):2122-2130. 13. Nadeu F, Clot G, Delgado J, et al. Clinical impact of the subclonal architecture and mutational complexity in chronic lymphocytic leukemia. Leukemia. 2018;32(3):645653. 14. Agathangelidis A, Ljungström V, Scarfò L, et al. Highly similar genomic landscapes in monoclonal b-cell lymphocytosis and ultrastable chronic lymphocytic leukemia with low frequency of driver mutations. Haematologica. 2018;103(5):865-873. 15. Brazdilova K, Plevova K, Skuhrova Francova H, et al. Multiple productive IGH rearrangements denote oligoclonality even in immunophenotypically monoclonal CLL. Leukemia. 2018;32(1):234-236. 16. Plevova K, Francova HS, Burckova K, et al. Multiple productive immunoglobulin heavy chain gene rearrangements in chronic lymphocytic leukemia are mostly derived from independent clones. Haematologica. 2014;99(2):329-338. 17. Hamblin TJ, Davis Z, Gardiner A, Oscier DG, Stevenson FK. Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood. 1999;94(6):1848-1854. 18. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood. 1999;94(6):18401847. 19. Seifert M, Sellmann L, Bloehdorn J, et al. Cellular origin and pathophysiology of chronic lymphocytic leukemia. J Exp Med. 2012;209(12):2183-2198. 20. Stamatopoulos K, Agathangelidis A, Rosenquist R, Ghia P. Antigen receptor stereotypy in chronic lymphocytic leukemia. Leukemia. 2017;31(2):282-291. 21. Jaramillo S, Agathangelidis A, Schneider C, et al. Prognostic impact of prevalent chronic lymphocytic leukemia stereotyped subsets: analysis within prospective clinical trials of the German CLL Study Group (GCLLSG). Haematologica. 2019 Dec 26. [Epub ahead of print] 22. Dühren-von Minden M, Übelhart R, Schneider D, et al. Chronic lymphocytic leukaemia is driven by antigen-independent cell-autonomous signalling. Nature. 2012;489(7415):309-312. 23. Minici C, Gounari M, Übelhart R, et al. Distinct homotypic B-cell receptor interactions shape the outcome of chronic lymphocytic leukaemia. Nat Commun. 2017;8(1): 15746.

24. Maity PC, Bilal M, Koning MT, et al. IGLV321∗01 is an inherited risk factor for CLL through the acquisition of a single-point mutation enabling autonomous BCR signaling. Proc Natl Acad Sci U S A. 2020;117(8):4320-4327. 25. Stamatopoulos B, Smith T, Crompot E, et al. The light chain IgLV3-21 defines a new poor prognostic subgroup in chronic lymphocytic leukemia: results of a multicenter study. Clin Cancer Res. 2018;24(20):5048-5057. 26. Oakes CC, Seifert M, Assenov Y, et al. DNA methylation dynamics during B cell maturation underlie a continuum of disease phenotypes in chronic lymphocytic leukemia. Nat Genet. 2016;48(3):253-264. 27. Queirós AC, Villamor N, Clot G, et al. A Bcell epigenetic signature defines three biologic subgroups of chronic lymphocytic leukemia with clinical impact. Leukemia. 2015;29(3):598-605. 28. Wojdacz TK, Amarasinghe HE, Kadalayil L, et al. Clinical significance of DNA methylation in chronic lymphocytic leukemia patients: results from 3 UK clinical trials. Blood Adv. 2019;3(16):2474-2481. 29. Herndon TM, Chen SS, Saba NS, et al. Direct in vivo evidence for increased proliferation of CLL cells in lymph nodes compared to bone marrow and peripheral blood. Leukemia. 2017;31(6):1340-1347. 30. ten Hacken E, Burger JA. Microenvironment interactions and B-cell receptor signaling in chronic lymphocytic leukemia: Implications for disease pathogenesis and treatment. Biochim Biophys Acta. 2016;1863 (3):401413. 31. Herreros B, Rodríguez-Pinilla SM, Pajares R, et al. Proliferation centers in chronic lymphocytic leukemia: the niche where NF-κB activation takes place. Leukemia. 2010;24(4): 872-876. 32. Herishanu Y, Pérez-Galán P, Liu D, et al. The lymph node microenvironment promotes Bcell receptor signaling, NF-kappaB activation, and tumor proliferation in chronic lymphocytic leukemia. Blood. 2011;117(2):563574. 33. López-Guerra M, Xargay-Torrent S, Rosich L, et al. The g-secretase inhibitor PF03084014 combined with fludarabine antagonizes migration, invasion and angiogenesis in NOTCH1-mutated CLL cells. Leukemia. 2015;29(1):96-106. 34. López-Guerra M, Xargay-Torrent S, Fuentes P, et al. Specific NOTCH1 antibody targets DLL4-induced proliferation, migration, and angiogenesis in NOTCH1-mutated CLL

haematologica | 2020; 105(9)

CLL pathogenesis and management cells. Oncogene. 2020;39(6):1185-1197. 35. Arruga F, Bracciamà V, Vitale N, et al. Bidirectional linkage between the B-cell receptor and NOTCH1 in chronic lymphocytic leukemia and in Richter’s syndrome: therapeutic implications. Leukemia. 2020;34(2):462-477. 36. Puente XS, Pinyol M, Quesada V, et al. Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature. 2011;475(7354):101-105. 37. Hanna BS, Öztürk S, Seiffert M. Beyond bystanders: myeloid cells in chronic lymphocytic leukemia. Mol Immunol. 2019;110 (1):77-87. 38. Riches JC, Davies JK, McClanahan F, et al. T cells from CLL patients exhibit features of Tcell exhaustion but retain capacity for cytokine production. Blood. 2013;121(9): 1612-1621. 39. Ramsay AG, Evans R, Kiaii S, Svensson L, Hogg N, Gribben JG. Chronic lymphocytic leukemia cells induce defective LFA-1-directed T-cell motility by altering Rho GTPase signaling that is reversible with lenalidomide. Blood. 2013;121(14):2704-2714. 40. Hanna BS, Roessner PM, Yazdanparast H, et al. Control of chronic lymphocytic leukemia development by clonally-expanded CD8+ T-cells that undergo functional exhaustion in secondary lymphoid tissues. Leukemia. 2019;33(3):625-637. 41. Llaó Cid L, Hanna BS, Iskar M, et al. CD8+ T-cells of CLL-bearing mice acquire a transcriptional program of T-cell activation and exhaustion. Leuk Lymphoma. 2020;61(2): 351-356. 42. Kipps TJ, Stevenson FK, Wu CJ, et al. Chronic lymphocytic leukaemia. Nat Rev Dis Prim. 2017;316096. 43. Wiestner A. The role of B-cell receptor inhibitors in the treatment of patients with chronic lymphocytic leukemia. Haematologica. 2015;100(12):1495-1507. 44. Mockridge CI, Potter KN, Wheatley I, Neville LA, Packham G, Stevenson FK. Reversible anergy of sIgM-mediated signaling in the two subsets of CLL defined by VH-gene mutational status. Blood. 2007;109(10):4424-4431. 45. Moore VDG, Brown JR, Certo M, Love TM, Novina CD, Letai A. Chronic lymphocytic leukemia requires BCL2 to sequester prodeath BIM, explaining sensitivity to BCL2 antagonist ABT-737. J Clin Invest. 2007;117(1):112-121. 46. Juliusson G, Oscier DG, Fitchett M, et al. Prognostic subgroups in B-cell chronic lymphocytic leukemia defined by specific chromosomal abnormalities. N Engl J Med. 1990;323(11):720-724. 47. Döhner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343(26):1910-1916. 48. Haferlach C, Dicker F, Schnittger S, Kern W, Haferlach T. Comprehensive genetic characterization of CLL: a study on 506 cases analysed with chromosome banding analysis, interphase FISH, IgV(H) status and immunophenotyping. Leukemia. 2007;21 (12):2442-2451. 49. Baliakas P, Iskas M, Gardiner A, et al. Chromosomal translocations and karyotype complexity in chronic lymphocytic leukemia: a systematic reappraisal of classic cytogenetic data. Am J Hematol. 2014;89(3):249-255. 50. Huh YO, Abruzzo L V., Rassidakis GZ, et al. The t(14;19)(q32;q13)-positive small B-cell leukaemia: a clinicopathologic and cytogenetic study of seven cases. Br J Haematol.

haematologica | 2020; 105(9)

2007;136(2):220-228. 51. Nadeu F, Diaz-Navarro A, Delgado J, Puente XS, Campo E. Genomic and epigenomic alterations in chronic lymphocytic leukemia. Annu Rev Pathol Mech Dis. 2020;15(1):149177. 52. Edelmann J, Holzmann K, Miller F, et al. High-resolution genomic profiling of chronic lymphocytic leukemia reveals new recurrent genomic alterations. Blood. 2012;120 (24):4783-4794. 53. Delgado J, Salaverria I, Baumann T, et al. Genomic complexity and IGHV mutational status are key predictors of outcome of chronic lymphocytic leukemia patients with TP53 disruption. Haematologica. 2014;99 (11):e231-e234. 54. Thompson PA, Stingo F, Keating MJ, et al. Outcomes of patients with chronic lymphocytic leukemia treated with first-line idelalisib plus rituximab after cessation of treatment for toxicity. Cancer. 2016;122(16): 2505-2511. 55. Baliakas P, Jeromin S, Iskas M, et al. Cytogenetic complexity in chronic lymphocytic leukemia: definitions, associations, and clinical impact. Blood. 2019;133(11): 12051216. 56. Baliakas P, Puiggros A, Xochelli A, et al. Additional trisomies amongst patients with chronic lymphocytic leukemia carrying trisomy 12: the accompanying chromosome makes a difference. Haematologica. 2016;101(7):299-302. 57. Martínez-Trillos A, Pinyol M, Navarro A, et al. Mutations in TLR/MYD88 pathway identify a subset of young chronic lymphocytic leukemia patients with favorable outcome. Blood. 2014;123(24):3790-3796. 58. Giménez N, Martínez-Trillos A, Montraveta A, et al. Mutations in the RAS-BRAF-MAPKERK pathway define a specific subgroup of patients with adverse clinical features and provide new therapeutic options in chronic lymphocytic leukemia. Haematologica. 2019;104(3):576-586. 59. Stilgenbauer S, Schnaiter A, Paschka P, et al. Gene mutations and treatment outcome in chronic lymphocytic leukemia: results from the CLL8 trial. Blood. 2014;123(21):32473254. 60. Malcikova J, Tausch E, Rossi D, et al. ERIC recommendations for TP53 mutation analysis in chronic lymphocytic leukemia update on methodological approaches and results interpretation. Leukemia. 2018;32(5): 1070-1080. 61. Sutton L-A, Ljungström V, Enjuanes A, et al. Comparative analysis of targeted next-generation sequencing panels for the detection of gene mutations in chronic lymphocytic leukemia: an ERIC multi-center study. Haematologica. 2020 Apr 9. [Epub ahead of print]. 62. Rossi D, Khiabanian H, Spina V, et al. Clinical impact of small TP53 mutated subclones in chronic lymphocytic leukemia. Blood. 2014;123(14):2139-2147. 63. Ojha J, Ayres J, Secreto C, et al. Deep sequencing identifies genetic heterogeneity and recurrent convergent evolution in chronic lymphocytic leukemia. Blood. 2015;125(3):492-498. 64. Amin NA, Seymour E, Saiya-Cork K, Parkin B, Shedden K, Malek SN. A quantitative analysis of subclonal and clonal gene mutations before and after therapy in chronic lymphocytic leukemia. Clin Cancer Res. 2016;22(17):4525-4535. 65. Landau DA, Sun C, Rosebrock D, et al. The evolutionary landscape of chronic lymphocytic leukemia treated with ibrutinib target-

ed therapy. Nat Commun. 2017;8(1):2185. 66. Blombery P, Anderson MA, Gong JN, et al. Acquisition of the recurrent Gly101Val mutation in BCL2 confers resistance to venetoclax in patients with progressive chronic lymphocytic leukemia. Cancer Discov. 2019;9(3):342-353. 67. Lampson BL, Brown JR. Are BTK and PLCG2 mutations necessary and sufficient for ibrutinib resistance in chronic lymphocytic leukemia? Expert Rev Hematol. 2018;11(3):185-194. 68. Zhao X, Lwin T, Silva A, et al. Unification of de novo and acquired ibrutinib resistance in mantle cell lymphoma. Nat Commun. 2017;8(1):14920. 69. Agarwal R, Chan YC, Tam CS, et al. Dynamic molecular monitoring reveals that SWI–SNF mutations mediate resistance to ibrutinib plus venetoclax in mantle cell lymphoma. Nat Med. 2019;25(1):119-129. 70. Pan R, Ruvolo V, Mu H, et al. Synthetic lethality of combined Bcl-2 inhibition and p53 activation in AML: mechanisms and superior antileukemic efficacy. Cancer Cell. 2017;32(6):748-760.e6. 71. Landau DA, Clement K, Ziller MJ, et al. Locally disordered methylation forms the basis of intratumor methylome variation in chronic lymphocytic leukemia. Cancer Cell. 2014;26(6):813-825. 72. Ott CJ, Federation AJ, Schwartz LS, et al. Enhancer architecture and essential core regulatory circuitry of chronic lymphocytic leukemia. Cancer Cell. 2018;34(6):982995.e7. 73. Mallm J, Iskar M, Ishaque N, et al. Linking aberrant chromatin features in chronic lymphocytic leukemia to transcription factor networks. Mol Syst Biol. 2019;15(5):e8339. 74. Pastore A, Gaiti F, Lu SX, et al. Corrupted coordination of epigenetic modifications leads to diverging chromatin states and transcriptional heterogeneity in CLL. Nat Commun. 2019;10(1):1874. 75. Rodríguez D, Bretones G, Quesada V, et al. Mutations in CHD2 cause defective association with active chromatin in chronic lymphocytic leukemia. Blood. 2015;126(2):195202. 76. Tsagiopoulou M, Chapaprieta V, DuranFerrer M, et al. Chronic lymphocytic leukemias with trisomy 12 show a distinct DNA methylation profile linked to altered chromatin activation. Haematologica. 2020 Feb 27. [Epub ahead of print]. 77. Landgren O, Albitar M, Ma W, et al. B-cell clones as early markers for chronic lymphocytic leukemia. N Engl J Med. 2009;360(7):659-667. 78. Criado I, Rodríguez-Caballero A, Gutiérrez ML, et al. Low-count monoclonal B-cell lymphocytosis persists after seven years of follow up and is associated with a poorer outcome. Haematologica. 2018;103(7):11981208. 79. Rawstron AC, Bennett FL, O’Connor SJM, et al. Monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia. N Engl J Med. 2008;359(6):575-583. 80. Rossi D, Spina V, Gaidano G. Biology and treatment of Richter syndrome. Blood. 2018;131(25):2761-2772. 81. Del Giudice I, Chiaretti S, Tavolaro S, et al. Spontaneous regression of chronic lymphocytic leukemia: clinical and biologic features of 9 cases. Blood. 2009;114(3):638-646. 82. Barrio S, Shanafelt TD, Ojha J, et al. Genomic characterization of high-count MBL cases indicates that early detection of driver mutations and subclonal expansion are predictors of adverse clinical outcome.

2215

J. Delgado et al. Leukemia. 2017;31(1):170-176. 83. Winkelmann N, Rose-Zerilli M, Forster J, et al. Low frequency mutations independently predict poor treatment-free survival in early stage chronic lymphocytic leukemia and monoclonal B-cell lymphocytosis. Haematologica. 2015;100(6):e237-e239. 84. Gruber M, Bozic I, Leshchiner I, et al. Growth dynamics in naturally progressing chronic lymphocytic leukaemia. Nature. 2019;570(7762):474-479. 85. Montserrat E, Sanchez‐Bisono J, Viñolas N, Rozman C. Lymphocyte doubling time in chronic lymphocytic leukaemia: analysis of its prognostic significance. Br J Haematol. 1986;62(3):567-575. 86. Hallek M, Cheson BD, Catovsky D, et al. iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL. Blood. 2018;131(25):2745-2760. 87. Al-Sawaf O, Robrecht S, Bahlo J, et al. Richter transformation in chronic lymphocytic leukemia (CLL)—a pooled analysis of German CLL Study Group (GCLLSG) front line treatment trials. Leukemia. 2020 Mar 17. [Epub ahead of print]. 88. Fabbri G, Khiabanian H, Holmes AB, et al. Genetic lesions associated with chronic lymphocytic leukemia transformation to Richter syndrome. J Exp Med. 2013;210(11):22732288. 89. Kadri S, Lee J, Fitzpatrick C, et al. Clonal evolution underlying leukemia progression and Richter transformation in patients with ibrutinib-relapsed CLL. Blood Adv. 2017;1(12):715-727. 90. Miller CR, Ruppert AS, Heerema NA, et al. Near-tetraploidy is associated with Richter transformation in chronic lymphocytic leukemia patients receiving ibrutinib. Blood Adv. 2017;1(19):1584-1588. 91. Kwok M, Oldreive C, Rawstron AC, et al. Integrative analysis of spontaneous CLL regression highlights genetic and microenvironmental interdependency in CLL. Blood. 2020;135(6):411-428. 92. Rai KR, Sawitsky A, Cronkite EP, Chanana AD, Levy RN, Pasternack BS. Clinical staging of chronic lymphocytic leukemia. Blood. 1975;46(2):219-234. 93. Binet JL, Auquier A, Dighiero G, et al. A new prognostic classification of chronic lymphocytic leukemia derived from a multivariate survival analysis. Cancer. 1981;48(1):198206. 94. Baliakas P, Hadzidimitriou A, Sutton L-A, et al. Recurrent mutations refine prognosis in chronic lymphocytic leukemia. Leukemia. 2015;29(2):329-336. 95. Rossi D, Rasi S, Spina V, et al. Integrated mutational and cytogenetic analysis identifies new prognostic subgroups in chronic lymphocytic leukemia. Blood. 2013;121(8): 1403-1412. 96. Baliakas P, Moysiadis T, Hadzidimitriou A, et al. Tailored approaches grounded on immunogenetic features for refined prognostication in chronic lymphocytic leukemia. Haematologica. 2019;104(2):360369. 97. International CLL-IPI working groups. An international prognostic index for patients with chronic lymphocytic leukaemia (CLLIPI): a meta-analysis of individual patient data. Lancet Oncol. 2016;17(6):779-790. 98. da Cunha-Bang C, Christiansen I, Niemann CU. The CLL-IPI applied in a populationbased cohort. Blood. 2016;128(17):21812183. 99. Molica S, Shanafelt TD, Giannarelli D, et al.

2216

The chronic lymphocytic leukemia international prognostic index predicts time to first treatment in early CLL: Independent validation in a prospective cohort of early stage patients. Am J Hematol. 2016;91(11):10901095. 100. Gentile M, Shanafelt TD, Rossi D, et al. Validation of the CLL-IPI and comparison with the MDACC prognostic index in newly diagnosed patients. Blood. 2016;128(16):2093-2095. 101. Delgado J, Doubek M, Baumann T, et al. Chronic lymphocytic leukemia: a prognostic model comprising only two biomarkers (IGHV mutational status and FISH cytogenetics) separates patients with different outcome and simplifies the CLL-IPI. Am J Hematol. 2017;92(4):375-380. 102. Condoluci A, Terzi di Bergamo L, Langerbeins P, et al. International prognostic score for asymptomatic early-stage chronic lymphocytic leukemia. Blood. 2020;135(21): 1859-1869. 103. Dighiero G, Maloum K, Desablens B, et al. Chlorambucil in indolent chronic lymphocytic leukemia. French Cooperative Group on Chronic Lymphocytic Leukemia. N Engl J Med. 1998;338(21):1506-1514. 104. Herling CD, Cymbalista F, Groß-OphoffMüller C, et al. Early treatment with FCR versus watch and wait in patients with stage Binet A high-risk chronic lymphocytic leukemia (CLL): a randomized phase 3 trial. Leukemia. 2020 Feb 18. [Epub ahead of print]. 105. Langerbeins P, Bahlo J, Rhein C, et al. Ibrutinib versus placebo in patients with asymptomatic treatment-naïve early stage CLL: primary endpoint results of the phase 3 double-blind randomized CLL12 trial. Hematol Oncol. 2019;37(Suppl 2):38-40. 106. 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. 107. Shanafelt TD, Wang X V., Kay NE, et al. Ibrutinib–rituximab or chemoimmunotherapy for chronic lymphocytic leukemia. N Engl J Med. 2019;381(5):432-443. 108. Woyach JA, Ruppert AS, Heerema NA, et al. Ibrutinib regimens versus chemoimmunotherapy in older patients with untreated CLL. N Engl J Med. 2018;379(26):25172528. 109. Moreno C, Greil R, Demirkan F, et al. Ibrutinib plus obinutuzumab versus chlorambucil plus obinutuzumab in first-line treatment of chronic lymphocytic leukaemia (iLLUMINATE): a multicentre, randomised, open-label, phase 3 trial. Lancet Oncol. 2019;20(1):43-56. 110. Burger JA, Sivina M, Jain N, et al. Randomized trial of ibrutinib vs ibrutinib plus rituximab in patients with chronic lymphocytic leukemia. Blood. 2019;133(10): 1011-1019. 111. Diop F, Moia R, Favini C, et al. Biological and clinical implications of BIRC3 mutations in chronic lymphocytic leukemia. Haematologica. 2020;105(2):448-456. 112. Tausch E, Schneider C, Robrecht S, et al. Prognostic and predictive impact of genetic markers in patients with CLL treated with obinutuzumab and venetoclax. Blood. 2020 Mar 23. [Epub ahead of print]. 113. Fischer K, Al-Sawaf O, Bahlo J, et al. Venetoclax and obinutuzumab in patients with CLL and coexisting conditions. N Engl J Med. 2019;380(23):2225-2236. 114. Sharman JP, Egyed M, Jurczak W, et al. Acalabrutinib with or without obinutuzum-

ab versus chlorambucil and obinutuzmab for treatment-naive chronic lymphocytic leukaemia (ELEVATE TN): a randomised, controlled, phase 3 trial. Lancet. 2020;395(10232):1278-1291. 115. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-42. 116. Farooqui MZH, Valdez J, Martyr S, et al. Ibrutinib for previously untreated and relapsed or refractory chronic lymphocytic leukaemia with TP53 aberrations: a phase 2, single-arm trial. Lancet Oncol. 2015;16(2): 169-176. 117. O’Brien S, Jones JA, Coutre SE, et al. Ibrutinib for patients with relapsed or refractory chronic lymphocytic leukaemia with 17p deletion (RESONATE-17): a phase 2, open-label, multicentre study. Lancet Oncol. 2016;17(10):1409-1418. 118. 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. 119. Chanan-Khan A, Cramer P, Demirkan F, et al. Ibrutinib combined with bendamustine and rituximab compared with placebo, bendamustine, and rituximab for previously treated chronic lymphocytic leukaemia or small lymphocytic lymphoma (HELIOS): a randomised, double-blind, phase 3 study. Lancet Oncol. 2016;17(2):200-211. 120. Roberts AW, Davids MS, Pagel JM, et al. Targeting BCL2 with venetoclax in relapsed chronic lymphocytic leukemia. N Engl J Med. 2016;374(4):311-322. 121. Stilgenbauer S, Eichhorst B, Schetelig J, 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):768-778. 122. Jones JA, Mato AR, Wierda WG, 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. 123. Coutre S, Choi M, Furman RR, et al. Venetoclax for patients with chronic lymphocytic leukemia who progressed during or after idelalisib therapy. Blood. 2018;131(15):1704-1711. 124. Seymour JF, Kipps TJ, Eichhorst B, et al. Venetoclax-rituximab in relapsed or refractory chronic lymphocytic leukemia. N Engl J Med. 2018;378(12):1107-1120. 125. Furman RR, Sharman JP, Coutre SE, et al. Idelalisib and rituximab in relapsed chronic lymphocytic leukemia. N Engl J Med. 2014;370(11):997-1007. 126. Sharman JP, Coutre SE, Furman RR, et al. Final results of a randomized, phase III study of rituximab with or without idelalisib followed by open-label idelalisib in patients with relapsed chronic lymphocytic leukemia. J Clin Oncol. 2019;37(16):13911402. 127. Kreuzer KA, Furman RR, Stilgenbauer S, et al. The impact of complex karyotype on the overall survival of patients with relapsed chronic lymphocytic leukemia treated with idelalisib plus rituximab. Leukemia. 2020;34(1):296-300. 128. Anderson MA, Tam C, Lew TE, et al. Clinicopathological features and outcomes of progression of CLL on the BCL2 inhibitor venetoclax. Blood. 2017;129(25):3362-3370. 129. Maddocks KJ, Ruppert AS, Lozanski G, et al. Etiology of ibrutinib therapy discontinuation and outcomes in patients with chronic lymphocytic leukemia. JAMA. Oncol

haematologica | 2020; 105(9)

CLL pathogenesis and management 2015;1(1):80-87. 130. Flinn IW, Hillmen P, Montillo M, et al. The phase 3 DUO trial: duvelisib vs ofatumumab in relapsed and refractory CLL/SLL. Blood. 2018;132(23):2446-2455. 131. Ghia P, Pluta A, Wach M, et al. ASCEND: phase III, randomized trial of acalabrutinib versus idelalisib plus rituximab or bendamustine plus rituximab in relapsed or refractory chronic lymphocytic leukemia. J Clin Oncol. 2020 May 27. [Epub ahead of print] 132. Awan FT, Schuh A, Brown JR, et al. Acalabrutinib monotherapy in patients with chronic lymphocytic leukemia who are intolerant to ibrutinib. Blood Adv 2019; 3(9):1553-1562. 133. Gribben JG. How and when I do allogeneic transplant in CLL. Blood. 2018;132(1):31-39. 134. Kharfan-Dabaja MA, Kumar A, Hamadani M, et al. Clinical practice recommendations for use of allogeneic hematopoietic cell transplantation in chronic lymphocytic leukemia on behalf of the Guidelines Committee of the American Society for

haematologica | 2020; 105(9)

Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2016;22(12): 2117-2125. 135. Dreger P, Ghia P, Schetelig J, et al. High-risk chronic lymphocytic leukemia in the era of pathway inhibitors: integrating molecular and cellular therapies. Blood. 2018;132(9): 892-902. 136. Bair SM, Porter DL. Accelerating chimeric antigen receptor therapy in chronic lymphocytic leukemia: the development and challenges of chimeric antigen receptor T-cell therapy for chronic lymphocytic leukemia. Am J Hematol. 2019;94(S1):S10-S17. 137. Sutton LA, Young E, Baliakas P, et al. Different spectra of recurrent gene mutations in subsets of chronic lymphocytic leukemia harboring stereotyped B-cell receptors. Haematologica. 2016;101(8):959967. 138. Burger JA, Chiorazzi N. B cell receptor signaling in chronic lymphocytic leukemia. Trends Immunol. 2013;34(12):592-601. 139. Burger JA, Quiroga MP, Hartmann E, et al.

High-level expression of the T-cell chemokines CCL3 and CCL4 by chronic lymphocytic leukemia B cells in nurselike cell cocultures and after BCR stimulation. Blood. 2009;113(13):3050-3058. 140. Dilillo DJ, Weinberg JB, Yoshizaki A, et al. Chronic lymphocytic leukemia and regulatory B cells share IL-10 competence and immunosuppressive function. Leukemia. 2013;27(1):170-182. 141. Lewinsky H, Barak AF, Huber V, et al. CD84 regulates PD-1/PD-L1 expression and function in chronic lymphocytic leukemia. J Clin Invest. 2018;128(12):5479-5488. 142. De Matteis S, Molinari C, Abbati G, et al. Immunosuppressive Treg cells acquire the phenotype of effector-T cells in chronic lymphocytic leukemia patients. J Transl Med. 2018;16(1):172. 143. Paggetti J, Haderk F, Seiffert M, et al. Exosomes released by chronic lymphocytic leukemia cells induce the transition of stromal cells into cancer-associated fibroblasts. Blood. 2015;126(9):1106-1117.

2217

REVIEW ARTICLE Ferrata Storti Foundation

Molecular heterogeneity of pyruvate kinase deficiency Paola Bianchi and Elisa Fermo

Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico Milano, UOC Ematologia, UOS Fisiopatologia delle Anemie, Milan, Italy

ABSTRACT

Haematologica 2020 Volume 105(9):2218-2228

ed cell pyruvate kinase (PK) deficiency is the most common glycolytic defect associated with congenital non-spherocytic hemolytic anemia. The disease, transmitted as an autosomal recessive trait, is caused by mutations in the PKLR gene and is characterized by molecular and clinical heterogeneity; anemia ranges from mild or fully compensated hemolysis to life-threatening forms necessitating neonatal exchange transfusions and/or subsequent regular transfusion support; complications include gallstones, pulmonary hypertension, extramedullary hematopoiesis and iron overload. Since identification of the first pathogenic variants responsible for PK deficiency in 1991, more than 300 different variants have been reported, and the study of molecular mechanisms and the existence of genotype-phenotype correlations have been investigated in-depth. In recent years, during which progress in genetic analysis, next-generation sequencing technologies and personalized medicine have opened up important landscapes for diagnosis and study of molecular mechanisms of congenital hemolytic anemias, genotyping has become a prerequisite for accessing new treatments and for evaluating disease state and progression. This review examines the extensive molecular heterogeneity of PK deficiency, focusing on the diagnostic impact of genotypes and new acquisitions on pathogenic non-canonical variants. The recent progress and the weakness in understanding the genotype-phenotype correlation, and its practical usefulness in light of new therapeutic opportunities for PK deficiency are also discussed.

Correspondence:

Pyruvate kinase enzyme

PAOLA BIANCHI, paola.bianchi@policlinico.mi.it Received: May 7, 2020. Accepted: July 3, 2020. Pre-published: July 23, 2020. doi:10.3324/haematol.2019.241141 ©2020 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.

2218

Pyruvate kinase (PK) is an allosterically regulated glycolytic enzyme that catalyzes the irreversible conversion of phosphoenolpyruvate to pyruvate, with the synthesis of one molecule of ATP. Since mature red blood cells totally depend on the ATP generated by glycolysis for maintaining cell integrity and function, PK plays a crucial role in erythrocyte metabolism; insufficient energy production may impair red blood cell homeostasis, leading to premature removal of PK-deficient erythrocytes from the circulation by the spleen.1,2 A secondary decrease in PK activity has been observed in the presence of reduced red cell membrane surface (as in hereditary spherocytosis3) or in acquired hematologic conditions (e.g., acute myeloid leukemias, or myelodysplastic syndromes),4,5 suggesting a functional relationship between structural membrane integrity and PK activity, and a wider involvement of glycolytic enzymes in cell control.4,5 The three-dimensional structures of a number of prokaryotic and eukaryotic PK have been solved to a high resolution, showing that in almost all organisms, functional PK is a homotetramer of approximately 200-240 kDa.6-9 Each subunit contains four domains, namely a small N-terminal helical domain (residues 1-84); an A domain with (β/a)8 barrel topology (residues 85-159 and 263-431); a β-stranded B domain (residues 160-262), inserted between helix a3 and strand β of the A domain, and a C domain with a+β topology (residues 432-574)10,11 (Figure 1A, B). The active site is located between the A and B domains, whereas the C domain contains the binding site for fructose 1,6 bisphosphate.12 Subunit interactions at the interfaces between the A domains and the C domains, as well as A/B and A/C interdomain interactions within one subunit are considered to be key determinants of the allosteric response of the enzyme. PK is quite a stable protein, and can last haematologica | 2020; 105(9)

Molecular heterogeneity of PK deficiency

the entire lifespan of erythrocytes.13 Pathological mutations causing PK deficiency can be localized in any of the protein domains, with major clusters in specific regions, such as the interface between the A and C domains, the A/A′ intersubunit interface, the hydrophobic core of the A domain, and the fructose 1,6 bisphosphate-binding site6,10,14,15 (Figure 1C). Several human PK mutants have been produced as recombinant forms and biochemically characterized10,11,13,16,17 showing that amino acid substitutions can affect thermostability, catalytic efficiency, and response to the allosteric effector.

Diagnosis of pyruvate kinase deficiency The diagnostic workup for PK deficiency is based on the patient’s personal and family medical history and clinical examination, and on several laboratory investigations, including the spectrophotometric assay of red blood cell PK activity.18 Molecular analysis of the PKLR gene is necessary to confirm the diagnosis, and overcomes the limitations of the enzymatic test, which may give false positive results in the case of heterozygous carriers, or false negative results in the case of recent transfusion, or an increased number of reticulocytes. Therefore, recently

published guidelines and recommendations conclude that enzyme analyses and DNA studies are complementary techniques for diagnosing PK deficiency.19 With the advent of next-generation sequencing (NGS) techniques, the PKLR gene is usually included in panels designed for diagnosing hereditary hemolytic anemias,20-24 allowing detection of an increasing number of cases, thus reducing misdiagnosis, and highlighting the extreme phenotypic variability of PK deficiency25-27 (Table 1).

Gene and variants The PKLR gene, located on chromosome 1q21, consists of 12 exons and is approximately 9.5 kb in size.31 The gene encodes for the liver (L) and erythrocyte (R) isoforms of the enzyme according to tissue-specific promoters;31,32 ten exons are shared by the two isoforms, while exons 1 and 2 are specifically transcribed to the PK-R and PK-L mRNA, respectively. The cDNA encoding PK-R is 2060 bp long and codes for 574 amino acids (Figure 1A). In the R-type promoter region, two CAC boxes and four GATA motifs are located within 270 bp from the translational initiation codon; the proximal 120 bp region has basal promoter activity and the region from -120 to -270 works as a powerful enhancer in erythroid cells.31

Figure 1. PKLR gene and red cell pyruvate kinase structure. (A) The PKLR gene, its chromosomal localization, extension and intron/exon organization. Numbering and mutations are usually reported in the literature using the RPK cDNA sequence of the PKLR gene, with the A of the initiation ATG being assigned number +1 (Transcript refseq ID NM_000298.5). (B) Structural domains of the human PK-R monomer, the N-terminal domain is reported in yellow, A-domain in red, B-domain in light blue and C-domain in green. The corresponding amino acids are reported below. *Represents the localization of residues directly involved in the allosteric site and catalytic center (yellow) and in the fructose 1,6 bisphosphate (FBP) activator (red). (C) Ribbon representation of the human erythrocyte pyruvate kinase monomer (left) in complex with the substrate and the allosteric activator fructose-1,6-diphosphate (red and purple) and tetramer based on the crystal structure described by Valentini et al.10 Circles indicate the A’A’ and the A/C subunit interfaces.

haematologica | 2020; 105(9)

2219

P. Bianchi and E. Fermo

The number of known pathogenic variants is continuously increasing. In a recent inventory, Canu et al.33 reported 260 mutations in the PKLR gene; the Human Genome Mutation Database (HGMD) reports 290 pathogenic variants (update March 2020); a detailed inventory of PKLR variants is also available in the PKLR Leiden Open Variation Database (https://databases.lovd.nl/shared/genes/ PKLR), including a more specific data collection (e.g., congress presentations and unpublished results). The HGMD does not yet include the results obtained in a single analysis of 257 patients with PK deficiency enrolled in the Pyruvate Kinase Deficiency Natural History Study (PKD NHS), a multicenter, international study; 127 different pathogenic variants were detected, comprising 84 missense and 43 non-missense variants (including 20 stop-gain variants, 11 affecting splicing, 5 large deletions, 4 in-frame indels, and 3 promoter variants).34,35 A similar distribution is observed by stratifying variants reported by the HGMD according to the type of mutations (Figure 2). A list of the more commonly detected mutations and variants with geographical distribution and ethnic background is reported in Table 2; similar information for the rarer/unique variants is available in mutation databases. Molecular analysis of the PKLR gene by Sanger sequencing usually covers the entire coding region, flanking intronic sequences and the erythroid-specific promoter. NGS analysis allows more extensive sequencing than the Sanger method (generally including entire coding and intronic flanking regions, promoter, 3’ upstream, and 5’ downstream regions) and can give information on the presence of large indels. Other techniques (e.g., multiplex ligation-dependent probe amplification or assays of copy number variations, comparative genomic hybridization arrays or digital polymerase chain reaction) can also be used to this latter purpose. Variants are usually reported in the literature using the RPK cDNA sequence of the PKLR gene, with the A of the initiation ATG being assigned number +1 (Transcript refseq ID NM_000298.5). American College of Medical Genetics and Genomics (ACMG) guidelines should be followed to interpret and assess sequence variants.42

Promoter and enhancer variants Only a few pathogenic variants have been identified in the promoter region, mostly clustering at two functionally important sequences, such as the consensus binding motif for GATA-1 at nucleotides c.-69 to -74,43-45 and a regulatory element (PKR-RE1) whose core CTCTG extends from nucleotides c.-87 to -83.46,47 The variant c.-72A>G, located in the GATA-1 motif, was found to be associated with low mRNA expression, and to be responsible for severe anemia when present in the homozygous state.15,45 Other variants have been reported with uncertain pathogenic significance: the variant c.-109C>T described by Pissard et al.,48 while not directly modifying any known binding site for a transacting factor, was found to be located within a region displaying basic promoter activity, very close to the region described as an erythroid enhancer49 (Figure 3). At the moment a clear disease-causing association of variants located in the enhancer region is not well established. Some reported variants in this regions, such as c.148C>T,49 or deletion reported at nucleotides c.-249delA or c.-248delT, do not seem to affect the expression of the gene, thus are considered non-pathogenic.46

Coding region variants The large majority of pathogenic variants are located in the coding region. Mutations are distributed throughout the PKLR gene and affect all exons (Figure 4). Most of them (about 66%) are missense mutations (Figure 2). Not every mutation detected by DNA sequencing can be immediately classified as a disease-causing variant, and should be considered ‘variants of unknown clinical significance’ until their pathogenic nature is confirmed by functional analysis such as PK enzymatic assays, western blotting, reverse transcriptase polymerase chain reaction analysis, or gene reporter assays.50 This is especially important when patients’ samples are not accompanied by complete clinical and laboratory information. Most variants in PK deficiency affect residues critical to the structure and/or function of the enzyme. However, it is not possible to establish a direct relationship between the severity of a pathogenic variant and its position; most patients in fact are compound heterozygous for two

Table 1. Recent studies performed by next-generation sequencing technologies in patients with hemolytic anemias.

Reference

Method

N. of genes analyzed

N. of cases studied with CHA

PKD diagnosis

New diagnosis and number and type of mismatched diagnoses

15 28 29 30

t-NGS t-NGS WES t-NGS

35 55 n.a. 76

36 43 4 21a

2 8 4 6

t-NGS

21b

25 23

t-NGS t-NGS

34 and 71 33

74c 57

7 3

2 new PKD 8 new PKD 4 new PKD 3 new PKD 2 CDA→ PKD 1 DBA→PKD 4 new PKD 2 CDA→PKD 7 CDA→PKD 2 new PKD 1 CDA→PKD

Number of genes included in the panel, number of cases analyzed in each study and cases diagnosed with pyruvate kinase deficiency are shown. Next-generation sequencing analysis allowed modification of a previous diagnosis; the number and the type of mismatched diagnosis is reported in the last column. aAll transfusion-dependent patients. bNo diagnosis despite extensive laboratory investigations. cSuspected diagnosis of congenital dyserythropoietic anemia. CHA: chronic hemolytic anemias; PKD: pyruvate kinase deficiency; t-NGS: targeted next-generation sequencing; WES: whole-exome sequencing; n.a.: not available; CDA: congenital dyserythropoietic anemia; DBA: Diamond-Blackfan anemia.

2220

haematologica | 2020; 105(9)

Molecular heterogeneity of PK deficiency

mutations and it is therefore difficult to determine the severity of any one individual variant in critical regions of the gene.

Splice site variants Splice site variants have been reported in all exon/intron boundary sequences. Most of these variants affect the Âą1 or Âą2 nucleotides of the donor/acceptor sequences, and are consequently considered to have a drastic effect on splicing, causing unstable and degraded mRNA. Care should be taken in the interpretation of more internal variants that require functional analysis before

defining their pathogenicity; this is the case of rare PKLR variants such as c.1269+5G>A,48 c.507-20C>A14, c.100+10G>A and c.375+10G>T,48 considered by authors to affect the splicing only basing on in silico analysis. It is worth noting that some missense mutations in the coding region may also affect splicing, in particular when located in the first/last nucleotides of the exons, e.g., c.507G>A,17 c.694G>T,48,51 c.1269G>C,52 or the c.1436G>A variant (p.R479H),38 located in the last nucleotide of exon 10, typically but not exclusively found in the Amish community. The deleterious effect of these variants should always be considered in genetic counseling, or in evaluat-

Figure 2. Type of PKLR pathogenic variants. (A) The type of PKLR pathogenic variants (n=290) reported in the Human Genome Mutation Database (March 2020). (B) The type of PKLR mutations (n=127) reported in a series of 257 patients with pyruvate kinase deficiency.35 (C) Genotypes in a series of 177 unrelated patients with pyruvate kinase deficiency.35

haematologica | 2020; 105(9)

2221

P. Bianchi and E. Fermo

ing the impact of genotype in clinical trials (see the section “From genotype to new therapies”).

Deep intronic variants An increasing number of pathogenic deep intronic mutations has been described across different disease conditions,53 and these mutations have been considered to justify the number of PK-deficient patients in whom it is not possible to find molecular defects. Among the 278 participants initially enrolled in the PKD NHS, 21 were considered ineligible for the study because of the inability to demonstrate two pathogenic variants even after excluding the large deletion analyzed by long-range polymerase chain reaction.34,35 This aspect has important implications, obviously from the diagnostic point of view, but also with regards to the possibility to access new specific therapies, as further discussed in the next paragraphs.26,54 Laboratory

confirmation of deep intronic variants is often difficult and may require specific testing such as loss of heterozygosity by analyzing an allele-specific cDNA, or the more complex minigene construct approach.55 In a recent study a deep intronic mutation (c.283+109C>T in intron 2) was detected by whole exome sequencing in compound heterozygosity with the missense mutation p.G332S, and was considered responsible for creating an alternative splicesome by in silico analysis; rapid mRNA degradation was confirmed by the observation of loss of heterozygosity of the p.G332S variant at the cDNA level.56 In another study, 13 PK-deficient related individuals with one or no pathogenic variants identified in the PKLR gene were analyzed by whole exome sequencing or whole genome sequencing. Five patients had an alternative diagnosis with mutations in GATA1, KIF23, and

Table 2. Most common mutations of the PKLR gene, ethnic distribution and allelic frequency.

Mutation

Effect

Exon

Mutation type

Geographical distribution/ethnicity

Allelic frequency (gnomAD Exome)

Ref.

c.1529G>A c.1456C>T c.1468C>T c.721G>T c.994G>A c.992 A>G

p.R510Q p.R486W p.R490W p.E241* p.G332S p.D331G

11 11 11 7 8 8

Missense Missense Missense Nonsense Missense Missense

Nothern EU, USA Southern EU, India Japan Caucasian Caucasian India

rs113403872 rs116100695 rs200133000 rs201953584 rs773626254 rs1443439423

0.000358 0.00305 0.000127 0.0000485 0.0000557 0.00000398

34 34,36 37 16 16 36

Missense/splicing

rs118204085

0.000127

36,38

7 i10-i11 i3-10

Missense Large deletion Large deletion

Pennsylvania Amish, Indian African Roma community Vietnamese

rs147689373 na na

0.000694 na na

39 40 17,41

Pathogenic variants with strong ethnic association c.1436G>A c.829G>A c.1437-518_1618+440del c.283+1914_1434del

p.R479H/abnormal splicing p.E277K 1149 bp deletion 5006 bp deletion

Figure 3. DNA sequence of the erythroid-specific PKLR promoter region. Conserved sequences between human and rat PKR promoter are underlined, the black arrow indicates the PK-R transcriptional start site. Yellow boxes indicate the GATA-1 motif, the green boxes identify the CAC/Sp1 motif and the blue box identifies the PKR-RE regulatory element. Colored arrows indicate motif direction. Mutated nucleotides reported in the literature associated with pyruvate kinase deficiency are indicated in red and reported in more detail in the box.

2222

haematologica | 2020; 105(9)

Molecular heterogeneity of PK deficiency

PIEZO1 genes, whereas in five other cases, whole genome sequencing identified different intronic variants, all predicted to perturb normal mRNA processing and confirmed by minigene assays.55

in this case, the authors postulated that the causative mutation resides in a novel, unidentified locus, responsible for upregulating PKLR gene expression.60

Other genes associated with pyruvate kinase deficiency Large insertions/deletions Large indels are rare, possibly due to the technical difficulties in identifying them. The most frequent is the deletion of the 1149 bp characteristic of the Roma community, which leads to skipping of exon 11.40 A very large deletion of 5006 bp that results (at the cDNA level) in the loss of exons 4 to 11 (c.283+1914_c.1434del5006) has been described in patients of Vietnamese origin.17,41 A large homozygous insertion of 367 bp (c.939_940ins367) containing an Alu element (AluYb9)57 was identified in two unrelated children with severe transfusion-dependent hemolytic anemia, from the Middle East. Other variants have been reported although the exact cut-off point has not been identified (3 of the 5 different large deletions identified within the PKD NHS).35 As for deep intronic variants, the search for large indels may add to the costs of analysis and require techniques not always available; however, it should always be considered when one mutation or no mutations at all are detected in a patient with clinical and biochemical diagnosis of PK deficiency. The eventuality of a large deletion at the heterozygous level should always be taken into account in patients carrying homozygous pathogenic variants, in whom the allelic transmission has not been confirmed through analysis of the parents.

Inherited pyruvate kinase hyperactivity Inherited hyperactivity of red blood cell PK (OMIM 102900) has been reported in only three families with apparent asymptomatic conditions with different etiologies.58-60 This rare condition was attributed in the past to a heterozygous mutation in the PKLR gene61 or to the persistent expression of the fetal isozyme PK-M2.59 More recently, there was a report of a family characterized solely by the increased expression of a kinetically normal PKR, in the absence of mutations in PKLR codifying and regulatory regions as well as variations in PKLR copy number, and exclusion of co-segregation with the PKLR locus;

KLF1 is a transcription factor involved in terminal erythroid differentiation, and regulates many of the genes implicated in red cell enzyme deficiencies, including PKLR. Decreased PK activity in the absence of PKLR mutations has been reported in patients who were compound heterozygotes for KLF1 variants, possibly leading to misdiagnosis of these cases with PK deficiency;62 these patients displayed severe, transfusion-dependent neonatal anemia with a broad spectrum of red cell morphological abnormalities and a remarkable persistence of fetal and embryonic globin synthesis.

Geographical distribution of PKLR variants PK deficiency has a worldwide geographical distribution. A careful literature review established that the prevalence of clinically-diagnosed PK deficiency was likely between 3.2 and 8.5 cases per million in Western populations, while the prevalence of diagnosed and undiagnosed PK deficiency could possibly be as high as 51 cases per million.63-65 The causes of this variability can be explained by the rarity of the disorder, the high percentage of undiagnosed cases and the absence of disease registries and specific population studies.65 There is a high frequency of PK deficiency in Middle East and sub-Saharan Africa populations, possibly due to selective pressure from malaria;39,66 some reports have, in fact, suggested that PK deficiency provides protection against infection and replication of P. falciparum in murine models and in ex vivo experiments with red cells from PKdeficient patients.67-69 Most of the molecular variants reported are private, and it is thus difficult to define a geographical distribution. Despite this, the most frequent mutations in PK deficiency are distributed with a strong ethnic and regional background; in addition, some mutations have a high frequency in specific populations as a result of a founder effect (Table 2).

Figure 4. Distribution of variants along the PKLR gene and pyruvate kinase structural domains. Distribution of unique pathogenic variants reported in the Human Genome Mutation Database along exons (left side), and distribution of affected residues in the four different structural domains (right side). aa: amino acid; N: Nterminal domain.

haematologica | 2020; 105(9)

2223

P. Bianchi and E.â&#x20AC;&#x2C6;Fermo

Mutations and clinical phenotype: the genotype-phenotype correlation The broad spectrum of clinical presentations reported in PK deficiency reflects the extensive molecular heterogeneity,16,26,34 and the search for a correlation between the genotype and the phenotype has been the matter of study for many years. The genotype-phenotype correlation has been investigated in clinical studies and by in vitro production and characterization of recombinant mutant proteins of the human enzyme,10,13,16,70 showing that patients with severe phenotypes more commonly carry nonsense mutations or missense pathogenic variants affecting the active site or stability of the PK protein.15,16,36,71 A recent analysis evaluated the genotype-phenotype correlation in the PKD NHS.35 In addition to the volume of patients/data collected (257 patients, 177 of them unrelated), analysis of this cohort had the great advantage of homogeneous data collection. Mutation types were classified according to previous approaches as missense (M) or non-missense (NM) (including nonsense, frameshift, splicing mutations, large deletions, in-frame indels, and promoter variants); patients with NM/NM mutations were found to have a more severe phenotype, with lower hemoglobin levels after splenectomy, a higher number of transfusions throughout their lifetime, a higher rate of iron overload, and a higher rate of splenectomy, when compared with patients with M/M or M/NM PKLR mutations. This categorization has some obvious limitations; in fact, although it is easy to predict the effects of a nonsense variant, because, independently of its nature, it results in protein degradation, predicting the effect of missense variants is more complex, and must take into account the effects on functional properties and the stability of the mutated protein. Studies on the biochemical characterization of recombinant mutant PK enzymes have actually warned against predictions of the effects of missense mutations simply based on the location and the nature of the replaced residues; as an example, the two most frequently reported mutations p.R486W and p.R510Q both affect arginine residues located at the A/C interface, but result in substantially different effects. The p.R486W substitution leads to an enzyme with moderately altered kinetic parameters, but does not affect protein stability, whereas the p.R510Q replacement is likely to disrupt a local network of hydrogen bonds and ultimately results in protein instability and altered allosteric responsiveness to ATP inhibition.10,13 The structural architecture of the PK molecule contributes greatly to the heterogeneity of biochemical properties of the abnormal variants; in fact, the majority of patients with PK deficiency are compound heterozygous for two missense mutations, and may therefore have several different combinations of tetramers, each with distinct kinetic, allosteric and structural properties. In addition, it is known that in patients with identical genotypes other genetic or environmental factors may affect the phenotype. This has been observed in a large number of patients homozygous for the p.R510Q mutant reported in three studies.16,34,72 In all series, variability in the severity of the disease and the well-being of the patients was observed, even within the same family. Patients displayed a wide range of hemoglobin levels 2224

(4.9-12.2 g/dL16 and 6.7-11.5 g/dL34) with a broad spedium of ages at diagnosis (0-56 and 0-47 years), but similar rates of splenectomy (44% and 37%). Phenotypic variability within the same family has been confirmed by the analysis of 88 siblings from 38 families:35 with intraclass correlations ranging from 0.4-0.61; about the same degree of similarity has been found either within or between sibling clusters for hemoglobin, total bilirubin, splenectomy, and cholecystectomy.35 Finally, PK-deficient patients usually tolerate anemia well, so the decision to transfuse or treat a patient is based on how the patient feels rather than on an arbitrary hemoglobin threshold.16 This is in part justified by the increased 2,3-diphosphogylcerate level typically found in these patients; as an important regulator of the oxygen affinity of hemoglobin, 2,3-diphosphogylcerate may enhance oxygen delivery.74 Al-Samkari et al.26 reported an illustrative case: despite continued severity of anemia after splenectomy, a PK-deficient patient did not require blood transfusion, maintaining a normal social life into adulthood. Quality of life assessments, including the Functional Assessment of Chronic Illness Therapy Fatigue subscale [FACIT-F, final score of 48 (score range 052)] and the Functional Assessment of Cancer Therapy [FACT-G, score of 96 (score range 0-104)], confirmed the patientâ&#x20AC;&#x2122;s good quality of life.

Epigenetic factors and co-inheritance Other causes of variability of clinical expression in PK deficiency could depend on possible individual differences in metabolic or proteolytic activity, which may modulate the basic effect of the mutations on ineffective erythropoiesis74,75 or differences in splenic function, and on the ability to compensate for the enzyme deficiency by overexpressing isozymes or using alternative pathways.16 In addition, other factors, such as genetic background, concomitant functional polymorphisms of other enzymes, post-translational or epigenetic modifications, and co-inheritance of other diseases may greatly contribute to the phenotypic heterogeneity and complications. Patients with PK deficiency usually develop secondary iron overload with a multifactorial pathogenesis, involving chronic hemolysis, ineffective erythropoiesis, and transfusion therapy;76-78 HFE mutations p.C282Y and p.H63D have been proposed as additional risk factors.79,80 Similarly, the co-inheritance of the UGT1A1 TA promoter polymorphism may contribute to the occurrence of gallstones, which are detected with increased frequency after the first decade of life in PK-deficient patients.81,82 The concomitance of PK deficiency and other hereditary anemias, such as glucose-6-phosphate dehydrogenase deficiency, hemoglobinopathies, and red blood cell membrane defects, has been reported on rare occasions, with variable contributions of the different diseases to the severity of hemolysis, and should always be considered when interpreting clinical severity.82,83 The number of these reports has grown in recent years due to the increased use of NGS technologies, allowing identification of multiple disease-associated variants in patients affected by congenital hemolytic anemias and complex patterns of inheritance.25,84,30 Heterozygosity for a mutation in the PKLR gene may haematologica | 2020; 105(9)

Molecular heterogeneity of PK deficiency

accompany other red cell diseases, confounding the hematologic pattern and sometimes making the diagnosis challenging; likewise, concomitant causes of anemia may explain some patients with decreased PK activity and only one mutation detected upon molecular analysis. In a series of 56 French patients diagnosed with PK deficiency by enzymatic assay and submitted to a molecular diagnosis, 17 cases were reported to carry a heterozygote PK mutation; in three of them, an association with other defects was found, namely a membrane defect, a hexokinase deficiency and a glucose-6-phosphate dehydrogenase deficiency.48 In these cases, complete hematologic investigation and molecular characterization of the involved genes are needed to clarify the correct diagnosis, also in view of therapeutic approaches and genetic counseling.85,86 Co-inheritance of heterozygous HbS and PK deficiency (either in the homozygous or heterozygous state)87-89 may induce sickling and worsening of phenotype. In the reported cases, the increase of intraerythrocytic 2,3diphosphogylcerate concentration induced by the PK deficiency resulted in a decreased oxygen affinity which favored sickling. The possible contribution of a heterozygous PK deficiency to modifying the clinical expression of a membrane defect or other congenital anemias is still debated. Some authors excluded a synergetic effect between carriership for PK deficiency in patients with concomitant hereditary spherocytosis,85,90 while others reported that the co-inheritance of heterozygous PK deficiency was associated with an aggravation of the phenotype in two families, one affected by hereditary spherocytosis91 and the other by congenital dyserythropoietic anemia associated with a GATA1 mutation.92

From genotype to new therapies The treatment of PK deficiency is based on supportive measures, including blood transfusions, splenectomy, and managing complications. The only curative treatment is hematopoietic stem cell transplantation (HSCT); however, due the risk of graft-versus-host disease, this should be considered only in severe cases or when it represents the only realistic therapeutic option. New therapeutic options that range from a small molecule PK activator to gene therapy are being developed, and may change the way of treating PK deficiency in the future. In this context, a confirmed diagnosis is crucial to have access to these new therapies, and consequently genotyping is becoming a need for most patients; moreover, it may influence the outcome of the treatment and must therefore be taken into account when directing the patient to possible therapies.

Hematopoietic stem cell transplantation van Straaten et al. recently evaluated the indications, procedures employed and outcomes of HSCT in the series of all the patients with PK deficiency treated between 1996 and 2015 (16 patients from Europe and Asia, no patients resulted as being treated in the USA in that period).93 Two additional cases were recently reported.94,95 The analysis of the genotypes of the treated patients showed a great heterogeneity, and surprisingly, no prevalence of nonsense pathogenic variants. Despite this, as reported in Table 3, most of the missense variants in this series affected amino acid residues that participate directly in the allosteric and catalytic binding site of the enzyme, supporting the observation of genotype-phenotype analysis. As expected, no correlation was observed

Table 3. Patients who have undergone hematopoietic stem cell transplantation and their genotype.

Pt 1 Pt 2 Pt 3 Pt 4 Pt 5 Pt 6 Pt 7 Pt 8 Pt 9 Pt 10 Pt 11 Pt 12 Pt 13 Pt 14 Pt 15 Pt 16 Ref 94 Ref 95

Sex

Country

Genotype

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

Asia EU Asia EU Asia EU EU EU EU Asia EU Asia Asia EU Asia Asia China Japan

Unknown p. [E241*; R532W] p.[K348N; R359H] p.[E241*; R488Q] p. [R40Q; D339N] p. [M377fs; M377fs] p.[G165V; R510Q] p.[G511E; E538*] p.[I494T; R559*] p.[V283A; I314T] p.[K541fs; K541fs] p.[D221Y; I314T] p.[V283A;V283A] p.[D331Q;D339H] c.[1270-3C>A];p.[G540*] c.[1270-3C>A];p.[G540*] p.[I314T; I314T] p.[Pro145Hisfs;Pro145Hisfs]

Mutation effect Unknown Nonsense Missense 2 Nonsense Missense Nonsense Missense Missense Missense Missense Nonsense Missense 3 Missense Missense 5 Nonsense Nonsense Missense 3 Nonsense

Unknown Missense 1 Missense 2 Missense Missense 3 Nonsense Missense 4 Nonsense Nonsense Missense 3 Nonsense Missense 3 Missense Missense 3 Nonsense Nonsense Missense 3 Nonsense

Splenectomy

Age at HCST

No Yes No No No Yes Yes Yes No No Yes No No Yes Yes No No Yes

5y 15 1 y 7 mo 3y 2 y 6 mo 17 y 39 y 7y 6y 1 y 6 mo 10 y 9y 1 y 6 mo 41 y 11 y 8y Unknown 32 y

Year 1996 2002 2009 2009 2009 2010 2011 2013 2013 2013 2014 2014 2015 2015 Unknown Unknown Unknown Unknown

Outcome Alive Deceased Alive Alive Alive Deceased Deceased Alive Deceased Alive Deceased Alive Alive Alive Alive Alive Alive Alive

Structure ref 10,11 47 10 16

10 10,17 10,14

Missense variants falling in â&#x20AC;&#x153;strategicâ&#x20AC;? functional amino acid residues or associated with documented thermo-unstable variants are reported in bold. 1Directly involved in the fructose 1,6 bisphosphate activator. 2Directly involved in the substrate and cation binding sites. 3Residues directly involved in the allosteric site and catalytic center. 4Highly unstable. 5Proximity of the substrate-binding site.

haematologica | 2020; 105(9)

2225

P. Bianchi and E.â&#x20AC;&#x2C6;Fermo

between the type of mutations and the outcome of the treatment.

Allosteric activator (AG-348) AG-348 is an allosteric activator of PK-R that binds in a pocket at the dimer-dimer interface, distinct from the allosteric activator fructose 1,6 bisphosphate binding domain, inducing the active R-state conformation of the PK-R tetramer. Preclinical studies showed that AG-348 enhanced activity in vitro in wild-type PK and in a broad spectrum of PKLR mutations; this finding was consistent with the known binding site for AG-348, which is distinct from the areas of the most common PKLR mutations.11 Data from phase I and phase II studies demonstrated that the glycolytic pathway is activated upon treatment with AG-348, and that 54% of PK-deficient subjects experienced a rise in hemoglobin, all of whom had at least one missense mutation.54,96 It has therefore been hypothesized that a minimal level of full-length PK protein is required for enzyme activation, excluding patients carrying two nonsense variants from the potential benefits of the treatment.54 In an AG-348 clinical trial, evaluating p.R479H as a splicing variant other than a simple amino acid substitution caused an increase of the percentage of patients with a hemoglobin response from 48% to 54%.54 A more recent study investigated the effect of ex vivo treatment with AG-348 on enzyme activity, thermostability, protein levels and ATP in PK-deficient red cells from 15 patients with different genotypes, including the most frequently reported variants in Caucasian p.R486W and p.R510Q;97 the overall results showed a mean 1.8-fold increase in PK activity and a 1.5-fold increase in ATP levels. Protein analyses suggested that a sufficient level of protein is required for cells to respond to AG-348 treatment, as previously reported.54 Interestingly, the thermostability of PK was also found to be significantly improved upon ex vivo treatment with AG-348, but with a high variability in response among the different genotypes; this was particularly evident in PK patients carrying the common mutation p.R510Q, which is known to affect catalytic activity only slightly, but to be highly unstable.13 Overall, these data demonstrated that the clinical utility of AG-348 in PK-deficient patients is influenced by the type of mutations, and that variability in the response can also be increased by the compound heterozygosity that is present in most patients. Prospective studies in patients across a broader range of genotypes and disease severity are required to identify patients who can benefit most from the treatment.

References 1. Nathan DG, Oski FA, Miller DR, et al. Lifespan and organ sequestration of the red cells in pyruvate kinase deficiency. N Engl J Med. 1968;278(2):73-81. 2. Aisaki K, Aizawa S, Fujii H, et al. Glycolytic inhibition by mutation of pyruvate kinase gene increases oxidative stress and causes apoptosis of a pyruvate kinase deficient cell line. Exp Hematol. 2007;35(8):1190-1200. 3. Andres O, Loewecke F, Morbach H, et al. Hereditary spherocytosis is associated with

2226

Gene therapy Ex vivo gene therapy for hematologic genetic disorders is becoming a reality in clinical practice. This gene therapy strategy is based on an autologous transplant in which the infused cells are genetically corrected ex vivo. There are several ongoing clinical trials on the use of gene therapy for rare anemias (Î˛-thalassemia, sickle cell disease, Fanconi anemia) and hereditary metabolic diseases.98,99 Preclinical studies focused on the treatment of PK deficiency by gene therapy have been successfully performed.100-102 Autologous cells corrected with a lentiviral vector carrying a codon optimized version of the wild-type cDNA sequence of the PK-R gene have been demonstrated to be able to compensate the disease phenotype in a murine model, without any adverse effect related to the procedure.102 This procedure has been designated as an orphan drug by the European Medicines Agency (EU/3/14/1330; https://goo.gl/T4N6mO) and by the U.S. Food and Drug Administration (DRU-20165168).103 An open-label, phase 1 gene therapy study consisting of autologous hematopoietic stem and progenitor cells transduced ex vivo with a lentiviral vector encoding for the PK enzyme has been approved and recently opened. PK-deficient patients with confirmed genotype and severe, transfusion-dependent anemia despite splenectomy may be eligible for enrollment (www.clinicaltrial.gov NCT04105166).

Conclusions One of the clear advantages of NGS technologies is the availability of molecular testing for rare diseases in many laboratories, resulting in increased awareness of rare congenital conditions, in the dramatically increased number of molecular variants, in the reduced time of diagnosis and number of misdiagnoses. However, the huge amount of data obtained should be interpreted in the light of knowledge of the pathogenic basis of diseases and always supported by functional studies: on top of the molecular lesion itself, the effect of mutations on the expression and functionality of the protein is known for only a few variants. In addition, the study of compensatory effects of other metabolic pathways and cellular involvement (e.g., membrane channel activities, membrane stability) in response to energy depletion will offer new insights into the interpretation of the effect of PKLR mutations and phenotype.

decreased pyruvate kinase activity due to impaired structural integrity of the red blood cell membrane. Br J Haematol. 2019;187(3): 386-395. 4. Boivin P, Galand C, Hakim J, et al. Acquired red cell pyruvate kinase deficiency in leukemias and related disorders. Enzyme. 1975;19(5-6):294-299. 5. Lin G, Xie Y, Liang X, Wu X, et al. Study on red cell enzymes and isoenzymes in patients with leukemia and myelodysplastic syndromes. Zhonghua Xue Ye Xue Za Zhi. 1997;18(7):350-353. 6. Mattevi A, Valentini G, Rizzi M, et al.

Crystal structure of Escherichia coli pyruvate kinase type I: molecular basis of the allosteric transition. Structure. 1995;3:729741. 7. Muirhead H, Clayden DA, Barford D, et al. The structure of cat muscle pyruvate kinase. EMBO J. 1986;5:475-481. 8. Larsen TM, Laughlin LT, Holden HM, et al. Structure of rabbit muscle pyruvate kinase complexed with Mn2+, K+, and pyruvate. Biochemistry. 1994;33(20):6301-6309. 9. Valentini G, Chiarelli LR, Fortin R, Speranza ML, Galizzi A, Mattevi A. The allosteric regulation of pyruvate kinase. J Biol Chem.

haematologica | 2020; 105(9)

Molecular heterogeneity of PK deficiency 2000;275(24):18145-18152. 10. Valentini G, Chiarelli LR, Fortin R, et al. Structure and function of human erythrocyte pyruvate kinase. Molecular basis of nonspherocytic hemolytic anemia. J Biol Chem. 2002;277(26):23807-23814. 11. Kung C, Hixon J, Kosinski PA, et al. AG-348 enhances pyruvate kinase activity in red blood cells from patients with pyruvate kinase deficiency. Blood. 2017;130(11):13471356. 12. Jurica MS, Mesecar A, Heath PJ, et al. The allosteric regulation of pyruvate kinase by fructose-1,6-bisphosphate. Structure. 1998; 6:195-210. 13. Wang C, Chiarelli LR, Bianchi P, et al. Human erythrocyte pyruvate kinase: characterization of the recombinant enzyme and a mutant form (R510Q) causing nonspherocytic hemolytic anemia. Blood. 2001;98(10): 3113-3120. 14. Kedar P, Hamada T, Warang P, et al. Spectrum of novel mutations in the human PKLR gene in pyruvate kinase-deficient Indian patients with heterogeneous clinical phenotypes. Clin Genet. 2009;75(2):157162. 15. Svidnicki MCCM, Santos A, Fernandez JAA, et al. Novel mutations associated with pyruvate kinase deficiency in Brazil. Rev Bras Hematol Hemoter. 2018;40(1):5-11. 16. Zanella A, Fermo E, Bianchi P, et al. Pyruvate kinase deficiency: the genotype-phenotype association. Blood Rev. 2007;21(4):217-231. 17. Fermo E, Bianchi P, Chiarelli LR, et al. Red cell pyruvate kinase deficiency: 17 new mutations of the PK-LR gene. Br J Haematol. 2005;129(6):839-846. 18. Beutler E. Red Cell Metabolism: A Manual of Biochemical Methods. New York: Grune & Stratton, Inc.; 1984 19. Bianchi P, Fermo E, Glader B, et al; with the endorsement of EuroBloodNet, the European Reference Network in Rare Hematological Diseases. Addressing the diagnostic gaps in pyruvate kinase deficiency: consensus recommendations on the diagnosis of pyruvate kinase deficiency. Am J Hematol. 2019;94(1):149-161. 20. Agarwal AM, Nussenzveig RH, Reading NS, et al. Clinical utility of next-generation sequencing in the diagnosis of hereditary haemolytic anaemias. Br J Haematol. 2016;174(5):806-814. 21. Del Orbe Barreto R, Arrizabalaga B, De la Hoz AB, et al. Detection of new pathogenic mutations in patients with congenital haemolytic anaemia using next-generation sequencing. Int J Lab Hematol. 2016;38(6):629-638. 22. Iwasaki T, Yamamoto T, Muramatsu H, et al. Clinical impact of captured-based targeted sequencing in diagnosis of congenital hemolytic anaemia. Rinsho Ketsueki. 2016;57:1489 23. Roy NB, Wilson EA, Henderson S. A novel 33-gene targeted resequencing panel provides accurate, clinical-grade diagnosis and improves patient management for rare inherited anaemias. Br J Haematol. 2016;175(2):318-330. 24. Fermo E, Vercellati C, Marcello AP, et al. Use of next generation sequencing panel to clarify undiagnosed cases of hereditary hemolytic anaemias. Blood. 2017;130 (Supplement 1):3480. 25. Russo R, Andolfo I, Manna F, Gambale A, et al. Multi-gene panel testing improves diagnosis and management of patients with hereditary anemias. Am J Hematol. 2018;93(5):672-682.

haematologica | 2020; 105(9)

26. Al-Samkari H, van Beers EJ, Kuo KHM, et al. The variable manifestations of disease in pyruvate kinase deficiency and their management. Haematologica. 2020;105(9):22292239. 27. Shefer Averbuch N, Steinberg-Shemer O, Dgany O, et al. Targeted next generation sequencing for the diagnosis of patients with rare congenital anemias. Eur J Haematol. 2018;101(3):297-304. 28. Jamwal M, Aggarwal A, Palodhi A, et al. Next-generation sequencing-based diagnosis of unexplained inherited hemolytic anemias reveals wide genetic and phenotypic heterogeneity. J Mol Diagn. 2020;22(4):579590. 29. Qin L, Nie Y, Chen L, et al. Novel PLKR mutations in four families with pyruvate kinase deficiency. Int J Lab Hematol. 2020;42(2):e84-e87. 30. Kedar PS, Harigae H, Ito E, et al. Study of pathophysiology and molecular characterization of congenital anemia in India using targeted next-generation sequencing approach. Int J Hematol. 2019;110(5):618626. 31. Kanno H, Fujii H, Miwa S. Structural analysis of human pyruvate kinase L-gene and identification of the promoter activity in erythroid cells. Biochem Biophys Res Commun. 1992;188(2):516-523. 32. Noguchi T, Tamada K, Inoue H, et al. The Land R-type isozymes of rat pyruvate kinase are produced from a single gene by use of different promoters. J Biol Chem. 1987;262(29):14366-14371. 33. Canu G, De Bonis M, Minucci A, Capoluongo E. Red blood cell PK deficiency: an update of PK-LR gene mutation database. Blood Cells Mol Dis. 2016;57:100-109. 34. Grace RF, Bianchi P, van Beers EJ, et al. Clinical spectrum of pyruvate kinase deficiency: data from the Pyruvate Kinase Deficiency Natural History Study. Blood. 2018;131(20):2183-2192. 35. Bianchi P, Fermo E, Lezon-Geyda K, et al. Genotype-phenotype correlation and molecular heterogeneity in pyruvate kinase deficiency. Am J Hematol. 2020;95(5):472482. 36. Warang P, Kedar P, Ghosh K et al. Molecular and clinical heterogeneity in pyruvate kinase deficiency in India. Blood Cells Mol Dis. 2013;51(3):133-137. 37. Zanella A, Bianchi P. Red cell pyruvate kinase deficiency: from genetics to clinical manifestations. Baillieres Best Pract Res Clin Haematol. 2000;13(1):57-81. 38. Kanno H, Ballas SK, Miwa S, et al. Molecular abnormality of erythrocyte pyruvate kinase deficiency in the Amish. Blood. 1994;83(8):2311-2316. 39. Machado P, Manco L, Gomes C, et al. Pyruvate kinase deficiency in sub-Saharan Africa: identification of a highly frequent missense mutation (G829A;Glu277Lys) and association with malaria. PLoS One. 2012;7:e47071. 40. Baronciani L, Beutler E. Molecular study of pyruvate kinase deficient patients with hereditary nonspherocytic hemolytic anemia. J Clin Invest. 1995;95(4):1702-1709. 41. Costa C, Albuisson J, Le TH, et al. Severe hemolytic anemia in a Vietnamese family, associated with novel mutations in the gene encoding for pyruvate kinase. Haematologica. 2005;90(1):25-30. 42. Richards S, Aziz N, Bale S, et al. ACMG Laboratory Quality Assurance Committee.Standards and guidelines for the interpretation of sequence variants: a joint

consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-424. 43. Marcello AP, Vercellati C, Fermo E, et al. A case of congenital red cell pyruvate kinase deficiency associated with hereditary stomatocytosis. Blood Cells Mol Dis. 2008;41(3): 261-262. 44. Coutinho R, Bento C, Almeida H, et al. Complex inheritance of chronic haemolytic anaemia. Br J Haematol. 2009;144(4):615616. 45. Manco L, Ribeiro ML, Maximo V, et al. A new PKLR gene mutation in the Râ&#x20AC;?type promoter region affects the gene transcription causing pyruvate kinase deficiency. Br J Haematol 2000;110(4):993-997. 46. van Wijk R, van Solinge WW, Nerlov C, et al. Disruption of a novel regulatory element in the erythroidâ&#x20AC;?specific promoter of the human PKLR gene causes severe pyruvate kinase deficiency. Blood. 2003;101(4):15961602. 47. Kager L, Minkov M, Zeitlhofer P, et al. Two novel missense mutations and a 5bp deletion in the erythroid-specific promoter of the PKLR gene in two unrelated patients with pyruvate kinase deficient transfusiondependent chronic nonspherocytic hemolytic anemia. Pediatr Blood Cancer. 2016;63(5):914-916. 48. Pissard S, Max-Audit I, Skopinski L, et al. Pyruvate kinase deficiency in France: a 3year study reveals 27 new mutations. Br J Haematol. 2006;133(6):683-689. 49. de Vooght KM, van Wijk R, van Wesel AC, et al. Characterization of the -148C>T promoter polymorphism in PKLR. Haematologica 2008;93(9):1407-1408. 50. Gallagher PG, Glader B. Diagnosis of pyruvate kinase deficiency. Pediatr Blood Cancer. 2016;63(5):771-772. 51. Titapiwatanakun R, Hoyer JD, Crain K, et al. Relative red blood cell enzyme levels as a clue to the diagnosis of pyruvate kinase deficiency. Pediatr Blood Cancer. 2008;51(6): 819-821. 52. Zanella A, Bianchi P, Baronciani L, et al. Molecular characterization of PK-LR gene in pyruvate kinase-deficient Italian patients. Blood. 1997;89(10):3847-3852. 53. Gallagher PG, Maksimova Y, Lezon-Geyda K, et al. Aberrant splicing contributes to severe a-spectrin-linked congenital hemolytic anemia. J Clin Invest. 2019;129 (7):2878-2887. 54. Grace RF, Rose C, Layton M, et al. Safety and efficacy of mitapivat in pyruvate kinase deficiency. N Engl J Med. 2019;381(10):933944. 55. Lezon-Geyda K, Rose MJ, McNaull MA, et al. PKLR intron splicing-associated mutations and alternate diagnoses are common in pyruvate kinase deficient patients with single or no PKLR coding mutations. Blood. 2018;132 (Suppl 1):3607. 56. Bagla S, Bhambhani K, Gadgeel M, et al. Compound heterozygosity in PKLR gene for a previously unrecognized intronic polymorphism and a rare missense mutation as a novel cause of severe pyruvate kinase deficiency. Haematologica. 2019;104(9):e428e431. 57. Lesmana H, Dyer L, Li X, et al. Alu element insertion in PKLR gene as a novel cause of pyruvate kinase deficiency in Middle Eastern patients. Hum Mutat. 2018;39(3): 389-393. 58. Zurcher C, Loos JA, Prins HK. Hereditary high ATP content of human erythrocytes.

2227

P. Bianchi and E. Fermo Folia Haematol Int Mag Klin Morphol Blutforsch. 1965;83(4):366-376. 59. Max-Audit I, Rosa R, Marie J. Pyruvate kinase hyperactivity genetically determined: metabolic consequences and molecular characterization. Blood. 1980;56(5):902-909. 60. van Oirschot BA, Francois JJ, van Solinge WW, et al. Novel type of red blood cell pyruvate kinase hyperactivity predicts a remote regulatory locus involved in PKLR gene expression. Am J Hematol. 2014;89(4):380384. 61. Beutler E, Westwood B, van Zwieten R, et al. G‐>T transition at cDNA nt 110 (K37Q) in the PKLR (pyruvate kinase) gene is the molecular basis of a case of hereditary increase of red blood cell ATP. Hum Mutat. 1997;9(3):282-285. 62. Viprakasit V, Ekwattanakit S, Riolueang S, et al. Mutations in Kruppel-like factor 1 cause transfusion-dependent hemolytic anemia and persistence of embryonic globin gene expression. Blood. 2014;123(10):1586-1595. 63. Beutler E, Gelbart T. Estimating the prevalence of pyruvate kinase deficiency from the gene frequency in the general white population. Blood. 2000;95(11):3585-3588. 64. Carey PJ, Chandler J, Hendrick A, et al. Prevalence of pyruvate kinase deficiency in northern European population in the north of England. Northern Region Haematologists Group. Blood. 2000;96(12): 4005-4006. 65. Secrest MH, Storm M, Carrington C, et al. Prevalence of pyruvate kinase deficiency: a systematic literature review. Eur J Haematol. 2020 Apr 12. [Epub ahead of print] 66. van Bruggen R, Gualtieri C, Iliescu A, et al. Modulation of malaria phenotypes by pyruvate kinase (PKLR) variants in a Thai population. PLoS One. 2015;10:e0144555. 67. Min-Oo G, Fortin A, Tam MF, et al. Pyruvate kinase deficiency in mice protects against malaria. Nat Genet. 2003;35(4):357362. 68. Ayi K, Min-Oo G, Serghides L, et al. Pyruvate kinase deficiency and malaria. New Eng J Med. 2008;358(17):1805-1810. 69. Qidwai T, Jamal F, Singh S. Exploring putative molecular mechanisms of human pyruvate kinase enzyme deficiency and its role in resistance against Plasmodium falciparum malaria. Interdiscipl Sci. 2014;6(2):158-166. 70. van Wijk R, Huizinga EG, van Wesel AC, et al. Fifteen novel mutations in PKLR associated with pyruvate kinase (PK) deficiency: structural implications of amino acid substitutions in PK. Hum Mutat. 2009; 30(3):446453. 71. Jaouani M, Manco L, Kalai M, et al. Molecular basis of pyruvate kinase deficiency among Tunisians: description of new mutations affecting coding and noncoding regions in the PKLR gene. Int J Lab Hematol. 2017;39(2):223-231. 72. Christensen RD, Yaish HM, Johnson CB, et al. .Six children with pyruvate kinase deficiency from one small town: molecular characterization of the PK-LR gene. J Pediatr. 2011;159(4):695-697. 73. Lakomek M, Winkler H, Pekrun A, et al. Erythrocyte pyruvate kinase deficiency. The influence of physiologically important metabolites on the function of normal and defective enzymes. Enzyme Protein.

2228

1994;48(3):149-163. 74. Finkenstedt A, Bianchi P, Theurl I, et al. Regulation of iron metabolism through GDF15 and hepcidin in pyruvate kinase deficiency. Br J Haematol. 2009;144(5):789793. 75. Goto T, Ubukawa K, Kobayashi I, et al. ATP produced by anaerobic glycolysis is essential for enucleation of human erythroblasts. Exp Hematol. 2019;72:14-26.e1 76. van Beers EJ, van Straaten S, Morton DH, et al. Prevalence and management of iron overload in pyruvate kinase deficiency: report from the Pyruvate Kinase Deficiency Natural History Study. Haematologica. 2019;104(2):e51-e53. 77. Zanella A, Berzuini A, Colombo MB, et al. Iron status in red cell pyruvate kinase deficiency: study of Italian cases. Br J Haematol. 1993;83(3):485-490. 78. Mojzikova R, Koralkova P, Holub D, et al. Iron status in patients with pyruvate kinase deficiency: neonatal hyperferritinaemia associated with a novel frameshift deletion in the PKLR gene (p.Arg518fs), and low hepcidin to ferritin ratios. Br J Haematol. 2014;165(4):556-563. 79. Zanella A, Bianchi P, Iurlo A, et al. Iron status and HFE genotype in erythrocyte pyruvate kinase deficiency: study of Italian cases. Blood Cells Mol Dis. 2001;27(3):653-661. 80. Rider NL, Strauss KA, Brown K, et al. Erythrocyte pyruvate kinase deficiency in an old-order Amish cohort: longitudinal risk and disease management. Am J Hematol. 2011;86(10):827-834. 81. Grace RF, Zanella A, Neufeld EJ, et al. Erythrocyte pyruvate kinase deficiency: 2015 status report. Am J Hematol. 2015;90 (9):825-830. 82. Perseu L, Giagu N, Satta S, et al. Red cell pyruvate kinase deficiency in Southern Sardinia. Blood Cells Mol Dis. 2010;45(4): 280-283. 83. Branca R, Costa E, Rocha S, et al. Coexistence of congenital red cell pyruvate kinase and band 3 deficiency. Clin Lab Haematol. 2004;26(4):297-300. 84. Christensen RD, Yaish HM, Nussenzveig RH, Agarwal AM. Siblings with severe pyruvate kinase deficiency and a complex genotype. Am J Med Genet A. 2016;170(9):24492452. 85. Vercellati C, Marcello AP, Fermo E, et al. A case of hereditary spherocytosis misdiagnosed as pyruvate kinase deficient hemolytic anemia. Clin Lab. 2013;59(3-4):421-424. 86. Fermo E, Vercellati C, Marcello AP, et al. Hereditary xerocytosis due to mutations in piezo1 gene associated with heterozygous pyruvate kinase deficiency and beta-thalassemia trait in two unrelated families. Case Rep Hematol. 2017;2017:2769570. 87. Cohen-Solal M, Préhu C, Wajcman H, et al, A new sickle cell disease phenotype associating Hb S trait, severe pyruvate kinase deficiency (PK Conakry), and an alpha2 globin gene variant (Hb Conakry). Br J Haematol. 1998;103(4):950-956. 88. Alli N, Coetzee M, Louw V, et al. Sickle cell disease in a carrier with pyruvate kinase deficiency. Hematology. 2008;13(6):369-372. 89. Manco L, Vagace JM, Relvas L, et al. Chronic haemolytic anaemia because of pyruvate kinase (PK) deficiency in a child heterozy-

gous for haemoglobin S and no clinical features of sickle cell disease. Eur J Haematol. 2010;84(1):89-90. 90. Zarza R, Moscardo M, Alvarez R, et al. Coexistence of hereditary spherocytosis and a new red cell pyruvate kinase variant: PK Mallorca. Haematologica. 2000;85(3):227232. 91. van Zwieten R, van Oirschot BA, Veldthuis M, et al. Partial pyruvate kinase deficiency aggravates the phenotypic expression of band 3 deficiency in a family with hereditary spherocytosis. Am J Hematol. 2015;90 (3):E35-E39. 92. Pereira J, Gonzalez A, Vagace J, et al. Congenital dyserythropoietic anemia associated to a GATA1 mutation aggravated by pyruvate kinase deficiency. Ann Hematol. 2016;95(9):1551-1553. 93. van Straaten S, Bierings M, Bianchi P, et al. Worldwide study of hematopoietic allogeneic stem cell transplantation in pyruvate kinase deficiency. Haematologica. 2018;103 (2):e82-e86. 94. He Y, Luo J, Lei Y, et al. A novel PKLR gene mutation identified using advanced molecular techniques. Pediatr Transplant. 2018;22 (2): e13143. 95. Shimomura M, Doi T, Nishimura S, et al. Successful allogeneic bone marrow transplantation using immunosuppressive conditioning regimen for a patient with red blood cell transfusion dependent pyruvate kinase deficiency anemia. Hematol Rep. 2020;12 (1):8305. 96. Yang H, Merica E, Chen Y, et al. Phase 1 single and multiple-ascending-dose randomized studies of the safety, pharmacokinetics, and pharmacodynamics of AG-348, a firstin-class allosteric activator of pyruvate kinase R, in healthy volunteers. Clin Pharmacol Drug Dev. 2019;8(2):246-259. 97. Rab MAE, van Oirschot BA, Kosinski PA, et al. AG-348 (Mitapivat), an allosteric activator of red blood cell pyruvate kinase, increases enzymatic activity, protein stability, and ATP levels over a broad range of PKLR genotypes. Haematologica. 2020 Jan 23. [Epub ahead of print] 98. Dunbar CE, High KA, Joung JK, et al. Gene therapy comes of age. Science. 2018;359 (6372):eaan4672. 99. Magrin E, Miccio A, Cavazzana M. Lentiviral and genome-editing strategies for the treatment of β-hemoglobinopathies. Blood. 2019;134(15):1203-1213. 100. Meza NW, Alonso-Ferrero ME, Navarro S, et al. Rescue of pyruvate kinase deficiency in mice by gene therapy using the human isoenzyme. Mol Ther. 2009;17(12):20002009. 101. Meza NW, Quintana-Bustamante O, Puyet A, et al. In vitro and in vivo expression of human erythrocyte pyruvate kinase in erythroid cells: a gene therapy approach. Hum Gene Ther. 2007;18(6):502-514. 102. Garcia-Gomez M, Calabria A, Garcia-Bravo M, et al. Safe and efficient gene therapy for pyruvate kinase deficiency. Mol Ther. 2016;24(7):1187-1198. 103. Quintana-Bustamante O, Fañanas-Baquero S, Orman I, et al. Gene editing of PKLR gene in human hematopoietic progenitors through 5' and 3' UTR modified TALEN mRNA. PLoS One. 2019;14(10):e0223775.

haematologica | 2020; 105(9)

REVIEW ARTICLE

The variable manifestations of disease in pyruvate kinase deficiency and their management

Ferrata Storti Foundation

Hanny Al-Samkari,1 Eduard J. van Beers,2 Kevin H.M. Kuo,3 Wilma Barcellini,4 Paola Bianchi,4 Andreas Glenthøj,5 María del Mar Mañú Pereira,6 Richard van Wijk,7 Bertil Glader8 and Rachael F. Grace9

Division of Hematology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA; 2Van Creveldkliniek, University Medical Centre Utrecht, University of Utrecht, Utrecht, the Netherlands; 3Division of Hematology, University of Toronto, University Health Network, Toronto, Ontario, Canada; 4UOS Ematologia, Fisiopatologia delle Anemie, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Milan, Italy; 5Department of Hematology, Herlev and Gentofte Hospital, Herlev, Denmark; 6Translational Research in Rare Anaemia Disorders, Vall d'Hebron Institut de Recerca (VHIR), Barcelona, Spain; 7 Department of Clinical Chemistry & Hematology, University Medical Center Utrecht, Utrecht University, Utrecht, the Netherlands; 8Lucile Packard Children’s Hospital, Stanford University School of Medicine, Palo Alto, CA, USA and 9Dana/Farber Boston Children's Cancer and Blood Disorders Center, Harvard Medical School, Boston, MA, USA. 1

Haematologica 2020 Volume 105(9):2229-2239

ABSTRACT

yruvate kinase deficiency (PKD) is the most common cause of chronic hereditary non-spherocytic hemolytic anemia and results in a broad spectrum of disease. The diagnosis of PKD requires a high index of suspicion and judicious use of laboratory tests that may not always be informative, including pyruvate kinase enzyme assay and genetic analysis of the PKLR gene. A significant minority of patients with PKD have occult mutations in non-coding regions of PKLR which are missed on standard genetic tests. The biochemical consequences of PKD result in hemolytic anemia due to red cell pyruvate and ATP deficiency while simultaneously causing increased red cell 2,3-diphosphoglycerate, which facilitates oxygen unloading. This phenomenon, in addition to numerous other factors such as genetic background and differences in splenic function result in a poor correlation between symptoms and degree of anemia from patient to patient. Red cell transfusions should, therefore, be symptom-directed and not based on a hemoglobin threshold. Patients may experience specific complications, such as paravertebral extramedullary hematopoiesis and chronic debilitating icterus, which require personalized treatment. The decision to perform splenectomy or hematopoietic stem cell transplantation is nuanced and depends on disease burden and long-term outlook given that targeted therapeutics are in development. In recognition of the complicated nature of the disease and its management and the limitations of the PKD literature, an international working group of ten PKD experts convened to better define the disease burden and manifestations. This article summarizes the conclusions of this working group and is a guide for clinicians and investigators caring for patients with PKD.

Introduction Pyruvate kinase deficiency (PKD) is the most common cause of chronic hereditary non-spherocytic hemolytic anemia, with a prevalence reported to be between 1:20,000 and 1:300,000 in Caucasian populations and a higher prevalence in areas in which malaria is endemic.1-4 Despite its well-described global geographical distribution, its incidence in large areas of the world remains unknown. Following the identification by Selwyn and Dacie in 1954 of a patient with hereditary hemolytic anemia whose erythrocytes were not rescued by glucose but were rescued by adenosine triphosphate (ATP) in an ex vivo hemolysis assay, it was haematologica | 2020; 105(9)

Correspondence: HANNY AL-SAMKARI hal-samkari@mgh.harvard.edu Received: October 21, 2019. Accepted: January 20, 2020. Pre-published: March 12, 2020. doi:10.3324/haematol.2019.240846 ©2020 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.

2229

H. Al-Samkari et al.

recognized that glycolytic defects could lead to hemolytic anemia.5 PKD was first described as a cause of hereditary hemolytic anemia in 1961 by Valentine and colleagues.6 Pyruvate kinase (PK) catalyzes the conversion of phosphenolpyruvate to pyruvate and is the rate-limiting enzyme in erythrocyte energy production, which is an exclusively anaerobic process. As only one of two glycolytic enzymes that generates ATP, homozygous or compound heterozygous loss-of-function mutations in the PKLR gene, which encodes erythrocyte PK, result in erythrocyte ATP shortage. This ATP deficiency presumably results in a reduced capacity to maintain the red cell membrane and diminished erythrocyte deformability, resulting in a shortened lifespan and destruction in the spleen. Because patients experience mainly extravascular hemolysis, splenectomy frequently improves anemia.7 A striking difference between PKD and most other hereditary hemolytic anemias is the dramatic post-splenectomy reticulocytosis characteristic of PKD, which typically increases 50% or more over pre-splenectomy counts. Reticulocytes have a much higher ATP requirement as compared with mature red cells but can rely on oxidative phosphorylation for energy. However, in the hypoxic splenic environment, PKdeficient reticulocytes must rely on glycolysis, which does not meet the ATP needs, leading to hemolysis.8-10 In the absence of the spleen, reticulocyte survival increases. The clinical manifestations of PKD are heterogenous and the severity varies considerably, from fetal hydrops with intrauterine demise to incidentally discovered asymptomatic fully compensated hemolysis to a severe transfusion-dependent anemia from birth through old age.11-13 Patientsâ&#x20AC;&#x2122; symptoms often do not correlate with the severity of anemia, adding an additional dimension of complexity in disease management.14 Complications include iron overload, pulmonary hypertension, endocrinopathies, osteoporosis and bone fractures, extramedullary hematopoiesis, gallstones, and lower extremity ulcers, among others.11 PKD is characterized by a high prevalence of iron overload regardless of whether patients require regular transfusions, necessitating monitoring in all patients and frequent institution of iron chelation therapy.15 Beyond splenectomy and hematopoietic stem cell transplantation, treatment has been largely supportive from the time PKD was first described.16,17 This is rapidly changing. A small molecule allosteric activator of PK is under development, with safety and efficacy demonstrated in phase I and II trials;18,19 it is currently being investigated in phase III trials. Following on the success of gene therapy in animal PKD models20 and humans with thalassemia and other hematologic disorders,21,22 a clinical trial of gene therapy for PKD was launched in 2019. Now more than ever before, it is critical for hematologists to elucidate the specific diagnoses of patients with congenital hemolytic anemias, including PKD, and institute proper interventions. The promise of effective targeted therapies has greatly revived interest in PKD, but published data remain limited and no evidence-based guidelines for the management of these patients exist. This article, developed and written by ten international experts in PKD, reviews the manifestations and spectrum of disease in patients and highlights the most common, most important, and most challenging presentations of this disease. To develop this report, this PKD Burden of Disease working group had multiple focused online discussions over 2230

the course of 1 year and an in-person meeting in San Diego (CA, USA). Given that many providers have limited experience caring for patients with PKD, hypothetical case presentations based on the collective PKD patient care experience of the working group are included to introduce each section and illustrate the key signs, symptoms, and complications which outline the scope and spectrum of disease in PKD.

Diagnostic challenges The newly-diagnosed adult patient CASE: A 29-year old man with a history of ulcerative colitis, inflammatory anemia, and cholecystitis (cholecystectomy performed at the age of 27) is referred for further workup of hemolytic anemia with a negative direct antiglobulin test. He underwent colectomy at the age of 17 for ulcerative colitis. Following colectomy, his chronic anemia (hemoglobin 9.0-11.0 g/dL) improved to a new baseline of 11.5-12.5 g/dL. He has long-standing hyperbilirubinemia that has been attributed to Gilbert syndrome, although genetic confirmation of this has not been performed. Currently his laboratory evaluation is remarkable for the following: hemoglobin 10.9 g/dL, mean cell volume 106 fL, absolute reticulocyte count 1,030 Ă&#x2014; 109/L, indirect bilirubin 3.7 mg/dL, lactate dehydrogenase 476 U/L, and undetectable haptoglobin. The direct antiglobulin test is negative. A peripheral blood film is nonspecific and hemoglobin electrophoresis is unremarkable. Testing for paroxysmal nocturnal hemoglobinuria, erythrocyte membrane defects, and glucose-6-phosphate dehydrogenase deficiency is negative. PK activity is 1.2 U/g hemoglobin (reference range, 6.7-14.3 U/g hemoglobin). A diagnosis of PKD is confirmed by genetic testing revealing compound heterozygosity for two missense mutations in the PKLR gene: c.1091G>A (p.Gly364Asp) and c.1529G>A (p.Arg510Gln). Patients with PKD who are regularly transfused or have severe hemolytic anemia are typically diagnosed in the neonatal period or in early childhood. Patients who have never or rarely been transfused, however, are frequently diagnosed as adults.23 These patients typically have mildto-moderate anemia that can be misdiagnosed as thalassemia trait (without genetic confirmation), iron deficiency (often in a menstruating woman), more common hemolytic anemias (e.g. hereditary spherocytosis), or inflammatory anemia. Patients can also evade early diagnosis because of normal or near normal hemoglobin concentrations in the setting of well-compensated hemolysis. Complications resulting from PKD, such as hyperbilirubinemia, gallstones, and iron overload, can often be mistaken as discrete diagnostic entities attributed to other causes, such as Gilbert syndrome or hereditary hemochromatosis. Recognition of even the mildest forms of disease is important for several reasons. Although the frequency of certain complications is highest in patients with two drastic PKLR mutations, many complications, including hemolysis, gallstones, iron overload, and aplastic crises are not uncommon in patients with two missense mutations and mild anemia.11 Folic acid supplementation is recommended in essentially all patients to prevent deficiency due to rapid cell turnover; recognition of increased folate requirements is particularly important in women of childbearing age. Iron overload is common even in patients who never receive red cell transfusion due to chronic hemolysis15 and may lead to haematologica | 2020; 105(9)

Manifestations of disease in PKD

Figure 1. Paravertebral extramedullary hematopoietic masses in pyruvate kinase deficiency. (A) Multiple transverse sections demonstrating paravertebral masses (arrows) in close proximity to nerve roots. (B) Sagittal section demonstrating a large paravertebral mass (bounded by arrows) extending from the vertebra.

liver dysfunction, cardiac dysfunction, or endocrinopathies if not addressed. A small but significant proportion of patients can develop extramedullary hematopoiesis.11 Extramedullary hematopoietic masses in PKD are often paravertebral,24-26 as illustrated in Figure 1. These masses can enlarge over time and can result in nerve root compression resulting in neurological compromise, including paralysis, if left untreated.27,28 Although evidence is lacking, some members of the working group treat patients with progressing paravertebral masses with chronic red cell transfusion to suppress the growth of the masses. Extramedullary hematopoietic masses can also be mistaken for malignant tumors, especially in undiagnosed patients. As in other hemolytic anemias, patients with PKD can develop lowerextremity ulcers, usually medial in association with the medial malleolus,29 which can be slow to heal or even fail to heal (Figure 2). There are no data to guide the management of leg ulcers in PKD, so ulcers are managed similarly to those seen in sickle cell disease or thalassemia.30,31 Once PKD has been diagnosed, screening and regular monitoring for complications from chronic hemolysis should be initiated, since many complications, such as iron overload, can be asymptomatic.15 The type and frequency of screening vary between institutions, but the screening is usually directed at complications that carry high morbidity if untreated. These include iron overload, extramedullary hematopoiesis, osteopenia, osteoporosis, gallstones, and pulmonary hypertension. Table 1 shows our consensus approach to screening in patients with PKD. haematologica | 2020; 105(9)

Figure 2. Lower extremity non-healing ulcer in an adult with pyruvate kinase deficiency. Note the location posterior to the medial malleolus.

2231

H. Al-Samkari et al.

Complexity of enzymatic and genetic diagnosis of pyruvate kinase deficiency CASE: A 24-month old boy previously diagnosed with hereditary xerocytosis and glucose-6-phosphate dehydrogenase deficiency is seen for a second opinion. He has had severe transfusion-dependent hemolytic anemia since birth. On evaluation, PK enzyme activity is in the normal range, but with a low PK/hexokinase ratio. There is no family history of anemia. On evaluation of the patient’s parents, the patient’s mother has PK enzyme activity approximately 50% of normal and a low PK/hexokinase ratio and the patent’s father has normal PK enzyme activity and PK/hexokinase ratio. Genetic testing is performed and the patient is found to be a compound heterozygote for c.1091G>A (p.Gly364Asp) and c.1529G>A (p.Arg510Gln) mutations in exons 8 and 11 of PKLR, respectively. As these are known pathogenic mutations, the diagnosis of PKD is confirmed. PKLR analysis of the patient’s mother showed the c.1091G>A mutation. Surprisingly,

PKLR analysis of the patient’s father revealed two wildtype alleles and no evidence of a PKLR c.1529G>A mutation (paternity was confirmed via additional testing). Further testing was performed to identify a possible somatic mutation in PKLR in the father. Massive parallel sequencing of the region encompassing nucleotide c.1529 of PKLR was performed on paternal genomic DNA isolated from peripheral blood, a buccal swab, urine, and semen. Sequencing of the same region was performed on genomic DNA isolated from peripheral blood from the patient, his mother and paternal grandparents. The results of this testing are shown in Table 2. The presence of the c.1529G>A mutation was demonstrated in DNA of all tissues tested from the father. DNA from the paternal grandparents showed absence of the c.1529G>A mutation, thereby confirming the post-zygotic origin of the c.1529G>A mutation in the father. Altogether, these results confirmed that the father is a mosaic for the c.1529G>A mutation in PKLR. PKD should be suspected in patients with an unex-

Table 1. Our consensus approach to routine screening of the patient with pyruvate kinase deficiency.

Condition

Recommended Screening

Cholelithiasis

Consider screening for gallstones with interval ultrasound examinations; discuss cholecystectomy with patient if gallstones are observed

Viral infections

Iron overload

Cholecystectomy should be considered at the time of splenectomy in patients undergoing splenectomy even in the absence of gallstones due to the high rate of gallstones in patients with pyruvate kinase deficiency (both in general and following splenectomy);40 patients should be counseled on the risk of intrahepatic cholestasis Human immunodeficiency virus and hepatitis C annually is reasonable if receiving blood transfusions in certain countries (risk varies depending on country) Baseline parvovirus B19 serology if parvovirus serological status unknown Non-transfused or minimally transfused patients: ferritin yearly; R2 or T2* magnetic resonance imaging of liver and heart once in early adulthood or when ferritin >500 μg/L with follow-up interval dependent on findings

Regularly transfused patients: ferritin every 6 months (every 3 months if receiving chelation therapy); R2 or T2* magnetic resonance imaging yearly Osteopenia 25-hydroxy-vitamin D levels yearly; if low, replete and re-check after 8 weeks of vitamin D repletion DEXA baseline in early adulthood, with follow-up interval dependent on findings Endocrinopathies Thyroid-stimulating hormone, sex hormones and fructosaminea yearly; can forego screening or increase screening interval in patients with no evidence of iron overload Pulmonary hypertension Echocardiogram once after the age of 30 or prior to pregnancy; otherwise perform only for concerning symptoms Extramedullary Perform imaging only for concerning symptoms; have a high index of suspicion for paravertebral extramedullary hematopoiesis hematopoiesis in cases of neuropathy or unexplained pain Fructosamine should be used instead of hemoglobin A1c for diabetes mellitus screening in patients with hemolytic anemias.

Table 2. Results of massive parallel sequencing of the region encompassing nucleotide c.1529 of PKLR on genomic DNA isolated from peripheral blood, buccal swab, urine, and semen from members of the family of the patient in the case described in “Complexity of enzymatic and genomic diagnosis in pyruvate kinase deficiency”

Individual

Patient Mother Father

Paternal grandfather Paternal grandmother

Sample

PKLR genotype by Sanger sequencing for nucleotide c.1529

Peripheral blood Peripheral blood Peripheral blood Urine Buccal swab Semen Peripheral blood Peripheral blood

c.1529G>A/wt wt/wt wt/wt wt/wt wt/wt c.1529G>A/wt wt/wt wt/wt

Massive parallel sequencing PKLR genotype Mutated allele for nucleotide c.1529 frequency (%) c.1529G>A/wt wt/wt c.1529G>A/wt c.1529G>A/wt c.1529G>A/wt c.1529G>A/wt wt/wt wt/wt

48.6 0 5.1 10.3 16.73 21.76 0 0

wt: wildtype.

2232

haematologica | 2020; 105(9)

Manifestations of disease in PKD

plained chronic Coombs-negative hemolytic anemia following an unrevealing standard hemolytic anemia evaluation.23 Tests to be considered in this situation are outlined in Table 3, and our stepwise approach to the diagnosis of a patient with chronic unexplained hemolytic anemia is detailed in Figure 3. The first test to perform in cases of suspected PKD is the PK enzyme assay. While it is a useful and inexpensive screening test, PK enzyme activity assays have a number of limitations. Recent transfusion can result in falsely normal enzyme activity.17,23 PK enzyme activity is red cell age-dependent, and so cases of suspected PKD with a normal PK enzyme assay should also be evaluated further via calculation of the PK/hexokinase ratio.32 This ratio relates PK activity to the activity of hexokinase, another red cell age-dependent enzyme used as an internal standard. A decreased PK/hexokinase ratio is suspicious for PKD and should be investigated further with genetic testing.23 Of clinical relevance, PK enzyme activity has not been shown to correlate with disease severity.6, 11 In addition, this case demonstrates the complexity of genetic diagnosis in PKD. Over 300 functional mutations in PKLR have been described.33,34 While most cases can be diagnosed with standard PKLR exon sequencing of the patient, occasionally more comprehensive analysis are necessary to confirm the diagnosis or allow for family planning in subsequent pregnancies. This includes screening for parental somatic mutations in cases in which an apparent de novo mutation is identified.35 In more routine cases, in which a couple is planning to conceive and one partner is a confirmed carrier of PKD (as occurs in all children of affected individuals), testing of the partner with an unconfirmed PKLR mutation status should be through sequencing rather than testing PK enzyme activity, as many carriers will have normal PK activity. An analysis of PK enzyme activity in 31 family studies (mother, father, and patient) performed by three of the authors (RvW, PB,

MMP, unpublished observation) revealed eight cases in which a carrier had normal PK activity. Therefore, in this setting enzymatic testing alone to identify possible carriers is inadequate and genetic testing is advised.

Occult mutations in pyruvate kinase deficiency CASE: A 47-year old woman with PKD on regular transfusions following splenectomy and cholecystectomy is referred for confirmation of her diagnosis after genetic testing revealed only heterozygosity for the c.460G>A (p.Ala154Thr) mutation. She was initially diagnosed with PKD over 30 years ago and despite splenectomy in childhood has required regular transfusions (2 units every 3 weeks). PK enzyme activity measured immediately prior to transfusion is 2.5 U/g hemoglobin (reference range, 6.714.3 U/g hemoglobin). The c.460G>A mutation was inherited from the patientâ&#x20AC;&#x2122;s mother who had normal PK activity (8.1 U/g hemoglobin) and a normal PK/hexokinase ratio of 9.0. No other relatives are available for analysis. To exclude the possibility of other causes of hereditary hemolytic anemia, a targeted next-generation sequencing gene panel analysis was performed, which did not identify additional mutations in 46 genes associated with hereditary hemolytic anemias. Given the severity of her disease, she was interested in considering a clinical trial of PKD-directed treatment but was excluded because she had only one mutated PKLR allele. While it is important to confirm an enzymatic diagnosis of PKD with genetic testing,23 up to 10% of patients with a diagnosis of PKD will have only one mutation identified on standard PKLR exome sequencing.36 It is recognized that PKLR intron splicing-associated mutations are sometimes found in patients with single or no PKLR coding mutations.37 In one study of 13 kindreds with hereditary non-spherocytic hemolytic anemia and either single or no identified mutations in PKLR coding regions, whole genome sequencing identified five kindreds with unique PKLR deep intronic mutations predicted to perturb normal

Table 3. Advanced diagnostic workup of a patient with unidentified Coombs-negative hemolytic anemia to be considered after basic testing (peripheral blood film, hemoglobin electrophoresis, etc.). Often many of these tests will be indicated in such a patient and may be obtained in a stepwise fashion to diagnose the underlying disorder. Our consensus stepwise approach is given in Figure 1.

Test

Situations to Consider

Pyruvate kinase enzyme assay, other glycolytic enzyme testing, pyrimidine 5â&#x20AC;&#x2122;-nucleotidase testing

Chronic/lifelong hemolytic anemia with bland peripheral blood film, possible splenomegaly; certain glycolytic defects may present with other pathological features (e.g. severe neuromuscular symptoms in triosephosphate isomerase deficiency, myopathy in phosphofructokinase deficiency) Basophilic stippling of red cells common in pyrimidine 5â&#x20AC;&#x2122;-nucleotidase deficiency Evidence of autoimmune hemolysis (e.g. spherocytes, microspherocytes) but standard direct antiglobulin test negative for IgG and C3 Possible red cell membrane disorders (e.g. hereditary spherocytosis, hereditary elliptocytosis, hereditary xerocytosis) or congenital dyserythropoietic anemias

Extended direct antiglobulin testing (e.g. IgA detection, elution) Osmotic gradient ektacytometry, osmotic fragility test, eosin-5-maleimide binding, genetic testing for red cell membrane mutations Flow cytometry for CD55 and CD59, multiparameter fluorescent aerolysin-based flow cytometry Ceruloplasmin ADAMTS13 activity Genetic testing for unstable hemoglobin variants Targeted next-generation sequencing panels haematologica | 2020; 105(9)

Hemolysis with thrombosis and/or other cytopenias

Screening for Wilson disease in a patient with unidentified Coombs-negative hemolytic anemia and psychiatric/neurological symptoms, evidence of liver dysfunction, and/or Kayser-Fleischer rings Concomitant thrombocytopenia in a patient with chronic unidentified hemolysis is concerning for congenital thrombotic thrombocytopenic purpura (Upshaw-Schulman syndrome) Chronic/lifelong hemolytic anemia with bland peripheral blood film or basophilic stippling of red cells, possible splenomegaly Cases with unexplained hemolytic anemia after comprehensive hematologic testing or when hematologic testing cannot be performed accurately (i.e. in recently-transfused patients, neonates, or shipped samples) 2233

H. Al-Samkari et al.

mRNA processing.38 Additional mutations in other regions involved in PKLR gene expression may also play a role in the pathogenesis of PKD.39 In addition to lack of diagnosis, patients with only one identifiable mutation in PKLR may be excluded from clinical trials, which use the presence of two mutated PKLR alleles for eligibility. This is what

occurred in the case described above: the patient described clearly has a clinical and enzymatic diagnosis of PKD, but standard exome sequencing-based genetic testing has failed to elucidate her genotype. Such patients with high suspicion of PKD as the cause of chronic hemolysis (e.g. persistently low PK enzyme activity) should be

Figure 3. Our consensus, stepwise approach to laboratory workup of a patient with chronic hemolytic anemia. The initial workup includes hemolysis testing performed routinely. The second-pass workup is intended to rule out relatively common inherited entities (including hemoglobinopathies not identified in the initial workup) as well as paroxysmal nocturnal hemoglobinuria, particularly relevant if the patient presents in adulthood. The third-pass workup allows for identification of pyruvate kinase deficiency and red cell membrane abnormalities not diagnosed in prior steps. If this three-step workup is unrevealing, additional testing is recommended to diagnose particularly rare inherited and acquired causes of hemolytic anemia. The diagnostician may narrow or broaden the workup at each step as appropriate and as testing is available; for example, molecular PKLR and KLF1 testing can be reasonably performed earlier in the workup. Additionally, the clinician should be aware that many specialized tests are poorly standardized between laboratories. aDeficiency may result in a hemolytic picture due to ineffective erythropoiesis; folate may be low in chronic hereditary anemias due to rapid cell turnover. bAllows identification of most hemoglobinopathies. TTP: thrombotic thrombotic thrombocytopenic purpura; DIC: disseminated intravascular coagulopathy; PNH: paroxysmal nocturnal hemoglobinuria; DAT: direct antiglobulin test; ADAMTS13: a disintegrin and metalloproteinase with thrombospondin motifs 13.

2234

haematologica | 2020; 105(9)

Manifestations of disease in PKD

sent for additional genetic analyses including non-coding intronic region sequencing and copy number variation analyses.

Impact of disease manifestations on treatment decisions Splenectomy and hematopoietic stem cell transplantation in pyruvate kinase deficiency CASE: A 6-year old boy with genetically confirmed PKD presents to the hematology clinic for a discussion regarding his management, particularly for consideration of splenectomy. This is prompted by the complications of his disease thus far. In his lifetime he has received a total of ten red blood cell transfusions during episodes of symptomatic anemia in the setting of increased hemolysis with viral infections. Supportive care for PKD is focused on transfusions and/or splenectomy. Approximately 60% of PKD patients are splenectomized, with a median hemoglobin rise of 1.6 g/dL.11 Splenectomy is often performed in patients with significant anemia, transfusion burden, and/or massive splenomegaly. The potential benefits and risks of splenec-

tomy must be weighed against those of continued anemia and/or red blood cell transfusions. The risks of splenectomy, including post-splenectomy infections and thrombotic complications, have been extensively reviewed elsewhere.40-43 The rate of venous thrombosis in PKD is 10%, which is similar to the rate in other hemolytic disorders, while the reported rate of post-splenectomy sepsis may be as high as 7%, which is higher than that reported in other cohorts of splenectomized individuals.36 The risk of sepsis is highest in the first year after surgery and in young children but is life-long. The decision to proceed with splenectomy should be delayed until after the age of 5 years, but this timing and the risk of post-splenectomy sepsis must be weighed against the risks of iron loading in patients treated with regular transfusions. Early referral to hematology and surgery for evaluation is advised to assess the pros and cons of splenectomy. The majority of patients with PKD will experience an improvement in hemolysis with both an increased hemoglobin and decreased transfusion burden after splenectomy; however, patients have continued compensated hemolysis with an ongoing risk of severe anemia due to increased hemolysis with infections, aplastic crises associated with parvovirus, and bilirubin-related gallbladder dis-

Figure 4. The glycolytic pathway. Deficiency of pyruvate kinase results in diminished ATP production as well as buildup of pathway intermediates proximal to pyruvate kinase, most notably 2,3-diphosphoglycerate. Modified with permission from Grace and Glader.57 Glucose-6-P: glucose-6-phosphate; Fructose-6-P: fructose-6-phosphate; Fructose 1,6-DP: fructose 1,6-diphosphate; DHAP, dihydroxyacetone phosphate; G3P: glucose-3-phosphate; 1,3-DPG: 1,3-diphosphoglycerate; 2,3-DPG, 2,3diphosphoglycerate; 3-PG: 3-phosphoglycerate; PEP: phosphoenolpyruvate. Blue: enzymes in glycolytic pathway that correlate with the more common glycolytic enzymopathies.

haematologica | 2020; 105(9)

2235

H. Al-Samkari et al.

ease.9,36,44 Approximately 15% of patients with PKD will be regularly transfused despite splenectomy.36 Although response to splenectomy is difficult to predict, those patients who continue to require transfusions despite splenectomy tend to have two drastic (e.g. non-missense) PKLR mutations and a higher pre-splenectomy rate of hemolysis.36 The decision to proceed with splenectomy is more common in US centers, whereas in European centers, more patients with PKD remain on transfusions and iron chelation with intact spleens. This practice variation likely reflects provider and patient preferences and values assigned to the potential risks and benefits of splenectomy. International expert guidelines recommend splenectomy in patients with PKD who receive regular transfusions or do not tolerate their anemia. Cholecystectomy may be recommended at the time of splenectomy given the persistent risk of bilirubin gallstones, but patients should be counseled about the risk of intrahepatic cholestasis and other complications.40 Full splenectomy is recommended given the reported poor hemoglobin response to partial splenectomy.45 Pre- and post-splenectomy vaccinations are critical for preventing sepsis, and antibiotic prophylaxis should be considered according to country-specific guidelines but is often considered lifelong. Post-splenectomy thromboprophylaxis can be considered in the immediate postoperative period in patients with additional thrombotic risk factors. If thromboprophylaxis is given, low-dose aspirin can be considered in adults with additional risk factors (including advanced age, a history of thrombosis, or cigarette smoking) until the platelet count is <500Ă&#x2014;109/L and in children until the platelet count is <1,000Ă&#x2014;109/L. Patients with a mildly low hemoglobin without symptoms of anemia, even with concurrent infections, can clearly avoid splenectomy. However, the management of many patients, like the one in this case, is not straightforward. He has received a considerable number of transfusions, possibly each time he has had a viral infection, and he likely has a high hemolytic rate and low hemoglobin at baseline. He would benefit from a post-splenectomy rise in hemoglobin, which could help him to avoid transfusions with illnesses going forward. Inquiring about agespecific signs and symptoms of fatigue will help to determine whether a chronically low hemoglobin is associated with daily symptoms, an additional reason to support moving forward with full splenectomy and cholecystectomy in this patient. Further complicating this discussion is the potential for future targeted treatments for PKD, including mitapivat (AG-348), an oral small molecule allosteric activator of PK currently in phase III clinical trials in adults. If the patient has at least one missense PKLR mutation, consideration could be given to delaying splenectomy while awaiting potential regulatory approval of this treatment. Off-target mild aromatase inhibition may, however, delay the future availability of this treatment until adulthood.46 Curative treatment with stem cell transplantation has been pursued in a small number of patients with only 16 transplanted patients reported worldwide. These patients had a high rate of grade IV graft-versus-host disease and only a 74% cumulative survival at 2 years; however, they had variable ages at transplant, were given different types of transplants, and were treated with a variety of conditioning regimens.16 Transplant could be considered for the 2236

patient in this case, particularly with an HLA-matched unaffected sibling. In the rare case that transplantation is under serious consideration, splenectomy and its associated infectious and thrombotic risks should be avoided. Given that patients with two drastic PKLR mutations and a higher hemolytic rate are less likely to benefit from splenectomy or mitapivat, consideration should be given to a stem cell transplant and the availability of matched donors prior to splenectomy. However, given the available data regarding the risks of transplant versus supportive care and the treatments in development, including gene therapy, an approach of ongoing supportive care with splenectomy and/or regular red cell transfusions is recommended rather than stem cell transplantation in the majority of patients, including the patient in this case.

Transfusion in pyruvate kinase deficiency CASE: AB (age 38) and BB (age 35) are brothers with PKD diagnosed by enzyme assay and confirmed by genetic testing. Both had splenectomies in childhood, and subsequently had baseline hemoglobin levels of 8-9 g/dL with 20-35% reticulocytes. In early childhood both brothers had similar burdens of fatigue and activity limitation attributable to PKD. In adulthood, the brothers have had divergent paths owing to decisions regarding regular transfusion. AB is relatively sedentary and is transfused once or twice a year, only for worsening symptoms of anemia. He is unemployed. He is concerned about iron overload and refuses regular transfusion with iron chelation due to side effects of chelation therapy in the past. He, therefore, resists red blood cell transfusions until he is clinically more symptomatic. BB, in contrast, is an active, working, married man with four children. In order to maintain this active lifestyle, he is regularly transfused to keep his hemoglobin concentration above 10-11 g/dL. To prevent iron overload with regular transfusions, he takes an oral iron chelation drug. Patients with PKD frequently live with chronic fatigue, poor exercise tolerance, and suboptimal work productivity.47 These symptoms may improve if the hemoglobin concentration is raised consistently with regular transfusions. The decision to transfuse a patient with PKD is based on the patientâ&#x20AC;&#x2122;s tolerance of anemia and demands of their lifestyle rather than an arbitrary hemoglobin threshold. In the above cases, two brothers with the same baseline hemoglobin and symptom level, the decision to regularly transfuse one but not the other was based on the impact of their disease on their quality of life. For the regularlytransfused sibling, red cell transfusions allow him to work, support his family, and live an active lifestyle; by contrast, his brother prioritizes avoiding the side effects of iron chelation and opts for on-demand transfusions only.

The asymptomatic patient with severe anemia CASE: An 8-year-old child with transfusion-dependent congenital anemia of unclear etiology is diagnosed with PKD following enzyme assay and confirmatory genetic testing, which demonstrated the presence of two deleterious non-missense mutations, c.721G>T (p.Glu241X) and c.284-2A>C (abnormal splicing), predicted to result in a complete absence of the PK-R isoform. During the course of his diagnostic workup, his 2,3diphosphoglycerate levels were measured at over three times the upper limit of the reference range. He underwent splenectomy with resolution of transfusion dependency and a post-splenectomy baseline hemoglobin of 7.0 g/dL. Despite continued severe haematologica | 2020; 105(9)

Manifestations of disease in PKD

anemia following splenectomy, the patient excelled at school and extracurricular activities. He has maintained a normal social life into adulthood, married, and had children. As a 40-year old man, he underwent quality of life assessments, including the Functional Assessment of Chronic Illness Therapy Fatigue subscale [FACIT-F, final score of 48 (score range 0-52)] and the Functional Assessment of Cancer Therapy [FACT-G, score of 96 (score range 0-104)], confirming an excellent quality of life. Despite his lack of complaints or need for transfusion, he was offered routine PKD screening (as detailed in Table 1) and was found to have a ferritin of 670 ng/mL. Subsequent liver magnetic resonance imaging demonstrated an average liver iron content of 8.9 mg/g dry weight (reference range, 0.17-1.8 mg/g dry weight). After a discussion of the risks of iron overload he was initiated on chelation therapy with deferasirox. This case illustrates several important points about the spectrum of disease in PKD, most notably the poor correlation between symptoms and severity of anemia. In the case described, the patient’s lack of symptoms despite his low hemoglobin was evident both from his high level of functioning and the results of two well-validated quality of life instruments. Patients may be less symptomatic than with other types of anemia for a given hemoglobin level, or occasionally entirely asymptomatic despite severe anemia.48 This is in part because 2,3-diphosphogylcerate, an important regulator of the oxygen affinity of hemoglobin, is increased in PKD, which may enhance oxygen delivery.17,49 This increase occurs due to the metabolic block in the glycolytic pathway resulting in upstream accumulation of glycolytic intermediates (Figure 4). The resulting P50 and other patient-related factors, such as older age and concurrent medical problems, play a role in symptom burden. This is important to recognize as unnecessary transfusion of patients who are clinically well can be harmful due to iron loading. The 2,3-diphosphogylcerate level obtained during the workup of the patient’s anemia prior to the diagnosis of PKD is biochemical evidence consistent with what is observed clinically regarding his excellent tolerance of anemia. Although correlations have been demonstrated between the presence of two non-missense mutations and both more severe disease and worse health-related quality of life,11 genotype may poorly predict phenotype in some patients with PKD.50 The clinical manifestations of the disease are caused by the complex interactions of the PKLR genetic background, concomitant functional polymorphisms of other enzymes, posttranslational or epigenetic modifications, ineffective erythropoiesis and differences in splenic function.50 Varying degrees of compensatory expression of the PK isozyme normally expressed in leukocytes, PK-M2, in PK-R-deficient erythrocytes appears to influence the clinical severity of PKD as well.51 Similar cases describing relatively asymptomatic PKD patients with even lower hemoglobin levels have been published,52 and indeed a better-than-expected tolerance of anemia in many patients likely explains the not uncommon delay observed in time to diagnosis.53 Further studies are needed to characterize the relationship between a given patient’s 2,3-diphosphogylcerate levels, P50 oxygen dissociation curve, clinical phenotype, genotype, and patient-reported quality of life in PKD and in other congenital hemolytic anemias. Lastly, this case highlights the occult iron overload common in PKD. Iron overload regularly occurs even in nevertransfused patients over a lifetime of chronic hemolysis, haematologica | 2020; 105(9)

necessitating regular evaluation of iron status in all PKD patients.15 In the completely asymptomatic patient who may only visit the clinic annually, assessment of iron status and institution of appropriate chelation therapy is easy to overlook, potentially leading to development of endorgan damage.

Special situations Pregnancy in pyruvate kinase deficiency CASE: A 24-year old woman (gravida 1, parity 0) with a history of iron-deficiency anemia due to menorrhagia (hemoglobin 9.5-11.0 g/dL) and a clinical diagnosis of Gilbert syndrome presents in the 12th week of pregnancy with anemia and profound fatigue. On evaluation, physical examination is remarkable only for mild pallor and jaundice and a palpable spleen tip. Laboratory testing demonstrates hemoglobin 8.3 g/dL, mean corpuscular volume 99 fL, absolute reticulocyte count 225x109/L, haptoglobin <20 mg/dL (reference range 30-200 mg/dL), lactate dehydrogenase 267 U/L (reference range 135-214 U/L), negative direct antiglobulin testing, and unremarkable red cell morphology on a peripheral blood film. Pyruvate kinase enzyme activity was reduced (3.9 U/g Hb [11-15.6]) and genetic testing revealed the presence of two missense mutations: c.1456C>T (p.Arg486Trp) and c.994G>A (p.Gly332Ser). Other congenital and acquired hemolytic conditions were ruled out by normal osmotic fragility, eosin-5-maleimide binding, ektacytometry, erythrocyte membrane protein content, hemoglobin electrophoresis, and CD55/59 antigen expression. During pregnancy the patient was closely followed by hematology and obstetrics staff. Because of severe fatigue, she was given a total of 15 units of packed erythrocytes during pregnancy (on average, a transfusion every 2 weeks) and four additional units at the time of delivery because of postpartum bleeding complications. Following delivery of a preterm but healthy infant, the mother’s anemia improved to a hemoglobin of 9.5 g/dL. Pregnancy is among the known triggers of hemolysis in PKD, along with infections and erythropoietic stressors. In addition to increased hemolysis, hydremia of pregnancy may further exacerbate anemia in pregnant women. Little is known about fertility and pregnancy in PKD, except for a tendency towards an increased frequency of preterm births and miscarriages, and an increased transfusion need.54,55 Folate supplementation is mandatory and we advise a minimum of 1 mg daily during pregnancy. Fetal growth should be carefully monitored. 2,3-diphosphogylcerate levels result in a rightward shift of the oxygen dissociation curve of hemoglobin and symptoms may not correlate well with severity of anemia. As in non-pregnant patients, the decision to transfuse is guided by the patient’s symptoms, rather than hemoglobin levels, and by fetal growth as measured by ultrasound. Given that splenectomy partially ameliorates the anemia in PKD, anticipation of pregnancy in a young woman with PKD may be considered as a factor in favor of splenectomy prior to pregnancy.40

Stigmatizing Icterus in pyruvate kinase deficiency CASE: A 14-year old boy with PKD who has been splenectomized arrives at the clinic for routine follow-up. He is able to be physically active to his satisfaction despite his anemia (baseline hemoglobin 9.0 g/dL). His primary complaint during this visit, as was the case in his past several visits, is severe jaundice of the 2237

H. Al-Samkari et al.

skin and scleral icterus. He has been the victim of considerable bullying by other children in his new school over the past year due to the yellow tinge of his skin and eyes. His mother believes it is making him depressed and his grades have fallen over the same period. His total bilirubin is 3.9 mg/dL. He is initiated on phenobarbital 15 mg daily, resulting in a reduction of his bilirubin to 2.0 mg/dL and significant improvement in the icterus. At his next visit, the patient remarks that the bullying has stopped. Patients with congenital hemolytic anemia often live with chronic icterus. This is not uncommonly a significant psychological burden and social stigma for patients with PKD with a considerable impact on quality of life. Schoolaged children and adolescents may face bullying and other negative social consequences. It is important for clinicians to recognize this and address the impact on their patients’ quality of life. Depending upon the degree of impact on a patient’s well-being, treatment may be indicated to address icterus. Low-dose, off-label phenobarbital (15-30 mg daily) could be considered as this agent induces hepatic bilirubin metabolism, although the potential side effects of this medication should be carefully explained to the patient and family.56 In especially severe cases, a regular exchange transfusion regimen can be considered to reduce bilirubin levels and alleviate icterus, although this approach was not uniformly agreed upon by all members of the working group.

drops precipitously and transfusion is frequently required. Similarly, patients can develop aplastic crises secondary to parvovirus B19, which should be suspected in the setting of a hemoglobin drop with a significant reduction in the normally robust reticulocytosis. This possibility can be evaluated with parvovirus B19 serology (including IgM testing) or polymerase chain reaction testing. The described scenario of a patient who is liberated from transfusions following splenectomy in childhood but decides to start regular transfusions once again in later adulthood due to anemia-related symptoms demonstrates a common but poorly-studied phenomenon in PKD: a reduced tolerance to the same degree of anemia as patients age. This phenomenon was universally recognized by working group members but has not been formally described in the literature. In this patient, an evaluation for cardiopulmonary disease, such as the pulmonary hypertension that can complicate PKD, is warranted.11,25 A decline in cardiopulmonary function also frequently occurs due to non-PKD-related disease or from age-related factors, which can reduce the patient’s ability to compensate for the degree of anemia. Cases such as this one emphasize that transfusion independence in PKD may come and go throughout the patient’s lifespan.

Conclusions Disease burden over the lifespan CASE: A 54-year old woman with PKD presents for routine follow-up. She was diagnosed in childhood and was transfusiondependent until she underwent splenectomy at the age of 8. This alleviated her need for routine transfusions and she settled at a hemoglobin of 8.0-9.0 g/dL. Between the age of 8 and 49, she required red cell transfusions only five times: in the post-operative setting following cholecystectomy, during pregnancy, and during three acute infectious episodes that resulted in dramatic drops in her hemoglobin and hospitalization. During her late 40s, she noticed slowly worsening exercise tolerance despite no change in her baseline hemoglobin and at the age of 49 began receiving intermittent transfusions to treat fatigue. By the age of 52, she required regular transfusion of two units of red cells every 8 weeks to raise her hemoglobin above 10.0 g/dL. Over time, her tolerance of anemia has diminished further, resulting in a gradual reduction of the interval between transfusions. Currently she requires two units of red cells every 6 weeks. She does not smoke and has no underlying cardiopulmonary disease, including no evidence of pulmonary hypertension. Patients with PKD are susceptible to hemolytic crises, typically occurring in the setting of acute viral or bacterial infection.41 Hemolytic crises may also be precipitated by pregnancy, surgery, or other major physiological stressors. During a hemolytic crisis, the hemoglobin concentration

References 1. Beutler E, Gelbart T. Estimating the prevalence of pyruvate kinase deficiency from the gene frequency in the general white population. Blood. 2000;95(11):3585-3588. 2. Carey PJ, Chandler J, Hendrick A, et al. Prevalence of pyruvate kinase deficiency in northern European population in the north of England. Northern Region Haematologists Group. Blood. 2000;96(12):4005-4006.

2238

While the management of PKD may change in the near future with the promise of therapeutic advances on the horizon, recognition and diagnosis of the disease by hematologists and institution of proper monitoring and supportive treatments will remain important. This requires an understanding of the manifestations of disease as they relate to diagnosis, treatment, and impact on the patient’s quality of life. With few distinctive clinical signs to suggest the disease, both a high index of suspicion and understanding of the limitations of diagnostic testing are needed to properly diagnose patients. Recognition that transfusions are symptom-directed and that iron overload and other complications are common regardless of the severity of the anemia is critical for proper care of PKD patients. The spectrum of disease in PKD is broad with symptoms varying between patients and within patients over time with age, supporting an individualized approach to monitoring and treatment. Acknowledgments HA is the recipient of the National Hemophilia FoundationShire Clinical Fellowship Award and the Harvard KL2/Catalyst Medical Research Investigator Training Award and the American Society of Hematology Scholar Award.

3. Machado P, Manco L, Gomes C, et al. Pyruvate kinase deficiency in sub-Saharan Africa: identification of a highly frequent missense mutation (G829A;Glu277Lys) and association with malaria. PLoS One. 2012;7(10):e47071. 4. Ayi K, Min-Oo G, Serghides L, et al. Pyruvate kinase deficiency and malaria. N Engl J Med. 2008;358(17):1805-1810. 5. Selwyn JG, Dacie JV. Autohemolysis and other changes resulting from the incubation

in vitro of red cells from patients with congenital hemolytic anemia. Blood. 1954;9(5): 414-438. 6. Valentine WN, Tanaka KR, Miwa S. A specific erythrocyte glycolytic enzyme defect (pyruvate kinase) in three subjects with congenital non-spherocytic hemolytic anemia. Trans Assoc Am Physicians. 1961;74:100110. 7. Necheles TF, Finkel HE, Sheehan RG, Allen DM. Red cell pyruvate kinase deficiency.

haematologica | 2020; 105(9)

Manifestations of disease in PKD

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

The effect of splenectomy. Arch Intern Med. 1966;118(1):75-78. Mentzer WC Jr., Baehner RL, SchmidtSchonbein H, Robinson SH, Nathan DG. Selective reticulocyte destruction in erythrocyte pyruvate kinase deficiency. J Clin Invest. 1971;50(3):688-699. Nathan DG, Oski FA, Miller DR, Gardner FH. Life-span and organ sequestration of the red cells in pyruvate kinase deficiency. N Engl J Med. 1968;278(2):73-81. Matsumoto N, Ishihara T, Nakashima K, Miwa S, Uchino F. Sequestration and destruction of reticulocyte in the spleen in pyruvate kinase deficiency hereditary nonspherocytic hemolytic anemia. Nihon Ketsueki Gakkai Zasshi. 1972;35(4):525-537. Grace RF, Bianchi P, van Beers EJ, et al. Clinical spectrum of pyruvate kinase deficiency: data from the Pyruvate Kinase Deficiency Natural History Study. Blood. 2018;131(20):2183-2192. Ferreira P, Morais L, Costa R, et al. Hydrops fetalis associated with erythrocyte pyruvate kinase deficiency. Eur J Pediatr. 2000;159 (7):481-482. Hennekam RC, Beemer FA, Cats BP, Jansen G, Staal GE. Hydrops fetalis associated with red cell pyruvate kinase deficiency. Genet Couns. 1990;1(1):75-79. van Beers E, Kuo K, Morton D, et al. Health related quality of life and fatigue in patients with pyruvate kinase deficiency. Blood. 2018;132(Suppl 1):4807. van Beers EJ, van Straaten S, Morton DH, et al. Prevalence and management of iron overload in pyruvate kinase deficiency: report from the Pyruvate Kinase Deficiency Natural History Study. Haematologica. 2019;104(2): e51-e53. van Straaten S, Bierings M, Bianchi P, et al. Worldwide study of hematopoietic allogeneic stem cell transplantation in pyruvate kinase deficiency. Haematologica. 2018;103 (2):e82-e86. Grace RF, Zanella A, Neufeld EJ, et al. Erythrocyte pyruvate kinase deficiency: 2015 status report. Am J Hematol. 2015;90(9):825-830. Yang H, Merica E, Chen Y, et al. Phase 1 single- and multiple-ascending-dose randomized studies of the safety, pharmacokinetics, and pharmacodynamics of AG-348, a firstin-class allosteric activator of pyruvate kinase R, in healthy volunteers. Clin Pharmacol Drug Dev. 2019;8(2):246-259. Grace RF, Rose C, Layton DM, et al. Safety and efficacy of mitapivat in pyruvate kinase deficiency. N Engl J Med. 2019;381(10):933944. Garcia-Gomez M, Calabria A, Garcia-Bravo M, et al. Safe and efficient gene therapy for pyruvate kinase deficiency. Mol Ther. 2016;24(7):1187-1198. Thompson AA, Walters MC, Kwiatkowski J, et al. Gene therapy in patients with transfusion-dependent beta-thalassemia. N Engl J Med. 2018;378(16):1479-1493. Yamaguti-Hayakawa GG, Ozelo MC. Gene therapy: paving new roads in the treatment of hemophilia. Semin Thromb Hemost. 2019;45(7):743-750. Bianchi P, Fermo E, Glader B, et al. Addressing the diagnostic gaps in pyruvate kinase deficiency: consensus recommendations on the diagnosis of pyruvate kinase deficiency. Am J Hematol. 2019;94(1):149161. Hipkins R, Thompson J, Naidoo P, Bishay E, Djearaman M, Pratt G. Images in haematol-

haematologica | 2020; 105(9)

25.

26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

38.

39.

ogy. Paravertebral extramedullary haemopoiesis associated with pyruvate kinase deficiency. Br J Haematol. 2009;147 (3):275. Bachmeyer C, Khalil A, Kerrou K, Girot R, Gounant V. Idiopathic pulmonary arterial hypertension in a patient with pyruvate kinase deficiency and paravertebral extramedullary hematopoiesis. Ann Hematol. 2009;88(6):603-605. Plensa E, Tapia G, Junca J, Perez R, Castella E, Marti S. Paravertebral extramedullary hematopoiesis due to pyruvate kinase deficiency. Haematologica. 2005;90(Suppl): ECR32. Goh DH, Lee SH, Cho DC, Park SH, Hwang JH, Sung JK. Chronic idiopathic myelofibrosis presenting as cauda equina compression due to extramedullary hematopoiesis: a case report. J Korean Med Sci. 2007;22(6):10901093. Gouliamos A, Dardoufas C, Papailiou I, Kalovidouris A, Vlahos L, Papavasiliou C. Low back pain due to extramedullary hemopoiesis. Neuroradiology. 1991;33(3):284285. Sawhney H, Weedon J, Gillette P, Solomon W, Braverman A. Predilection of hemolytic anemia-associated leg ulcers for the medial malleolus. Vasa. 2002;31(3):191-193. Aessopos A, Kati M, Tsironi M, Polonifi E, Farmakis D. Exchange blood transfusions for the treatment of leg ulcerations in thalassemia intermedia. Haematologica. 2006; 91(5 Suppl):ECR11. Alavi A, Kirsner RS. Hemoglobinopathies and leg ulcers. Int J Low Extrem Wounds. 2015;14(3):213-216. Jansen G, Koenderman L, Rijksen G, Punt K, Dekker AW, Staal GE. Age dependent behaviour of red cell glycolytic enzymes in haematological disorders. Br J Haematol. 1985;61 (1):51-59. Warang P, Kedar P, Ghosh K, Colah R. Molecular and clinical heterogeneity in pyruvate kinase deficiency in India. Blood Cells Mol Dis. 2013;51(3):133-137. Pissard S, Max-Audit I, Skopinski L, et al. Pyruvate kinase deficiency in France: a 3year study reveals 27 new mutations. Br J Haematol. 2006;133(6):683-689. Manu-Pereira Mdel M, Gonzalez-Roca E, van Solinge WW, et al. Pyruvate kinase deficiency and severe congenital hemolytic anemia in a double heterozygous patient with paternal transmission of an early germ-line de novo mutation. Am J Hematol. 2015;90 (12):E217-219. Grace RF, Bianchi P, van Beers EJ, et al. The clinical spectrum of pyruvate kinase deficiency: data from the Pyruvate Kinase Deficiency Natural History Study. Blood. 2018;131(20): 2183-2192. Bagla S, Bhambhani K, Gadgeel M, Buck S, Jin JP, Ravindranath Y. Compound heterozygosity in PKLR gene for a previously unrecognized intronic polymorphism and a rare missense mutation as a novel cause of severe pyruvate kinase deficiency. Haematologica. 2019;104(9):e428-e431. Lezon-Geyda K, Rose MJ, McNaull MA, et al. Pklr intron splicing-associated mutations and alternate diagnoses are common in pyruvate kinase deficient patients with single or no Pklr coding mutations. Blood. 2018;132(Suppl 1):3607. van Oirschot BA, Francois JJ, van Solinge WW, et al. Novel type of red blood cell pyruvate kinase hyperactivity predicts a remote regulatory locus involved in PKLR gene

40.

41.

42.

43.

44.

45.

46.

47.

48. 49.

50.

51.

52.

53.

54.

55.

56. 57.

expression. Am J Hematol. 2014;89(4):380384. Iolascon A, Andolfo I, Barcellini W, et al. Recommendations regarding splenectomy in hereditary hemolytic anemias. Haematologica. 2017;102(8):1304-1313. Grace RF, Mark Layton D, Barcellini W. How we manage patients with pyruvate kinase deficiency. Br J Haematol. 2019;184(5):721-734. Kristinsson SY, Gridley G, Hoover RN, Check D, Landgren O. Long-term risks after splenectomy among 8,149 cancer-free American veterans: a cohort study with up to 27 years follow-up. Haematologica. 2014;99(2):392-398. Rodeghiero F, Ruggeri M. Short- and longterm risks of splenectomy for benign haematological disorders: should we revisit the indications? Br J Haematol. 2012;158(1):16-29. Zanella A, Fermo E, Bianchi P, Valentini G. Red cell pyruvate kinase deficiency: molecular and clinical aspects. Br J Haematol. 2005;130(1):11-25. Sandoval C, Stringel G, Weisberger J, Jayabose S. Failure of partial splenectomy to ameliorate the anemia of pyruvate kinase deficiency. J Pediatr Surg. 1997;32(4):641-642. Kung C, Hixon J, Kosinski PA, et al. AG-348 enhances pyruvate kinase activity in red blood cells from patients with pyruvate kinase deficiency. Blood. 2017;130(11):13471356. Grace RF, Cohen J, Egan S, et al. The burden of disease in pyruvate kinase deficiency: patients' perception of the impact on healthrelated quality of life. Eur J Haematol. 2018;101(6):758-765. Oski FA. Clinical consequences of enzyme deficiencies in the erythrocyte. Ann Clin Lab Sci. 1971;1(2):177-183. Lakomek M, Winkler H, Pekrun A, et al. Erythrocyte pyruvate kinase deficiency. The influence of physiologically important metabolites on the function of normal and defective enzymes. Enzyme Protein. 1994;48(3):149-163. Zanella A, Fermo E, Bianchi P, Chiarelli LR, Valentini G. Pyruvate kinase deficiency: the genotype-phenotype association. Blood Rev. 2007;21(4):217-231. Diez A, Gilsanz F, Martinez J, PerezBenavente S, Meza NW, Bautista JM. Lifethreatening nonspherocytic hemolytic anemia in a patient with a null mutation in the PKLR gene and no compensatory PKM gene expression. Blood. 2005;106(5):1851-1856. Aksu T, Yarali N, Fermo E, et al. A case with pyruvate kinase deficiency remarkably sensitive to heat. J Pediatr Hematol Oncol. 2018;40(7):e458-e460. Aydin Koker S, Oymak Y, Bianchi P, et al. A new variant of PKLR gene associated with mild hemolysis may be responsible for the misdiagnosis in pyruvate kinase deficiency. J Pediatr Hematol Oncol. 2019;41(1):e1-e2. Amankwah KS, Dick BW, Dodge S. Hemolytic anemia and pyruvate kinase deficiency in pregnancy. Obstet Gynecol. 1980;55(3 Suppl):42S-44S. Wax JR, Pinette MG, Cartin A, Blackstone J. Pyruvate kinase deficiency complicating pregnancy. Obstet Gynecol. 2007;109(2 Pt2):553-555. Capron JP, Erlinger S. Barbituates and biliary function. Digestion. 1975;12(1):43-56. Grace RF, Glader B. Red blood cell enzyme disorders. Pediatr Clin North Am. 2018;65 (3):579-595.

2239

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2240-2249

Red Cell Biology & its Disorders

XPO1 regulates erythroid differentiation and is a new target for the treatment of β-thalassemia

Flavia Guillem,1,2,3 Michaël Dussiot,1,2,3* Elia Colin,1,2,3* Thunwarat Suriyun,1,2,3 Jean Benoit Arlet,1,2,3,4 Nicolas Goudin,5 Guillaume Marcion,6,7 Renaud Seigneuric,6,7 Sebastien Causse,6,7 Patrick Gonin,8 Marc Gastou,3,8,9 Marc Deloger,10 Julien Rossignol,1,11,12 Mathilde Lamarque,1,2,3 Zakia Belaid Choucair,1,2 Emilie Fleur Gautier,3,13 Sarah Ducamp,3,13 Julie Vandekerckhove,1 Ivan C. Moura,1,2,3† Thiago Trovati Maciel,1,2,3 Carmen Garrido,6,7,14 Xiuli An,15 Patrick Mayeux,3,13 Narla Mohandas,15 Geneviève Courtois1,2,3# and Olivier Hermine1,2,3,11#

INSERM UMR 1163, CNRS ERL 8254, Laboratory of Cellular and Molecular Mechanisms of Hematological Disorders and Therapeutical Implications, Paris, France; 2 Imagine Institute, Université Paris Descartes, Sorbonne Paris-Cité et Assistance Publique-Hôpitaux de Paris, Hôpital Necker, Paris, France; 3Laboratory of Excellence GRex, Paris, France; 4Service de Médecine Interne, Faculté de Médecine Paris Descartes, Sorbonne Paris-Cité et Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France; 5US24, Cell Imaging Platform, Necker Federative Structure of Research (SFR-Necker), Paris, France; 6INSERM, Unité Mixte de Recherche 866, Equipe Labellisée Ligue Contre le Cancer and Association pour la Recherche contre le Cancer, and Laboratoire d’Excellence Lipoprotéines et Santé (LipSTIC), Dijon, France; 7Faculty of Medicine and Pharmacy, University of Burgundy, Dijon, France; 8 Gustave Roussy, Université Paris-Saclay, Plateforme d'Evaluation Préclinique-UMS 3655/US23, Villejuif, France; 9Université Paris 7 Denis Diderot-Sorbonne Paris Cité, Paris, France; 10Institut Curie, PSL Research University, INSERM, U 900, MINES, ParisTech, Paris, France; 11Service d’Hématologie, Faculté de Médecine Paris Descartes, Sorbonne Paris-Cité et Assistance Publique-Hôpitaux de Paris Hôpital Necker, Paris, France; 12Département d'Hématologie, Gustave Roussy, Université Paris-Saclay, Villejuif, France; 13Institut Cochin, INSERM U1016, CNRS UMR8104, Université Paris Descartes, and Plateforme de Proteomique Paris 5 (3P5), Paris, France; 14Centre Anticancéreux George François Leclerc, Dijon, France and 15Red Cell Physiology Laboratory, New York Blood Center, New York, NY, USA 1

*MD and EC contributed equally to this work; #GC and OH contributed equally to this work as co-senior authors. †This article is dedicated to the memory of Ivan C. Moura who passed away during its preparation.

Correspondence: OLIVIER HERMINE ohermine@gmail.com Received: November 22, 2018. Accepted: November 19, 2019. Pre-published: November 21, 2019. doi:10.3324/haematol.2018.210054 ©2020 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.

2240

ABSTRACT

-thalassemia major (β-TM) is an inherited hemoglobinopathy caused by a quantitative defect in the synthesis of β-globin chains of hemoglobin, leading to the accumulation of free a-globin chains that aggregate and cause ineffective erythropoiesis. We have previously demonstrated that terminal erythroid maturation requires a transient activation of caspase-3 and that the chaperone Heat Shock Protein 70 (HSP70) accumulates in the nucleus to protect GATA-1 transcription factor from caspase-3 cleavage. This nuclear accumulation of HSP70 is inhibited in human β-TM erythroblasts due to HSP70 sequestration in the cytoplasm by free a-globin chains, resulting in maturation arrest and apoptosis. Likewise, terminal maturation can be restored by transduction of a nuclear-targeted HSP70 mutant. Here we demonstrate that in normal erythroid progenitors, HSP70 localization is regulated by the exportin-1 (XPO1), and that treatment of β-thalassemic erythroblasts with an XPO1 inhibitor increased the amount of nuclear HSP70, rescued GATA-1 expression and improved terminal differentiation, thus representing a new therapeutic option to ameliorate ineffective erythropoiesis of β-TM. haematologica | 2020; 105(9)

XPO1 is a target to treat β-thalassemia

Introduction Erythropoiesis is a multistep process leading to red cell production from hematopoietic stem cells. We have demonstrated that entry into the terminal maturation stage of erythroblasts, marked by important morphological modifications, requires a transient activation of caspase-3.1 However, in this process, some caspase-3 targets including GATA1, the transcription factor involved in the expression of erythroid genes like erythropoietin (EPO) receptor, Glycophorin A or globin chains, remains uncleaved. The chaperone Heat Shock Protein 70 (HSP70), which is constitutively expressed during erythroid differentiation, accumulates in the nucleus and protects GATA-1 transcription factor from caspase-3 cleavage, allowing continued terminal maturation.2 We have demonstrated that this nuclear accumulation of HSP70 did not occur in human β-TM erythroblasts, resulting in GATA1 cleavage and thus erythroblast maturation arrest and apoptosis. Indeed, in β-TM, HSP70 is sequestrated in the cytoplasm by the excess of free a-globin chains preventing its nuclear localization.3 Likewise, terminal maturation in human β-TM could be restored by transduction of a nuclear-targeted HSP70 mutant. Thus, any regulation enabling an increase in HSP70 nuclear concentration would be an interesting therapeutic strategy to correct ineffective erythropoiesis of β-TM. However, mechanisms involved in HSP70 trafficking during human erythroid differentiation are still unknown. Thus, we focused our work on elucidating those mechanisms. We have previously shown by both confocal microscopy and immunoblot analyses that, in erythroblasts derived from CD36+ cells, EPO starvation induced a dramatic decrease in both nuclear localization of HSP70 and as a consequence GATA-1 expression. Addition of the nuclear export inhibitor Leptomycin B (LMB)4,5 prevented this phenomenon.2 These previous observations strongly suggested that HSP70 localization could be regulated mainly by its export. Exportins are proteins of the β-karyopherin group that mediate export from the nucleus to the cytoplasm of cellular proteins (cargos) or mRNAs, using the RanGTP/GDP gradient. There are seven known exportins expressed in human cells (XPO1 to XPO7) and all show preferential cargo specificity.6 It has been recently reported that the erythroid-specific isoform of exportin-7 (XPO7) was the most abundant exportin expressed at the mRNA level in very late erythropoiesis, and was involved in chromatin condensation and enucleation in murine erythroid differentiation.7 However, since XPO7 is expressed and plays a role at the later stage of erythroid differentiation, it might not be considered as a good candidate to be an exporter of HSP70. In order to decipher which exportin(s) was/were involved in HSP70 nuclear export in erythroid progenitor, we have analyzed the expression profile of the seven different human exportins from transcriptomic and proteomic data from erythroid cells at different differentiation stages of maturation, and showed functionally that XPO1 is involved in HSP70 export and could be considered as a good target to ameliorate ineffective erythropoiesis of β-TM.

from cord blood from healthy donors. This study was performed according to the Declaration of Helsinki with the approval from the ethics committee of our institution [Comité de Protection des Personnes (CPP) “Ile de France II”]. All patients gave written informed consent. In the first step of culture (“cell expansion”), isolated CD34+ progenitors (Miltenyi CD34 Progenitor Cell Isolation Kit) were grown in the presence of 100 ng/mL IL-6, 10 ng/mL IL3, and 100 ng/mL SCF for 7 days. At day 7, erythroid progenitors were switched to the second phase of culture, which allows the differentiation and maturation of erythroblasts: cells were cultured in the presence of 10 ng/mL IL-3, 50 ng/mL SCF, and 2 U/mL EPO in IMDM (Gibco cell culture) supplemented with 15% BIT 9500 (Stem Cell Technologies), as previously described. Solid KPT-251 was reconstituted in DMSO. Cells were treated with KPT-251 at 100 or 1000 nM, or with DMSO alone as control condition.

Cell differentiation Erythroid differentiation was assessed by various methods. First, morphological analysis after May-Grunwald-Giemsa (MGG) staining was used. Cells were examined under Zeiss microscope axioplan 2, Camera Qimaging QICAM FAST 1394. The number of immature (proerythroblasts + basophilic), polychromatophilic, and mature (orthochromatic erythroblasts + reticulocytes) erythroblasts was assessed in each experiment by counting approximately 200 cells in consecutive fields and expressed as a percentage of total cells. Additionally, differentiation was assessed by calculating a terminal maturation index (TMI) on MGG, defined by the number of orthochromatic erythroblasts + reticulocytes*100/ number of polychromatophilic cells per slide. This allowed us to better characterize the maturation arrest at the polychromatophilic stage, which is known to be a hallmark of β-TM ineffective erythropoiesis,8 and its modulation. Flow cytometry analysis was also performed at several different times during erythroid culture after double labeling with Band3 and 4-integrin, two optimal surface markers to differentiate highly mature erythroblasts.9 PE-conjugated anti-Band3 (PE-BRIC6 conjugate, Bristol Institute for Transfusion Sciences, UK) and APC-conjugated anti- 4-integrin (Miltenyi Biotec) antibodies were used for flow cytometry (FACS Gallios). This double labeling allowed us to assess terminal erythroblastic differentiation and purify cell populations by cell sorting (BD FACS Aria II SORP). High Band3 (high band3, low a4-integrin), and low Band3 (low band3, high 4-integrin) gates were defined, and cells from each gate were sorted and analyzed morphologically after MGG staining.

Statistical analysis Statistical analyses were performed with GraphPadPrism (Version 4.0 GraphPad software). Data are expressed as the mean±standard deviation or standard error of the mean. Paired t-test, or Anova Dunn’s, Dunnett’s, and Tukey’s multiple comparison tests were used as appropriate. P<0.05 was considered statistically significant; lower P-values are indicated in the figure legends. Further information is available in the Online Supplementary Appendix.

Results Methods Erythroid liquid culture and KPT-251 treatment

XPO1 can interact with HSP70 nuclear export sequence

Erythroid cells were generated in vitro from peripheral blood circulating CD34+ cells from adult patients with β0-thalassemia major (β-TM), which were collected before routine transfusion or

In our previous work,2 we demonstrated that HSP70 export mediated by EPO deprivation could be repressed by treatment with the nuclear export inhibitor LMB.

haematologica | 2020; 105(9)

2241

F. Guillem et al. A

Figure 1 (previous page). In human erythroid progenitors, HSP70 is exported from the nucleus by an XPO1-dependent mechanism. Expression profiles for the seven different exportins (XPO1-XPO7) during terminal erythroid differentiation in human: mRNA expression from proerythroblast stage (ProE) to orthochromatic stage (Ortho). Values are extracted from public data and presented as log of reads per kilobase of transcript per million reads (logRPKM); EB: early basophile; LB: late basophile; Poly: polychromatophilic, (A) and protein expression from progenitor stage (ProG) to orthochromatic stage (Ortho). Values are presented as mean of protein copies per cell. ProG1: BFU-E; ProG2: CFUE; Baso1: early basophile; Baso2: late basophile; Poly: polychromatophilic (B). Data are representative of three independent experiments. (C) Putative XPO1 specific leucin-rich NES in the protein sequence of human HSP70 (NP_005336) at position L394-L403. The interactions between purified XPO1 and WT HSP70 as well as XPO1 with the nucleartargeted HSP70 mutant (S400A) were analyzed using BLI. WT HSP70 exhibits a much higher signal (i.e. affinity) for the ligand XPO1 compared to the nuclear HSP70 protein bearing a mutation in the NES residue S400A. Data are representative of two independent experiments. (D) Proximity of HSP70 and XPO1 proteins was analyzed in CD36+ erythroid progenitors derived from cord blood, by Duolink assay, using anti-XPO1 and anti-HSP70 antibodies (or anti-GATA1 for negative control). Red spots indicate <40 nm proximity between cellular-bound antibodies. Nuclei are stained with DAPI (blue). Images have been observed by confocal microscopy (x63 oil objective, scale bar= 5 μm). Data are representative of three independent experiments. (E) HSP70 and XPO1 direct interaction was demonstrated by CoIP experiments. HSP70 and XPO1 immunoblot detection is shown in total lysate (TL), in eluate from HSP70 IP and from IgG Control (IgG CTL) IP. The data are representative of three independent experiments in human erythroïd cells. (F) Erythroid progenitors from β-thalassemia major (β-TM) patient at day 2 of CD36+ culture were transduced with a shRNA specific for XPO1 or a sh scramble (shCTL). Both constructions express GFP. GFP+ cells were sorted and stained with anti-HSP70 or anti-XPO1 antibodies, and DAPI. XPO1 and HSP70 nuclear expression (mean pixel) were analyzed at day 2 following transduction, by ImageStream. In addition, HSP70 nuclear translocation was evaluated by measuring the similarity score between HSP70 and DAPI nuclear stainings. Data are presented as mean ±standard error of mean (SEM). On average, 30,000 events were collected in all experiments. P-values are determined by paired t-test. ***P<0.0001. Three illustrative images (ImageStream) of shCTL and shXPO1 conditions are presented. Cells were probed for HSP70 expression and run on the ImageStream. Bright field (white), DAPI (purple), HSP70 (green), and HSP70/DAPI composite (scale bar=7 μm). Data are representative of six independent experiments, n=2 different β-TM patients with n=2 different shRNA XPO1.

2242

haematologica | 2020; 105(9)

XPO1 is a target to treat β-thalassemia

Here, we validate these previous findings using confocal microscopy and ImageStream analysis for a more quantitative approach. In normal culture conditions of cord blood-derived CD36+ erythroid progenitors at day 3, a 2hour treatment with LMB at 20nM induced an increase in HSP70 nuclear accumulation (Online Supplementary Figure S1A and B) resulting in an increase in HSP70 nuclear/cytoplasmic ratio. In order to determine which specific exportin was involved in HSP70 export, we first analyzed expression of the seven exportins during erythropoiesis. All of them were expressed at the mRNA and protein levels during human erythroid differentiation as shown by transcriptomic and proteomic analysis10,11 (Figure 1A and B). Based on the proteomic database (Figure 1B), we found that exportin-2 (XPO2) is the most expressed exportin at the protein level in human erythroid progenitors. However, since primary function of XPO2 is to mediate re-export of importin-a from the nucleus to the cytoplasm, it is unlikely to be the candidate exportin involved in HSP70 nuclear export. In contrast, XPO1, in addition to exporting RNAs, mediates export of a broad range of cargo proteins bearing a leucine-rich Nuclear Export Sequence (NES),12,13 that include a large variety of tumor suppressor proteins (e.g. p53, p21, FOXO) and thereby fulfills all the criteria to be a good candidate in exporting HSP70. We performed in silico analysis of published data for GATA-1 chromatin immuno-precipitation (ChIP) in

Table 1. GATA-1 chromatin immunoprecipitation (ChIP) peak scores in XPO1 gene and three known erythroid genes.

Gene name

Genomic peak location

Peak score

XPO1

Intragenic intron Intragenic intron Intragenic intron Intragenic intron Intragenic intron Immediate down. Intron Promoter Enhancer Intergenic Intergenic Intergenic Intergenic Intragenic exon Immediate down. Intron Enhancer Enhancer Enhancer Promoter Promoter Enhancer Enhancer Promoter

53 695.35 76.79 94.13 231.83 62.86 292.11 110.24 66.87 288.01 73.56 89.17 101.5 3100 72.94 186.49 55.94 898.4 105.72 1259.62 313.4 55.51

EPOR

HBA1

HBA2

Annotated peaks extracted from GATA-1 ChIP public data on human erythroid progenitors14 analyzed with Galaxy software. Peak scores and locations for XPO1 and the known erythroid genes EPOR (erythropoietin receptor), HBA1 (Hemoglobin a-1) and HBA2 (Hemoglobin a-2). Peak score in MACS output corresponds to -10*log10pvalue of peak region. P-value corresponds to the probability that the candidate peak (reads enrichment) is random.

haematologica | 2020; 105(9)

human erythroid cells,14 and found twelve annotated peaks for GATA-1 in the promoter of the XPO1 gene suggesting that it might be an erythroid regulated gene (Table 1). In addition, XPO1 was the second most expressed exportin in erythroblasts, and concomitantly with HSP70 nuclear accumulation, its expression was down-regulated both at the mRNA and protein levels along terminal differentiation (black line in Figure 1A and B). This hypothesis was further supported by in silico analysis showing that human HSP70 protein sequence contains a putative leucine-rich NES at position L394L400 that could interact with XPO1 (Figure 1C, NES human HSP70 protein WT). In order to validate this hypothesis, we first showed that XPO1 interacted with HSP70 in vitro by protein-protein interaction experiments using BioLayer interferometry (BLI) (Figure 1C, red line). Furthermore, we observed a decreased binding affinity between XPO1 and HSP70 with the mutant Serine 400 Alanine (HSP70S400A) in the putative leucine-rich NES (Figure 1C, blue line). This mutant HSP70 shows restricted nuclear localization when transduced in erythroid cells.3 Thus, these results validate the importance of putative NES in the interaction of HSP70 with XPO1, and in the cellular localization of HSP70. Next, using proximity ligation assay (PLA/Duolink assay) (Figure 1D) and performing CoIP experiments in normal conditions (Figure 1E), we obtained additional evidence that this interaction is direct and occurs in vivo in CD36+ derived erythroid progenitors. Interestingly, the fraction of XPO1 protein immunoprecipitating with HSP70 upon EPO deprivation is increased as shown by CoIP experiments (data not shown).

XPO1 is the exporter of HSP70 in human erythroid progenitors To document the functional role of XPO1 in human erythropoiesis, and its putative target to treat β-TM, we first repressed XPO1 expression in human β-TM erythroblasts. For this purpose, CD36+ cells from a β-TM patient were transduced with lentiviruses expressing a XPO1 specific shRNA (shXPO1) or a sh scramble used as control (shCTL). Two days after transduction, ImageStream quantification showed that the cells transduced with shXPO1 efficiently repressed the expression of XPO1 [mean pixel nuclear XPO1 536.4±3.517 (shCTL) vs. 520.0±2.767 (shXPO1) (P<0.0001] (Figure 1F and Online Supplementary Figure 2A). Similar results were obtained in erythroblasts from cord blood and peripheral blood stem cells (data not shown). In agreement with our hypothesis, XPO1 repression was associated with a significant increase in HSP70 nuclear translocation [similarity score 1.090±0.012 for shCTL vs. 1.378±0.011 for shXPO1 (P<0.0001)] and in HSP70 nuclear accumulation [mean pixel nuclear HSP70 139.1±1.176 for shCTL vs. 159.2±1.173 for shXPO1 (P<0.0001), corresponding to a 14.5% increase] (Figure 1F). Similar results were obtained using two different shRNA specifically targeting XPO1. As a control of efficacy, P53 that is a well-known target of XPO1, was also significantly increased in the nucleus (Online Supplementary Figure 2B and C). Taken together, these data suggest that in human erythroid cells, HSP70 interacts with XPO1 through a leucine-rich NES region, which is required for its nuclear export and may thus be a good target to improve ineffective erythropoiesis of β-TM. 2243

F. Guillem et al. A

Figure 2. KPT-251 treatment has low effect on cell proliferation and cell death. Cell death and proliferation curves analysis of β-thalassemia major (β-TM) (A and B) and cord blood (C and D) erythroid progenitors, assessed by blue trypan staining at 24, 48 and 72 hours (H) of treatment with KPT100nM, KPT1000nM, or DMSO (control). Daily mean percentage±standard deviation (SD) of dead cells (n=5 independent experiments for β-TM and n=3 independent experiments for cord blood). Daily mean±standard deviation of cell proliferation (n=6 independent experiments for β-TM and n=3 independent experiments for cord blood). P-values are determined by ANOVA Dunnett’s multiple comparison test **P<0.01, NS: not significant.

Chemical inhibition of XPO1 using KPT-251 treatment increases the amount of nuclear HSP70 and GATA-1 in β-TM erythroid progenitors

We repressed XPO1 activity using KPT-251 (Merck Millipore), a Selective Inhibitor of Nuclear Export (SINE) that specifically inhibits the formation of XPO1-cargo complex by interacting with the NES binding-groove of XPO1. KPT-251 treatment induces minimal toxicity in non-cancerous hematopoietic cells both in vitro and in vivo,15-19 with almost no effect on cell survival (Figure 2A and C) and a decrease in cell proliferation but not significant (Figure 2B and D). This is in contrast to LMB treatment, an inhibitor of nuclear export which is highly toxic for erythroid progenitors (data not shown).We exposed erythroid progenitors from β-TM patients, at day 4 of CD36+ day of culture, to KPT-251 at 100nM, 1000nM, or to DMSO (control) for 72 hours (h). As evidenced by immunoblot analyses at day 7 of CD36+ culture, the treatment resulted in a significant dose-dependent decrease in the amount of XPO1 protein compared to control (DMSO) in cytoplasmic extracts (CE) (Figure 3A). XPO1 protein decrease following KPT treatment was also observed in erythroid progenitors derived from cord blood, and significant XPO1 protein decrease can be observed in erythroid progenitors from 24 h of treatment (data not shown). This finding is consistent with previous studies using different primary cancer cells and cell lines.20,21 XPO1 repression by the treatment probably acts through a proteasome-dependent mechanism as reported earlier for KPT-185 and KPT-330, two other members of the KPT family.22,23 In β-TM erythroblasts, western blot 2244

analysis showed the decrease in XPO1 protein amount is associated with a significant dose-dependent increase in HSP70 protein amount in nuclear extracts (NE) (Figure 3A). Nuclear GATA1 amounts and cytoplasmic HSP70 amounts remain unchanged. These results were further confirmed by confocal analyses (Figure 3B); after a 72-h exposure to KPT-251, increases were observed in HSP70 nuclear concentration [mean fluorescence intensity (MFI) nuclear HSP70 48.27±4.6 (control), 58.85±5.2 (KPT 100 nM) (NS) and 79.06±9.0 (KPT 1000nM) (P<0.01)] and in HSP70 nuclear/cytoplasmic (N/C) ratio [MFI ratio 0.24±0.02 (control), 0.35±0.02 (KPT 100nM) (P<0.01) and 0.37±0.03 (KPT 1000nM) (P<0.01)]. Consistent with an increase in nuclear location of HSP70, GATA-1 nuclear concentration was also increased as assessed by confocal analyses [MFI nuclear GATA-1 48.53±5.9 (control), 44.32±2.9 (KPT 100 nM) (NS) and 74.17±2.8 (KPT 1000nM) (P<0.01)] (Figure 3B). Quantification by ImageStream (Figure 3C and D) further confirmed the increase in HSP70 nuclear accumulation by KPT-251 treatment [mean pixel nuclear HSP70 95.22±0.40 (control), 98.27±0.36 (KPT 100nM) (P<0.01) and 108.2±0.38 (KPT 1000nM) (P<0.01)], and of the HSP70 N/C ratio [mean pixel ratio 1.551±0.02 (control), 1.654±0.02 (KPT 100nM) (P<0.01) and 1.647±0.02 (KPT 1000nM) (P<0.01)] and GATA-1 nuclear expression [mean pixel nuclear GATA-1 730.2±2.1 (control), 756.2±2.2 (KPT100nM) (P<0.01) and 816.5±2.1 (KPT 1000nM) (P<0.01)] induced by KPT treatment, in a dose-dependent manner. In addition, as determined by similarity score, the fraction of erythroblasts with a nuclear translocation of HSP70 was increased folhaematologica | 2020; 105(9)

XPO1 is a target to treat β-thalassemia

Figure 3. KPT-251 treatment increases the amount of nuclear HSP70 and GATA-1 in β-thalassemia major (β-TM) erythroid progenitors. Erythroblasts derived from β-TM peripheral blood cells were treated at day 4 of CD36+ cell culture with 100nM, 1000nM of KPT-251, or with DMSO (control) for 72 hours (H). All data were analyzed at day 7 of CD36+cell culture (72 hours of treatment). (A) Immunoblot from 10 μg of nuclear extracts (NE) and 30 μg of cytoplasmic extracts (CE), (representative of three independent experiments performed on two different β-TM patient cell cultures). Graph shows optical relative quantity values of XPO1, HSP70 and GATA1 proteins normalized to that of HSP90 for CE and to that of HDAC2 for NE. Conditions KPT 100nM and 1000nM are normalized to that of DMSO condition. Absence of cytoplasmic proteins contamination in nuclear extracts is evidenced by the absence of HSP90 in NE. (B) Graph shows nuclear mean fluoresence intensity (MFI) of HSP70 and GATA-1, and HSP70 nuclear/cytoplasmic (N/C) ratio of MFI in treated (KPT 100 and 1000) and control (DMSO) cells determined by confocal microscopy images analyses. Data are presented as mean±standard error of mean (SEM) (for a minimum of 30 cells per condition), and are normalized on area. Pvalues are determined by ANOVA Dunnett’s multiple comparison test. Data are representative of three independent experiments. (C) HSP70 and GATA-1 nuclear expression (mean pixel), HSP70 N/C ratio (mean pixel) and HSP70 nuclear translocation (similarity score) were analyzed by ImageStream. Data are presented in histograms as mean±SEM. P-values are determined by ANOVA Dunnett’s multiple comparison test (representative of three independent experiments). On average, 30,000 events were collected in all experiments. (D) Three illustrating images of ImageStream experiments. Cells were probed for HSP70 and GATA-1 expression and run on the ImageStream. Bright field (white), HSP70 (green), GATA-1 (red), DAPI (purple) and HSP70/DAPI composite (scale bar=7μm). Respective similarity score±SEM are indicated under each group of images. *P<0.05; **P<0.01.

haematologica | 2020; 105(9)

2245

F. Guillem et al.

Figure 4. KPT-251 treatment improves terminal erythroid maturation in β-thalassemia major (β-TM) erythroid progenitors in vitro. β-TM erythroid progenitors were treated at day 4 of CD36+ cell culture with 100nM, 1000nM of KPT-251, or with DMSO (control) for 72 hours (day 7 CD36+ cell culture). (A) The mean fluorescence intensity (MFI) of Band 3 was analyzed by flow cytometry after 72 hours of treatment. MFI were normalized on DMSO condition. P-values are determined by ANOVA Dunnett’s multiple comparison test **P<0.01, NS: not significant (n=8 independent experiments, n=3 different β-TM samples). (B) (Left) Percentage of high Band3 cell population under the different treatment conditions and (Right) absolute number of high Band3 cells, normalized to DMSO treatment. P-values are determined by ANOVA Dunn’s multiple comparison test, **P<0.01, NS: not significant, n=8 independent experiments, n=3 β-TM patients. (C) Representative flow cytometry plots (a4-integrin and Band3 staining) of β-TM erythroid progenitors treated with KPT1000nM or DMSO. Strategy for cell sorting purification of high Band3 (red box) and low Band3 (blue box) erythroblasts populations after 72 hours (H) of KPT1000nM or DMSO treatment (day 7 CD36+ cell culture). A representative morphological analysis (x25 oil objective, scale bar= 10 μm) of purified cells from each gate by May-Grünwald-Giemsa staining. Corresponding graph showing the percentage of mature cells (orthochromatic erythroblasts + reticulocytes) contained in low Band3 and in high Band3 gates (n=3 independent experiments, n=2 different β-TM patients). P-values are determined by paired t-test ***P<0.001. (D) Proportions (%) of immature, polychromatophilic (PolyC), and mature (orthochromatic erythroblasts + reticulocytes) cells after 72 hours of treatment with KPT100nM, KPT 1000nM or DMSO. NS: not significant. (E) Corresponding TM index for the different conditions of treatment. P values are determined by ANOVA Dunnett’s multiple comparison test *P<0.05, NS: not significant (n=5 independent experiments).

2246

haematologica | 2020; 105(9)

XPO1 is a target to treat β-thalassemia

Figure 5. Schematic illustration of the molecular mechanisms modulated by KPT treatment in β-thalassemia major (β-TM) erythroid progenitors compared to β-TM and normal erythroid progenitors in normal conditions. Schematic representation of molecular mechanisms in normal erythroid progenitor (EP), β-TM EP, and β-TM EP treated with KPT (cells on the right). The big arrow on the top represents the direction of differentiation progression. The decrease in XPO1 protein expression is represented by the pink triangle. During early differentiation stages (cell on the left), XPO1 exports HSP70 from the nucleus to the cytoplasm, while XPO1 expression is high. The entry of HSP70 being constant and mediated by Hikeshi, HSP70 is localized both in the cytoplasm and the nucleus at this stage. In normal EP, along differentiation, while XPO1 expression decreases, HSP70 accumulates in the nucleus until caspase-3 activates, corresponding to basophilic stage. Nuclear HSP70 protects GATA1 from caspase-3 cleavage to enable terminal maturation. In β-TM EP, at the stage of caspase-3 activation, HSP70 is trapped in the cytoplasm by the excess of free a-globin chains and can not protect GATA1 from cleavage. This results in maturation arrest at the polychromatophilic stage. In β-TM EP treated with KPT, XPO1 activity is repressed. This allows nuclear retention of the small amount of HSP70 that managed to get into the nucleus despite cytoplasm trapping by α chains. At the moment of caspase-3 activation, HSP70 is present in sufficient amount to protect GATA1 and enable an improvement in β-TM EP terminal maturation.

lowing KPT-251 treatment [similarity score 0.7483±0.006 (control), 0.8587±0.005 (KPT100nM) (P<0.01) and 0.9872±0.005 (KPT1000nM) (P<0.01)]. To further demonstrate that XPO1 is indeed the main protein involved in HSP70 nuclear export, we tested its role in HeLa cells, which upon heat shock at 43°C exhibited nuclear HSP70 localization as a consequence of both an increase in HSP70 nuclear inflow rate due to an increase of Hikeshi expression and a reduction in nuclear outflow by an unknown mechanism.24 After a 6-h recovery phase at 37°C, the outflow rate increases and HSP70 progressively re-localizes in the cytoplasm.24 To demonstrate the role of XPO1 in the nuclear export of HSP70, HSP70 outflow following heat shock was analyzed by ImageStream with or without XPO1 repression (24 h KPT251 pre-treatment at 1000nM). As expected, heat shock induced increased nuclear HSP70 localization and after 6 h of recovery at 37°C, HSP70 exited the nucleus, which was delayed when cells have been pre-treated with KPT-251 as compared to control treated cells (Online Supplementary haematologica | 2020; 105(9)

Figure S3). This observation suggests that XPO1 is the main exportin of HSP70.

Chemical inhibition of XPO1 using KPT-251 treatment ameliorates erythroid terminal differentiation of β-TM erythroid progenitors with low cytotoxicity As expected, this nuclear accumulation of both HSP70 and GATA-1 was associated with an increase in terminal erythroid differentiation of β-TM erythroblasts as assessed by flow cytometry analysis showing a significant increase in total Band3 MFI9 [MFI normalized on DMSO treated cells used as a control: DMSO=1, KPT100nM=1.06±0.04 (NS), and KPT1000nM=1.48±0.09 (P<0.01)] (Figure 4A). Consistent with an increase in terminal erythroid maturation, the fraction of cells expressing high Band3 was also increased by KPT treatment in a dose-dependent manner [2.62%±0.39 (control), 3.13%±0.53 (KPT100nM) (NS) and 9.27%±1.44 (KPT 1000nM) (P<0.01)] (Figure 4B, left). The absolute number of high Band3 cells was also increased (Figure 4B, right). To ensure that this high Band3 pool rep2247

F. Guillem et al.

resented orthochromatic erythroblasts and reticulocytes, as previously described,3 we sorted them by FACS based on their Band3 and 4-integrin expression (Figure 4C). As expected, mature erythroblast content (orthochromatic erythroblasts+reticulocytes) was enriched in high Band3 gate (62.23%±5.31) compared to low Band3 gate (22.83%±3.49) (P=0.0004). We further confirmed that in β-TM progenitors, KPT treatment alleviated the maturation arrest at the polychromatophilic stage as it decreased the proportion of polychromatophilic cells and increased the proportion of mature cells (Figure 4D). More importantly, KPT treatment significantly increased the terminal maturation index (TMI), determined as the ratio of orthochromatic erythroblasts+reticulocytes *100/ polychromatophilic cells, at the dose of 1000nM KPT [TMI normalized to control DMSO: DMSO=1, KPT100nM=1.8±0.5 (NS) KPT1000nM=3.7±1.05, (P<0.05)] (Figure 4E). Finally, we show that the significant increase in TMI and in absolute number of mature cells at 1000nM is associated with a high increase in HbF amount in the pool of mature cells, which is not observed at lower doses of treatment. This effect was observed with a greater magnitude at higher dose (1500nM). However at this dose, inhibition of cell proliferation and apoptosis were higher (Online Supplementary Figure S4). In Figure 5, we schematically illustrate the molecular mechanisms modulated by KPT treatment in β-TM erythroid progenitors compared to β-TM and normal erythroid progenitors in normal conditions.

Discussion Taken together, our data demonstrate that XPO1 is the main nuclear exporter of HSP70 in various cell types, and it may participate in the regulation of HSP70 in human erythroblasts during normal and pathological erythropoiesis. Decreased expression of XPO1 along erythroid differentiation provides an explanation for the augmentation of nuclear localization of HSP70 during normal erythroid maturation. Interestingly, proteins of the Ran machinery (RAN, RANBP1, RCC1, NUTF1 AND RAN-

References 1. Zermati Y, Garrido C, Amsellem S, et al. Caspase activation is required for terminal erythroid differentiation. J Exp Med. 2001; 193(2):247-254. 2. Ribeil J-A, Zermati Y, VandekerckhoveJ, et al. Hsp70 regulates erythropoiesis by preventing caspase-3-mediated cleavage of GATA-1. Nature. 2007;445(7123):102-105. 3. Arlet J-B, Ribeil J-A, Guillem F, et al. HSP70 sequestration by free a-globin promotes ineffective erythropoiesis in β-thalassaemia. Nature. 2014;514(7521):242-246. 4. Kudo N, Wolff B, Sekimoto T, et al. Leptomycin B inhibition of signal-mediated nuclear export by direct binding to CRM1. Exp Cell Res. 1998;242(2):540-547. 5. Nishi K, Yoshida M, Fujiwara D, Nishikawa M, Horinouchi S, Beppu T. Leptomycin B

2248

6. 7.

GAP1) essential for XPO1 activity are also expressed during erythroid differentiation and also down-regulated at the time of caspase activation (data not shown). Furthermore, we demonstrate that inhibition of XPO1 by KPT-251 treatment improves terminal maturation of βTM erythroblasts by preventing GATA-1 degradation by caspase-3. Thus, XPO1 inhibitors could be added to the armamentarium of therapeutic options of β-TM to correct ineffective erythropoiesis and its pathological consequences (e.g. iron overload and extramedullary hematopoiesis), to increase hemoglobin level, to reduce transfusion burden and ultimately to increase overall survival. Interestingly, KPT compounds are currently under clinical development and exhibit high efficacy and safety profile in cancer therapy. More recently, a new generation of KPT compounds with reduced toxicity is being developed (clinicaltrials.gov identifier: NCT02649790) and as such, if shown to be not toxic in the long-term, they could be used in β-TM as well. In addition, maintaining HSP70 in the nucleus by blocking XPO1 could be useful in other pathologies of erythropoiesis where HSP70 is not localized in the nucleus, such as myelodysplastic syndrome25 and congenital erythroblastopenia,26 but also in other diseases such as some neurodegenerative diseases, in which protein aggregates may also prevent HSP70 nuclear location. Acknowledgments The authors would like to thank the Department of Biotherapy and the Maternity at the Necker Hospital (Paris, France) for providing blood samples. This work was supported by the French National Research Agency ANR-13-BSV1-0029-05 “HSPathies”, LabEx GRex financial support ANR-11-IDEX-0, the French National Research Agency (ANR) under the program “Investissements d’Avenir” (LabExGRex and LipSTIC), the Institut National du Cancer (INCa), the Ligue Nationale Contre le Cancer (“Label of Excellence”), Ministère de l'Enseignement Supérieur et de la Recherche, FEDER and Région Bourgogne. Funding This program has received a state subsidy managed by the National Research Agency under the "Investments for the Future" program bearing the reference ANR-01-A0-IAHU.

targets a regulatory cascade of crm1, a fission yeast nuclear protein, involved in control of higher order chromosome structure and gene expression. J Biol Chem. 1994; 269(9):6320-6324. Lui K, Huang Y. RanGTPase: a key regulator of nucleocytoplasmic trafficking. Mol Cell Pharmacol. 2009;1(3):148-156. Hattangadi SM, Martinez-Morilla S, Patterson HC, et al. Histones to the cytosol: exportin 7 is essential for normal terminal erythroid nuclear maturation. Blood. 2014; 124(12):1931-1940. Mathias LA, Fisher TC, Zeng L, et al. Ineffective erythropoiesis in beta-thalassemia major is due to apoptosis at the polychromatophilic normoblast stage. Exp Hematol. 2000;28(12):1343-1353. Hu J, Liu J, Xue F,et al. Isolation and functional characterization of human erythroblasts at distinct stages: implications for

10.

11.

12.

13.

14.

understanding of normal and disordered erythropoiesis in vivo. Blood. 2013; 121(16): 3246-3253. An X, Schulz VP, Li J, et al. Global transcriptome analyses of human and murine terminal erythroid differentiation. Blood. 2014;1 23(22):3466-3477. Gautier E-F, Ducamp S, Leduc M, et al. Comprehensive Proteomic Analysis of Human Erythropoiesis. Cell Rep. 2016; 16(5):1470-1484. Fukuda M, Asano S, Nakamura T, et al. CRM1 is responsible for intracellular transport mediated by the nuclear export signal. Nature. 1997;390(6657):308-311. Xu D, Farmer A, Chook YM. Recognition of nuclear targeting signals by Karyopherin-β proteins. Curr Opin Struct Biol. 2010; 20(6):782-790. Pinello L, Xu J, Orkin SH, Yuan G-C. Analysis of chromatin-state plasticity iden-

haematologica | 2020; 105(9)

XPO1 is a target to treat Î˛-thalassemia

15.

16.

17.

18.

tifies cell-type-specific regulators of H3K27me3 patterns. Proc Natl Acad Sci U S A. 2014;111(3):344-353. Etchin J, Sun Q, Kentsis A, et al. Antileukemic activity of nuclear export inhibitors that spare normal hematopoietic cells. Leukemia. 2013;27(1):66-74. Lapalombella R, Sun Q, Williams K, et al. Selective inhibitors of nuclear export show that CRM1/XPO1 is a target in chronic lymphocytic leukemia. Blood. 2012; 120(23): 4621-4634. Etchin J, Sanda T, Mansour MR, et al. KPT330 inhibitor of CRM1 (XPO1)-mediated nuclear export has selective anti-leukaemic activity in preclinical models of T-cell acute lymphoblastic leukaemia and acute myeloid leukaemia. Br J Haematol. 2013;161(1):117-127. Kojima K, Kornblau SM, Ruvolo V, et al. Prognostic impact and targeting of CRM1

haematologica | 2020; 105(9)

19.

20.

21.

22.

in acute myeloid leukemia. Blood. 2013; 121(20):4166-4174. Walker CJ, Oaks JJ, Santhanam R, et al. Preclinical and clinical efficacy of XPO1/CRM1 inhibition by the karyopherin inhibitor KPT-330 in Ph+ leukemias. Blood. 2013;122(17):3034-3044. Wettersten HI, Landesman Y, Friedlander S, Shacham S, Kauffman M, Weiss RH. Specific inhibition of the nuclear exporter exportin-1 attenuates kidney cancer growth. PloS One. 2014;9(12):1-15. Ranganathan P, Yu X, Na C, et al. Preclinical activity of a novel CRM1 inhibitor in acute myeloid leukemia. Blood. 2012;120(9):1765-1773. Gravina GL, Mancini A, Sanita P, et al. KPT-330, a potent and selective exportin-1 (XPO-1) inhibitor, shows antitumor effects modulating the expression of cyclin D1 and survivin [corrected] in prostate cancer

models. BMC Cancer. 2015;15:941-960. 23. Tai Y-T, Landesman Y, Acharya C, et al. CRM1 inhibition induces tumor cell cytotoxicity and impairs osteoclastogenesis in multiple myeloma: molecular mechanisms and therapeutic implications. Leukemia. 2014;28(1):155-165. 24. Zeng X-C, Bhasin S, Wu X, et al. Hsp70 dynamics in vivo: effect of heat shock and protein aggregation. J Cell Sci. 2004;117(Pt 21):4991-5000. 25. Frisan E, Vandekerckhove J, de Thonel A, et al. Defective nuclear localization of Hsp70 is associated with dyserythropoiesis and GATA-1 cleavage in myelodysplastic syndromes. Blood. 2012;119(6): 1532-1542. 26. Gastou M, Rio S, Dussiot M, et al. The severe phenotype of Diamond-Blackfan anemia is modulated by heat shock protein 70. Blood Adv. 2017;1(22):1959-1976.

2249

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2250-2261

Granulocyte Biology & its Disorders

Impaired microRNA processing in neutrophils from rheumatoid arthritis patients confers their pathogenic profile. Modulation by biological therapies Ivan Arias de la Rosa,1* Carlos Perez-Sanchez,2* Patricia Ruiz-Limon,3* Alejandra Patiño-Trives,1 Carmen Torres-Granados,1 Yolanda Jimenez-Gomez,1 Maria del Carmen Abalos-Aguilera,1 Irene Cecchi,4 Rafaela Ortega,1 Miguel Angel Caracuel,1 Jerusalem Calvo-Gutierrez,1 Alejandro EscuderoContreras,1 Eduardo Collantes-Estevez,1# Chary Lopez-Pedrera1# and Nuria Barbarroja#1,5 *IAR, CPS and PRL contributed equally as co-first authors.

#ECE, CLP and NB contributed equally as co-senior authors.

Rheumatology service, Maimonides Institute for Research in Biomedicine of Cordoba (IMIBIC)/Reina Sofia Hospital/University of Cordoba, Cordoba, Spain; 2Deparment of Medicine, University of Cambridge, School of Clinical Medicine, Addenbroke’s Hospital, Cambridge Institute for Medical Research, Cambridge, UK; 3Biomedical Research Institute (IBIMA), Service of Endocrinology and Nutrition, Malaga Hospital Complex (Virgen de la Victoria), Malaga, Spain; 4Department of Clinical and Biological Sciences, Center of Research of Immunopathology and Rare Diseases-Coordinating Center of Piemonte and Valle d’Aosta Network for Rare Diseases, Turin, Italy and 5CIBER Fisiopatología de la Obesidad y Nutrición (CIBEROBN), Instituto de Salud Carlos III, Madrid, Spain 1

ABSTRACT

Correspondence: NURIA BARBARROJA nuria.barbarroja.exts@juntadeandalucia.es Received: August 22, 2018. Accepted: January 10, 2020. Pre-published: January 16, 2020. doi:10.3324/haematol.2018.205047 ©2020 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.

2250

he aim of this study was to investigate the microRNA (miRNA) expression pattern in neutrophils from rheumatoid arthritis (RA) patients and its contribution to their pathogenic profile and to analyze the effect of specific autoantibodies or inflammatory components in the regulation of miRNA in RA neutrophils and its modulation by biological therapies. Neutrophils were isolated from paired peripheral blood (PB) and synovial fluid samples of 40 patients with RA and from PB of 40 healthy donors. A miRNA array was performed using nCounter technology. Neutrophils from healthy donors were treated in vitro with antibodies to citrullinated protein antigens isolated from RA patients and tumor necrosis factor-a (TNF-a) or interleukin-6. A number of cytokines and chemokines were analyzed. In vitro treatments of RA-neutrophils with tocilizumab or infliximab were carried out. Transfections with pre-miRNA and DICER downregulation experiments were further performed. RA-neutrophils showed a global downregulation of miRNA and genes involved in their biogenesis, alongside with an upregulation of various potential mRNA targets related to migration and inflammation. Decreased levels of miRNA and DICER correlated with autoimmunity, inflammation and disease activity. Citrullinated protein antigens and TNF-a decreased the expression of numerous miRNA and their biogenesis-related genes, increasing their potential mRNA targets. Infliximab reversed those effects. Transfections with pre-miRNA-223, -126 and -148a specifically modulated genes regulating inflammation, survival and migration whereas DICER depletion influenced the inflammatory profile of neutrophils. Taken together RA-neutrophils exhibited a global low abundance of miRNA induced by autoantibodies and inflammatory markers, which potentially contributed to their pathogenic activation. miRNA biogenesis was significantly impaired in RAneutrophils and further associated with a greater downregulation of miRNA mainly related to migration and inflammation in synovial fluid neutrophils. Finally, anti-TNF-a and anti-interleukin-6 receptor treatments can modulate miRNA levels in the neutrophils, minimizing their inflammatory profile. haematologica | 2020; 105(9)

Impaired microRNA processing in RA neutrophils

Introduction Several immune cells including T and B lymphocytes, macrophages, synovial fluid (SF) fibroblast and neutrophils are known to be relevant in the rheumatoid arthritis (RA) pathogenesis.1 Among them, RA neutrophils are activated cells, characterized by a prolonged lifespam, increased migratory capacity and production of inflammatory molecules and reactive oxygen species (ROS). In severe acute inflammation, SF accumulates a great number of these cells in a more activated state, promoting cartilage destruction and joint damage.2 Antibodies to citrullinated protein antigens (ACPA) are currently considered the most specific autoantibodies in RA, being related to the activity of the disease and poorer prognosis.3 ACPA have been shown to be able to induce neutrophils to produce high levels of inflammatory mediators, ROS and to generate NETosis.2,4 Epigenetic modifications contribute to the development of RA, affecting disease susceptibility and severity.5,6 Among them, several microRNA (miRNA) have been linked to the chronic inflammation in RA.5 MiRNA are short noncoding RNA present in all multi-cellular organisms involved in a broad range of cellular processes. They cause posttranscriptional and posttranslational gene silencing, by recognizing a specific sequence of mRNA, binding to it and inhibiting its translation into protein.7 MiRNA is first transcribed into long primary miRNA of several kb in length (pri-miRNA) and this pri-miRNA is then processed by Drosha into a precusor miRNA (premiRNA) of appoximately 70-nucleotide. The pre-miRNA is transported out of the nucleus by exportin 5 (XPO-5) and is then processed by DICER into a mature double stranded miRNA of approximately 22 nucleotides. The RNA-induced silencing complex (RISC) (composed of the transactivation-responsive RNA-binding protein [TRBP] and argonaute [AGO]) removes the complementary strand. DICER then binds to RISC, forming the core of RISC-loading complex. DICER is considered a crucial factor in miRNA processing since its presence is necessary for the stimulation of RNA processing by AGO.8,9 Functional miRNA is able to bind to the 3’-untranslated region (UTR) of the target mRNA, causing mRNA cleavage or translational repression.10 Several studies, mainly conducted on lymphocytes, monocytes, macrophages and SF fibroblasts, have reported that the role of various miRNAs in the pathogenesis of RA is critical for the increased expression of inflammatory cytokines and prolonged cell survival.5,11 We undertook this study to evaluate the miRNA profile and the proteins involved in miRNA processing in circulating and SF neutrophils from RA patients, in order to gain an insight of its role in the different activation states of these cells. The effects of ACPA or inflammatory components and biological therapies on the expression of miRNA in neutrophils was further assessed.

Methods For details see the Online Supplemental Materials and Methods.

RA patients and healthy donors Forty RA patients and 40 healthy donors (HD) were included in

haematologica | 2020; 105(9)

this study. RA patients fulfilled at least four 1987 American College of Rheumatology (ACR) criteria and achieved a total score of 6 or greater according to 2010 criteria. The patients were under the following treatment regimes: corticosteroids (50.0%), leflunomide (42.5%), hydroxychloroquine (45.0%), NSAID (80.0%) and methotrexate (65%). All patients were tested for the presence of ACPA and rheumatoid factor (RF) by clinical laboratory routine analysis. All participants enrolled were Caucasian, recruited at the Department of Rheumatology, and gave their written informed consent approved by the ethical committee of the Reina Sofia Hospital (Cordoba, Spain). Clinical details of the RA patients and HD are shown in Table 1. Peripheral blood (PB) was withdrawn from all the RA patients and the HD. SF from RA patients was obtained through arthrocentesis. The study design is displayed on a flow chart (Online Supplementary Figure S1).

Isolation of neutrophils from PB and SF Neutrophils from PB of HD and paired SF and PB samples of RA patients were isolated (after centrifugation to obtain buffy coat and osmotic lysis of the pellet) by immunomagnetic positive selection with human anti-CD15 microbeads (Miltenyi Biotec S.L, Bergisch Gladbach, Germany) using AUTOMACs (Miltenyi Biotec).12

miRNA expression profiling The nCounter miRNA Assay (NanoString Technology, Seattle, WA, USA) detects simultaneously 800 human miRNA in each sample. 100 ng of RNA, pooled samples of neutrophils from PB of 10 HD, neutrophils from PB of 10 RA patients and neutrophils from SF of 10 RA patients were prepared by ligating a specific DNA tag (miR-tag) onto the 3 end of each mature miRNA followed by 16-20 hour hybridization (at 65ºC) to nCounter Reporter and Capture probes. The rest of the protocol was performed following the manufacturer’s recommendations (NanoString Technologies, Seattle, WA, USA). Data were normalized by the geometric mean of top 100 miRNA detected using the nSolver software. This miRNA array was performed in pooled samples of the 10 RA patients that best represented the mean values of age, sex, disease activity, evolution time and autoimmunity of the clinical validation cohort.

IgG-ACPA isolation from RA patients Immunoglobulin G (IgG) from serum of five different RA patients with high titers of ACPA and negative for RF (enriched IgG-ACPA) and 5 HD (IgG- normal human serum [NHS]) were isolated using HiTrap protein G HP columns (GE Healthcare, Chicago, Il, USA).

In vitro treatments of neutrophils Neutrophils purified from five RA patients (taking Diseasemodifying antirheumatic drugs and not taking any biological therapies) were pre-treated with FCRII blocking reagent (Miltenyi Biotec, Bergisch Gladbach, Germany) for 15 min and subsequently incubated with infliximab (IFX) at 100 g/mL or tocilizumab (TCZ) at 20 g/mL for 6 hours. The selection of these patients allowed the isolation of assumingly activated neutrophils, leading to an increased expression of inflammatory cytokines, in order to demonstrate the effects of miRNA transfection by proving the reduction in expression levels of those molecules. Neutrophils purified from five HD were treated in vitro with IgG-NHS or enriched IgG-ACPA (500 ug/mL), tumor necrosis factor-a (TNF-a) and interleukin-6 (IL-6) (10 ng/mL) for 6 hours. Samples were processed for RT-PCR analyses.

2251

I.A. de la Rosa et al. Table 1. Clinical details of rheumatoid arthritis patients and healthy donors.

Clinical parameters Woman/Men (n/n) Age (years) Disease evolution (years) RF+ (%) ACPA+ (%) Obesity (%) Diabetes (%) Hypertension (%) Smoker (%) DAS28 Cholesterol (mg/mL) HDL-Cholesterol (mg/mL) LDL-Cholesterol (mg/mL) Triglycerides (mg/mL) ESR (mm/h) CRP (mg/mL) Treatments Corticosteroids (%) Hydroxychloroquine (%) NSAIDs (%) Methotrexate (%)

RA patients (n=40)

HD controls (n=40)

26/14 55.54 ± 4.03 9.20 ± 2.23 57.50 65.00 8.00 10.00 12.50 15.00 3.75 ± 0.20 202.15 ± 12.65 43.16 ± 2.68 136.81 ± 11.01 99.50 ± 10.16 15.75 ± 2.89 10.23 ± 2.73

22/18 48.50 ± 8.45 3.00 5.00 6.00 12.00 200.20 ± 15.60 53.63 ± 12.75 115.75 ±17.40 88.50 ± 25.39 4.65 ± 1.20 1.55 ± 0.22

50.00 45.00 80.00 70.00

0.035 0.033 0.048

0.01 0.03

Values are mean ± standard deviation (SD). HDL: high density lipoprotein; LDL: low density lipoprotein; DAS: disease activity score; ACPA: antibodies to citrullinated protein antigens; RA: rheumatoid arthritis; HD: healthy donor; RF: rheumatoid factor; ESR: erythrocyte sedimentation rate; CRP: C reactive protein; NSAIDS: non-steroideal anti-inflammatory drugs.

Results Global decrease in miRNA levels of neutrophils from RA patients Among the 800 miRNA analyzed, levels of 133 miRNA were detected in neutrophils. Using an above two-fold change cut-off, 94 miRNA were reduced in PB-RA neutrophils compared to PB-HD, and three of them were elevated (Online Supplementary Table S1 and Figure 1A). Additionally, SF neutrophils showed 34 miRNA even further reduced compared to its paired PB sample (fold change cut-off change above two-fold) (Figure 1B). Ingenuity Pathway Analysis (QIAGEN IPA) uncovered the main enriched biological functions and pathways in which these miRNA are involved, which include immune disease, inflammatory response and connective disorders (Figure 1D).

Low abundance of miRNA levels in RA neutrophils might be due to a defect in the miRNA processing Eight altered miRNA were identified by QIAGEN IPA as the main regulators of proteins involved in the abnormal activation of neutrophils in RA, including miRNA -126, 148a, -29c, let-7b, -30c, -17, -21 and 223 (Figure 2). The expression of these miRNA was validated in all the samples separately and a technical validation was performed separately in the 10 samples previously used for the pool. In addition, a clinical validation was carried out separately in the 30 remaining samples (Online Supplementary Figure S2). The levels of most of the selected miRNA were significantly reduced in PB-RA neutrophils compared to PB-HD neutrophils. A greater reduction in the expression of miR148a, miR-29c and let-7b in the SF paired samples was 2252

observed (Figure 3A). In addition, there was not significant difference in the reduced levels of miRNA among patients treated or not treated with methotrexate (Online Supplementary Figure S3). There was a significant reduction in the expression of genes involved in the miRNA processing (DICER and AGO-1) in neutrophils from PB-RA patients compared to PB-HD. Of note, DICER, AGO-1, AGO-2 and XPO-5 were diminished in neutrophils from the SF of RA patients (Figure 3B).

Bioinformatic identification and expression of the putative targets of reduced miRNA in RA neutrophils Seven putative mRNA targets were chosen based on their recognized role in the pathogenesis of RA, being key factors in inflammation (TNF-a, IL-1β, IL-6R), cell adhesion (VEGF-A), migration (IL-8) and survival (STAT3 and AKT). These targets were significantly upregulated in PBRA neutrophils (Figure 3C). A greater alteration was observed in SF neutrophils. Using enrichment analysis of those selected targets, enriched pathways mainly related to inflammatory processes were revealed. This included a broad range of secondary chemokines and cytokines which are indirectly connected with the eight selected mRNA targets, amplifying the inflammatory cascade (Online Supplementary Figure S4). Thus, a human cytokine array was performed in neutrophils from RA patients (PB and SF) and HD (PB). Neutrophils from PB of RA patients showed increased protein expression of CCL5, CD40L, CXCL1, CXCL2, IL1ra, IL-16, IL-18, IL-32a, PAI-1 and TREM-1 compared to HD (Figure 3D). A differential proteome profile was observed in neutrophils from SF of RA patients compared haematologica | 2020; 105(9)

Impaired microRNA processing in RA neutrophils

Figure 1. Global downregulation of miRNA expression profile in rheumatoid arthritis neutrophils. (A) Forty-one miRNA altered in neutrophils from PB-RA patients compared to neutrophils from PB-HD using a fold change cut-off >5. (B) Thirty-four miRNA more reduced in RA synovial neutrophils compared to its paired PB sample (fold change cut-off >2). (C) Heat-map of the differentially expressed miRNA profile in PB-HD, PB-RA and SF-RA. (D) Functional classification of the altered miRNA using Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, CA, USA; https://analysis.ingenuity.com). The analysis included only the functions and pathways with average IPA score >2 indicated as –log (P-value). Threshold bar indicates the cut-off point of significance (P>0.05), using Fisher’s exact test. miRNA: microRNA; RA: rheumatoid arthritis; PB: peripheral blood; HD: healthy donor; SF: synovial fluid.

to HD and PB paired samples (MIP-1 /1 , CCL5, CD40L, C5/5a, CXCL1, CXCL12, ICAM-1, IL-1 , IL-1ra, IL-8, IL13, IL-16, IL-18, MIF and TREM-1) (Figure 3D).

eight selected mRNA targets (CCL1, CCL2, MIP-1a/β, CCL5, CD40L, CXCL1, CXCL12, G-CSF, GM-CSF, IFN-g and IL-8) (Figure 4E).

Reduced levels of miRNA in RA neutrophils are related to autoimmunity, clinical and serological parameters

Inflammatory mediators decrease the expression of miRNA in neutrophils, which might be restored by IFX or TCZ

Decreased levels of both, miRNA and DICER significantly correlated with the activity of the disease, levels of ACPA and clinical inflammatory markers. Elevated serum levels of TNF-a correlated with low levels of DICER. However, there was not association between the levels of miRNA and serum TNF-a (Figure 4A).

ACPA reduces miRNA levels in healthy neutrophils Enriched IgG-ACPA downregulated the expression of the eight selected miRNA in healthy neutrophils (Figure 4B). Accordingly, a significant reduction of genes involved in the miRNA biogenesis was observed (Figure 4C). Enriched IgG-ACPA also increased the expression of the selected mRNA targets (Figure 4D). Finally, enriched IgGACPA promoted a significant upregulation of secondary chemokines and cytokines indirectly related with the haematologica | 2020; 105(9)

TNF-a and IL-6 levels were significant elevated in serum from RA patients; a further increase was observed in SF from those RA patients (Figure 5A). In vitro, TNF-a downregulated the levels of the eight selected miRNA alongside with a decrease in the expression of DICER and AGO-2 in HD neutrophils (Figure 5B). Treatment with IL-6 reduced the levels of miR-126, let-7b, miR-17, AGO1 and AGO2 (Figure 5C). We observed that fresh neutrophils from RA patients had significantly higher levels of TNF-a and IL-6 mRNA compared to freshly isolated neutrophils from HD (Online Supplementary Figure S5A). In addition, after 6 hours of in vitro culture, levels of TNF-a and IL-6 were elevated in the culture media of RA neutrophils (Online Supplementary Figure S5B). 2253

I.A. de la Rosa et al.

Figure 2. Bioinformatic identification of deregulated miRNA and protein targets related to the pathogenic profile of neutrophils in rheumatoid arthritis. In silico studies were performed to identify eight altered miRNA using QIAGEN’s Ingeunity Pathway Analysis (IPA, QIAGEN Redwood City, USA; https://analysis.ingenuity.com) as the main regulators of proteins involved in the abnormal activation of neutrophils in RA: inflammation, migration and cell survival. By using the tool miRNA target filter of IPA, a network including the selected miRNA and mRNA targets experimentally observed and predicted with high confidence, was generated. miRNA: microRNA; RA: rheumatoid arthritis.

In vitro treatment of active neutrophils purified from RA patients with IFX restored the low levels of the eight selected miRNA while TCZ only up-regulated the expression of miR-148a (Figure 5D). Accordingly, IFX upregulated the expression levels of AGO1. Regarding mRNA targets, IFX reduced the mRNA expression of VEGF-A, TNF-a, IL-1β, IL-8 and STAT3. Treatment with TCZ also diminished the expression of various of these mRNA targets such as IL-1β, IL-8, IL-6R and STAT3 (Figure 5E). In addition, protein release of TNF-a, IL-8 and IL-1β was reversed in RA neutrophils after treatments with both, IFX and TCZ (Figure 5F).

Overexpression of miR-126, miR-148a and miR-223 in RA neutrophils decreases specific targets involved in inflammation, migration and cell survival We selected three downregulated miRNA to evaluate their role in migration, proinflammatory profile and cell survival of RA neutrophils: miR-223 was the most abundant in neutrophils, miR-148a and miR-126 have several potential and demonstrated mRNA targets involved in inflammation. As visible in Figure 6, miR-126, miR-148a and miR-223 overexpression led to a downregulation of their specific mRNA targets: miR-126 overexpression induced a significant downregulation of VEGF-A protein expression (Figure 6A), and miR-223 overexpression promoted a significant decrease in the protein expression of IL-8 and IL-1 . Interestingly, miR-223 transfection induced 2254

a significant increase of VEGF mRNA and protein (Figure 6C). On the other hand, miR-148a overexpression reduced gene and protein expression levels of TNF-a (Figure 6B).

DICER downregulation in neutrophils might exacerbate their inflammatory profile Using a low number of lentiviral particles, 25% of DICER expression was inhibited in HL-60 neutrophil-like cells (Figure 6D). This reduction promoted a significant decrease in all the selected miRNA (Figure 6E). Protein levels of a number of cytokines and chemokines were significantly upregulated in neutrophils after DICER downregulation (Figure 6F).

Discussion This study reports for the first time the altered miRNA expression profile in neutrophils from RA patients, describing a defect in the miRNA processing machinery responsible for a global low abundance of miRNA, mediated by ACPA and inflammatory mediators, promoting the high inflammatory profile of these cells in RA. Several miRNA have been shown to be increased in the PB and in inflamed joints in RA patients, correlated to disease activity and to promoting the production of inflammatory mediators involved in the synovitis.13,14 Likewise, haematologica | 2020; 105(9)

Impaired microRNA processing in RA neutrophils

Figure 3. Low abundance of miRNA levels in rheumatoid arthritis neutrophils might be due to a defect in the miRNA processing. (A) Validation of the miRNA array in PB-HD neutrophils (n=40), PB-RA neutrophils (n=40) and SF-RA neutrophils (n=40). (B) Expression of genes related to miRNA biogenesis machinery in PB-HD (n=40), PB-RA (n=40) and SF-RA neutrophils (n=40). (C) Gene expression of putative mRNA targets of the selected miRNA in PB-HD (n=40), PB-RA (n=40) and SFRA neutrophils (n=40). (D) Proteome profile of chemokines and cytokines in neutrophils from PB-HD (n=10), PB-RA (n=10) and SF-RA neutrophils (n=10). MiR: microRNA; RA: rheumatoid arthritis; PB: peripheral blood; HD: healthy donor; SF: synovial fluid; AGO-1: argonaute-1; AGO-2: argonaute-2; XPO-5: exportin-5; VEGF-A: vascular endothelial growth factor A; TNF-α: tumor necrosis factor-alpha; IL-1β: interleukin-1 beta; IL-8: interleukin-8; IL-6R: interleukin-6 receptor; STAT-3: signal transducer and activator of transcription 3; AKT: protein kinase B; MIP-1a/1β: macrophage inflammatory protein 1 alpha/1 beta; CCL5: chemokine (C-C motif) ligand 5; CD40 ligand: cluster differentiation 40 ligand; C5/5a: complement component C5/5a; CXCL-1: chemokine (C-X-C motif) ligand 1; CXCL12: chemokine (C-X-C motif) ligand 12; ICAM-1: intercellular adhesion molecule 1; IL-1ra: interleukin-1 receptor antagonist; IL-13: interleukin 13; IL-16: interleukin 16; IL-18: interleukin 18; IL-32a: interleukin 32 alpha; MIF: macrophage migration inhibitory factor; PAI-1: plasminogen activator inhibitor-type 1; TREM-1: triggering receptor expressed on myeloid cells 1. Data are presented as mean ± standard error of the mean (SEM); a: significant differences vs. PB-HD P<0.05; b: significant differences vs. PB-RA P<0.05.

haematologica | 2020; 105(9)

2255

I.A. de la Rosa et al.

Figure 4. Reduced levels of miRNA in rheumatoid arthritis neutrophils are related to ACPA. (A) Correlation studies of reduced miRNA and DICER expression levels in neutrophils from peripheral blood with clinical parameters such as, ACPA levels, DAS28, CRP and inflammatory markers including TNF-a and IL-6 serum cytokine levels. (B) In vitro effect of enriched IgG-ACPA in miRNA expression in healthy neutrophils. (C) In vitro effect of enriched IgG-ACPA in the expression of genes related to the miRNA biogenesis in healthy neutrophils. (D) In vitro effect of enriched IgG-ACPA in the mRNA expression of potential mRNA targets of the validated miRNA. (E) Proteome profile of chemokines and cytokines in neutrophils treated with IgG-NHS and enriched IgG-ACPA. ACPA: antibodies to citrullinated protein antigens; IgGNHS: immunoglobulin G from normal human serum; miR: microRNA; HD: healthy donor; DAS28: disease activity score 28; CRP: C-reactive protein; TNF-a: tumor necrosis factor-alpha; IL-6: interleukin-6; AGO-1: argonaute-1; AGO-2: argonaute-2; XPO-5: exportin-5; VEDF-A: vascular endothelial growth factor A; IL-1β: interleukin1 beta; IL-8: interleukin-8; IL-6R: interleukin-6 receptor; STAT-3: signal transducer and activator of transcription 3; AKT: protein kinase B; CCL1: chemokine (C-C motif) ligand 1; CCL2: chemokine (C-C motif) ligand 2; MIP-1a/1β: macrophage inflammatory protein 1 alpha/1 beta; CCL5: chemokine (C-C motif) ligand 5; CD40 ligand: cluster differentiation 40 ligand; CXCL-1: chemokine (C-X-C motif) ligand 1; CXCL12: chemokine (C-X-C motif) ligand 12; G-CSF: granulocyte colony-stimulating factor; GM-CSF: granulocyte-macrophage colony-stimulating factor; IFN-g: interferon gamma; IL-1a: interleukin-1 alpha; PAI-1: plasminogen activator inhibitor-type 1. Data are presented as mean ± standard error of the mean (SEM) (n= 5); a: significant differences P<0.05.

2256

haematologica | 2020; 105(9)

Impaired microRNA processing in RA neutrophils

Figure 5. Modulation of the expression of miRNA and genes involved in their processing in healthy donor neutrophils by inflammatory components - in vitro effects of TCZ and IFX in RA active neutrophils. (A) Serum and synovial levels of TNF-α and IL-6 (40 HD-serum, 40 RA-serum and 40 RA-SF). a: significant differences vs. HD serum P<0.05; b: significant differences vs. RA serum P<0.05. (B) Expression of miRNA and genes involved in their biogenesis machinery in HD neutrophils treated with TNF-a. (C) Expression of miRNA and genes involved in their biogenesis machinery in HD neutrophils treated with IL-6. (D) Expression of miRNA and genes involved in their processing in RA active neutrophils treated in vitro with TCZ or IFX. (E) Gene expression of putative mRNA targets of the validated miRNA in RA active neutrophils treated in vitro with TCZ or IFX. (F) Protein levels of putative mRNA targets of the validated miRNA in RA active neutrophils treated in vitro with TCZ or IFX. miRNA: microRNA; HD: healthy donor; TCZ: tocilizumab; IFX: infliximab; RA: rheumatoid arthritis; AGO-1: argonaute-1; AGO-2: argonaute-2; XPO-5: exportin-5; VEGFA: vascular endothelial growth factor; TNF-a: tumor necrosis factor-alpha; IL-1β: interleukin-1 beta; IL-8: interleukin-8; IL-6R: interleukin-6 receptor; STAT-3: signal transducer and activator of transcription 3; AKT: protein kinase B; Non-t: non-treated. Data are presented as mean ± standard error of the mean (SEM) of five independent experiments; a: significant differences respective non-treated P<0.05.

haematologica | 2020; 105(9)

2257

I.A. de la Rosa et al. A

Figure 6. miR-126, miR-148a and miR-223 regulate the expression of specific mRNA targets involved in inflammation, migration and cell survival on neutrophils. DICER downregulation induces an inflammatory profile in neutrophils. (A) Overexpression of miR-126 decreased the expression of VEGF in RA active neutrophils. (B) Overexpression of miR-148a decreased the expression of TNF-a in RA active neutrophils. (C) Overexpression of miR-223 decreased the expression of IL-8 and Il-1β and increased VEGF in RA active neutrophils. (D) DICER expression levels after lentivirus transfection by PCR and Western blot. (E) Expression levels of miRNA in DICER downregulated neutrophils. (G) Protesome array of cytokines and chemokines in HL60 neutrophil-like cells after downregulation of DICER by lentiviral transfection. MiRNA: microRNA; RA: rheumatoid arthritis; VEGF-A: vascular endothelial growth factor A; TNF-a: tumor necrosis factor-alpha; IL-8: interleukin-8; IL-1β: ιnterleukin-1 beta; CCL1: chemokine (C-C motif) ligand 1; CCL2: chemokine (C-C motif) ligand 2; MIP-1a/1β: macrophage inflammatory proteins alpha and beta; CCL5: chemokine (C-C motif) ligand 5; IL-1ra: interleukin 1 receptor antagonist; IL-6: interleukin-6; IL-13: interleukin-13; IL-16: interleukin-16; IL-17a: interleukin-17a; MIP: macrophage migration inhibitory factor; PAI-1: plasminogen activator inhibitor-1; Scr: scrambled; L-Scr: lentivirus-scrambled; L-DICER: lentivirus-DICER. Data are presented as mean ± standard error of the mean (SEM) of two independent experiments; a: significant differences P<0.05.

2258

haematologica | 2020; 105(9)

Impaired microRNA processing in RA neutrophils

various single nucleotide polymorphism have been studied in miRNA in RA.15 Our study shows a global downregulation of the miRNA expression in RA neutrophils, more marked in SF neutrophils, suggesting that it might contribute to the abnormal activated profile of these cells in the synovium. In this sense, global downregulation of miRNA has been shown in human alveolar macrophages induced by cigarette smoking, responsible for the changes in gene expression associated with the disease.16 In autoimmune disorders, we recently described a global downregulation of the miRNA levels in neutrophils from patients with systemic lupus erythematosus and antiphospholipid syndrome, which may indicate that chronic inflammation and/or autoimmunity is associated with a reduction of miRNA in neutrophils.17 In the present study, we demonstrate that either ACPA or inflammatory mediators, especially TNF-a can modulate the miRNA expression profile, through a reduction of several proteins involved in its processing, which might be translated into an increase of genes that might be involved in inflammation, cell survival and migration. Up to date, no study has reported the effect of ACPA on the expression of miRNA. In our hands, the reduced levels of miRNA and DICER in RA neutrophils correlated with elevated levels of ACPA. Accordingly, our in vitro studies demonstrated a direct involvement of these autoantibodies in the deregulation of various miRNA and their specific protein targets- globally related to the pathogenesis of RA. We further demonstrated that the global downregulation of the miRNA expression in RA neutrophils was associated, at least partially, with the reduced levels of DICER. A recent study suggested the role of DICER in neutrophils differentiation, where the DICER inhibition attenuated the activation of autophagy, a process that is needed for proper neutrophil differentiation.12 DICER plays a crucial role in miRNA biogenesis. Thus, it has been suggested that mRNA and protein levels of DICER must be strictly controlled since small changes can initiate various pathological processes.18 Here, we prove a novel role for DICER in neutrophils, showing that a small reduction of protein levels can induce a proinflammatory profile in neutrophils by downregulating several miRNA and, hence, a number of putative targeted cytokines and chemokines. Currently, little is known about the miRNA regulating neutrophil function. Several miRNA have recently been involved in the neutrophil development and function and in various pathological states, including miRNA-155, miRNA-34a, miRNA-223, miRNA-142, miRNA-452 and miRNA-466L.19 Overexpression of miRNA-155 and miRNA-34a in neutrophils from patients with myelodysplastic syndrome has been shown to contribute to an alteration of the migration.20 In addition, decreased levels of both, the miRNA-145 and the miRNA-143 have been shown in acute myelogenous leukemia, which are responsible for the blockade of the differentiation process of the neutrophils.21 Alongside with previous evidence, here we show that miRNA-223 is one of the most abundant miRNA on neutrophils.22 It has recently been demonstrated that the miRNA-223 is an important regulator blocking the infiltration of neutrophils in alcoholic hepatic disease.22 Supporting a role for this miRNA in the infiltration capacity of the neutrophils, in the present work we demonhaematologica | 2020; 105(9)

strated how the overexpression of the miRNA-223 in neutrophils of RA patients reduced specifically the expression of IL-1 and IL-8, molecules involved in inflammation and migration. The role of the miRNA-126 in the vascular integrity has been also evidenced.23 We observed reduced levels of the miRNA-126 in RA neutrophils, while its induced overexpression in RA neutrophils significantly reduced the levels of VEGF, pointing to the role of this miRNA in neutrophil adhesion and migration. Multiple functions have been attributed to miRNA148a in several diseases and low levels of miRNA-148 have been shown to be related to less survival time and increased recurrence risk in bladder cancer.24 In addition, miRNA-148 has been related to innate and adaptive immune responses.25 Our data is in agreement with these studies, since we found reduced levels of the miRNA-148 in RA neutrophils, associated to increased levels of TNF, a key inflammatory protein driving the RA disease. Other miRNA found decreased in RA neutrophils, such as miRNA-21, Let-7 and miRNA-30, have previously been reported to be altered in different types of tumors, thus playing a relevant role in tumorigenesis, invasion and metastasis of cancer cells.26-29 In addition, Let-7 and miRNA-17 regulate the T-cell response.27,30 Finally, a recent study demonstrated that the levels of all the members of the miRNA29 family were decreased in PB mononuclear cells and CB34+ cells of the bone marrow of acute myelogenous leukemia patients. The normalization of their levels partially inhibited the abnormal proliferation of the blasts, blocked the myeloid differentiation and repressed the apoptosis.31 TNF-a and IL-6 are key inflammatory effectors in RA, whose levels are elevated in RA serum and further increased in RA SF. We found a marked effect of TNF-a on neutrophils, lowering the expression of genes related to miRNA processing (including DICER and AGO-1) and downregulating the eight miRNA selected. By contrast, IL-6 had less effect but was still able to reduce the levels of miRNA-126, let-7b and miRNA-17 alongside with the expression of AGO-1 and AGO-2. Treatment of active RA neutrophils with IFX or TCZ reduced the inflammatory profile, by downregulating the gene expression of VEGFA, TNF-a, IL-1Î˛, IL8, IL6R and STAT3. However, only IFX was able to restore the global levels of selected miRNA, alongside with genes involved in their processing in RA neutrophils, an effect that might be expected after the efficciency of TNF-a observed to reduce the miRNA levels, thus suggesting that IFX might specifically minimize the abnormal profile of the RA neutrophils through the inhibition of TNF-a, which directly acts by reducing the expression of miRNA. In agreement with these results, we recently demonstrated that in vivo treatment with anti-TNF-a drugs for six months regulated the levels of several miRNA in the plasma of RA patients. Moreover, miRNA-23 and miRNA-223 were identified as potential biomarkers of therapy effectiveness.32 Altogether, our study shows that neutrophils from RA patients have a defect in the miRNA biogenesis machinery, are more marked in SF neutrophils, and are induced by ACPA and inflammatory mediators. This defect might be directly associated with the abnormal neutrophil activation, thus increasing their proinflammatory profile, observable by the higher expression of a number of chemokines and cytokines. 2259

I.A. de la Rosa et al.

Among the miRNA altered in RA neutrophils, we demonstrate that the miRNA-223, miRNA-126 and miRNA-148 are involved in the modulation of genes involved in processes such as migration, inflammation and cell survival in neutrophils. Finally, biological therapies would be able to improve miRNA processing, upregulating the levels of miRNA, which might reduce the activation of the neutrophil. Beyond the regulation of miRNA in RA neutrophils, there are likely other epigenetic mechanisms that potentially contribute to the abnormal activation of these cells in RA context, such as chromatin modification. This study has several limitations: it is a cross-sectional study in which consecutive patients from standard clinical practice were recruited. These patients were treated with standard therapy, including immunosuppressants, at the time of the sample and clinical detail collection. Thus, the effects of specific treatments on the expression levels of miRNA or the molecules involved in their biogenesis in neutrophils was not analyzed. The isolation of neutrophils with anti-CD15 microbeads is a further potential limitation of this work, however, up to date no unique method for the isolation of neutrophil isolation has been accepted worldwide and differing techniques have shown either some activation or functional impairment of the cells and the presence of small amounts of contaminating cells. Choosing an adequate method to isolate neutrophils from SF is thus challenging. In our hands, after testing a variety of potential priming/ activation techniques and the percentage of contaminating cells, the isolation of

References 1. Deng GM, Lenardo M. The role of immune cells and cytokines in the pathogenesis of rheumatoid arthritis. Drug Discovery Today: Disease Mechanisms. 2006; 3(2):163-168. 2. Barbarroja N, Pérez-Sanchez C, RuizLimón P, et al. Anticyclic citrullinated protein antibodies are implicated in the development of cardiovascular disease in rheumatoid arthritis. Arterioscler Thromb Vasc Biol. 2014;34(12):2706-2716. 3. Song, YW, Kang EH. Autoantibodies in rheumatoid arthritis: rheumatoid factors and anticitrullinated protein antibodies. QJM. 2010;103(3):139-146. 4. Corsiero E, Pratesi F, Prediletto E, Bombardieri M, Migliorini P. NETosis as source of autoantigens in rheumatoid arthritis. Front Immunol. 2016;7:485. 5. Araki, Y., and Mimura, T. The mechanisms underlying chronic inflammation in rheumatoid arthritis from the perspective of the epigenetic landscape. J Immunol Res. 2016;2016:6290682. 6. Ceribelli A, Yao B, Dominguez-Gutierrez PR, Nahid MA, Satoh M, Chan EK. MicroRNAs in systemic rheumatic diseases. Arthritis Res Ther. 2011;13(4):229. 7. Fabian MR, Sonenberg N, Filipowicz W. Regulation of mRNA translation and stability by microRNAs. Ann Rev Biochem. 2010;79:351-379.

2260

neutrophils with anti-CD15 microbeads was chosen as a suitable approach to obtain sufficiently numbers of inactivated neutrophils. Nevertheless, a consensus on the selection of the right isolation method that allows the comparison of neutrophil paired samples from SF and PB is still needed. Acknowledgments We thank all the patients for their kind participation in this study. Funding This work was supported by grants from the Instituto de Salud Carlos III (CP15/00158, PI17/01316 and PI18/00837), cofinanciado por el fondo europeo de desarrollo regional de la Union Europea, ‘una manera de hacer Europa’, Spain, the Regional Health System (ref. PI 0285 2017), and the Spanish Inflammatory and Rheumatic Diseases Network (RIER, RD16/0012/0015). CL-P was supported by contracts from both the Junta de Andalucia and Spanish Ministry of Science and Innovation (Ramon y Cajal). NB was supported by Ministry of Health postdoctoral fellowship (Miguel Servet Programme). YJ-G was supported by a contract from the University of Cordoba [co-funded by the Research Plan of the University of Cordoba and the Operating Program of the European Regional Development Funds (ERDF) for Andalusia. CL-P was supported by a contract from the Junta de Andalucia] (Nicolas Monardes Programme). YJ-G was supported by a contract from the University of Cordoba (co-funded by the Research Plan of the University of Cordoba and the Operating Program of the European Regional Development Funds [ERDF]) for Andalusia).

8. Chendrimada TP, Gregory RI, Kumaraswamy E. TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing. Nature. 2005; 436(7051):740-744 9. Gregory RI, Chendrimada TP, Cooch N, Shiekhattar R. Human RISC couples microRNA biogenesis and posttranscriptional gene silencing. Cell. 2005;123(4):631640 10. Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell. 2009; 136(2):215–233. 11. Rajasekhar M, Olsson AM, Steel KJ, et al. MicroRNA-155 contributes to enhanced resistance to apoptosis in monocytes from patients with rheumatoid arthritis. J Autoimmun. 2017;79:53-62. 12. Zahler S, Kowalski C, Brosig A, Kuppat C, Becker BF, Gerlach, E. The function of neutrophils isolated by a magnetic antibody cell separation technique is not altered in comparison to a density centrifugation method. J Immunol Methods. 1997;200(12):173-179. 13. Duroux-Richard I, Jorgensen C, Apparailly F. What do microRNAs mean for rheumatoid arthritis? Arthritis Rheum. 2011; 64(1):11-20. 14. Kurowska-Stolarska M, Alivernini S, Ballantine L, et al. MicroRNA-155 as a proinflammatory regulator in clinical and experimental arthritis. Proc Natl Acad Sci U S A. 2011;108(27):11193-11198. 15. Yang B, Zhang JL, Shi YY, et al. Association

16.

17.

18.

19.

20.

21.

22.

study of single nucleotide polymorphisms in pre-miRNA and rheumatoid arthritis in a Han Chinese population. Mol Biol Rep. 2011;38(8):4913-4919. Graff JW, Powers LS, Dickson AM, et al. Cigarette smoking decreases global microRNA expression in human alveolar macrophages. PLoS One. 2012;7(8):e44066. Pérez-Sánchez C, Aguirre MA, Ruiz-Limón P, et al. Atherothrombosis-associated microRNAs in Antiphospholipid syndrome and Systemic Lupus Erythematosus patients. Sci Rep. 2016;6:31375. Kurzynska-Kokorniak A, Koralewska N, Pokornorwska M, et al. The many faces of Dicer: the complexity of the mechanisms regulating Dicer gene expresssion and enzyme activities. Nucleic Acids Res. 2015; 43(9):4365-4380. Gurol T, Zhou W, Deng Q. MicroRNAs in neutrophils: potential next generation therapeutics for inflammatory ailments. Immunol Rev. 2016;273(1):29-47. Cao M, Shikama Y, Kimura H, et al. Mechanisms of impaired neutrophil migration by microRNAs in myelodysplastic syndromes. J Immunol. 2017;198(5):1887-1899. Batliner J, Buehrer E, Fey M, and Tschan M. Inhibition of the miR-143/145 cluster attenuated neutrophil differentiation of APL cells. Leuk Res. 2012;36(2):237-240. Li M, He Y, Zhou Z, et al. MicroRNA-223 ameliorates alcoholic liver injury by inhibiting the IL-6-p47 oxidative stress pathway in neutrophils. Gut. 2017;66(4):705-715.

haematologica | 2020; 105(9)

Impaired microRNA processing in RA neutrophils

23. Chen H, Li L, Wang S, et al. Reduced miRNA-126 expression facilitates angiogenesis of gastric cancer through its regulation on VEGF-A. Oncotarget. 2014;5(23): 11873-85. 24. Miao C, Zhang J, Zhao K, et al. The significance of microRNA-148/152 family as a prognostic factor in multiple human malignancies: a meta-analysis. Oncotarget. 2017; 8(26):43344-43355. 25. Chen Y, Song YX, Wang ZN. The microRNA-148/152 family: multi-faceted players. Mol Cancer. 2013;12:43. 26. Canfrán-Duque A, Rotllan N, Zhang X, et al. Macrophage deficiency of miR-21 promotes apoptosis, plaque necrosis, and vas-

haematologica | 2020; 105(9)

cular inflammation during atherogenesis. EMBO Mol Med. 2017;9(9):1244-1262. 27. Wells AC, Daniels KA, Angelou CC, et al. Modulation of let-7 miRNAs controls the differentiation of effector CD8 T cells. Elife. 2017;6. 28. Liu Y, Zhou Y, Gong X, Zhang C. MicroRNA-30a-5p inhibits the proliferation and invasion of gastric cancer cells by targeting insulin-like growth factor 1 receptor. Exp Ther Med. 2017;14(1);173-180. 29. Hu H, Li H, He Y. MicroRNA-17 downregulates expression of the PTEN gene to promote the occurrence and development of adenomyosis. Exp Ther Med. 2017; 14(4):3805-3811.

30. Nandakumar P, Tin A, Grove ML, et al. MicroRNAs in the miR-17 and miR-15 families are downregulated in chronic kidney disease with hypertension. PLoS One. 2017;12(8):e0176734. 31. Gong J, Yu J, Lin H, et al. The role, mechanism and potentially therapeutic application of microRNA-29 family in acute myeloid leukemia. Cell Death Differ. 2013; 21(1):100-112. 32. Castro-Villegas C, Pérez-Sánchez C, Escudero A, et al. Circulating miRNAs as potential biomarkers of therapy effectiveness in rheumatoid arthritis patients treated with anti-TNFa. Arthritis Res Ther. 2015; 17:49.

2261

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2262-2272

Myeloproliferative Neoplasms

Ruxolitinib and interferon-a2 combination therapy for patients with polycythemia vera or myelofibrosis: a phase II study Anders Lindholm Sørensen,1,2 Stine Ulrik Mikkelsen,3,4 Trine Alma Knudsen,1 Mads Emil Bjørn,5 Christen Lykkegaard Andersen,3,6 Ole Weis Bjerrum,3 Nana Brochmann,1 Dustin Andersen Patel,1,2 Lise Mette Rahbek Gjerdrum,7 Daniel El Fassi,5 Torben A. Kruse,8 Thomas Stauffer Larsen,9 Hans Torben MouritsAndersen,10 Claus Henrik Nielsen,2 Christina Ellervik,5,11,12 Niels Pallisgaard,7 Mads Thomassen,8 Lasse Kjær,1 Vibe Skov1 and Hans Carl Hasselbalch1

Department of Hematology, Zealand University Hospital, Roskilde, Denmark; 2Institute for Inflammation Research, Center for Rheumatology and Spine Diseases, Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark; 3Department of Hematology, Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark; 4Biotech Research and Innovation Center, University of Copenhagen, Copenhagen, Denmark; 5 Herlev University Hospital, Copenhagen, Denmark; 6Department of Public Health, Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark; 7 Department of Pathology, Zealand University Hospital, Roskilde, Denmark; 8 Department of Clinical Genetics, Odense University Hospital, Odense, Denmark; 9 Department of Hematology, Odense University Hospital, Odense, Denmark; 10 Department of Hematology, South-West Jutlandic Hospital, Esbjerg, Denmark; 11 Department of Laboratory Medicine, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA and 12Data and Development Support, Region Zealand, Sorø, Denmark 1

ABSTRACT

Correspondence: ANDERS LINDHOLM SØRENSEN anderslindholmsorensen@hotmail.com Received: August 18, 2019. Accepted: December 20, 2019. Pre-published: January 16, 2020. doi:10.3324/haematol.2019.235648 ©2020 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.

2262

e report the final 2-year end-of-study results from the first clinical trial investigating combination treatment with ruxolitinib and low-dose pegylated interferon-a2 (PEG-IFNa2). The study included 32 patients with polycythemia vera and 18 with primary or secondary myelofibrosis; 46 patients were previously intolerant of or refractory to PEGIFNα2. The primary outcome was efficacy, based on hematologic parameters, quality of life measurements, and JAK2 V617F allele burden. We used the 2013 European LeukemiaNet and International Working GroupMyeloproliferative Neoplasms Research and Treatment response criteria, including response in symptoms, splenomegaly, peripheral blood counts, and bone marrow. Of 32 patients with polycythemia vera, ten (31%) achieved a remission which was a complete remission in three (9%) cases. Of 18 patients with myelofibrosis, eight (44%) achieved a remission; five (28%) were complete remissions. The cumulative incidence of peripheral blood count remission was 0.85 and 0.75 for patients with polycythemia vera and myelofibrosis, respectively. The Myeloproliferative Neoplasm Symptom Assessment Form total symptom score decreased from 22 [95% confidence interval (95% CI):, 16-29] at baseline to 15 (95% CI: 10-22) after 2 years. The median JAK2 V617F allele burden decreased from 47% (95% CI: 33-61%) to 12% (95% CI: 6-22%), and 41% of patients achieved a molecular response. The drop-out rate was 6% among patients with polycythemia vera and 32% among those with myelofibrosis. Of 36 patients previously intolerant of PEG-IFNa2, 31 (86%) completed the study, and 24 (67%) of these received PEG-IFNa2 throughout the study. In conclusion, combination treatment improved cell counts, reduced bone marrow cellularity and fibrosis, decreased JAK2 V617F burden, and reduced symptom burden with acceptable toxicity in several patients with polycythemia vera or myelofibrosis. #EudraCT2013-003295-12. haematologica | 2020; 105(9)

Ruxolitinib and interferon-a for PV and MF

Introduction This is the first clinical trial investigating combination treatment with ruxolitinib and pegylated interferon-a2 (PEG-IFNa2) for patients with chronic Philadelphia-negative myeloproliferative neoplasms. Hydroxyurea is the most frequently used cytoreductive agent, but some patients are intolerant of or resistant to this drug.1-3 PEGIFNa2 is another first-line treatment option, but its clinical use is limited by toxicity.3-12 Ruxolitinib reduces symptom burden in patients with polycythemia vera (PV) and primary or secondary myelofibrosis (MF), but the clinical benefit in patients with PV is still controversial.3,13-19 Ruxolitinib may increase the efficacy and tolerability of PEG-IFNa2,20,21 and we have reported promising results from a 1-year interim analysis, with an overall remission rate of 9% in PV patients and 39% in MF patients.22 PEG-IFNa2 effectively normalizes blood cell counts and may prevent disease-related thromboembolic complications in patients with essential thrombocythemia, PV, and MF.4-8,10,11,23-26 Furthermore, an anti-clonal activity of PEGIFNa2 has been shown by reductions of the Januskinase-2 (JAK2) V617F allele burden and sustained deep molecular, hematologic, and histological remission in some patients, even several years after cessation of treatment.5,6,23,27-29 However, some patients do not respond adequately to PEG-IFNa2, and treatment is limited by frequent toxicities and high discontinuation rates, of 10–50%, due to adverse events.5-8,10-12 In particular, inflammatory side effects such as fever, flu-like symptoms, fatigue, and autoimmune thyroiditis are troublesome.5-7,10,11 JAK1-2 inhibitor treatment with ruxolitinib reduces disease-related symptoms and splenomegaly, and it may prolong survival in patients with MF.13,14,16,18,19 Furthermore, ruxolitinib reduces elevated blood cell counts and symptom burden in patients with PV.15,17 Inflammation is a critical element of myeloproliferative neoplasms, and the effect of ruxolitinib seems to be mediated mainly by antiinflammatory mechanisms.20 Moreover, a reduction in the JAK2 V617F allele burden has been observed.30 Adding ruxolitinib to PEG-IFNa2 treatment may increase the efficacy and tolerability of PEG-IFNa2 by reducing inflammation.21,22 Combination treatment may also have a synergistic effect on the malignant clone, and reduced dosage of both drugs may lead to fewer side-effects compared with monotherapies.20 The rationales for the combination treatment have previously been described in detail.21 We report the 2-year end-of-study results from the phase II COMBI study assessing efficacy and safety of combination treatment with ruxolitinib and PEG-IFNa2.

to the 2008 World Health Organization criteria were considered eligible, if they had evidence of active disease, defined as one or more of the following: need for phlebotomy, white blood cell count ≥10x09/L, platelet count ≥400x109/L, constitutional symptoms, pruritus, symptomatic splenomegaly, and previous thrombosis. Key exclusion criteria were: Eastern Cooperative Oncology Group performance status ≥3, severe comorbidity, white blood cell count <1.5x109/L, and platelet count <100x109/L. Patients were initially treated with PEG-IFNa2a [Pegasys®; Genentech (Roche), South San Francis-co, CA, USA] 45 μg/week or PEG-IFNa2b (PegIntron®; Merck Sharp & Dohme, Hertfordshire, UK) 35 μg/week subcutaneously and ruxolitinib (Jakavi®; Novartis, Basel, Switzerland) 5-20 mg BID orally depending on platelet count. Study visits included documentation of adverse events, fullscale hematology, blood biochemistry investigations, determination of the Myeloproliferative Neoplasm Symptom Assessment Form (MPN-SAF) score,31 and assessment of compliance by research staff. Bone marrow biopsies were done at baseline and after 1 and 2 years of treatment. Spleen size was measured as the longest diameter by sonography. The proportions of JAK2 V617F and CALR-mutated alleles were quantified using a high-sensitivity real-time quantitative polymerase chain reaction method on whole blood.32,33

Endpoints The primary outcome was efficacy, based on hematologic parameters, quality of life measurements, and the JAK2 V617F burden. The 2013 European LeukemiaNet and International Working Group-Myeloproliferative Neoplasms Research and Treatment response criteria were used to assess efficacy.34,35 In brief, for patients with PV, a complete remission (CR) required resolution of disease-related symptoms and hepatosplenomegaly, peripheral blood count remission (PBCR), no progression of the disease, no hemorrhagic or thrombotic events, and bone marrow histological remission (BMHR). A partial remission (PR) required all the above except BMHR. For patients with MF, a CR required resolution of disease-related symptoms and hepatosplenomegaly, and PBCR. A PR required resolution of disease-related symptoms and hepatosplenomegaly, and either PBCR or BMHR. Clinical improvement was defined as an anemia response or symptoms response. Progressive disease was defined as significantly increased splenomegaly or leukemic transformation.34,35 Molecular response (MR) was classified as either complete (CMR), defined as eradication of a pre-existing abnormality, or partial (PMR), defined as a ≥50% decrease in JAK2 V617F allele burden from baseline in patients with a baseline allele burden ≥20%.

Statistical analyses Methods Study design The COMBI study (#EudraCT2013-003295-12) was an investigator-initiated, multicenter, open-label, single-arm phase II study; we included 50 patients: 32 with PV and 18 with MF. The study was conducted between 2014 and 2018 at three sites in Denmark and approved by the Danish Regional Science Ethics Committee and the Danish Medicines Agency. It was done under the principles of the Declaration of Helsinki. Patients gave written informed consent to their participation. For further details on the methods, see the Online Supplementary Methods section. Patients aged ≥18 years with a diagnosis of PV or MF according haematologica | 2020; 105(9)

We did the statistical analyses using R.3.2.3. Response rates are presented with descriptive statistics. All patients initiating the study treatment were evaluated for response, similar to the modified intention-to-treat principle used for randomized, controlled trials. Statistical analyses are described in the Online Supplementary Methods section. P values <0.05 were considered statistically significant.

Results Patients’ characteristics In total, we included 51 patients in the study, and 50 patients initiated combination treatment; 32 had PV, and 2263

A.L. Sørensen et al.

19 had MF (Table 1). One patient with MF developed acute myeloid leukemia and died shortly after inclusion, before initiation of the combination treatment. The patient was excluded from further analysis. The MF patients were in an early stage of disease, with three (17%) having intermediate-2 risk and none having highrisk disease based on the Dynamic International Prognostic Scoring System-Plus (DIPSS+) score. Notably, 47 (94%) of the patients had previously been treated with PEG-IFNa2; 31 (66%) were intolerant, ten (21%) were refractory, and six (13%) were both. Moreover, 27 (54%) had been treated with hydroxyurea, and 25 (50%) had been treated with both hydroxyurea and PEG-IFNa2.

Response evaluations Of 32 patients with PV, ten (31%) achieved a remission; three (9%) achieved CR, and seven (22%) achieved PR (Figure 1A). Three additional patients achieved BMHR but did not fulfill the criteria for remission; one patient had an increase in total symptom score (TSS); one patient had elevated platelet count, and one patient had both. One patient had progressive disease due to the development of post-PV MF. Three PV patients dropped out and were classified as having no response. One patient chose not to have a bone marrow biopsy done after 2 years, and five patients had an unsuccessful biopsy. Two of these were in PR. Of 18 patients with MF, eight (44%) achieved remission; five (28%) achieved CR, and three (17%) achieved PR. Moreover, two (12%) had clinical improvement, both having symptoms response with a ≥50% reduction in TSS (Figure 1A). One additional patient achieved BMHR and PBCR but suffered from grade 1 fatigue and was defined as having stable disease. Two patients had stable disease. Five patients dropped out but did not have progressive disease.

Peripheral blood cell count remission Of 32 PV patients, five fulfilled the criteria for PBCR at baseline. For the 27 patients who did not have PBCR at baseline, the median time to PBCR was 1 month, and the cumulative incidence of PBCR after 2 years was 0.85 (Figure 1B). Of the five patients with PBCR at baseline, four had PBCR at 2 years. The median duration of the first PBCR was 14 months (Online Supplementary Figure S1), but 12 of the 13 patients who lost PBCR achieved it again during the study period. The proportion of PV patients in PBCR during the study is shown in Figure 1C. Of 14 PV patients in need of phlebotomies within 3 months before inclusion, four needed phlebotomies during the trial, three of whom required just one phlebotomy. Two additional patients required phlebotomies during the trial. Of 18 MF patients, three fulfilled the criteria for PBCR at baseline. For the 15 patients who did not have PBCR at baseline, the median time to PBCR was 3 months, and the cumulative incidence of PBCR after 2 years was 0.73 (Figure 1B). Of the three patients with PBCR at baseline, two had PBCR after 2 years. The median duration until the first PBCR was 5 months (Online Supplementary Figure S1), but seven of the eight patients who lost PBCR achieved it again during the study period. The proportion of MF patients in PBCR during the study is shown in Figure 1D. Hematocrit, white blood cell count, and platelet count 2264

were all significantly reduced after 2 weeks, and throughout the study period (Figure 1E-G); there were no significant differences between patients with PV and those with MF.

Patient-reported quality-of-life outcomes During the 2 years of treatment, we observed a statistically significant reduction in MPN-SAF TSS from baseline at all time points except at 1 and 2 years (Figure 2A). The median TSS was reduced from 22 [95% confidence interval (95% CI): 16-29] at baseline to 15 (95% CI: 10-22) after 2 years. Compared with patients not achieving remission, patients in remission after 2 years had significantly (P<0.05) larger reductions in TSS at several time points (figure 2B). The median TSS in patients achieving remission was reduced from 17 (95% CI: 10-27) at baseline to 7 (95% CI: 4-13) after 2 years. The median TSS in patients not achieving remission was 26 (95% CI: 18-37) at baseline and 25 (95% CI: 16-36) after 2 years. The following items of the TSS were significantly reduced at more than half of the time points, compared with baseline: early satiety (P<0.05), night sweats (P<0.01), itching (P<0.01) and weight loss (P<0.001) (Figure 2C). We found no significant difference in TSS change between patients with MF and patients with PV.

Table 1. Patients’ characteristics at baseline.

Polycythemia vera (n=32) Age (years), median (IQR) 57 (49, 67) Sex, male (%) 19 (59) Post-ET or -PV myelofibrosis, n (%) Years since diagnosis, median (IQR) 6.9 (2.9, 10.0) DIPSS+ score for myelofibrosis patients Low, n (%) Intermediate-1, n (%) Intermediate-2, n (%) High-risk PV, n (%) 21 (66) Prior thrombosis, n (%) 9 (28) Constitutional symptoms, n (%) 16 (50) Need of phlebotomies in last 3 months, n (%) 14 (44) Prior cytoreductive treatment PEG-IFNa2, n (%) 30 (94) HU, n (%) 21 (66) PEG-IFNa2 and HU, n (%) 19 (59) Anagrelide, n (%) 8 (25) JAK2 V617F positive, n (%) 32 (100) JAK2 V617F allele burden (%), median (IQR) 40 (19, 79) Palpable spleen, n (%) 6 (19) Spleen size by US (cm), median (IQR) 14.0 (12.1, 17.3) MPN-SAF TSS, median (IQR) 21.0 (5.5, 35.0) Hemoglobin (g/dL), median (IQR) 13.3 (12.7, 13.9) Hematocrit, median (IQR) 0.42 (0.41, 0.44) WBC count (x 109/L), median (IQR) 8.4 (5.4, 12.7) Platelet count (x 109/L), median (IQR) 401 (251, 592)

Myelofibrosis (n=18) 59 (51, 67) 10 (56) 5 (27.8) 6.2 (2.6, 8.0) 6 (33) 9 (50) 3 (17) 2 (11) 8 (44) 0 (0) 17 (94) 6 (33) 6 (33) 5 (28) 12 (67) 45 (23, 62) 3 (17) 14.0 (12, 18.5) 24.0 (18.5–41.0) 12.8 (12.0, 13.4) 0.40 (0.35, 0.41) 7.8 (5.4, 10.5) 419 (381, 464)

IQR: interquartile range; ET: essential thrombocythemia; PV: polycythemia vera; DIPSS+: Dynamic International Prog-nostic Scoring System-Plus; PEG-IFNa2: pegylated interferon-a2; HU, hydroxyurea; US: ultrasonography; MPN-SAF TSS, Myeloproliferative Neoplasm Symptom Assessment Form total symptom score; WBC: white blood cell.

haematologica | 2020; 105(9)

Ruxolitinib and interferon-a for PV and MF

Spleen size In nine patients, five with PV and four with MF, palpable splenomegaly was present at baseline. After 2 years, four of these patients no longer had palpable splenomegaly, two had reduced palpable splenomegaly, one had increased palpable splenomegaly, and two had discontinued treatment. In one of the two patients who dropped out, palpable splenomegaly was reduced before drop-out; in the other, it increased. Spleen size, measured by sonography, was reduced during treatment (P<0.001)

(Online Supplementary Figure S2), and after 2 years of therapy, spleen size was reduced by 10% (95% CI: 6-15%). The reduction was statistically significant for PV patients at all time points, but not for MF patients at 6 months and subsequently (Online Supplementary Figure S2).

Molecular response We observed statistically significant reductions in the JAK2 V617F allele burden at all time points (P<0.001) (Figure 3A). The median JAK2 V617F allele burden after 2

Figure 1. Response evaluations and peripheral blood cell remission. (A) Response evaluations after 2 years of treatment in patients with polycythemia vera (PV) or myelofibrosis (MF). Responses are classified as complete remission (CR), partial remission (PR), no response (NR), progressive disease (PD), drop-outs (DO), and, for MF patients, clinical improvement (CI) and stable disease (SD) (B) Kaplan-Meier plot depicting the cumulative incidence of peripheral blood count remission (PBCR) in patients not in PBCR at baseline (n=42). (C, D) Distribution of participants in PBCR and the number of patients dropping out of the protocol, patients with PV, and patients with MF. (E-G) Estimated change, with 95% confidence interval, in hematocrit (HCT) (E), white blood cell count (WBC) (F), and platelet count (PLT) (G), using generalized linear mixed models.

haematologica | 2020; 105(9)

2265

A.L. SĂ¸rensen et al.

years was 12% (95% CI: 6-22%) compared with 47% (95% CI: 33-61%) at baseline. In 33 of 44 JAK2 V617Fmutated patients, measurements of the JAK2 V617F allele burden were available from within the year before inclusion. There were no significant differences in the JAK2 V617F allele burden before inclusion compared with baseline. Of the 44 patients with the JAK2 V617F mutation, one (2%) achieved CMR, and 17 (39%) achieved PMR within 2 years of treatment (Figure 2B). Setting the limit of CMR at 1%, four (9%) patients achieved CMR, and 16 (36%) achieved PMR within 2 years of treatment (Online Supplementary Figure S3). Further analyses were done with a limit of CMR 1-5% (Online Supplementary Table S1). We observed no difference in the JAK2 V617F allele burden change between patients with PV and patients with MF (Online Supplementary Figure S4). The JAK2 V617F allele

burden of patients who did not complete 2 years of treatment (n=5) is shown in Online Supplementary Figure S5; two of these patients had reductions in JAK2 V617F allele burden. Of four patients with a CALR mutation, one had a decrease in allele burden during treatment, two had an increase, and one dropped out of the study before a second measurement could be done (Online Supplementary Figure S6). Patients achieving remission after 2 years had a statistically significant greater reduction in JAK2 V617F allele burden compared with patients not achieving remission (Figure 2C). Of the 39 patients with the JAK2 V617F mutation who completed the study, 15 had a MR and 24 did not; 11 (73%) patients with a MR achieved remission, while five (21%) without a MR achieved remission

Figure 2. Change in Myeloproliferative Neoplasm Symptom Assessment Form (MPN-SAF) total symptom score and individual symptoms. (A) Median total symptom score (TSS) with 95% confidence interval (95% CI). (B) The median TSS with 95% CI in participants with remission after 2 years and participants without remission. (C) Individual symptom scores during treatment. Significance is defined as a statistically significant reduction compared with baseline at more than half of the time points. All analyses were done using generalized linear mixed models comparing baseline values with measurements during treatment. *P<0.05; **P<0.01; ***P<0.001

2266

haematologica | 2020; 105(9)

Ruxolitinib and interferon-a for PV and MF

(P=0.003). Likewise, 14 (93%) patients with a MR were in PBCR, while 12 (50%) without MR were in PBCR (P=0.014). Moreover, patients with a MR at 2 years had a higher rate of PBCR (P=0.025) with a median time to PBCR of 1 month, compared with 6 months for patients without a MR (Figure 2D).

Drop-out and toxicity Of the 32 PV patients who initiated treatment, two (6%) dropped out within 2 years. Of 18 MF patients, six (33%) dropped out. In total, the drop-out rate was 16%. Three patients had an inadequate response; four had adverse events; among these, two had neuropsychiatric adverse events, one had multiple infections and gastrointestinal bleeding and requested to be taken off protocol (Online Supplementary Table S2). Eight (25%) PV patients and one

(6%) MF patient discontinued PEG-IFNa2 treatment while continuing treatment with ruxolitinib, and one PV (3%) patient discontinued ruxolitinib treatment. In total, 17 (34%) patients, ten (31%) with PV and seven (39%) with MF, discontinued PEG-IFNa2; ten (59%) of these due to side-effects likely related to PEG-IFNa2 (Table 2). Of the 50 patients, 32 (64%) had the dosage of ruxolitinib reduced, and 35 (70%) had the dosage of PEG-IFNa2 reduced. The prevalence of adverse events, including diseaserelated and study drug-unrelated events, observed in â&#x2030;Ľ10% of the patients and all grade 3 or 4 adverse events, are presented in Table 2. Four thromboembolic events occurred in three patients, two patients with PV and one with MF: the events were myocardial infarction in one patient, ischemic stroke in one patient, and both retinal

Figure 3. Change in JAK2-V617F allele burden and molecular response. (A) Median JAK2 V617F allele burden (JAK2) with 95% confidence interval (95% CI), using a generalized linear mixed model to compare baseline with time points during treatment. (B) Waterfall plot over the best relative reduction in JAK2 with an indication of complete molecular response (CMR), partial molecular response (PMR), and no molecular response (NMR) with baseline value below or above 20%. (C) Median JAK2 and 95% CI in patients in partial or complete remission after 24 months and patients not in remission with stars representing a statistically significant difference in change from baseline between groups. (D) Cumulative incidence of peripheral blood cell remission (PBCR) in patients achieving a molecular response after 2 years and patients not achieving a molecular response. *P<0.05; **P<0.01; ***P<0.001

haematologica | 2020; 105(9)

2267

A.L. Sørensen et al.

artery occlusion and multiple small pulmonary emboli in one PV patient. Grade 1 or 2 hematologic adverse events, and grade 3 anemia in patients with MF, were frequent. We only recorded a few other grade 3 or 4 hematologic adverse events. With regards to non-hematologic adverse events, fatigue, influenza-like symptoms, pruritus, upper airway infection, joint pain, muscle pain, and elevated lactate dehydrogenase were observed in ≥30% of patients.

Six (12%) patients had grade 3 or 4 pneumonia, and one (2%) had a grade 3 or 4 urinary tract infection. Two (4%) patients had herpes zoster infection grade 1 or 2. In 11 (22%) patients, mood alterations, including depressive symptoms, agitation, anxiety, and memory impairment, were observed. No autoimmune thyroiditis occurred in our study, but one patient was diagnosed with Sjögren syndrome. All the drug-related adverse events are known

Table 2. Adverse events reported in ≥ 10% of patients and all grade 3 or 4 adverse events.

Overall (n=50) Drop-outs, n (%) 8 (16%) Total discontinuation of PEG-IFNa2, n (%)a 17 (34%) Adverse events Grade 1–2, Grade 3–4, N. (%) N. (%) Hematologic, n (%) Anemia Thrombocytopenia Leukopenia Non-hematologic, n (%) Cardiac ischemia Palpitations Arterial hypertension Pulmonary embolism Fatigue Night sweats Influenza-like symptomsb Fever Pruritus Local injection site reactionsb Dyspepsia Nausea Abdominal pain Gastrointestinal bleedingc Upper airway infection Pneumonia Urinary tract infection Weight gain Joint pain Muscle pain Mood alterationd Neuropathy: sensory Dizziness CNS ischemiae Syncope Headache Dyspnea Acute renal failure Proteinuria Benign bladder tumor Elevated alanine transaminase Elevated lactate dehydrogenase

Polycythemia vera (n=32) 2 (6%) 10 (31%) Grade 1–2, Grade 3–4, N. (%) N. (%)

Myelofibrosis (n=18) 6 (32%) 7 (39%) Grade 1–2, Grade 3–4, N. (%) N. (%)

38 (76.0) 12 (24.0) 22 (44)

7 (14.0) 2 (4.0) 1 (2.0)

23 (71.9) 9 (28.1) 15 (46.9)

1 (3.1) 2 (6.2) 1 (3.1)

15 (83.3) 3 (16.7) 7 (38.9)

6 (33.3) 0 (0.0) 0 (0.0)

0 (0.0) 5 (10.0) 2 (4.0) 0 (0.0) 23 (46.0) 14 (28.0) 15 (30.0) 9 (18.0) 20 (40.0) 10 (20.0) 6 (12.0) 14 (28.0) 6 (12.0) 0 (0.0) 23 (46.0) 3 (6.0) 4 (8.0) 7 (14.0) 19 (38.0) 21 (42.0) 11 (22.0) 12 (24.0) 11 (22.0) 0 (0.0) 0 (0.0) 12 (24.0) 8 (16.0) 0 (0.0) 0 (0.0) 0 (0.0) 7 (14.0) 17 (34.0)

1 (2.0) 0 (0.0) 3 (6.0) 1 (2.0) 1 (2.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 3 (6.0) 0 (0.0) 6 (12.0) 1 (2.0) 0 (0.0) 2 (4) 1 (2.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (3.8) 1 (2.0) 0 (0.0) 0 (0.0) 1 (2.0) 1 (2.0) 1 (2.0) 0 (0.0) 0 (0.0)

0 (0.0) 1 (5.6) 2 (6.2) 0 (0.0) 15 (46.9) 10 (31.2) 11 (34.4) 6 (18.8) 17 (53.1) 6 (18.8) 3 (9.4) 9 (28.1) 3 (9.4) 0 (0.0) 15 (46.9) 3 (9.4) 2 (6.2) 5 (15.6) 14 (43.8) 16 (50.0) 8 (25.0) 7 (21.9) 6 (18.8) 0 (0.0) 0 (0.0) 7 (21.9) 5 (15.6) 0 (0.0) 0 (0.0) 0 (0.0) 6 (18.8) 12 (37.5)

1 (3.1) 0 (0.0) 3 (9.4) 1 (3.1) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 5 (15.6) 0 (0.0) 0 (0.0) 2 (6.2) 1 (3.1) 0 (0.0) 0 (0.0) 0 (0.0) 1 (3.1) 1 (3.1) 0 (0.0) 0 (0.0) 1 (3.1) 1 (3.1) 1 (3.1) 0 (0.0) 0 (0.0)

0 (0.0) 4 (12.5) 0 (0.0) 0 (0.0) 8 (44.4) 4 (22.2) 4 (22.2) 3 (16.7) 3 (16.7) 4 (22.2) 3 (16.7) 5 (27.8) 3 (16.7) 0 (0.0) 8 (44.4) 0 (0.0) 2 (11.1) 2 (11.1) 5 (27.8) 5 (27.8) 3 (16.7) 5 (27.8) 5 (27.8) 0 (0.0) 0 (0.0) 5 (27.8) 3 (16.7) 0 (0.0) 0 (0.0) 0 (0.0) 1 (5.3) 5 (26.3)

0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 1 (5.6) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 3 (16.7) 0 (0.0) 1 (5.6) 1 (5.6) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 1 (5.6) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)

Adverse events were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events (version 3.0). aThe total number includes patients who dropped out of the study. bRelated to pegylated interferon-a2 treatment. cOne patient had esophageal varices with bleeding, one had Mallory Weiss lesions, and one had rectal bleeding. dFive patients had depressive symptoms, three experienced agitation, two anxiety, and two had memory impairment while one had both anxiety and memory impairment. eOne patient had an ischemic stroke and a retinal artery occlusion. PEG-IFNa2: peglylated interferon-a2; CNS: central nervous system

2268

haematologica | 2020; 105(9)

Ruxolitinib and interferon-a for PV and MF

toxicities of either ruxolitinib or PEG-IFNa2. The drug doses and adverse events over time are shown in Figure 4. The highest number of adverse events was observed within the first 3 months of treatment.

Subgroup analyses Of 32 PV patients who initiated treatment, 20 completed the study per protocol. Of these 20 patients, ten (50%) achieved remission; three (15%) had CR, and seven (35%) had PR. Eight of the 20 patients (40%) had a MR. At 2 years, 16 (80%) were in PBCR. Of the 18 MF patients who initiated treatment, 12 completed the study per protocol. Of these 12 patients, eight (67%) achieved remission; five (42%) had CR, three (25%) had PR. Two (13%) had a clinical improvement. Six patients (50%) had a MR, and nine (75%) were in PBCR. Eight patients with the JAK2 V617F mutation discontinued PEG-IFNa2 and continued ruxolitinib as monotherapy. Their JAK2 V617F allele burdens are shown in Online Supplementary Figure S7. Of these patients, three had PBCR, and none had remission at 24 months. Three achieved MR; however, two of these continued PEGIFNa2 until 15 and 21 months.

We stratified the main results based on the reason for the prior discontinuation of PEG-IFNa2 (Online Supplementary Table S3). Patients who were previously intolerant of PEG-IFNa2 or naĂŻve to this treatment had more substantial reductions in the JAK2 V617F allele burden compared with patients previously refractory to PEGIFNa2 treatment (mean reductions 61%, 65%, and 34%, respectively). None of the three PEG-IFNa2-naĂŻve patients dropped out or discontinued PEG-IFNa2 treatment, and two achieved remission and MR. We observed no other notable differences.

Discussion In this study, we showed that a novel combination of ruxolitinib and low-dose PEG-IFNa2 is an effective treatment with acceptable toxicity for patients with PV or MF. We observed remission rates of 31% for patients with PV and 44% for patients with MF. Moreover, both groups had relatively high rates of sustained PBCR. Furthermore, we found statistically significant reductions in the MPN-SAF TSS, spleen size, and JAK2 V617F allele burden. The drop-

Figure 4. Adverse events and drug doses in patients with polycythemia vera (left) and myelofibrosis (right). The number of adverse events per patient during the study period with mean doses of pegylated interferona2 (PEG-IFNa2) and ruxolitinib (RUX). Gr. 1-2 non-hem AEs: grade 12 non-hematologic adverse events; gr. 3-4 non-hem AEs: grade 3-4 nonhematologic adverse events; All gr. hem AEs: All grade hematologic adverse events.

haematologica | 2020; 105(9)

2269

A.L. Sørensen et al.

out rate was comparable with that in previous studies of PEG-IFNa2 in interferon-naïve patients.4-8,10,11 The preliminary results from the phase I/II RUXOPEG study (ClinicalTrials.gov NCT02742324) in patients with more advanced MF support our findings.36 In that study, a decrease in spleen size, improvement in blood counts, and a reduction in JAK2 V617F allele burden were observed in ten evaluable patients. In total, 19% of PV patients and 33% of MF patients fulfilled the criteria for BMHR. These findings suggest a disease-modifying effect of the combination treatment.34,35 Profound effects of PEG-IFNa2 on bone marrow histology have been observed before, but usually after longer periods of treatment.28,29,37,38 Long-term treatment with ruxolitinib in patients with MF seems to decrease bone marrow fibrosis in only a small subset of patients.39 In a study investigating PEG-IFNa2 treatment in 30 patients with low- or intermediate-1 risk MF, two (7%) had CR, nine (30%) had PR, and four (13%) had clinical improvement after a median of 80.3 months.37 In our study, 28%, 17%, and 12% of MF patients achieved CR, PR, and clinical improvement, respectively, within 2 years. Normalization of blood cell counts may be critical in reducing the risk of thrombosis and improving survival.1,40,41 Therefore, a key finding of this study is the fast and sustained normalization of elevated blood cell counts. Similar high rates of hematologic response were observed previously with a starting dose of PEG-IFNa2 90 µg/week in patients with PV.5-7 In two phase II studies on PV patients, the median time to complete hematologic response was 2 and 3 months, respectively.5,6 In our study, the median time to PBCR was 1 month for PV patients. Similar to our study, most patients in one of those studies had received previous cytoreductive treatment.6 In the phase II study of ropeginterferon-a2b for PV patients, the median time to complete hematologic resposne was more than 1 year.42 Notably, baseline cell counts were higher in these studies than in ours, and fewer patients had received cytoreductive treatment before inclusion in two of the studies, which may explain the shorter time to response for PV patients in our study. Importantly, the response criteria are not identical in the studies, and comparisons between the studies should, therefore, be interpreted with caution. Ropeginterferon-a2b induced higher rates of blood cell count and spleen responses than did hydroxyurea after 2 years in an ongoing randomized controlled trial.43 Ruxolitinib monotherapy in patients with PV previously treated with hydroxyurea, in the RESPONSE I and II studies, resulted in PBCR rates of only 24% and 17% after 6 months.15,17 In comparison, 80% of PV patients in our study achieved PBCR within 6 months. The MPN-SAF TSS decreased significantly during combination treatment. Ruxolitinib has previously been shown to reduce symptom burden in patients with MF or PV with relatively few side-effects.14,15,17,19 In contrast, in a study comparing hydroxyurea with PEG-IFNa2, no reduction in TSS was observed in either group.44 The marked reduction in the JAK2 V617F allele burden and a 41% MR rate add further evidence of a selective effect of combination treatment on the malignant clone. Similar findings were made in patients treated with PEGIFNa2, and in the RESPONSE I study in which 34% achieved MR after a median of 25 months of ruxolitinib treatment.5,6,23,30 Notably, the baseline median JAK2 V617F allele burden was 83% in the RESPONSE I study com2270

pared with 47% in our study. It is likely that a higher proportion of patients in RESPONSE I was evaluable for PMR than in our study. Ruxolitinib treatment in patients with MF has resulted in markedly lower rates of MR.45 Importantly, most previous studies assessing the JAK2 V617F allele burden have a limit of detection of ≥1%.5,6.30,42,45 The assay used in our study has a sensitivity of ≤0.1% mutated alleles; this is important for comparisons of CMR and the total rate of MR. We found an association between MR and both remission and PBCR. Similarly, during ropeginterferon-a2b treatment, an association between MR and hematologic response was observed.42,43 Likewise, in a phase II study of PEG-IFNa2, 94% of patients with MR also had a hematologic response, and ten of 13 patients with complete bone marrow response, also achieved MR; seven had CMR.11,29 The clinical benefit of MR is still unclear, but higher JAK2 V617F allele burden may be associated with a higher risk of thrombotic events and progression to MF.41,46 The drop-out rate in our study was 16%; 8% stopped due to adverse events. Notably, the drop-out rate was 6% in PV patients and 32% in MF patients, respectively. In total, 31% of PV patients and 37% of MF patients discontinued PEG-IFNa2 when including patients who continued ruxolitinib monotherapy and drop-outs. This proportion was similar in the two groups. Similar rates were reported in previous studies on PEG-IFNa2 treatment.48,10,11 However, in our study, 94% had previously been treated with PEG-IFNa2; we therefore expected a high rate of PEG-IFNa2 discontinuation. Moreover, in an ongoing Danish study, the rate of PEG-IFNa2 discontinuation in treatment-naïve patients was approximately 50%.12 Consequently, our findings are still very encouraging. Ruxolitinib has been shown to increase the risk of infections, particularly in patients with MF.13,15,17,19,47 We observed a relatively high infection rate, with six patients (12%) developing grade 3 pneumonia. Importantly, patients with fever were hospitalized and treated with intravenous antibiotics, immediately for safety precautions, and thereby fulfilled the criteria for grade 3 or 4 adverse events. This may have led to an overestimation of grade 3 or 4 infectious adverse events. Indeed, only three of the six grade 3 or 4 cases of pneumonia registered were radiologically or microbiologically verified. Grade 1 or 2 anemia was recorded in 71.9% of PV patients. A high initial dose of ruxolitinib may account for this. The initial ruxolitinib dose used in this study was 20 mg BID for most patients compared with 10 mg BID in the RESPONSE studies, since the phase II study dose-finding study was published after preparation of the protocol.15,17 Accordingly, dosage reduction was frequent. The inclusion criteria used in this trial do not directly reflect the European LeukemiaNet guidelines for the initiation of cytoreductive treatment.3 Indeed, both high- and low-risk PV patients were included. This is similar to other clinical trials investigating PEG-IFNa2 treatment.5,6,42 The indications for administering PEG-IFNa2 are wider in Denmark than internationally, and a strategy of an early intervention targeting the malignant clone has previously been described.48 Indeed, combination treatment as an early intervention may induce deep clinical, histological and molecular remission more frequently than PEG-IFNa2 monotherapy.5,6,23,27-29 Achieving deep remission could lead to periods of observation without treatment or maintenance treatment with a low dose of PEG-IFNa2, which haematologica | 2020; 105(9)

Ruxolitinib and interferon-a for PV and MF

could greatly improve the quality of life of patients with myeloproliferative neoplasms. Moreover, the risk of thrombosis is greatest shortly after diagnosis, and aggressive initial treatment may reduce the risk of early thrombosis.49 To this end, a new phase II study investigating combination therapy in newly diagnosed PV patients has been initiated (#EudraCT2018-0041-50). One strength of our study is the investigation of bone marrow remissions, which is essential in assessing a possible disease-modifying effect of new treatments.34,35 Additionally, we used the validated MPN-SAF questionnaire to assess the symptom burden during the treatment. In patients with myeloproliferative neoplasms, in whom it is essential to balance the efficacy of a treatment with its toxicity, quality of life is a clinically relevant endpoint.34,35 Furthermore, we used a highly sensitive method to assess the JAK2 V617F allele burden.32 A limitation of our study is the relatively small population of MF patients. Accordingly, results in this group should be interpreted with caution. Moreover, the study lacked a dose-finding phase I part, which would have been relevant. Furthermore, the study was planned for 2 years of treatment, but hematologic and histological responses have been observed after several years of PEG-IFNa treat-

References 1. Alvarez-Larran A, Pereira A, Cervantes F, et al. Assessment and prognostic value of the European LeukemiaNet criteria for clinicohematologic response, resistance, and intolerance to hydroxyurea in polycythemia vera. Blood. 2012;119(6):1363-1369. 2. Parasuraman S, DiBonaventura M, Reith K, et al. Patterns of hydroxyurea use and clinical outcomes among patients with polycythemia vera in real-world clinical practice: a chart review. Exp Hematol Oncol. 2015;5:3. 3. Barbui T, Tefferi A, Vannucchi AM, et al. Philadelphia chromosome-negative classical myeloproliferative neoplasms: revised management recommendations from European LeukemiaNet. Leukemia. 2018;32(5):10571069. 4. Samuelsson J, Hasselbalch H, Bruserud O, et al. A phase II trial of pegylated interferon alpha-2b therapy for polycythemia vera and essential thrombocythemia: feasibility, clinical and biologic effects, and impact on quality of life. Cancer. 2006;106(11):2397-2405. 5. Kiladjian JJ, Cassinat B, Chevret S, et al. Pegylated interferon-alfa-2a induces complete hematologic and molecular responses with low toxicity in polycythemia vera. Blood. 2008;112(8):3065-3072. 6. Quintas-Cardama A, Kantarjian H, Manshouri T, et al. Pegylated interferon alfa2a yields high rates of hematologic and molecular response in patients with advanced essential thrombocythemia and polycythemia vera. J Clin Oncol. 2009;27 (32):5418-5424. 7. Gowin K, Thapaliya P, Samuelson J, et al. Experience with pegylated interferon alpha2a in advanced myeloproliferative neoplasms in an international cohort of 118 patients. Haematologica. 2012;97(10):15701573. 8. Ianotto JC, Boyer-Perrard F, Gyan E, et al.

haematologica | 2020; 105(9)

10.

11.

12.

13.

14.

15.

16.

ment, and the response after 2 years may underestimate the long-term clinical effect.28,29,37,38 Most patients on combination treatment at the end of the study continued treatment and post-hoc analyses with longer follow-up are being planned to assess the long-term response to combination treatment. In conclusion, combination treatment with ruxolitinib and low-dose PEG-IFNa2 improved peripheral blood cell counts, bone marrow cellularity and fibrosis along with symptom burden with acceptable toxicity in some patients with PV and proliferative MF. Most patients in the study were intolerant of or refractory to standard PEGIFNa2 treatment, and more than half had discontinued previous treatment with hydroxyurea, highlighting that this combination treatment is a viable choice for patients with few treatment options left. Acknowledgments The authors thank the patients and their families and the research staff. Danish Telemedicine A/S developed the software system for the online MPN-SAF questionnaire. We thank Associate Professor Julie Lyng Formann from the Section of Biostatistics, University of Copenhagen, for guidance in repeated measurements statistics.

Efficacy and safety of pegylated-interferon alpha-2a in myelofibrosis: a study by the FIM and GEM French cooperative groups. Br J Haematol. 2013;162(6):783-791. Kiladjian JJ, Giraudier S, Cassinat B. Interferon-alpha for the therapy of myeloproliferative neoplasms: targeting the malignant clone. Leukemia. 2016;30(4):776-781. Gowin K, Jain T, Kosiorek H, et al. Pegylated interferon alpha - 2a is clinically effective and tolerable in myeloproliferative neoplasm patients treated off clinical trial. Leuk Res. 2017;54:73-77. Masarova L, Patel KP, Newberry KJ, et al. Pegylated interferon alfa-2a in patients with essential thrombocythaemia or polycythaemia vera: a post-hoc, median 83 month follow-up of an open-label, phase 2 trial. Lancet Haematol. 2017;4(4):e165-e175. Knudsen TA, Hansen DL, Ocias LF, et al. Long-term efficacy and safety of recombinant interferon alpha-2 vs. hydroxyurea in polycythemia vera: preliminary results from the three-year analysis of the daliah trial - a randomized controlled phase III clinical trial. Blood. 2018;132(Suppl 1):580-580. Harrison C, Kiladjian JJ, Al-Ali HK, et al. JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N Engl J Med. 2012;366(9):787-798. Verstovsek S, Mesa RA, Gotlib J, et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med. 2012;366(9):799-807. Vannucchi AM, Kiladjian JJ, Griesshammer M, et al. Ruxolitinib versus standard therapy for the treatment of polycythemia vera. N Engl J Med. 2015;372(5):426-435. Al-Ali HK, Griesshammer M, le Coutre P, et al. Safety and efficacy of ruxolitinib in an open-label, multicenter, single-arm phase 3b expanded-access study in patients with myelofibrosis: a snapshot of 1144 patients in the JUMP trial. Haematologica. 2016;101(9):1065-1073.

17. Passamonti F, Griesshammer M, Palandri F, et al. Ruxolitinib for the treatment of inadequately controlled polycythaemia vera without splenomegaly (RESPONSE-2): a randomised, open-label, phase 3b study. Lancet Oncol. 2017;18(1):88-99. 18. Verstovsek S, Gotlib J, Mesa RA, et al. Longterm survival in patients treated with ruxolitinib for myelofibrosis: COMFORT-I and -II pooled analyses. J Hematol Oncol. 2017;10 (1):156. 19. Palandri F, Tiribelli M, Benevolo G, et al. Efficacy and safety of ruxolitinib in intermediate-1 IPSS risk myelofibrosis patients: results from an independent study. Hematol Oncol. 2018;36(1):285-290. 20. Koschmieder S, Mughal TI, Hasselbalch HC, et al. Myeloproliferative neoplasms and inflammation: whether to target the malignant clone or the inflammatory process or both. Leukemia. 2016;30(5):1018-1024. 21. Bjorn ME, Hasselbalch HC. Minimal residual disease or cure in MPNs? Rationales and perspectives on combination therapy with interferon-alpha2 and ruxolitinib. Expert Rev Hematol. 2017;10(5):393-404. 22. Mikkelsen SU, Kjaer L, Bjorn ME, et al. Safety and efficacy of combination therapy of interferon-alpha2 and ruxolitinib in polycythemia vera and myelofibrosis. Cancer Med. 2018;7(8):3571-3581. 23. Stauffer Larsen T, Iversen KF, Hansen E, et al. Long term molecular responses in a cohort of Danish patients with essential thrombocythemia, polycythemia vera and myelofibrosis treated with recombinant interferon alpha. Leuk Res. 2013;37(9):1041-1045. 24. Silver RT. Long-term effects of the treatment of polycythemia vera with recombinant interferon-alpha. Cancer. 2006;107(3): 451-458. 25. Marchioli R, Finazzi G, Landolfi R, et al. Vascular and neoplastic risk in a large cohort of patients with polycythemia vera. J Clin Oncol. 2005;23(10):2224-2232.

2271

A.L. SĂ¸rensen et al. 26. Barbui T, Carobbio A, Rumi E, et al. In contemporary patients with polycythemia vera, rates of thrombosis and risk factors delineate a new clinical epidemiology. Blood. 2014;124(19):3021-3023. 27. Quintas-Cardama A, Abdel-Wahab O, Manshouri T, et al. Molecular analysis of patients with polycythemia vera or essential thrombocythemia receiving pegylated interferon alpha-2a. Blood. 2013;122(6):893-901. 28. Utke Rank C, Weis Bjerrum O, Larsen TS, et al. Minimal residual disease after long-term interferon-alpha2 treatment: a report on hematological, molecular and histomorphological response patterns in 10 patients with essential thrombocythemia and polycythemia vera. Leuk Lymphoma. 2016;57 (2):348-354. 29. Masarova L, Yin CC, Cortes JE, et al. Histomorphological responses after therapy with pegylated interferon alpha-2a in patients with essential thrombocythemia (ET) and polycythemia vera (PV). Exp Hematol Oncol. 2017;6:30. 30. Vannucchi AM, Verstovsek S, Guglielmelli P, et al. Ruxolitinib reduces JAK2 p.V617F allele burden in patients with polycythemia vera enrolled in the RESPONSE study. Ann Hematol. 2017;96(7):1113-1120. 31. Emanuel RM, Dueck AC, Geyer HL, et al. Myeloproliferative neoplasm (MPN) symptom assessment form total symptom score: prospective international assessment of an abbreviated symptom burden scoring system among patients with MPNs. J Clin Oncol. 2012;30(33):4098-4103. 32. Larsen TS, Pallisgaard N, Moller MB, et al. Quantitative assessment of the JAK2 V617F allele burden: equivalent levels in peripheral blood and bone marrow. Leukemia. 2008;22(1):194-195. 33. Kjaer L, Cordua S, Holmstrom MO, et al. Differential dynamics of CALR mutant allele burden in myeloproliferative neoplasms during interferon alfa treatment.

2272

PLoS One. 2016;11(10): e0165336. 34. Tefferi A, Cervantes F, Mesa R, et al. Revised response criteria for myelofibrosis: International Working GroupMyeloproliferative Neoplasms Research and Treatment (IWG-MRT) and European LeukemiaNet (ELN) consensus report. Blood. 2013;122(8):1395-1398. 35. Barosi G, Mesa R, Finazzi G, et al. Revised response criteria for polycythemia vera and essential thrombocythemia: an ELN and IWG-MRT consensus project. Blood. 2013;121(23):4778-4781. 36. Kiladjian J-J, Soret-Dulphy J, Resche-Rigon M, et al. Ruxopeg, a multi-center Bayesian phase 1/2 adaptive randomized trial of the combination of ruxolitinib and pegylated interferon alpha 2a in patients with myeloproliferative neoplasm (MPN)-associated myelofibrosis. Blood. 2018;132 (Suppl 1):581-581. 37. Silver RT, Barel AC, Lascu E, et al. The effect of initial molecular profile on response to recombinant interferon-alpha (rIFNalpha) treatment in early myelofibrosis. Cancer. 2017;123(14):2680-2687. 38. Silver RT, Vandris K, Goldman JJ. Recombinant interferon-alpha may retard progression of early primary myelofibrosis: a preliminary report. Blood. 2011;117(24): 6669-6672. 39. Kvasnicka HM, Thiele J, Bueso-Ramos CE, et al. Long-term effects of ruxolitinib versus best available therapy on bone marrow fibrosis in patients with myelofibrosis. J Hematol Oncol. 2018;11(1):42. 40. Marchioli R, Finazzi G, Specchia G, et al. Cardiovascular events and intensity of treatment in polycythemia vera. N Engl J Med. 2013;368(1):22-33. 41. Barbui T, Finazzi G, Falanga A. Myeloproliferative neoplasms and thrombosis. Blood. 2013;122(13):2176-2184. 42. Gisslinger H, Zagrijtschuk O, BuxhoferAusch V, et al. Ropeginterferon alfa-2b, a

43.

44.

45.

46.

47.

48.

49.

novel IFNalpha-2b, induces high response rates with low toxicity in patients with polycythemia vera. Blood. 2015;126(15): 1762-1769. Gisslinger H, Klade C, Georgiev P, et al. Evidence for superior efficacy and disease modification after three years of prospective randomized controlled treatment of polycythemia vera patients with ropeginterferon alfa-2b Vs. HU/BAT. Blood. 2018;132(Suppl 1):579-579. Mesa RA, Kosiorek HE, Mascarenhas J, et al. Impact on MPN symptoms and quality of life of front line pegylated interferon alpha2a vs. hydroxyurea in high risk polycythemia vera and essential thrombocythemia: results of myeloproliferative disorders research consortium (MPD-RC) 112 global phase III trial. Blood. 2018;132(Suppl 1):3032-3032. Deininger M, Radich J, Burn TC, et al. The effect of long-term ruxolitinib treatment on JAK2p.V617F allele burden in patients with myelofibrosis. Blood. 2015;126(13):15511554. Vannucchi AM, Pieri L, Guglielmelli P. JAK2 allele burden in the myeloproliferative neoplasms: effects on phenotype, prognosis and change with treatment. Ther Adv Hematol. 2011;2(1):21-32. Verstovsek S, Passamonti F, Rambaldi A, et al. A phase 2 study of ruxolitinib, an oral JAK1 and JAK2 inhibitor, in patients with advanced polycythemia vera who are refractory or intolerant to hydroxyurea. Cancer. 2014;120(4):513-520. Hasselbalch HC, Holmstrom MO. Perspectives on interferon-alpha in the treatment of polycythemia vera and related myeloproliferative neoplasms: minimal residual disease and cure? Semin Immunopathol. 2019;41(1):5-19. Griesshammer M, Kiladjian JJ, Besses C. Thromboembolic events in polycythemia vera. Ann Hematol. 2019;98(5):1071-1082.

haematologica | 2020; 105(9)

ARTICLE

Acute Myeloid Leukemia

SETDB1 mediated histone H3 lysine 9 methylation suppresses MLL-fusion target expression and leukemic transformation

Ferrata Storti Foundation

James Ropa,1,2 Nirmalya Saha,1* Hsiangyu Hu,1* Luke F. Peterson,3 Moshe Talpaz3 and Andrew G. Muntean1

Department of Pathology, University of Michigan Medical School; 2Department of Computational Medicine and Bioinformatics, University of Michigan Medical School and 3 Department of Internal Medicine/Division of Hematology/Oncology, University of Michigan School of Medicine and Comprehensive Cancer Center, Ann Abor, MI, USA 1

*NS and HH contributed equally as co-second authors.

Haematologica 2020 Volume 105(9):2273-2285

ABSTRACT

pigenetic regulators play a critical role in normal and malignant hematopoiesis. Deregulation, including epigenetic deregulation, of the HOXA gene cluster drives transformation of about 50% of acute myeloid leukemia (AML). We recently showed that the histone 3 lysine 9 methyltransferase SETDB1 negatively regulates the expression of the proleukemic genes Hoxa9 and its cofactor Meis1 through deposition of promoter H3K9 trimethylation in MLL-AF9 leukemia cells. Here, we investigated the biological impact of altered SETDB1 expression and changes in H3K9 methylation on AML. We demonstrate that SETDB1 expression is correlated to disease status and overall survival in AML patients. We recapitulated these findings in mice, where high expression of SETDB1 delayed MLL-AF9 mediated disease progression by promoting differentiation of leukemia cells. We also explored the biological impact of treating normal and malignant hematopoietic cells with an H3K9 methyltransferase inhibitor, UNC0638. While myeloid leukemia cells demonstrate cytotoxicity to UNC0638 treatment, normal bone marrow cells exhibit an expansion of cKit+ hematopoietic stem and progenitor cells. Consistent with these data, we show that bone marrow treated with UNC0638 is more amenable to transformation by MLL-AF9. Next generation sequencing of leukemia cells shows that high expression of SETDB1 induces repressive changes to the promoter epigenome and downregulation of genes linked with AML, including Dock1 and the MLL-AF9 target genes Hoxa9, Six1, and others. These data reveal novel targets of SETDB1 in leukemia that point to a role for SETDB1 in negatively regulating pro-leukemic target genes and suppressing AML.

Introduction Epigenetic deregulation has emerged as an important contributor to oncogenesis and disease progression in a variety of malignancies, including leukemia.1 Deep sequencing has revealed that genes encoding epigenetic modifying proteins are mutated in over 70% of acute myeloid leukemia (AML) patients.2 H3K9me2/3 marks large regions of condensed transcriptionally inactive chromatin, such as pericentric heterochromatin.3 H3K9me2/3 also plays a functional role in the dynamic repression of genes in euchromatic regions of the genome.4 Two families of proteins are associated with deposition of H3K9 methylation: the SUV39 family and PRDM family. The SUV39 family of H3K9 methyltransferases consists of SUV39H1/2, EHMT1/2, and SETDB1/2.4 Our lab and others have previously demonstrated that members of the SUV39 family of H3K9 methyltransferases bind to the polymerase associated factor complex (PAF1c).5,6 SETDB1, G9a (EHMT2), and GLP (EHMT1), were identified in a proteomics study exploring the interactome of the PAF1c in AML.5 The PAF1c is an epigenetic regulator complex that physically haematologica | 2020; 105(9)

Correspondence: ANDREW G. MUNTEAN andrewmu@umich.edu Received: April 4, 2019. Accepted: September 25, 2019. Pre-published: September 26, 2019. doi:10.3324/haematol.2019.223883

ÂŠ2020 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.

2273

J. Ropa et al.

associates with RNA polymerase II (RNAPII) and both positively and negatively regulates gene transcription.7–11 In AML, the PAF1c is critical for the regulation of a proleukemic HOXA gene program in AML cells through the recruitment of MLL and MLL-fusion proteins to the HOXA locus via direct physical interactions.12–15 Hoxa9 and its co-factor MEIS1 are upregulated in about 50% of AML and associated with a poor patient prognosis.16 Given our recent data linking H3K9 methyltransferases with Hoxa9 and Meis1 repression along with altered H3K9me3 in AML patients compared to CD34+ cells,17 it is important to understand the epigenetic and biological impact of H3K9 methyltransferases on AML. SETDB1 is a H3K9 mono/di/tri-methyltransferase involved in heterochromatin regulation and euchromatic gene silencing.18 SETDB1 was shown to bind gene loci associated with development in mouse embryonic stem (ES) cells, such as the Hoxd cluster of genes.19 SETDB1 has been implicated as an oncogene in melanoma, breast cancer, liver cancer, and lung cancer.4 Importantly, Ceol et al. reported amplification of SETDB1 in melanoma results in aberrant binding and regulation of the HOXA locus.20 In contrast to these oncogenic roles for SETDB1, Avgustinova and colleagues report that depletion of the H3K9 methyltransferase G9a in squamous tumors leads to a delayed, but more aggressive phenotype due to expanded cancer progenitor pools with increased genomic instability.21 In the hematopoietic system, the methyltransferase activity of G9a is required for leukemogenesis due to a physical interaction with HOXA9.22 Importantly, loss of G9a has no effect on hematopoietic stem cells.22,23 SETDB1, however, is required for both hematopoietic stem and progenitor cell (HSPC) maintenance and leukemic stem cells.23 Further, Cuellar and colleagues show that SETDB1 mediated silencing of endogenous retroviral elements is required for the growth of AML cell lines.24 Together, these studies suggest that therapeutic targeting of SETDB1 may benefit AML patients. However, we recently demonstrated that SETDB1 negatively regulates the expression of the pro-leukemic Hoxa9 and Meis1 genes in MLL-AF9 transformed AML cells through association with the PAF1c, which localizes to Hoxa and Meis1 loci. The PAF1c-SETDB1 interaction mediates promoter H3K9me3 and repression of Hoxa9 and Meis1 expression.5 Further, SETDB1 expression is inversely correlated with HOXA9 and MEIS1 expression in AML patient samples.5 These data imply a more complex role for H3K9 methylation in AML similar to skin tumors whereby H3K9 methyltransferases display both oncogenic and suppressive roles.20,21 Thus, further investigation into the role of H3K9 methyltransferases in AML is required. Here we show that AML patients with higher expression of SETDB1 display a better prognosis, consistent with repression of HOXA9 and MEIS1 expression. SETDB1 overexpression induces cellular differentiation and delays disease onset in a mouse model of AML, recapitulating AML patient survival. We also investigated the utility of inhibiting H3K9 methyltransferases in AML cells and HSPC, demonstrating that inhibition of H3K9 methylation in HSPC leads to retention of self-renewal capacity in HSPC and more efficient transformation by the MLL-AF9 fusion protein. Finally, we show that SETDB1 regulates gene expression by inducing changes in the epigenetic landscape and chromatin accessibility at gene targets critical to leukemogenesis. 2274

Methods Patient sample data Data for patient gene expression relative to normal hematopoietic cells were mined from BloodPool on BloodSpot database.25 Bloodspot assigns AML patient samples to a closest normal hematopoietic counterpart using transcriptomic profiles.26 AML patient RNA sequencing (RNA-seq) and survival data were mined from The Cancer Genome Atlas.2

Cell line generation Cell lines were generated from C57Bl/6 (Taconic Farms) mouse bone marrow or from SETDB1floxed27 mice. Platinum-E viral packaging cells were transfected with the indicated constructs: MSCVneo-FLAG-MLL-AF9 (MA9), MSCVneo-FLAG-E2A-HLF (EHF), MSCVhygro-FLAG-EHMT2 (G9a) (Ge lab; Addgene #41721), MSCVpuro-HA-SETDB1 or empty vector (EV) controls. Cells were spinfected with viral supernatants and 5 ug/mL polybrene (Millipore), selected with 1 mg/mL G418 (Invitrogen) and 1 ug/mL puromycin (Invitrogen) or 200 ug/mL hygromycin (Invitrogen) and cultured in IMDM with 15% stem cell fetal bovine serum (Millipore), 1% penicillin/ streptomycin (Invitrogen), 10 ng/mL interleukin-3 (IL-3) and 100 ng/mL stem cell factor (R&D).

Mouse modelling Primary MLL-AF9 mouse leukemia cells were retrovirally transduced with MSCVpuro-HA-SETDB1 or EV and selected in 2 ug/mL puromycin for 4 days. 100,000 cells were injected intravenously into lethally irradiated (950 rads) C57Bl/6 mice. Mice were monitored for survival, moribund mice were euthanized, and bone marrow, spleen, and liver were harvested. Animal studies were approved by the University of Michigan’s Committee on Use and Care of Animals and Unit for Laboratory Medicine.

Quantitative PCR (qPCR) RNA was harvested using the Qiagen RNeasy mini plus kit. cDNA synthesis was performed using the SuperScript III kit (Invitrogen). qPCR was performed using fast SYBR-green mastermix (Thermo Fisher). Primers sequences are listed in the Online Supplementary Table S2.

Sequencing libraries preparation MA9+ empty vector (EV) and MA9+SETDB1 cells were harvested for next generation sequencing library preparation for RNAseq, ChIP-seq, and ATAC-seq. ChIP-seq antibodies were validated using Epicypher histone peptide arrays (Online Supplementary Figure S6). Library preparation details are in the Online Supplementary Materials and Methods.

Data availability Sequencing data is available via the Gene Expression Omnibus, accession GSE136850.

Results SETDB1 expression is correlated with AML patient prognosis Given our data linking H3K9 methyltransferases with HOXA9 and MEIS1 expression, we investigated their expression in AML patient sample data. Interestingly, SETDB1, SUV39H1, and SUV39H2 exhibit lower expression in AML patient samples when compared with normal hematopoietic cells, with median expression levels that haematologica | 2020; 105(9)

SETDB1 expression suppresses MLL-fusion driven AML

TOGA Overall Survival vs. Expression of H3K9 methyltransferases

E P value (n=170)

Figure 1. SETDB1 expression is low in acute myeloid leukemia (AML) and correlated with AML patient survival. (A) BloodPool microarray gene expression data mined from Bloodspot shows AML patient gene expression relative to the nearest normal hematopoietic counterpart for the indicated genes encoding H3K9 methyltransferases (n=1991) in a violin plot. Max probe was used for all genes. (B) BloodPool data for AML patients’ SETDB1 expression relative to nearest normal counterpart divided by karyotype (n=1,991) shown in a boxplot. Max probe was used for SETDB1 expression. (C) TCGA RNA-seq patient sample data is divided by “high” expression (above median) and “low” expression (below median) for the indicated gene encoding an H3K9 methyltransferase. Boxplot shows overall survival for each stratified gene (n=173). (D) Kaplan Meier curve showing overall survival of AML patients stratified by SETDB1 expression above (“high”) or below (“low”) median (n=173). (E) Multivariate analysis using Cox Hazard Proportion analysis to assess the hazard ratio associated with changing levels of SETDB1 expression. SETDB1 expression is expressed in log2 (RSEM), so the hazard ratio is associated with a two-fold change in SETDB1 expression. Age did not satisfy the model and was stratified by patients <60 years old and patients >60 years old (n=173). Statistics: log-rank followed by multiple testing correction using Benjamani-Hochberg false discovery rate (FDR) (C); log-rank (D); Cox Proportional Hazard Model (E); n: biological replicates; *: P/padj<0.05; RNA-seq: RNA sequencing.

are 66%, 56%, and 41% relative to their nearest normal hematopoietic counterparts, respectively (Figure 1A).25,26 The downregulation of these genes in AML samples was consistent regardless of AML karyotype (Figure 1B and Online Supplementary Figure S1A-B).25,26 We next tested whether H3K9 methyltransferase gene expression significantly correlated with patient survival using publicly available RNA-seq data2 and found that only SETDB1 expression significantly correlated with patient survival (Figure 1C). Median survival for patients with SETDB1 expression above the median was 26.3 months and 9.5 months in patients with SETDB1 expression below the median (Figure 1D).2 Univariate and multivariate analyses reveal that higher expression of SETDB1 is associated with a higher overall survival rate with a P<0.003 and a lower expected hazard ratio of 0.29 per two-fold change in expression (Figure 1D-E).2

SETDB1 or G9A expression reduces AML growth and colony formation We next explored the biological effects of SETDB1 expression on the transformation and growth of AML cells. First, we explored whether there is a difference in Setdb1 expression in mouse AML relative to normal HSPC by isolating lineage negative (Lin–) cKit+ cells from mouse haematologica | 2020; 105(9)

bone marrow. qPCR demonstrates that Setdb1 expression is reduced in murine MLL-AF9 and CALM-AF10 leukemias compared to normal HSPC, consistent with the patient sample data (Figure 2A). We next performed colony replating assays where Lin– mouse bone marrow cells were retrovirally co-transduced with the MLL-AF9 fusion oncogene with and without SETDB1 and plated in semi-solid methylcellulose. Overexpression of human SETDB1 significantly reduced MLL-AF9 mediated colony formation (Figure 2B-C) while not affecting the expression of the exogenous MLL-AF9 (Online Supplementary Figure S2A).24 Colony formation driven by a separate leukemic fusion protein, E2A-HLF, was also reduced in the presence of SETDB1, suggesting a general effect on AML transformation (Online Supplementary Figure S2B-C). Ex vivo proliferation assays demonstrate that overexpression of SETDB1 in MLL-AF9 or E2A-HLF transformed AML cells leads to a significant reduction in cellular proliferation (Figure 2D and Online Supplementary Figure S2D). Interestingly, we observed a strong selective pressure to reduce exogenous SETDB1 expression in cultured MLLAF9+SETDB1 cells (Online Supplementary Figure S2E) resulting in rescue of MLL-AF9 cellular proliferation (Online Supplementary Figure S2F). MLL-AF9 cells that overexpress SETDB1 undergo morphological changes consis2275

J. Ropa et al.

Figure 2. Overexpression of SETDB1 delays acute myeloid leukemia growth. (A) Quantitative PCR (qPCR) using primers for mSetdb1 to determine expression levels in primary MLL-AF9 (n=3) or CALM-AF10 (n=2) acute myeloid leukemia (AML) cells compared to Lin-cKit+ mouse bone marrow hematopoietic stem and progenitor cell (HSPC) (n=2 pools of 5 mice each). (B) Mouse Lin- bone marrow was retrovirally transduced with the indicated plasmid vectors and plated in methylcellulose. Colonies were counted after 7 days and re-plated, for a total of three rounds. Shown is one representative experiment of n=4, error bars are standard deviations of the technical replicates. (C) Representative Iodonitrotetrazolium chloride (INT) staining of colony assay plates for MA9 cells with or without SETDB1 overexpression. Images were taken on a BioRad ChemiDoc XRS+ at 1X magnification at the same time with the same contrast. (D) Lin- bone marrow cells were retrovirally transduced with MA9 in the presence or absence of SETDB1 overexpression, selected for 2 weeks, then proliferation was monitored by viable cell count daily. Shown is one representative experiment of n=4. Error bars are standard deviations of technical replicates. (E) Representative cytospin and Hema3 stained MA9 cells in the presence or absence of SETDB1 overexpression (n=3). (F) qPCR measurement of genes associated with differentiation in MA9 cells in the presence or absence of SETDB1 overexpression. Data are normalized to glyceraldehyde 3-phosphate dehydrogenase (Gapdh) relative to MA9+EV gene expression (delta delta ct), n=3. Error bars are standard deviation of delta delta ct. (G-H) Quantification of fluorescence-activated cell sorting (FACS) analysis to determine the relative amount of live, dead, and dying cells in (G) resting AML cells or (H) AML cells treated with Daunorubicin. Cells were stained with FITC-Annexin V and DAPI and run on an LSRII flow cytometer. Cells are defined as follows: AnnexinV+,DAPI+ = dead; AnnexinV-,DAPI- = alive; AnnexinV+,DAPI- = apoptotic. Shown are relative quantifications comparing the SETDB1 overexpression cells to control. Error bars indicate standard deviations. (G) n=3, (H) n=2. (I) Survival for MLL-AF9 AML mouse model. Primary MA9 cells were transduced with SETDB1 or EV control and selected for 4 days prior to tail vein injections in sublethally irradiated mice. Shown is the Kaplan Meier survival curve; MA9+EV n=14; MA9+SETDB1 n=8 (six censored mice did not develop AML). (J) Spleen weights of moribund or censored mice from MLL-AF9 AML mouse model. Error bars represent standard deviations. Above are representative images of spleens from euthanized mice. (K) qPCR using primers specific for the HA-SETDB1 exogenous construct from the AML mouse model to measure expression of the plasmid. RNA was harvested before injection or from moribund mice (n=2, 4 MA9+EV preinjection, moribund mice; n=2, 8 MA9+SETDB1 pre-injection, moribund mice). Error bars represent standard deviations. Statistics: significance was determined by: two-sample t-test comparing relative expression (Actb or Gapdh) of AML primary cells to Lin-cKit+ for Setdb1 expression(A) or MA9+SETDB1 to MA9+EV for four different genesâ&#x20AC;&#x2122; expression (F); generalized linear modeling followed by ANOVA where each MA9+SETDB1 replicate was paired to the MA9+EV control from the same biological replicate. Main effect is reported if there are no significant interactions (see statistical analysis in the Online Supplementary Materials and Methods). (B/D/G/H); Log-rank test (I). MA9: MLL-AF9; EV: empty vector control; n: biological replicates; *: P<0.05: Actb: beta actin; Gapdh: glyceraldehyde 3-phosphate dehydrogenase.

2276

haematologica | 2020; 105(9)

SETDB1 expression suppresses MLL-fusion driven AML

Figure 3. H3K9 methyltransferase inhibitor UNC0638 enhances hematopoietic stem and progenitor cell colony formation capacity. (A-B) Lin- mouse bone marrow was isolated and treated in culture for 4 days with the indicated dose of UNC0638. (A) Cells pretreated with UNC0638 were harvested, lysed in SDS loading buffer, and run on SDS-PAGE. Shown is a Western blot probed for the indicated antibodies. Below the H3K9me2/3 Western blots are numbers indicating the band densities normalized to the Western blot probing for total H3 and normalized to vehicle control. (C) Cells pretreated with UNC0638 were plated in methylcellulose and colonies were counted after 7 days, n=4. Colony numbers are shown relative to the vehicle/non-silencing control for each replicate, bar graphs represent the mean and error bars represent the standard deviation of these normalized values. (C) INT stained representative colonies from (B). (D) Isolated human CD34+ cells were treated for 4 days with the indicated doses of UNC0638 and plated in methylcellulose. Colonies were counted after 14 days, n=2. Colony numbers are shown relative to the vehicle/ non-silencing control for each replicate, bargraphs represent the mean and error bars represent the standard deviation of these normalized values. (E) INT stained representative colonies from (D). Statistics: significance was determined by generalized linear modeling followed by ANOVA where each treated group was paired to the vehicle treatment from the same biological replicate. Main effect is reported if there are no significant interactions (B/D) (See statistical analysis in Online Supplementary Materials and Methods); n: biological replicates; *: P<0.05: INT: 2-(p-iodophenyl)-3-(p-nitrophenyl)-5-phenyl tetrazolium chloride; HSPC: hematopoietic stem and progenitor cell.

tent with differentiation (Figure 2E). Genes associated with hematopoietic differentiation, including Id2, Cd80, Nab2, and Itgam have significantly increased expression upon overexpression of SETDB1 (Figure 2F). Additionally, overexpression of SETDB1 in MLL-AF9 cells leads to increased apoptosis (Figure 2G and Online Supplementary Figure S2G). We next examined whether SETDB1 expression is associated with reduced chemoresistance. MA9+SETDB1 do not exhibit reduced apoptosis after treatment with Daunorubicin at their IC50 (Figure 2H and Online Supplementary Figure S2H-I), suggesting SETDB1 expression is not associated with increased Daunorubicin sensitivity. Overexpression of another H3K9 methyltransferase, human G9A, also reduces colony formation and proliferation of MLL-AF9 cells (Online Supplementary Figure S3A-C) and induces morphological changes consistent with differentiation (Online Supplementary Figure S3D). These data haematologica | 2020; 105(9)

demonstrate that expression of multiple H3K9 methyltransferases reduces AML cell proliferation and colony forming potential and induces AML differentiation.

SETDB1 expression delays MLL-AF9 mediated AML To examine the effects of SETDB1 in vivo, we transplanted primary mouse MLL-AF9 AML cells retrovirally transduced with or without SETDB1 into sublethally irradiated syngeneic recipient mice and monitored survival. Consistent with AML patient data, overexpression of SETDB1 significantly delays MLL-AF9 mediated leukemogenesis in vivo (Figure 2I). All moribund mice from both the control MLL-AF9 group and MLL-AF9+SETDB1 group exhibited splenomegaly, leukemic infiltration in the liver (Figure 2J and Online Supplementary Figure S3E) and similar MLL-AF9 expression levels (Online Supplementary Figure S3F). We measured expression of exogenous SETDB1 in 2277

J. Ropa et al.

MLL-AF9 leukemic cells before injection and in bone marrow of moribund mice and observed a decrease in exogenous SETDB1 expression in 6 of 8 mice (Figure 2K). These data demonstrate that SETDB1 expression suppresses MLL-AF9 mediated leukemic progression in vivo.

H3K9 methylation impairs HSPC colony formation and suppresses leukemic transformation We examined how loss of Setdb1 affects normal HSPC

and AML growth and transformation. First, using inducible CreER-mediated knock-out of Setdb1, we confirmed that Setdb1 is required for MLL-AF9 cell growth23 and demonstrated that Setdb1 is also required for the growth of E2A-HLF leukemic cells (Online Supplementary Figure S4A-B). We next sought to determine the effect of reduced expression but not deletion of Setdb1 on AML cell growth. However, heterozygous deletion of Setdb1 in MLL-AF9-Setdb1fl/+-CreER cells did not reduce Setdb1

Figure 4. H3K9 methyltransferase inhibitor UNC0638 preserves primitive hematopoietic cells amenable to MLL-AF9 transformation. (A-G) Lin- mouse bone marrow was treated with the indicated doses of UNC0638 for 4 days. (A-D) Cells were stained with anti-CKIT conjugated to APC fluorophore (A-B) or anti-Cd11b conjugated to PE fluorophore (C-D). Flow cytometry was performed to analyze cKit+ or Cd11b– populations. (A/C) are representative flow plots for 0.75 μM UNC0638 treatments. B/D show the biological replicates for cKit+ populations (B) or Cd11b– populations (D) of treated cells relative to vehicle (n=3). Bar graphs represent the median and error bars show the range of these normalized values. (E) RNA was harvested after 4 days of treatment with UNC0638 and quantitative PCR (qPCR) was used to determine changes in Hoxa9 expression. Plotted are the biological replicates relative to vehicle. Bar graphs represent the median and error bars show the range of these normalized values. (F) After treatment with UNC0638, cells were spinfected with MigR1-MLL-AF9, which also expresses a green fluorescent protein (GFP) reporter. Cells were monitored for GFP expression by flow cytometry until 100% GFP was achieved. (G) Representative flow plots from different time points during the GFP monitoring experiment. For F-G: shown is one representative experiment of n=4. Statistics: significance was determined by two-sample t-test comparing relative expression (Gapdh) of treated cells compared to vehicle cells (E). Significance was determined by generalized linear modeling followed by ANOVA where each treated group was paired to the vehicle treatment from the same biological replicate. Main effect is reported if there are no significant interactions(C-D) (see statistical analysis in the Online Supplementary Materials and Methods); n: biological replicates; *: P<0.05; Gapdh: glyceraldehyde 3-phosphate dehydrogenase.

2278

haematologica | 2020; 105(9)

SETDB1 expression suppresses MLL-fusion driven AML

protein (Online Supplementary Figure S4C). Therefore, we utilized short hairpin RNA (shRNA) mediated knockdown of Setdb1 using a doxycycline inducible MLL-AF9 Tet-on cell line in competitive growth assays. Interestingly, our

results were inconclusive where two shRNA constructs with efficient knockdown displayed negative selection, while two less efficient shRNA showed a positive selection compared to MLL-AF9-Tet-on cells relative to the

Figure 5. SETDB1 overexpression downregulates oncogenic gene programs and upregulates differentiation gene programs in acute myeloid leukemia. (A) MA plot shows fold changes of genes in MA9+SETDB1/ MA9+EV versus the average expression of those genes in all samples. Red highlighted genes are significantly upregulated; blue highlighted genes are significantly downregulated; yellow highlighted genes are the Hoxa cluster of genes and the Hoxa9 cofactor Meis1 (n=3/ condition). (B) Gene set analysis (GSA) using CAMERA shows that genes that are upregulated by Hoxa9 and Meis1 are downregulated by SETDB1. (C) GSA analysis using CAMERA shows that genes that are upregulated in mature blood cells relative to primitive blood cells are upregulated by SETDB1. (D) DESeq2-normalized RNA-seq counts of genes associated with differentiation (n=3/ condition). Error bars represent the range of counts. (E) GSA analysis using ROAST shows that genes that are directly bound by MLL-AF9 are downregulated by SETDB1. (F) Overlap analysis of genes that are downregulated by SETDB1 in mouse MA9 cells and genes that have reduced promoter H3K9me3 in acute myeloid leukemia (AML) patient samples compared to normal human CD34+ cells. (G) Overlap analysis of genes that are downregulated by SETDB1 in mouse MA9 cells and genes that are upregulated upon SETDB1 knockdown by Crispr-Cas9 in human THP-1 cells. For both data sets in G, FDR of 0.1 was used as a cutoff for consistency with the previously published THP-1 dataset. Statistics: unless otherwise noted, significant gene expression changes are defined by DESeq2 algorithm with fold change >1.5 and padj<0.05 (A/D/F); false discovery rate (FDR) is calculated by CAMERA using Benjamani-Hochberg correction (B-C); Hypergeometric tests (F-G). For RNA-seq, n = 3 for each cell condition. EV: empty vector control; n: biological replicates; *: P/padj<0.05.

haematologica | 2020; 105(9)

2279

J. Ropa et al.

untreated control (Online Supplementary Figure S4D-E). This may suggest Setdb1 is maintained within a narrow expression window in AML cells. To test the effect of Setdb1 knock-down on HSPC colony formation we transduced Lin– bone marrow with shRNA targeting Setdb1. Our results show a trend for increased HSPC colony formation without changes in Hoxa9 or Meis1 expression following Setdb1 knockdown (Online Supplementary Figure S4F, data not shown). Because inhibition of H3K9 methyltransferases has been proposed as a therapeutic option to treat AML,22,23 we explored the effects of chemical inhibition of H3K9 methylation in normal and malignant hematopoietic cells. Without a selective SETDB1 small molecule inhibitor, we utilized the G9a inhibitor UNC0638, which addresses the function of H3K9 methyltransferases more broadly. Lin– mouse bone marrow cells treated with UNC0638 exhibit a reduction in both H3K9me2 and H3K9me3 (Figure 3A).28,29 Given our data demonstrating Hoxa9 and Meis1 are repressed by H3K9 methylation,5 we tested the effect of UNC0638 on HSPC self-renewal. Previous reports have demonstrated that mouse Lin–Ska1+cKit+ cells (LSK) are preserved in culture following treatment with UNC0638.30 To explore this further, Lin– cells were treated with increasing doses of UNC0638 for five days prior to plating in semi-solid methylcellulose in the presence of SCF and IL-3. Interestingly, treatment with UNC0638 significantly increased colony formation capacity of Lin– cells in a dosedependent manner (Figure 3B-C). Further, human CD34+ cells isolated from mobilized peripheral blood and treated with increasing doses of UNC0638 demonstrate increased colony formation capacity (Figure 3D-E), consistent with reports that UNC0638 preserves CD34+ cells in culture.31 Thus, chemical or genetic inhibition of H3K9 methyltransferases preserves self-renewal capacity of bone marrow cells. We also examined the effects of UNC0638 chemical inhibition of H3K9 methylation on AML cells. Consistent with previous studies, treatment with UNC0638 results in reduced cellular proliferation of MLL-AF9 cells (Online Supplementary Figure S5A).22 Since UNC0638 preserves self-renewal of HSPC (Figure 3), we asked whether inhibition of H3K9me alters MLL-AF9 mediated transformation of bone marrow cells. First, we found that UNC0638 treatment of Lin– bone marrow cells preserve more primitive cKit+ and Cd11b– populations (Figure 4A-D and Online Supplementary Figure S5B-C). Lin– cells display significantly increased Hoxa9 expression in response to UNC0638 treatment (Figure 4E). Lin–cKit+ (LK) and related primitive hematopoietic cell populations are more amenable to transformation than more differentiated subtypes.32 To explore whether this expansion of LK cells results in greater AML transformation capacity, we pretreated Lin– bone marrow cells with UNC0638 for four days then retrovirally transduced them with MigR1-MLL-AF9 and monitored green fluorescent protein (GFP) by flow cytometry. We observed a more rapid expansion of GFP+ MLLAF9 cells following treatment with UNC0638 compared to vehicle, with a 1.4-1.7 fold increase in GFP+ cells before both populations reached 100% GFP positivity (Figure 4FG and Online Supplementary Figure S5D). One of four replicates demonstrated a more rapid expansion of vehicle treated cells; however, GFP+ cells were increased two-fold in vehicle treated cells at day 1 suggesting this is attributable to differences in transduction rates. Our combined 2280

data suggests that H3K9 methyltransferases can suppress leukemic transformation and may point to a narrow window of H3K9 methylation that is optimal for leukemic transformation and cell growth.

SETDB1 regulates oncogenic gene programs in AML To explore the gene programs regulated by SETDB1 in AML, we performed RNA-sequencing experiments on MLL-AF9 cells overexpressing SETDB1. 2,285 genes are upregulated and 1,771 genes are downregulated by SETDB1 overexpression with a fold-change of 1.5 or more and an false discovery rate (FDR) of <0.05 (Figure 5A and Online Supplementary Table S4). Notably, many genes found in the Hoxa cluster were significantly downregulated, including Hoxa3, Hoxa5, Hoxa6, Hoxa9, and the Hoxa9 cofactor Meis1; while only one Hoxa gene was upregulated, the long non-coding RNA Hoxa11os (Figure 5A, highlighted genes). In fact, gene programs that are upregulated by forced expression of HOXA9 and MEIS1 in mouse cells are significantly downregulated by SETDB1 overexpression using gene set analysis33 (Figure 5B). We show significant upregulation of genes that exhibit increased expression in mature blood cells compared to HSPC, consistent with the differentiation observed upon SETDB1 overexpression in AML34,35 (Figure 5C-D and Figure 2F-G). Because SETDB1 binds the PAF1c, which is required for localization of MLL fusion proteins,5 we asked how SETDB1 expression affects direct targets of MLL-AF9. Interestingly, genes bound by regulation of MLL-AF9 were significantly downregulated upon SETDB1 overexpression,36 suggesting H3K9me3 regulation MLL-AF9 gene programs in leukemic cells (Figure 5E). 193 genes downregulated by SETDB1 overexpression are reported to have reduced promoter H3K9 methylation in AML relative to normal CD34+ cells,17 suggesting SETDB1 may be responsible for regulating a subset of these genes, including Kit, Cbl, Ptpn11, Six1, and other genes that are important in AML (Figure 5F and Online Supplementary Table S4). There is significant overlap between genes downregulated by SETDB1 and genes upregulated by Crispr/Cas9 mediated knockdown of SETDB1 in human THP-1 AML cells harboring an MLL-AF9 fusion (Figure 5E and Online Supplementary Table S4).24 This suggests SETDB1 regulates conserved pro-leukemic gene programs in leukemic cells including direct MLL fusion targets.

SETDB1 regulates the epigenome to affect changes in chromatin accessibility and gene expression We have shown that overexpression of SETDB1 in MLL-AF9 cells leads to global increases in H3K9me3.5 To understand the specific epigenomic changes induced by SETDB1 in AML, we performed ChIP-seq for H3K9me3 in MLL-AF9 and MLL-AF9+SETDB1 leukemic cells. We also performed ATAC-seq to assess changes in chromatin accessibility. We first explored differences at the HOXA locus due to its importance in a large subset of AML, including MLL leukemias.5,37 Overexpression of SETDB1 reduced chromatin accessibility and increased H3K9me3 at posterior Hoxa genes, which results in reduced transcription of Hoxa9 (Figure 6A). We sought to define whole genome epigenetic regulation mediated by SETDB1 but observed only 552 consensus H3K9me3 peaks (Online Supplementary Figure S7A and Online Supplementary Table S5), which included regions enriched for H3K9me3, such as the zinc finger protein cluster on chromosome 7 (Online haematologica | 2020; 105(9)

SETDB1 expression suppresses MLL-fusion driven AML

Supplementary Figure S7B). Repetitive elements may consume H3K9me3 sequencing depth making it difficult to map with stringency. We circumvented this by performing ChIP-Seq for H3K9ac, which is mutually exclusive of H3K9me3 and associated with gene activation, in contrast to H3K9me3.38 We performed H3K9ac ChIP-seq on MLLAF9 and MLL-AF9+SETDB1 cells and saw changes at over 6,000 promoter regions, including both increased and decreased H3K9ac signal (Online Supplementary Figure S7C and Online Supplementary Table S5). Gene ontology analy-

sis reveals that genes with decreased promoter H3K9ac in MLL-AF9+SETDB1 cells were associated with cell cycle and RNA binding, whereas genes with increased promoter H3K9ac were associated with signaling pathways and negative regulation of proliferation (Figure 6B). Overlapping downregulated genes with genes that exhibit reduced promoter ATAC-seq and H3K9ac ChIP-seq signal in MLL-AF9+SETDB1 cells and scoring by their combined fold changes reveals several interesting targets including Six1 and Mefv, which are implicated as biomarkers in

Figure 6. SETDB1 regulates the epigenetic landscape of acute myeloid leukemia oncogenes and biomarkers. (A) Sequencing tracks showing H3K9me3 ChIP-seq (top), ATAC-seq (middle), and RNA-seq (bottom) signals for MA9+EV (blue) or MA9+SETDB1 (red) cells. Shown here is the entire Hoxa cluster of genes and a closer view of Hoxa9 specifically. (B) Gene ontology analysis using DAVID to query biological process, molecular function, and Kegg pathway gene sets that are overrepresented in the following groups: H3K9ac ChIP-seq peaks that have significantly reduced (blue) or increased (red) signal intensity in MA9+SETDB1 compared to MA9+EV. (C) Overlap analysis of genes that are downregulated by SETDB1, genes where SETDB1 drives reduced promoter H3K9ac, and genes where SETDB1 drives reduced promoter ATAC-seq signal. (D) Top ten scoring genes when fold changes for the three datasets in C are summed. (E) Sequencing tracks showing H3K9ac ChIP-seq (top), ATAC-seq (middle), and RNA-seq (bottom) signals for MA9+EV (blue) or MA9+SETDB1 (red) cells. Shown here is the locus for Dock1. Statistics: differentially bound regions are defined by DiffBind false discovery rate (FDR) <0.05 (B/C/D); Significant gene expression changes are defined by DESeq2 algorithm with fold change >1.5 and padj<0.05 (C). For all ChIP-seq and ATAC-seq studies, n = 2 for each cell condition. AML: acute myeloid leukemia; EV: empty vector control; n: biological replicates; *: P/padj<0.05; RNA-seq: RNA sequencing; ChIP-seq: chromatin immunoprecipitation followed by sequencing; ATAC-seq: assay for transposaseaccessible chromatin coupled with next-generation sequencing.

haematologica | 2020; 105(9)

2281

J. Ropa et al. AML patients39,40 (Figure 6C-D and Online Supplementary Table S5). We also observed significant loss of promoter H3K9ac and chromatin accessibility and reduced gene expression of Dock1 (Figure 6E), which is a prognostic marker of AML that displays changes in DNA methylation in AML patient samples relative to normal HSPC.41,42 We identified several direct binding targets of MLL-AF9 that underwent epigenetic remodeling and expression

changes with overexpression of SETDB1. Thus, we asked whether SETDB1 impacts H3K79me2, which is deposited by DOT1L and associated with MLL-fusion proteins,36 using ChIP-seq for H3K79me2. We found a marked decrease of H3K9ac and ATAC-seq signal at H3K79me2 peaks in MLL-AF9+SETDB1 cells suggesting a role for H3K9 modifications in regulating genes marked with H3K79me2 (Figure 7A-B). To further explore the role of

Figure 7. SETDB1 regulates the epigenetic landscape of a subset of MLL-AF9 target genes. (A) Signal track for H3K9ac ChIP-seq centered around promoters of genes that have reduced H3K79me2 in MA9+SETDB1 relative to MA9+EV cells. (B) Signal track for ATAC-seq centered around promoters of genes that have reduced H3K79me2 in MA9+SETDB1 relative to MA9+EV cells. (C) Overlap analysis of genes that are downregulated by SETDB1 in acute myeloid leukemia (AML), have reduced H3K9ac ChIP-seq, ATAC-seq, or H3K79me2 ChIP-seq upon SETDB1 overexpression in AML, or MLL-AF9 direct binding targets (Bernt et al. 2011). (D) Sequencing tracks showing H3K79me2 ChIP-seq (top), H3K9ac ChIP-seq (second from top), ATAC-seq (third from top), and RNA-seq (bottom) signals for MA9+EV (blue) or MA9+SETDB1 (red) cells. Shown here is the Six1 gene locus. (E) ChIP-qPCR: anti-H3K9me3 or anti- immunoglobulinG (anti-IgG) immunoprecipitated DNA were subjected to quatitative PCR (qPCR) using primers for the promoter region of Six1 and Dock1. Shown is relative quantification compared to total anti-H3 immunoprecipitated DNA. n=2. (F) Anti-SETDB1 ChIP-seq track from the ENCODE project showing the SIX1 locus in K562 leukemia cells. Solid bars represent called peaks as reported by ENCODE. Statistics: differentially bound regions are defined by DiffBind false discovery rate (FDR) <0.05 (A/B/C); Significant gene expression changes are defined by DESeq2 algorithm with fold change >1.5 and padj<0.05 (C). For all ChIP-seq and ATAC-seq studies, n=2 for each cell condition. EV: empty vector control; n: biological replicates; *: P/padj<0.05. ChIP-seq: chromatin immunoprecipitation followed by sequencing; ATAC-seq: assay for transposase-accessible chromatin coupled with next-generation sequencing.

2282

haematologica | 2020; 105(9)

SETDB1 expression suppresses MLL-fusion driven AML

SETDB1 in regulating MLL-AF9 targets, we performed an overlap analysis for genes that, upon SETDB1 overexpression, are downregulated, lose promoter H3K9ac, compact chromatin, lose gene body H3K79me2 signal, and are bound by MLL-AF9.36 This defined a target list of genes that may be coregulated by SETDB1 and MLL-AF9. Included in this group is Gfi1, which has been shown to affect AML cell growth;43 Rap1gds1, a nucleotide exchange factor; Arid1b, a member of the SWI/SNF complex; and Six1, which promotes formation of leukemic stem cells40 (Figure 7C). Six1 is of particular interest given its role in promoting leukemogenesis and the striking reductions observed in H3K9ac, H3K79me2, ATAC-seq signal, and gene expression at this locus (Figure 7D). We investigated H3K9me3 levels at specific loci by performing ChIP-qPCR at the promoter of Six1 and Dock1. We detected increased H3K9me3 in MLL-AF9 cells overexpressing SETDB1

(Figure 7E). Additionally, ENCODE data from human K562 cells shows SETDB1 binding at SIX1 (Figure 7F), GFI1 (Figure S7D), RAP1GDS1, and ARID1B (data not shown), but not DOCK1 (data not shown), suggesting we identified both direct and indirect targets. To determine whether SETDB1 expression inhibits cell growth primarily through Hoxa9 repression, we asked whether HOXA9/MEIS1 driven colony formation is affected by SETDB1 overexpression. A modest but insignificant decrease in colony formation is observed in HOXA9/MEIS1 transformed cells following SETDB1 overexpression (Figure 8A). However, SETDB1 overexpression is accompanied by significant upregulation of exogenous Meis1 (Figure 8B and Online Supplementary Figure S8A-B) that may account for the modest effects. Thus, Hoxa9, Meis1, and their downstream targets are likely affected by SETDB1 to influence AML cell growth.

Figure 8. SETDB1 affects Hoxa9, Meis1 and downstream targets. (A) Mouse Lin- bone marrow was retrovirally transduced with the indicated plasmid vectors and plated in methylcellulose. Colonies were counted after 7 days and re-plated, for a total of three rounds. Shown is one representative experiment of n=3, error bars are standard deviations of the technical replicates. (B) Flow cytometry showing differences in mean fluorescence intensities in HOXA9/MEIS1 transformed cells that are overexpressing SETDB1 or empty vector (EV) control. (C) Working model for the proposed role of SETDB1/H3K9 methylation in acute myeloid leukemia (AML) initiation and maintenance. AML initiates from hematopoietic stem and progenitor cell (HSPC) with variable H3K9 methylation and maintain lower SETDB1 expression. After establishment of AML, inhibition of H3K9 methyltransferases leads to loss of retroviral silencing and cell death. Stabilization of SETDB1 or G9a leads to repressed Hox gene expression and relief of blocked differentiation. Statistics: generalized linear modeling followed by ANOVA where each MA9+SETDB1 replicate was paired to the MA9+EV control from the same biological replicate. Main effect is reported if there are no significant interactions (A) (See statistical analysis in the Online Supplementary Materials and Methods).

haematologica | 2020; 105(9)

2283

J. Ropa et al.

Taken together, our data demonstrate profound effects of SETDB1 on the epigenome and transcriptome that affects genes critical for AML.

Discussion Here we demonstrate that SETDB1 and H3K9 methylation suppresses AML disease progression in vivo through the repression of pro-leukemic genes including direct MLL-fusion protein targets. We found that AML patient samples exhibit lower SETDB1 expression compared to normal hematopoietic cells and that higher SETDB1 expression correlates strongly with better overall AML patient survival (Figure 1). We recapitulated these findings in mice where forced expression of SETDB1 in MLL-AF9 driven AML induces differentiation of AML cells and increases disease latency (Figure 2). These data suggest SETDB1 suppresses AML cell growth and self-renewal by relieving the block in differentiation. We attribute the phenotypes in AML cells to altered H3K9 methylation. We altered H3K9 methylation levels genetically (SETDB1 and G9a) and through small molecule inhibition (UNC0638). Similar to our results with SETDB1, manipulation of G9a suggests H3K9 methylation can suppress AML progression by promoting differentiation (Figures 3-4, and Online Supplementary Figure S3). Thus, H3K9 methylation may have a more general effect on AML initiation and progression. Interestingly, Lehnertz and colleagues reported G9a overexpression accelerated Hoxa9/Meis1 mediated leukemia in vivo. We found Hoxa9/Meis1 mediated transformation in vitro was inhibited by SETDB1, but to a lesser extent than MLL-AF9 (Figure 8). Unique experimental strategies or functions for SETDB1 and G9a may account for these differences.22 Consistent with a role for H3K9 methylation in suppressing hematopoietic transformation, deletion of the H3K9 methyltransferase, SUV39H1 (and to a lesser degree SUV39H2), leads to the development of B-cell lymphomas in mice.44 Additionally, SETDB2 resides in a region of chromosome 13 that is commonly deleted in chronic lymphocytic leukemia (CLL).45 Thus, H3K9 methylation is likely exquisitely regulated in hematopoietic cells and performs context dependent functions that require further investigation to fully understand its role in AML. Mechanistically, we found that SETDB1 is linked with altered H3K9 methylation and acetylation, decreased chromatin accessibility and transcriptional repression of critical AML oncogenes (Figures 5-7). These genes included several that have been implicated in myelodysplastic syndromes (MDS) and AML.39 We show SETDB1 regulates Dock1 expression, which is correlated with leukemic stem cell gene signatures and a poor prognosis in AML patients.41,46 We also observed that SETDB1 represses genes associated with AML, such as Hoxa9 and Six1, which are direct targets of MLL-AF9.16,37,40 Interestingly, Six1 was recently shown to be important for leukemic stem cell maintenance where depletion of Six1 leads to

2284

increased disease latency.47 This suggests repression of Six1 may contribute to SETDB1 mediated extension of leukemic disease latency. These data point to SETDB1 negatively regulating a pro-leukemic gene program, many of which are potential therapeutic targets. Thus, understanding the mechanisms regulating SETDB1 at the transcriptional and post-translational level may be a valuable therapeutic approach for AML. For example, miRNA29 is a critical mediator of SETDB1 expression.48 Another potential mediator of H3K9 methylation is the PAF1c. We showed SETDB1 binds to the PAF1c and mediates promoter H3K9me3 of the Hoxa9 and Meis1 loci.5 Further, we and others identified G9a and SUV39H1 as interacting partners of the PAF1c.5,6 Interestingly, SETDB1, G9a, GLP and SUV39H1 form a complex that directs H3K9 methylation to euchromatic gene promoters.49 Thus, the PAF1c may recruit H3K9 methyltransferases to specific targets to mediate gene repression. The PAF1c is a critical regulator of transcription of several pro-leukemic genes in AML cells through direct physical interaction with wild-type MLL and MLL-fusion proteins.13,15 It will be interesting to consider the biochemical interplay between H3K9 methyltransferases and MLL-fusion proteins with the PAF1c. Previous studies have demonstrated that SETDB1 and G9a are required for AML initiation and progression.22â&#x20AC;&#x201C;24 Our current data demonstrating that SETDB1 suppresses AML growth may suggest AML cells maintain a narrow SETDB1 expression level. We show increased SETDB1 expression induces differentiation of AML cells through H3K9me3 and repression of self-renewal genes. Conversely, loss of SETDB1 is detrimental to leukemic cells due to derepression of endogenous retroviral elements (ERV) and inhibition of HOXA9 transcriptional activity22,24 (Figure 8C). Given the essential role for SETDB1 in leukemia, small molecule inhibition of H3K9 methyltransferases has been proposed as a therapeutic option.22,23 However, a recent study shows depletion of G9a increased cancer progenitor cell populations that initiate a delayed but more aggressive disease state.21 Thus, it is critical to fully understand the effects of chemically inhibiting of H3K9 methylation as a treatment for AML. Further investigation into the roles of SETDB1, G9a and more generally H3K9 methylation levels will likely shed light on the precise role of these methyltransferases in normal and malignant hematopoiesis and determine the value of these epigenetic modifiers as therapeutic targets. Acknowledgments The authors thank Dr. Jianyong Shou for a SETDB1 expression construct and Dr. Schahram Akbarian for Setdb1fl/fl mice. We also thank Dr. Russell Ryan, Dr. Emmalee Adelman, Dr. Maria Figueroa, and Dr. Sami Malek for helpful discussion. Funding This work was supported by NIH grants R01-HL-136420 (AGM), T32CA140044 (JR) and P30CA046592 (University of Michigan Flow Cytometry Core).

haematologica | 2020; 105(9)

SETDB1 expression suppresses MLL-fusion driven AML

References 1. Flavahan WA, Gaskell E, Bernstein BE. Epigenetic plasticity and the hallmarks of cancer. Science. 2017;357(6348). 2. Cancer Genome Atlas Research Network, Ley TJ, Miller C, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013; 368(22):2059-2074. 3. Wang J, Jia ST, Jia S. New insights into the regulation of heterochromatin. Trends Genet. 2016;32(5):284-294. 4. Rao VK, Pal A, Taneja R. A drive in SUVs: from development to disease. Epigenetics. 2017;12(3):177-186. 5. Ropa J, Saha N, Chen Z, et al. PAF1 complex interactions with SETDB1 mediate promoter H3K9 methylation and transcriptional repression of Hoxa9 and Meis1 in acute myeloid leukemia. Oncotarget. 2018;9(31):22123-22136. 6. Yang Y-J, Han J-W, Youn H-D, Cho E-J. The tumor suppressor, parafibromin, mediates histone H3 K9 methylation for cyclin D1 repression. Nucleic Acids Res. 2010; 38(2):382-390. 7. Jaehning JA. The Paf1 complex: platform or player in RNA polymerase II transcription? Biochim Biophys Acta. 2010;1799(5-6):379388. 8. Van Oss SB, Cucinotta CE, Arndt KM. Emerging insights into the roles of the Paf1 complex in gene regulation. Trends Biochem Sci. 2017;42(10):788-798. 9. Chen FX, Xie P, Collings CK, et al. PAF1 regulation of promoter-proximal pause release via enhancer activation. Science. 2017;357(6357):1294-1298. 10. Krogan NJ, Dover J, Wood A, et al. The Paf1 complex is required for histone H3 methylation by COMPASS and Dot1p: linking transcriptional elongation to histone methylation. Mol Cell. 2003; 11(3):721-729. 11. Yu M, Yang W, Ni T, et al. RNA polymerase II-associated factor 1 regulates the release and phosphorylation of paused RNA polymerase II. Science. 2015;350(6266):13831386. 12. Serio J, Ropa J, Chen W, et al. The PAF complex regulation of Prmt5 facilitates the progression and maintenance of MLL fusion leukemia. Oncogene. 2018;37(4):450-460. 13. Milne TA, Kim J, Wang GG, et al. Multiple interactions recruit MLL1 and MLL1 fusion proteins to the HOXA9 locus in leukemogenesis. Mol Cell. 2010;38(6):853-863. 14. Muntean AG, Chen W, Jones M, Granowicz EM, Maillard I, Hess JL. MLL fusion protein-driven AML is selectively inhibited by targeted disruption of the MLL-PAFc interaction. Blood. 2013; 122(11):1914-1922. 15. Muntean AG, Tan J, Sitwala K, et al. The PAF complex synergizes with MLL fusion proteins at HOX loci to promote leukemogenesis. Cancer Cell. 2010;17(6):609-621. 16. Collins CT, Hess JL. Role of HOXA9 in leukemia: dysregulation, cofactors and essential targets. Oncogene. 2016;35(9):1090-1098. 17. MĂźller-Tidow C, Klein H-U, Hascher A, et al. Profiling of histone H3 lysine 9 trimethylation levels predicts transcription factor activity and survival in acute myeloid leukemia. Blood. 2010; 116(18):

haematologica | 2020; 105(9)

3564-3571. 18. Schultz DC, Ayyanathan K, Negorev D, Maul GG, Rauscher FJ. SETDB1: a novel KAP-1-associated histone H3, lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins. Genes Dev. 2002;16(8):919-932. 19. Bilodeau S, Kagey MH, Frampton GM, Rahl PB, Young RA. SetDB1 contributes to repression of genes encoding developmental regulators and maintenance of ES cell state. Genes Dev. 2009;23(21):2484-2489. 20. Ceol CJ, Houvras Y, Jane-Valbuena J, et al. The histone methyltransferase SETDB1 is recurrently amplified in melanoma and accelerates its onset. Nature. 2011;471 (7339):513-517. 21. Avgustinova A, Symeonidi A, Castellanos A, et al. Loss of G9a preserves mutation patterns but increases chromatin accessibility, genomic instability and aggressiveness in skin tumours. Nat Cell Biol. 2018;20(12):1400-1409. 22. Lehnertz B, Pabst C, Su L, et al. The methyltransferase G9a regulates HoxA9dependent transcription in AML. Genes Dev. 2014;28(4):317-327. 23. Koide S, Oshima M, Takubo K, et al. Setdb1 maintains hematopoietic stem and progenitor cells by restricting the ectopic activation of nonhematopoietic genes. Blood. 2016;128(5):638-649. 24. Cuellar TL, Herzner A-M, Zhang X, et al. Silencing of retrotransposons by SETDB1 inhibits the interferon response in acute myeloid leukemia. J Cell Biol. 2017; 216(11):3535-3549. 25. Bagger FO, Sasivarevic D, Sohi SH, et al. BloodSpot: a database of gene expression profiles and transcriptional programs for healthy and malignant haematopoiesis. Nucleic Acids Res. 2016;44(D1):D917-24. 26. Rapin N, Bagger FO, Jendholm J, et al. Comparing cancer vs normal gene expression profiles identifies new disease entities and common transcriptional programs in AML patients. Blood. 2014;123(6):894-904. 27. Jiang Y, Loh Y-HE, Rajarajan P, et al. The methyltransferase SETDB1 regulates a large neuron-specific topological chromatin domain. Nat Genet. 2017;49(8):1239-1250. 28. Kim Y, Lee H-M, Xiong Y, et al. Targeting the histone methyltransferase G9a activates imprinted genes and improves survival of a mouse model of Prader-Willi syndrome. Nat Med. 2017;23(2):213-222. 29. Vedadi M, Barsyte-Lovejoy D, Liu F, et al. A chemical probe selectively inhibits G9a and GLP methyltransferase activity in cells. Nat Chem Biol. 2011;7(8):566-574. 30. Ugarte F, Sousae R, Cinquin B, et al. Progressive chromatin condensation and H3K9 methylation regulate the differentiation of embryonic and hematopoietic stem cells. Stem Cell Rep. 2015;5(5):728-740. 31. Chen X, Skutt-Kakaria K, Davison J, et al. G9a/GLP-dependent histone H3K9me2 patterning during human hematopoietic stem cell lineage commitment. Genes Dev. 2012;26(22):2499-2511. 32. Krivtsov AV, Figueroa ME, Sinha AU, et al. Cell of origin determines clinically relevant subtypes of MLL-rearranged AML. Leukemia. 2013;27(4):852-860. 33. Hess JL, Bittner CB, Zeisig DT, et al. c-Myb is an essential downstream target for homeobox-mediated transformation of

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

hematopoietic cells. Blood. 2006;108(1): 297-304. Ivanova NB, Dimos JT, Schaniel C, Hackney JA, Moore KA, Lemischka IR. A stem cell molecular signature. Science. 2002;298(5593):601-604. Laslo P, Spooner CJ, Warmflash A, et al. Multilineage transcriptional priming and determination of alternate hematopoietic cell fates. Cell. 2006;126(4):755-766. Bernt KM, Zhu N, Sinha AU, et al. MLLrearranged leukemia is dependent on aberrant H3K79 methylation by DOT1L. Cancer Cell. 2011;20(1):66-78. Faber J, Krivtsov AV, Stubbs MC, et al. HOXA9 is required for survival in human MLL-rearranged acute leukemias. Blood. 2009;113(11):2375-2385. Ghare SS, Joshi-Barve S, Moghe A, et al. Coordinated histone H3 methylation and acetylation regulate physiologic and pathologic fas ligand gene expression in human CD4+ T cells. J Immunol. 2014;193(1):412421. Celik S, Tangi F, Oktenli C. Increased frequency of Mediterranean fever gene variants in multiple myeloma. Oncol Lett. 2014;8(4):1735-1738. Wang Q-F, Wu G, Mi S, et al. MLL fusion proteins preferentially regulate a subset of wild-type MLL target genes in the leukemic genome. Blood. 2011;117(25):6895-6905. Lee S-H, Chiu Y-C, Li Y-H, et al. High expression of dedicator of cytokinesis 1 (DOCK1) confers poor prognosis in acute myeloid leukemia. Oncotarget. 2017; 8(42):72250-72259. Jung N, Dai B, Gentles AJ, Majeti R, Feinberg AP. An LSC epigenetic signature is largely mutation independent and implicates the HOXA cluster in AML pathogenesis. Nat Commun. 2015;6:8489. Volpe G, Walton DS, Grainger DE, et al. Prognostic significance of high GFI1 expression in AML of normal karyotype and its association with a FLT3-ITD signature. Sci Rep. 2017;7(1):11148. Peters AH, Oâ&#x20AC;&#x2122;Carroll D, Scherthan H, et al. Loss of the Suv39h histone methyltransferases impairs mammalian heterochromatin and genome stability. Cell. 2001; 107(3):323-337. Falandry C, Fourel G, Galy V, et al. CLLD8/KMT1F is a lysine methyltransferase that is important for chromosome segregation. J Biol Chem. 2010; 285(26):20234-20241. Zhang Y, Li Y, Zhang L, Fang X, Zhen C, Wang X. Prognostic importance of DOCK1 transcript levels, and biologic insights from DOCK1 -associated gene and microrna expression signatures in de novo acute myeloid leukemia. Blood. 2017;130(Suppl 1):3952. Chu Y, Chen Y, Li M, et al. Six1 regulates leukemia stem cell maintenance in acute myeloid leukemia. Cancer Sci. 2019; 110(7):2200-2210. Wong C-M, Wei L, Law C-T, et al. Up-regulation of histone methyltransferase SETDB1 by multiple mechanisms in hepatocellular carcinoma promotes cancer metastasis. Hepatology. 2016;63(2):474-487. Fritsch L, Robin P, Mathieu JRR, et al. A subset of the histone H3 lysine 9 methyltransferases Suv39h1, G9a, GLP, and SETDB1 participate in a multimeric complex. Mol Cell. 2010;37(1):46-56.

2285

ARTICLE Ferrata Storti Foundation

Acute Myeloid Leukemia

ASLAN003, a potent dihydroorotate dehydrogenase inhibitor for differentiation of acute myeloid leukemia

Jianbiao Zhou,1,2 Jessie Yiying Quah,1* Yvonne Ng,1* Jing-Yuan Chooi,2* Sabrina Hui-Min Toh,1* Baohong Lin,3 Tuan Zea Tan,1 Hiroki Hosoi,1 Motomi Osato,1,4 Qihui Seet,5 A.G. Lisa Ooi,5 Bertil Lindmark,5 Mark McHale5 and Wee-Joo Chng1,2,3

Haematologica 2020 Volume 105(9):2286-2297

1 Cancer Science Institute of Singapore, National University of Singapore; 2Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore; 3 Department of Hematology-Oncology, National University Cancer Institute, NUHS; 4 Department of Pediatrics, National University of Singapore, Yong Loo Lin School of Medicine and 5ASLAN Pharmaceuticals, Singapore

*JYQ, YN, J-YC and SH-MT contributed equally to this work.

ABSTRACT

Correspondence: WEE-JOO CHNG mdccwj@nus.edu.sg Received: June 22, 2019. Accepted: November 5, 2019. Pre-published: November 7, 2019. doi:10.3324/haematol.2019.230482 ÂŠ2020 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.

2286

ifferentiation therapies achieve remarkable success in acute promyelocytic leukemia, a subtype of acute myeloid leukemia (AML). However, excluding acute promyelocytic leukemia, clinical benefits of differentiation therapies in AML are negligible except for those targeting mutant isocitrate dehydrogenase 1/2. Dihydroorotate dehydrogenase catalyzes the fourth step of the de novo pyrimidine synthesis pathway. ASLAN003 is a highly potent dihydroorotate dehydrogenase inhibitor that induces differentiation, as well as reducing cell proliferation and viability, of AML cell lines and primary AML blasts including chemoresistant cells. Apoptotic pathways are triggered by ASLAN003, and this drug also significantly inhibits protein synthesis and activates AP-1 transcription, contributing to its capacity to promote differentiation. Finally, ASLAN003 substantially reduces leukemic burden and prolongs survival in AML xenograft mice and AML patient-derived xenograft models. Notably, the drug has no evident effect on normal hematopoietic cells and exhibits excellent safety profiles in mice, even after a prolonged period of administration. Our results, therefore, suggest that ASLAN003 is an agent targeting dihydroorotate dehydrogenase with potential for use in the treatment of AML. ASLAN003 is currently being evaluated in a phase IIa clinical trial in patients with AML.

Introduction Acute myeloid leukemia (AML) cells originate from hematopoietic stem cells, but fail to differentiate into functional mature cells; instead, they are arrested at an early stage of differentiation.1-4 AML-M3 (according to the French-American-British classification), acute promyelocytic leukemia, is a unique subtype with a specific t(15;17) chromosomal translocation, resulting in the PML-RARA fusion gene.5 The introduction of all-trans retinoic acid, a vitamin A metabolite, and subsequently arsenic trioxide, transformed the clinical management of acute promyelocytic leukemia, turning a highly fatal disease into a definitively curable one that can be treated without the need for toxic chemotherapy.6,7 In contrast to their excellent effectiveness in acute promyelocytic leukemia, differentiation therapies have not been as effective in the other types of AML despite decades of intensive laboratory research and numerous clinical trials. The one exception to date is treatment targeting AML with mutated isocitrate dehydrogenase (IDH) 1 or 2.8,9 Enasidenib, a selective, non-competitive inhibitor of IDH2, induces differentiation of AML cells through reducing the oncometabolite 2hydroxyglutarate in mutated IDH2.10 Ivosidenib, an IDH1 inhibitor, also induces differentiation through a similar mechanism in mutated IDH1. The approval of haematologica | 2020; 105(9)

ASLAN003 for differentiation of AML

enasidenib and ivosidenib for relapsed/refractory AML with mutated IDH2 and mutated IDH1, respectively, by the USA Food and Drug Administration renewed enthusiasm for differentiation therapy. Pyrimidines and pyrimidine derivatives are the building blocks of both deoxyribonucleic acid (DNA) and ribonucleic acid (RNA), and protein glycosylation, which are the essential cellular components.11 Dihydroorotate dehydrogenase (DHODH) catalyzes the fourth enzymatic step in de novo pyrimidine biosynthesis, converting the ubiquinone-mediated oxidation of dihydroorotate to orotate.12,13 DHODH has been a therapeutic target for malaria, rheumatoid arthritis, and multiple sclerosis.14-16 Recently, an elegant study revealed an unexpected role of DHODH in the differentiation of AML blast cells.17 The DHODH inhibitor used in that study, brequinar, was originally discovered by Du Pont in 1985.17,18 However, clinical trials of brequinar in solid tumors demonstrated myelosuppression with predominant thrombocytopenia, which limit its potential use in AML.16,19 ASLAN003 (LAS186323) is a novel, bioavailable and potent small molecule DHODH inhibitor. The drug was discovered by Almirall, S.A. and global rights to the compound were granted to ASLAN Pharmaceuticals Singapore in 2012, which re-named it as ASLAN003. ASLAN003 is a potent inhibitor of human DHODH enzyme activity, with a half maximal inhibitory concentration (IC50) of 35 nM, and high plasma protein binding (>99%). In phase I single and multiple ascending dose clinical trials, ASLAN003 has been shown to be tolerated by healthy volunteers. In this study, we set out to investigate the effects of ASLAN003 on AML cell function in vitro and in vivo, as well as to elucidate the molecular mechanism of DHODH inhibition of AML cell differentiation.

Methods Cell lines and primary acute myeloid leukemia cells, drugs and chemicals Details on the cell lines and primary bone marrow (BM) cell culture, drugs and chemicals are provided in the Online Supplementary Methods.

Cell viability assays, western blot analysis, polymerase chain reaction, and FACS analysis Experiments were conducted as previously described.20,21 Details of the cell viability assays, western blot analysis, realtime quantitative reverse transcriptase polymerase chain reaction (qRT-PCR), and FACS analysis of myeloid cell surface antigens are available in the Online Supplementary Methods. The primer sequences are provided in Online Supplementary Table S1.

RNA sequencing and data analysis The experiments and subsequent analysis of KG-1 and MOLM14 cells treated with ASLAN003 or DMSO were performed as detailed in the Online Supplementary Methods.

Protein synthesis assays Click-iT assays were performed using an O-propargylpuromycin (OPP) Alexa Fluor® 488 Protein Synthesis Assay Kit from ThermoFisher (C10456) according to the manufacturer’s recommendation. MOLM-14 and KG-1 cells were exposed to ASLAN003 1 μM or 2 μM for 1 h before OPP 20 mM was added for 1 h. DMSO was used as a control. Cells were washed in icecold phosphate-buffered saline and then fixed and permeabilized prior to FACS analysis using a LSRII flow cytometer (BD Biosciences).

In vivo efficacy of ASLAN003 The in vivo efficacy of ASLAN003 was tested in a human AML cell line xenograft model and in human AML patient-derived xenograft (PDX) models. For the human AML cell line xenograft model, we used female NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ, NGS mice (4-6 weeks old), purchased from The Jackson Laboratory (Bar Harbor, ME, USA) through InVivos (Singapore). The animals were maintained in specific pathogen-free conditions. Exponentially growing THP-1 and MOLM-14 cells (3 × 106 each) were injected into mice via the tail vein. From the second day of inoculation of AML cells, the mice were administered either vehicle, as a control, or ASLAN003 50 mg/kg by oral gavage once daily in a 200 μL volume. For the human AML PDX models, the AML-14 PDX line was established from a patient with AML-M4 with a normal karyotype, while the AML-23 PDX line was established from a patient diagnosed with chronic myeloid leukemia in accelerated phase. The protocols were reviewed and approved by the Institutional Animal Care and Use Committee in compliance with guidelines on the care and use of animals for scientific purpose. More details are provided in the Online Supplementary Methods.

Statistical analysis A Student t-test (two-tailed paired) was used to examine the statistical difference for in vitro cell line experiments, and P values <0.05 were considered to be statistically significant. Data are presented as mean ± standard deviation (SD). Kaplan-Meier analyses were conducted using GraphPad Prism® version 7 (GraphPad Software; La Jolla, CA, USA) and statistical significance was calculated by the log-rank test (P<0.05).

Data availability RNA-sequencing data for MOLM-14 and KG-1 cells have been deposited in the Gene Expression Omnibus with accession number GSE128950.

Wright-Giemsa staining and nitro blue tetrazolium assay

Results

After exposure to ASLAN003 or dimethylsulfoxide (DMSO) for 96 h, AML cells (1 x 106) were harvested and distributed equally for Wright-Giemsa staining and nitro blue tetrazolium (NBT) reduction assay (details in the Online Supplementary Methods).

ASLAN003 inhibits cell proliferation and induces cell differentiation of acute myeloid leukemia cell lines

Assessment of mitochondrial membrane potential The mitochondrial membrane potential was determined using a JC-10 Assay Kit (Sigma, MAK160). Details are provided in the Online Supplementary Methods. haematologica | 2020; 105(9)

ASLAN003 was found to inhibit leukemic cell proliferation of THP-1, MOLM-14 and KG-1 with IC50 values of 152 nM, 582 nM, and 382 nM, respectively (Figure 1A). It is worth noting that cell viability was maintained at ~50% at ASLAN003 1 μM and higher (Figure 1A). This indicates that the mode of action of ASLAN003 differs from that of cytotoxic drugs, which cause increased cell death with higher doses. We also examined the selectivity of 2287

J. Zhou et al. A

B D

Figure 1. Effects of ASLAN003 on cell viability and differentiation in human acute myeloid leukemia cell lines. (A) Dose-response curves of ASLAN003 treatment for 48 h on cell viability of the acute myeloid leukemia (AML) cell lines THP-1, MOLM-14, and KG-1. The percentage of cell viability relative to that of dimethylsulfoxide (DMSO)-treated cells is shown. Data represent three independent replicates [mean ± standard deviation (SD)]. (B) FACS analysis of myeloid differentiation cell surface antigens CD11b and CD14 on ASLAN003-treated and DMSO-treated AML cell lines. The treatment time was 96 h. Data represent the mean ± SD of three replicates. **P<0.01; *P<0.05. (C) Representative images of Wright-Giemsa staining for morphological examination of AML cell lines treated with 100 nM ASLAN003 or DMSO for 96 h. The images were taken under an Olympus IX71 light microscope (Japan) with original magnification x 400 (objective lenses x 40). (D) Nitro blue tetrazolium (NBT) reduction assays for AML cells treated with 100 nM ASLAN003 or DMSO for 96 h. Cells positive for NBT-reducing activity, containing precipitated formazan particles, were counted. The bar graph shows the mean percentage of NBT+ cells in ten random 10 x (objective lenses) fields ± SD for each group. *P<0.001. (E) FACS analysis of CD11b+ cells after exposure of THP-1 cells to ASLAN003 or DMSO for 24 h and 48 h. The data display the mean ± SD of three different experiments. *P<0.001.

2288

haematologica | 2020; 105(9)

ASLAN003 for differentiation of AML

ASLAN003 and brequinar on AML cells over normal CD34+CD38+ BM cells obtained from healthy donors. CD34+CD38+ cells are considered as dividing myeloid progenitor cells. The IC50 values of ASLAN003 and brequinar were 5.22 μM and 2.87 μM, respectively (Online Supplementary Figure S1A). ASLAN003 and brequinar were on average 11-fold more active in AML cells than in normal CD34+CD38+ BM myeloid progenitor cells, suggesting a favorable therapeutic index for DHODH inhibitors. Next we examined the differentiation effects of ASLAN003 on THP-1, MOLM-14 and KG-1 cells. Treatment of these leukemic cells with ASLAN003 consistently resulted in a substantial increase of CD11b in all three cell lines. The CD11b+ population was increased by 86.7%, 63.9% and 86.5% in THP-1, MOLM-14 and KG-1 cells treated with 100 nM ASLAN003 after normalization to values for DMSO-treated samples (P<0.001) (Figure 1B, Online Supplementary Figure S1B). CD14+ cells were also significantly increased in MOLM-14, THP-1 and KG-1 cells following treatment with ASLAN003 (P<0.01) (Figure 1B, Online Supplementary Figure S1B). Secondly, cells treated with ASLAN003 displayed morphological changes with a lower nucleocytoplasmic ratio, condensed chromatin, and increased nuclear lobulation, which are characteristics of myeloid maturation (Figure 1C). Thirdly, we employed a NBT reduction assay to evaluate functional evidence of myeloid maturation. After 96 h of treatment with ASLAN003 100 nM, 95.2% of the THP-1 cells, 62.4% of the MOLM-14 cells, and 93.6% of the KG-1 cells were positive for NBT reduction (P<0.001). At concentrations of ASLAN003 as low as 50 nM, more than 50% of the THP-1 cells showed increased NBT reduction compared with DMSO controls (P<0.001) (Figure 1D). Furthermore, in a time-dependent experiment, following 24 or 48 h exposure to 100 nM ASLAN003, almost 100% of THP-1 cells became CD11b+ (Figure 1E, Online Supplementary Figure S1C). Taken together, these results suggest that ASLAN003 can rapidly induce differentiation of AML cells.

Differentiation effect of ASLAN003 and brequinar as well as uridine rescue Parallel experiments were carried out to compare the efficacy of ASLAN003 and brequinar. MOLM-14 cells were incubated with brequinar or ASLAN003 100 nM, a concentration similar to the half maximal effective concentration (EC50) of ASLAN003 in MOLM-14 cells, for 96 h. After normalization to the respective controls, MOLM14 cells treated with brequinar had 33.1% of CD11b+ cells, while those treated with the same dose of ASLAN003 had 63.9% of CD11b+ cells (Figure 2A, Online Supplementary Figure S1D). These results showed a nearly two-fold higher potency of ASLAN003 compared to brequinar (P<0.05). Because DHODH coverts dihydroorotate into orotate, which is a precursor of uridine, inhibition of DHODH leads to a diminished uridine pool in cells.16 In both MOLM-14 and THP-1 cells, ASLAN003-mediated differentiation was completely rescued by addition of 50 μM uridine (Figure 2B), with no significant further rescue detected in the presence of higher concentrations of uridine (100 μM and 150 μM) (data not shown). Interestingly, uridine also rescued the cell viability of ASLAN003-treated MOLM-14 and THP-1 cells (P<0.05) (Figure 2C). Overall, these data demonstrate that uridine could abrogate the effects of ASLAN003 on cell differentiation and haematologica | 2020; 105(9)

cell viability, implying the on-target specificity of ASLAN003.

ASLAN003 decreases viability and induces differentiation in primary acute myeloid leukemia blasts and myelodysplastic syndrome samples To confirm the clinical relevance of our observations in human AML cell lines, we examined the effect of ASLAN003 on cell viability and differentiation status in BM cells obtained from patients with de novo or relapsed AML and myelodysplastic syndrome (MDS). ASLAN003 displayed excellent potency in inducing differentiation and cell death in some primary AML blasts. For example, in patient UPN1 with AML-M1 with t(9;22) and a complex karyotype, exposure to ASLAN003 at the concentrations of 2 μM and 4 μM led, respectively, to 22% and 30% increases in CD11b+ cells, as well as 31% and 35% increases in CD14+ cells. Concomitantly, there were decreases of 18% and 27%, respectively, in cell viability. Furthermore, in patient UPN6 with AML-M2 with FLT3ITD and NPM1 mutations, following incubation with ASLAN003 2 μM and 4 μM for 96 h, we observed 18% and 23% more CD13/CD33 double positive cells, accompanying reduced cell viability. Among these tested samples, UNP5 with deletion of chromosome 7 had the most sensitive response, with the CD11b+ population increasing by 62% in response to ASLAN003 1 μM. Importantly, ASLAN003 was still effective in promoting differentiation and cell death of myeloid cells in relapsed AML (UNP13). Morphological analysis and NBT assays demonstrated the features of neutrophil differentiation in ASLAN003-treated AML blasts from selected cases (Online Supplementary Figure S2). In summary, the response of primary BM cells from AML patients to ASLAN003 was classified into three categories: sensitive if any of the myeloid markers CD11b, CD14, CD13 or CD33 increased ≥15%; moderately sensitive if the markers increased ≥5%, but <15%; and resistant if the markers did not increase or increased <5%. Among the AML samples, we observed six (43%) sensitive cases, six (43%) moderately sensitive cases and two (14%) resistant cases (Table 1). For BM samples from MDS patients, three cases (50%) were sensitive to ASLAN003 and three cases (50%) were moderately sensitive (Table 1). Thus, on the bases of these data, MDS cells appear to be sensitive to ASLAN003 treatment. No resistant cases were seen, but the number of cases tested was limited. Notably, ASLAN003, at the concentrations of 2 μM and 4 μM, was shown to have a negligible impact on cell viability and differentiation status of mononuclear cells from a healthy donor, suggesting that ASLAN003 is not toxic to normal hematopoietic cells (Table 1). Collectively, these experiments provide evidence that ASLAN003 treatment of primary cells obtained from either de novo or relapsed patients leads to myeloid differentiation and cell death.

Transcriptome analysis reveals the effects of ASLAN003 on apoptosis, differentiation, metabolism and translation initiation in acute myeloid leukemia cells To understand the impact of ASLAN003 on transcriptional networks in AML, we performed RNA-sequencing on DMSO- and ASLAN003-treated KG-1 and MOLM-14 cells. These cells shared 320 upregulated genes and 225 downregulated genes (posterior probability of differential 2289

J. Zhou et al.

expression: 0.95-1, fold-change cutoff: 1.5) (Figure 3A, Online Supplementary Table S2). We then manually classified the list of genes. Among the genes upregulated by ASLAN003, 27 (8.4%) were related to myeloid differentiation, 15 (4.7%) were cell surface antigens, and 8 (2.5%) were associated with apoptosis (Figure 3B). The downregulated gene list was particularly enriched with 49 ribosome family genes (19.6%) and 21 metabolism-related genes (8.4%) (Figure 3B). The gene expression changes of selected genes associated with apoptosis and myeloid differentiation were confirmed by qRT-PCR analysis (Figure

3C). Single-sample gene set enrichment analysis showed significant enrichment of “myeloid differentiation_up”, “hematopoietic stem cell_down”, “targets of HoxA9 and Meis1_down” signatures, and suppression of “pyrimidine ribonucleoside triphosphate metabolic process” (Figure 3D). These signatures were aligned with the observed effects of ASLAN003. The gene ontology term analysis revealed that upregulated genes were involved in cellular response to ”neutrophil degranulation”, “neutrophil mediated immunity”, “positive regulation of caspase activity”, “positive regulation of apoptosis”, “regulation of extrinsic

Figure 2. Comparing ASLAN003 versus brequinar and the effect of uridine on cell differentiation by ASLAN003. (A) Comparison between the effects of 100 nM ASLAN003 and 100 nM brequinar on differentiation of MOLM-14 cells. The absolute increase of CD11b+ cells was calculated based on the percentage of CD11b+ cells increased in treated samples compared to control samples exposed to dimethylsulfoxide (DMSO). Representative FACS plots are shown. (B, C) Supplementation of uridine blocks ASLAN003-induced differentiation (B) and cell death (C) in MOLM-14 and THP-1 cells. Leukemic cells were incubated with DMSO, ASLAN003, or ASLAN003 + 50 μM uridine. ASLAN003 was used at a concentration of 100 nM for MOLM-14 cells and 50 nM for THP-1 cells. The percentages of CD11b+ cells (B) or viable cells (C) are illustrated and representative FACS plots are shown. The data were based on triplicate different experiments (mean ± standard deviation). *P<0.05; n.s.: not significant.

2290

haematologica | 2020; 105(9)

ASLAN003 for differentiation of AML

apoptosis” and “positive regulation of myeloid cell differentiation”, while downregulated genes associated with ”co-translational protein targeting to membrane”, “rRNA metabolic process”, “ribosome biogenesis”, “translation” and “peptide biosynthetic process”, showed the most significant changes (Figure 3E). These data from transcriptome analysis suggest that ASLAN003 might impair protein synthesis and induce the differentiation and apoptosis transcriptional program in AML cells.

ASLAN003 triggers apoptotic pathways Fas (CD95, APO-1) belongs to the cell death receptor family. Upon binding to its ligand FasL, they form deathinducing signaling complex (DISC), initiating the extrinsic apoptosis cascade.22 Our RNA-sequencing data showed that the expression of FAS was upregulated after treatment (Figure 3B, C, Online Supplementary Table S2), suggesting that ASLAN003 could trigger the extrinsic apoptosis pathway. Indeed, western blot analysis revealed a significant increase in cleaved caspase 8, the extrinsic pathway-specific caspase, in ASLAN003-treated KG-1 and MOLM-14 cells (Figure 4A). Because DHODH is located on the inner mitochondrial membrane, we also assessed whether ASLAN003 could induce loss of mitochondrial membrane potential (ΔΨm), an indicator of the early phase of intrinsic (mitochondrial or BCL-2-regulated) apoptosis. In the DMSO-treated cells, the majority of cells had intact mitochondrial membranes. ASLAN003 was found to disrupt ΔΨm in a dose-depen-

dent manner in THP-1, MOLM-14 and KG-1 cells (Figure 4B). To investigate whether loss of ΔΨm is DHODH-specific, we conducted mitochondrial membrane potential assays in MOLM-14 cells treated with brequinar, another DHODH inhibitor and cytarabine, a chemotherapeutic agent. Both brequinar and cytarabine could induce loss of ΔΨm (Online Supplementary Figure S3), suggesting that loss of ΔΨm is not DHODH-specific, most likely being a marker of cell health. However, given the localization of DHODH, the effect of its inhibitors on mitochondrial membrane potential might be direct, while the effect of cytarabine might be indirect. As a result of loss of ΔΨm, increased leakage of cytochrome c from mitochondria into the cytosol, a characteristic of activation of the intrinsic apoptosis pathway, was also observed in ASLAN003treated cells (Figure 4A). ASLAN003 treatment also induced cleaved caspase-3 and -7 (Figure 4A). Altogether, these results suggest a role for both intrinsic and extrinsic pathways in ASLAN003-induced apoptosis.

ASLAN003 inhibits protein synthesis and induces differentiation of acute myeloid leukemia cells via activation of AP-1 transcription factors Transcriptome and gene ontology analysis showed a greater enrichment of protein translation-related genes and ribosome proteins among the genes downregulated by ASLAN003. Furthermore, gene expression of four members of the family of eukaryotic translation initiation factors (eIF), namely EEF1B2, EIF4B, EIF3L, and EEF1B2P3,

Table 1. Clinical characteristics of patients with acute myeloid leukemia and myelodysplastic syndrome and their responses to ASLAN003 in an ex vivo assay.

Response group

Diagnosis

Karyotype

FLT3

NPM1

Patient ID

Highly sensitive AML (43%) MDS (50%)

AML-M1 AML-M5 AML-M2 AML with MDS AML-M5 AML-M4 MDS MDS MDS

t(9;22) Normal +8 t(8;21) Normal -7 Normal Normal Complex

WT WT ITD NA NA WT NA NA NA

WT WT Mutant NA NA WT NA NA NA

UPN1 UPN4 UPN6 UPN7 UPN9 UPN5 UPN16 UPN17 UPN18

Moderately sensitive AML (43%) MDS (50%)

AML-M2 AML-M1 AML-M4 Relapsed AML AML-M4 AML-M4 MDS MDS MDS

Normal -9 Normal +13 Inv(16) Inv(16) Normal -7 -13, +8

TKD NA WT ITD TKD ITD NA WT NA

NA NA WT NA WT WT NA WT NA

UPN2 UPN8 UPN11 UPN12 UPN10 UPN13 UPN15 UPN19 UPN20

Resistant AML (14%)

AML-M5a AML-M1 Healthy donor

+8 +11 Normal

TKD WT WT

Mutant WT WT

UPN3 UPN14

The response of primary bone marrow cells to ASLAN003 was classified into three categories: sensitive if any of myeloid markers CD11b, CD14, CD13 or CD33 increased ≥15%; moderately sensitive if the markers increased ≥5%, but <15%; and resistant if the markers increased <5%, by FACS analysis. FLT3: fms-like tyrosine kinase; NPM: nucleophosmin-1; ID: identify; AML: acute myeloid leukemia; MDS: myelodysplastic syndromes; WT: wildtype; UPD: unique patient number; ITD: internal tandem duplication; NA; not available; TKD: tyrosine kinase domain.

haematologica | 2020; 105(9)

2291

2292

Figure 3. Transcriptome changes in acute myeloid leukemia cells induced by ASLAN003. (A) Venn diagrams showing gene expression data for MOLM-14 and KG-1 cells. The left diagram indicates the numbers of upregulated genes in each and both of the two cell lines and the right diagram represents the numbers of downregulated genes (percentage in brackets). (B) Functional classification of genes whose expression is altered in both MOLM-14 and KG-1 cell lines. (C) Real-time quantitative reverse transcriptase polymerase chain reaction confirmed the gene expression changes of five selected genes identified by RNA-sequencing experiments in MOLM-14 and KG-1 cells. The expression of each gene in dimethylsulfoxide (DMSO)treated cells was set at 1 (baseline). The experiments were performed in triplicate (mean Âą standard deviation). *P<0.001; **P<0.05. (D) Single-sample gene set enrichment analysis revealed significantly greater enrichment of myeloid differentiation up-regulation, hematopoietic stem cell downregulation, targets of HoxA9 and Meis1 downregulation in ASLAN003-treated MOLM-14 and KG-1 cells, as well as lower enrichment of pyrimidine ribonucleoside triphosphate metabolic process. Individual P values for each pathway in DMSO-treated versus two ASLAN003-treated cell lines are shown. (E) Gene ontology (GO) analysis of upand down-regulated genes shared between MOLM-14 and KG-1 cells. Bar graphs showed the enriched GO terms of biological processes and their corresponding P values (-log10).

J. Zhou et al.

haematologica | 2020; 105(9)

ASLAN003 for differentiation of AML

Figure 4. ASLAN003-induced apoptosis involves both mitochondrial and death receptor pathways. (A) Whole-cell lysates or cytosolic fractions from KG1 and MOLM-14 cells treated with dimethylsulfoxide (DMSO), 0.5 μM or 1 μM ASLAN003 for 48 h were used to measure apoptosis-related proteins by western blot analyses. (B) THP-1, MOLM-14 and KG1 cells were exposed to 1 μM or 2 μM ASLAN003 or DMSO as the control for 48 h and then stained with JC-10 and analyzed by flow cytometry for quantification of intrinsic mitochondrial membrane potential (MMP). Representative FACS plots are shown. The bars in the bar chart represent the means of the increased depolarization of MMP of the three cell lines after exposure to ASLAN003 in two independent experiments, the error bars denote the standard deviation.

haematologica | 2020; 105(9)

2293

J. Zhou et al.

were significantly decreased in ASLAN003-treated cells (Online Supplementary Table S2). The eIF proteins are essential factors for protein synthesis, hence we performed an assay to determine the effects of ASLAN003 on protein synthesis. Indeed, ASLAN003 inhibited protein synthesis, as demonstrated by the reduced incorporation of OPP at protein translation sites in both MOLM-14 and KG-1 cells (Figure 5A). We also investigated proteins involved in the control of the mRNA translation process. In agreement with the RNA-sequencing data, western blot analysis confirmed the downregulation of EIF4B, and RPL6 proteins (Figure 5B) As revealed by RNA-sequencing and qRT-PCR analysis, FOS and JUN were upregulated by ASLAN003. We, therefore, decided to delineate the role of activating protein 1 (AP-1) transcription factors in ASLAN003-mediated differentiation. Western blot analysis confirmed the dosedependent increase of c-FOS protein levels in response to exposure to ASLAN003 (Figure 5B). T-5224 is a selective, small-molecule inhibitor of AP-1.23 Of note, T-5224 did not decrease cell viability when applied at doses of up to 125 μM in MOLM-14 and KG-1 cells (Online Supplementary Figure S4A). We then evaluated the effect of T-5224 on ASLAN003-mediated differentiation. Addition of 20 μM T-5224 completely abolished the differentiation effect of ASLAN003 in KG-1 cells and dampened the effect by half in MOLM-14 cells (both P<0.001) (Figure 5C, Online Supplementary Figure S4B). T-5224 alone had a minimal

effect on differentiation (Figure 5C). These results therefore suggest that ASLAN003-mediated differentiation is facilitated, at least partially, via activation of AP-1 transcription factors.

Robust in vivo efficacy of ASLAN003 in multiple mouse xenograft acute myeloid leukemia models To determine in vivo efficacy of ASLAN003 in AML, we first used two mouse xenograft models of the human AML cell lines, MOLM-14, and THP-1. Treatment with ASLAN003 (50 mg/kg, once daily oral gavage) was well tolerated as evidenced by the fact that there were no significant differences in body weight, hemoglobin concentration or platelet counts between the vehicle control and treated groups in these two models (Online Supplementary Figure S5A). Survival was significantly prolonged in the ASLAN003-treated groups compared to the vehicle control groups in both xenograft models (P=0.03 and P<0.001) (Figure 6A). In the MOLM-14 xenograft model, ASLAN003 substantially reduced the number of disseminated tumors, but also the size of these tumors relative to those in controls (Online Supplementary Figure S5B). Interestingly, in the THP-1 xenograft model, we observed that the livers of control mice were swollen and the surfaces were covered by copious white dots, a manifestation of leukemic infiltration. In sharp contrast, the appearance and size of the livers remained largely normal in ASLAN003-treated mice bearing THP-1 cells (Online Supplementary Figure S5B). Taken

Figure 5. Effects of ASLAN003 on protein synthesis and AP-1 transcription factors. (A) MOLM-14 and KG-1 cells were incubated with ASLAN003 at the doses of 1 μM and 2 μM for 1 h before addition of O-propargyl-puromycin (OPP) reagent for 1 h, followed by flow cytometry analyses. The graph represents fold decreases in OPP labeling [means ± standard deviation (SD)] (n=3) in MOLM-14 and KG-1 cells, with values for dimethylsulfoxide (DMSO)-treated cells set at 1.0. Statistical comparisons between groups are shown (Student t-test). *P<0.05; **P<0.01. (B) Immunoblotting analysis of whole cell lysates extracted from KG-1 and MOLM-14 cells for markers as indicated. The treatment time was 48 h. GAPDH was used as a loading control. (C) MOLM-14 and KG-1 cells were treated with ASLAN003 alone, T5224 alone, the two drugs in combination, or DMSO as a control, for 96 h, and then subjected to FACS quantification of human CD11b antigen. The graphs show the percentage of the CD11b+ population (n=2, mean ± SD). Statistical comparisons between the effects of ASLAN003 as a single agent and in combination treatment are shown (**P<0.001). The percentage of CD11b+ cells were not statistically different between DMSO- and T-5224-treated samples (P>0.05).

2294

haematologica | 2020; 105(9)

ASLAN003 for differentiation of AML

together, the therapeutic effects of ASLAN003 were multiple: prolonged survival, and prevention of tumor dissemination and leukemic infiltration into organs. Next, we tested the in vivo effect of ASLAN003 on leukemic burden (human CD45+ cells) and differentiation (CD11b+ or CD14+ cells). The numbers of human CD45+ cells in BM, peripheral blood, spleen and liver were all significantly reduced in ASLAN003-treated mice compared to those in control mice in both models (Figure 6B). In

agreement with in vitro observations, FACS analysis showed significantly increased numbers of CD11b+ and CD14+ cells in BM of treated mice in both models (Figure 6B). Taken together, these experiments confirmed that ASLAN003 could induce differentiation and reduce leukemic burden in vivo. We also evaluated the therapeutic efficacy of ASLAN003 in patient-derived AML xenografts. For AML14, an indolent line, at the end of experiments all mice

Figure 6. The in vivo efficacy of ASLAN003 in xenograft models. (A) Kaplan-Meier survival curves of mice treated with either ASLAN003 or vehicle control for the MOLM-14 and THP-1 xenograft models. Mice were administered either ASLAN003 50 mg/kg or the same volume of the vehicle by daily oral gavage. Treatment was started 3 days after inoculation of leukemic cells. The number of mice in each group and log-rank P values are indicated. (B) The leukemic burdens in mouse bone marrow (BM), peripheral blood (PB), spleen and liver were compared for the ASLAN003-treated and vehicle-treated groups for MOLM-14 xenograft models (n=3) and THP-1 models (n=4). The number of human CD45+ cells was determined by FACS analysis as a surrogate marker for leukemic burden. FACS analysis was performed to assess the percentage of human specific CD11b+, CD14+ leukemic cells in BM and PB samples harvested from these mice. *P<0.05; **P<0.01. (C) FACS analysis of human CD45+ cells in BM samples harvested from mice with AML-14 patient-derived xenotransplants (PDX) treated with ASLAN003 or vehicle control (n=4, *P=0.039). (D) Kaplan-Meier survival curves of animals with AML-23 PDX receiving ASLAN003 or control treatment. (E) FACS analysis was performed to assess the percentage of human specific CD45+ in BM, PB, spleen and liver, as well as CD11b+, CD14+ leukemic cells in BM samples harvested from mice with AML23 PDX (n=3). *P<0.05; **P<0.01.

haematologica | 2020; 105(9)

2295

J. Zhou et al.

were alive and no obvious signs of disease were observed in either the control or ASLAN003 group. However, we found that the leukemic burden was significantly less in ASLAN003-treated PDX than in vehicle-treated PDX (P=0.04) (Figure 6C). The weight of the mice increased gradually and similarly in both groups during the treatment (P=0.42) (Online Supplementary Figure S5C). For AML-23, an aggressive PDX line, all animals with vehicletreated xenografts (n=9) succumbed to the disease within a month (median survival 20 days). In contrast, 50% of the animals with ASLAN003-treated PDX (n=8) were still active and alive at the end of experiments on day 37 (P=0.0002) (Figure 6D). Importantly, the percentages of human CD45+ leukemia cells were significantly reduced in BM, peripheral blood, spleen, and liver of the ASLAN003 group as compared to the control group. The in vivo differentiation effect of ASLAN003 was also confirmed by the observation of increased human CD11b+ and CD14+ cells in BM (Figure 6E). As for the AML-14 line, there was no statistical difference between the body weight of the AML-23 control group and the ASLAN003-treated group (P=0.73) (Online Supplementary Figure S5C). Overall, these data demonstrate that ASLAN003 treatment mediates therapeutic efficacy in AML-PDX by extending survival, reducing leukemic burden and inducing differentiation. Notably, ASLAN003 appears safe and well tolerated even after prolonged in vivo administration.

Discussion Recently, DHODH has been demonstrated to be a novel target for differentiation therapy in AML.24 Brequinar is the first DHODH inhibitor that has shown potency in inducing differentiation, but its clinical use is impeded by its hematologic toxicity and ineffectiveness in early trials in patients with solid tumors.19,25,26 Several other DHODH inhibitors have been described. PTC299, an inhibitor of VEGFA mRNA translation, has also been shown to target DHODH and to have broad activity against hematologic cancer cells in a preclinical setting.27 Isobavachalcone, a natural product, has been reported to target DHODH, resulting in apoptosis and differentiation of AML cell lines at a high concentration (10 μM) in vitro.28 BAY 2402234, a novel DHODH inhibitor, induces differentiation and inhibits proliferation in multiple AML subtypes and is currently being evaluated in a phase I trial in myeloid malignancies.29 In this study, we comprehensively characterize ASLAN003, a novel, potent DHODH inhibitor. ASLAN003 induces massive differentiation of AML cell lines, as well as primary AML and MDS cells. ASLAN003 triggers apoptotic pathways in AML cell lines. Multiple mechanisms may account for these effects of ASLAN003 on leukemia cells. In general, myeloid leukemia cells have a higher proliferation rate compared to normal myeloid cells, thus requiring more energy (ATP) and abundant amounts of precursors for many biosynthetic pathways. The de novo biosynthesis of pyrimidine provides multiple essential precursors for such pathways. By targeting DHODH, a key enzyme in pyrimidine biosynthesis, ASLAN003 significantly depletes pyrimidine nucleotides, leaving insufficient precursors for leukemia cells to biosynthesize DNA, RNA, and proteins. Consistently, our RNA-sequencing data revealed that a large family of genes 2296

associated with protein translation initiation was the top and the largest class downregulated by ASLAN003 treatment. We further experimentally validated that ASLAN003 inhibits protein synthesis in AML cells. A growing body of evidence supports a critical onco-addiction on active protein translation in AML cells. Aberrant protein translation contributes to arrested differentiation of myeloid cells and leukemogenesis.30,31 EIF4B, one member of the eIF family, is downregulated by ASLAN003. EIF4B has been found to stimulate translation of a particular set of genes with long, structured 5’-untranslated regions, such as MYC, BCL-2, and XIAP, which promote cell survival and proliferation.32 Ribavirin, which blocks the binding of eIF4E to mRNA, has been shown to induce complete or partial remission in some relapsed AML patients.33 Notably, our study has determined that ASLAN003meditated AP-1 activation is important for the reversal of the blocked differentiation of AML cells. Transcription factors for AP-1 comprise several families of protein dimers, mainly JUN (c-Jun, JunB and JunD), FOS (c-Fos, FosB, Fra1, and Fra2) and ATF (ATFa, ATF-2, and ATF-3).34 It is known that AP-1 transcription factors are implicated in the differentiation of leukemia cells.35 Deletion of JunB in transgenic mice causes leukemogenic stem cell expansion, resulting in a myeloproliferative disorder which resembles early human chronic myelogenous leukemia.36 Early studies demonstrated that cytarabine treatment induced differentiation of AML cells and enhanced JUN/AP-1 activity was observed.37 Overexpression of c-Fos overrides the blockage of differentiation mediated by c-Myc and potentiates interleukin-6-induced differentiation in AML cells.38 In agreement with these findings, our data indicate a vital role for AP-1 in ASLAN003-induced differentiation of AML. The differentiation effect of ASLAN003 is almost completely negated in KG-1 cells and partially abrogated in MOLM-14 cells by co-treatment with an AP-1 inhibitor, T-5224. In further support of its clinical relevance, ASLAN003 induced primary AML blast myelocytic differentiation and decreased viability. It is worth noting that these therapeutic effects were achieved not only in samples from patients with de novo AML, but also in samples from patients with relapsed disease. In our study, a once daily dose of ASLAN003 50 mg/kg for 2 to 11 weeks in two AML cell line xenografts and two PDX models of NSG mice did not affect the animals’ body weight, indicating that the drug is safe and well-tolerated. In summary, our study demonstrates that ASLAN003 is a novel, potent DHODH inhibitor characterized by antiAML efficacy in vitro and in vivo and remarkable tolerability. We also provide molecular mechanisms through which ASLAN003 exerts multiple actions, including induction of apoptotic pathways, inhibition of protein translation and activation of AP-1 transcription factors. Taken together, our findings support the further development of ASLAN003 for clinical use in AML, a disease for which novel therapies are much needed. ASLAN003 has been granted orphan drug designation for the treatment of AML by the Food and Drug Administration and is currently being evaluated in a phase IIa trial in AML (ClinicalTrials.gov: NCT03451084). Acknowledgments This work was supported by a research fund from ASLAN Pharmaceuticals and the Singapore National Research haematologica | 2020; 105(9)

ASLAN003 for differentiation of AML

Foundation and the Ministry of Education under the Research Center of Excellence Program to WJC and NMRC ClinicianScientist IRG Grant CNIG11nov38 (to JZ). WJC is also supported by a NMRC Clinician Scientist Investigator award.

References 1. Heidel FH, Mar BG, Armstrong SA. Selfrenewal related signaling in myeloid leukemia stem cells. Int J Hematol. 2011;94(2):109-117. 2. Mack EKM, Marquardt A, Langer D, et al. Comprehensive genetic diagnosis of acute myeloid leukemia by next-generation sequencing. Haematologica. 2019;104(2): 277-287. 3. Misaghian N, Ligresti G, Steelman LS, et al. Targeting the leukemic stem cell: the Holy Grail of leukemia therapy. Leukemia. 2009;23(1):25-42. 4. Zhou J, Chng WJ. Identification and targeting leukemia stem cells: the path to the cure for acute myeloid leukemia. World J Stem Cells. 2014;6(4):473-484. 5. Zelent A, Guidez F, Melnick A, Waxman S, Licht JD. Translocations of the RARalpha gene in acute promyelocytic leukemia. Oncogene. 2001;20(49):7186-7203. 6. Fasan A, Haferlach C, Perglerova K, Kern W, Haferlach T. Molecular landscape of acute promyelocytic leukemia at diagnosis and relapse. Haematologica. 2017;102(6):e222e224. 7. Wang ZY, Chen Z. Acute promyelocytic leukemia: from highly fatal to highly curable. Blood. 2008;111(5):2505-2515. 8. Nowak D, Stewart D, Koeffler HP. Differentiation therapy of leukemia: 3 decades of development. Blood. 2009;113 (16):3655-3665. 9. Tenen DG. Disruption of differentiation in human cancer: AML shows the way. Nat Rev Cancer. 2003 ;3(2):89-101. 10. Stein EM, DiNardo CD, Pollyea DA, et al. Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood. 2017;130(6):722-731. 11. Huang M, Graves LM. De novo synthesis of pyrimidine nucleotides; emerging interfaces with signal transduction pathways. Cell Mol Life Sci. 2003;60(2):321-336. 12. Reis RAG, Calil FA, Feliciano PR, Pinheiro MP, Nonato MC. The dihydroorotate dehydrogenases: past and present. Arch Biochem Biophys. 2017;632:175-191. 13. Evans DR, Guy HI. Mammalian pyrimidine biosynthesis: fresh insights into an ancient pathway. J Biol Chem. 2004;279(32):3303533038. 14. Bar-Or A, Pachner A, Menguy-Vacheron F, Kaplan J, Wiendl H. Teriflunomide and its mechanism of action in multiple sclerosis.

haematologica | 2020; 105(9)

This study is also partially supported by the RNA Biology Center at CSI Singapore, NUS, by funding with Singapore Ministry of Educationâ&#x20AC;&#x2122;s Tier 3 grants, grant number MOE2014-T3-1-006.

Drugs. 2014;74(6):659-674. 15. Lolli ML, Sainas S, Pippione AC, et al. Use of human dihydroorotate dehydrogenase (hDHODH) inhibitors in autoimmune diseases and new perspectives in cancer therapy. Recent Pat Anticancer Drug Discov. 2018;13(1):86-105. 16. Munier-Lehmann H, Vidalain PO, Tangy F, Janin YL. On dihydroorotate dehydrogenases and their inhibitors and uses. J Med Chem. 2013;56(8):3148-3167. 17. 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. 18. Dexter DL, Hesson DP, Ardecky RJ, et al. Activity of a novel 4-quinolinecarboxylic acid, NSC 368390 [6-fluoro-2-(2'-fluoro-1,1'biphenyl-4-yl)-3-methyl-4-quinolinecarb oxylic acid sodium salt], against experimental tumors. Cancer Res. 1985;45(11 Pt 1):5563-5568. 19. Madak JT, Bankhead A 3rd, Cuthbertson CR, Showalter HD, Neamati N. Revisiting the role of dihydroorotate dehydrogenase as a therapeutic target for cancer. Pharmacol Ther. 2019;195:111-131. 20. Zhou J, Lu X, Tan TZ, Chng WJ. X-linked inhibitor of apoptosis inhibition sensitizes acute myeloid leukemia cell response to TRAIL and chemotherapy through potentiated induction of proapoptotic machinery. Mol Oncol. 2018;12(1):33-47. 21. Zhou J, Bi C, Janakakumara JV, et al. Enhanced activation of STAT pathways and overexpression of survivin confer resistance to FLT3 inhibitors and could be therapeutic targets in AML. Blood. 2009;113(17):40524062. 22. Debatin KM, Stahnke K, Fulda S. Apoptosis in hematological disorders. Semin Cancer Biol. 2003;13(2):149-158. 23. Avouac J, Palumbo K, Tomcik M, et al. Inhibition of activator protein 1 signaling abrogates transforming growth factor betamediated activation of fibroblasts and prevents experimental fibrosis. Arthritis Rheum. 2012;64(5):1642-1652. 24. Sykes DB. The emergence of dihydroorotate dehydrogenase (DHODH) as a therapeutic target in acute myeloid leukemia. Expert Opin Ther Targets. 2018;22(11):893-898. 25. 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. 26. Arteaga CL, Brown TD, Kuhn JG, et al. Phase I clinical and pharmacokinetic trial of

27.

28.

29.

30. 31.

32.

33.

34. 35.

36.

37.

38.

brequinar sodium (DuP 785; NSC 368390). Cancer Res. 1989;49(16):4648-4653. Cao L, Weetall M, Trotta C, et al. Targeting of hematologic malignancies with PTC299, a novel potent inhibitor of dihydroorotate dehydrogenase with favorable pharmaceutical properties. Mol Cancer Ther. 2019;18 (1):3-16. Wu D, Wang W, Chen W, et al. Pharmacological inhibition of dihydroorotate dehydrogenase induces apoptosis and differentiation in acute myeloid leukemia cells. Haematologica. 2018;103(9):1472-1483. Christian S, Merz C, Evans L, et al. The novel dihydroorotate dehydrogenase (DHODH) inhibitor BAY 2402234 triggers differentiation and is effective in the treatment of myeloid malignancies. Leukemia. 2019;33(10):2403-2415. Girardi T, De Keersmaecker K. T-ALL: ALL a matter of translation? Haematologica. 2015;100(3):293-295. Topisirovic I, Guzman ML, McConnell MJ, et al. Aberrant eukaryotic translation initiation factor 4E-dependent mRNA transport impedes hematopoietic differentiation and contributes to leukemogenesis. Mol Cell Biol. 2003;23(24):8992-9002. Shahbazian D, Parsyan A, Petroulakis E, et al. Control of cell survival and proliferation by mammalian eukaryotic initiation factor 4B. Mol Cell Biol. 2010;30(6):1478-1485. Assouline S, Culjkovic B, Cocolakis E, et al. Molecular targeting of the oncogene eIF4E in acute myeloid leukemia (AML): a proof-ofprinciple clinical trial with ribavirin. Blood. 2009;114(2):257-260. Hess J, Angel P, Schorpp-Kistner M. AP-1 subunits: quarrel and harmony among siblings. J Cell Sci. 2004;117(Pt 25):5965-5973. Liebermann DA, Gregory B, Hoffman B. AP1 (Fos/Jun) transcription factors in hematopoietic differentiation and apoptosis. Int J Oncol. 1998;12(3):685-700. Passegue E, Wagner EF, Weissman IL. JunB deficiency leads to a myeloproliferative disorder arising from hematopoietic stem cells. Cell. 2004;119(3):431-443. Brach MA, Herrmann F, Kufe DW. Activation of the AP-1 transcription factor by arabinofuranosylcytosine in myeloid leukemia cells. Blood. 1992;79(3):728-734. Shafarenko M, Amanullah A, Gregory B, Liebermann DA, Hoffman B. Fos modulates myeloid cell survival and differentiation and partially abrogates the c-Myc block in terminal myeloid differentiation. Blood. 2004;103 (11):4259-4267.

2297

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2298-2307

Non-Hodgkin Lymphoma

Prognostic impact of somatic mutations in diffuse large B-cell lymphoma and relationship to cell-of-origin: data from the phase III GOYA study Christopher R. Bolen,1* Magdalena Klanova,2,3,4* Marek Trnény,2 Laurie H. Sehn,5 Jie He,6 Jing Tong,6 Joseph N. Paulson,7 Eugene Kim,7 Umberto Vitolo,8 Alice Di Rocco,9 Günter Fingerle-Rowson,4 Tina Nielsen,4 Georg Lenz10 and Mikkel Z. Oestergaard11

Bioinformatics and Computational Biology, Genentech Inc., South San Francisco, CA, USA; 21st Department of Medicine, Charles University General Hospital, Prague, Czech Republic; 3Institute of Pathological Physiology, 1st Faculty of Medicine, Charles University, Prague, Czech Republic; 4Pharma Development Clinical Oncology, F. Hoffmann-La Roche Ltd., Basel, Switzerland; 5British Columbia Cancer Centre for Lymphoid Cancer, Vancouver, British Columbia, Canada; 6Foundation Medicine Inc., Cambridge, MA, USA; 7Department of Biostatistics, Product Development, Genentech Inc., South San Francisco, CA, USA; 8 A.O. Universitaria Città della Salute e della Scienza di Torino, Dipartimento di Ematologia, Torino, Italy; 9Department of Cellular Biotechnologies and Hematology, Sapienza University, Rome, Italy; 10Department of Medicine A, Hematology, Oncology and Pneumology, University Hospital Münster, Münster, Germany and 11Oncology Biomarker Development, F. Hoffmann-La Roche Ltd. Basel, Switzerland 1

*CRB and MK contributed equally as co-first authors.

ABSTRACT

Correspondence: CHRISTOPHER R. BOLEN bolen.christopher@gene.com Received: June 14, 2019. Accepted: November 14, 2019. Pre-published: November 14, 2019. doi:10.3324/haematol.2019.227892 ©2020 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.

2298

iffuse large B-cell lymphoma (DLBCL) represents a biologically and clinically heterogeneous diagnostic category with well-defined cellof-origin (COO) subtypes. Using data from the GOYA study (clinicaltrials.gov identifier: NCT01287741), we characterized the mutational profile of DLBCL and evaluated the prognostic impact of somatic mutations in relation to COO. Targeted DNA next-generation sequencing was performed in 499 formalin-fixed paraffin-embedded tissue biopsies from previously untreated patients. Prevalence of genetic alterations/mutations was examined. Multivariate Cox regression was used to evaluate the prognostic effect of individual genomic alterations. Of 465 genes analyzed, 59 were identified with mutations occurring in at least 10 of 499 patients (≥2% prevalence); 334 additional genes had mutations occurring in ≥1 patient. Single nucleotide variants were the most common mutation type. On multivariate analysis, BCL2 alterations were most strongly associated with shorter progression-free survival (multivariate hazard ratio: 2.6; 95% confidence interval: 1.6-4.2). BCL2 alterations were detected in 102 of 499 patients; 92 had BCL2 translocations, 90% of whom had germinal center B-cell-like DLBCL. BCL2 alterations were also significantly correlated with BCL2 gene and protein expression levels. Validation of published mutational subsets revealed consistent patterns of co-occurrence, but no consistent prognostic differences between subsets. Our data confirm the molecular heterogeneity of DLBCL, with potential treatment targets occurring in distinct COO subtypes.

Introduction Diffuse large B-cell lymphoma (DLBCL) represents a biologically and clinically heterogeneous diagnostic category. Distinct DLBCL cell-of-origin (COO) subtypes, arising from different stages of normal B-cell development and with different prognostic outcomes, were identified almost two decades ago.1-3 Several studies have since described the landscape of recurrent somatic mutations in DLBCL and demonstrated the molecular uniqueness of the distinct COO subtypes, and recent studies have suggested clinically relevant genetic subgroups exist within each subhaematologica | 2020; 105(9)

Somatic mutations in DLBCL

type.4-9 While germinal center B-cell-like (GCB) DLBCL is characterized by frequent translocations of the BCL2 gene, a key regulator of the intrinsic apoptotic pathway, or mutations of the epigenetic modifiers, CREBBP and EZH2, these abnormalities are rare in activated B-cell-like (ABC) DLBCL.10 In contrast, mutations in genes encoding proteins implicated in B-cell receptor signaling and the NFκB pathway, such as CD79b or MYD88, or genes involved in regulation of the cell cycle such as CDKN2A, contribute to the molecular pathogenesis of ABC DLBCL.11-14 While the prognostic impact of the distinct COO subtypes has been confirmed in several studies,2,3,15,16 the influence of key genomic alterations on the clinical outcomes of DLBCL patients is less clear, particularly their added clinical prognostic value over the International Prognostic Index (IPI) and COO. Mutations of several genes, such as TP53, MYD88 or CDKN2A, have been shown to be associated with poor prognosis in DLBCL patients.11,17-19 Many of these alterations, such as loss of CDKN2A or mutations of MYD88, are significantly enriched within the prognostically inferior ABC subtype and their independent prognostic role needs to be confirmed. A recent observational study by Reddy et al.19 retrospectively explored 150 genetic drivers of DLBCL in 1,001 patients and developed a genomic risk model comprising genetic alterations, COO DLBCL subtype, IPI score, and dual MYC and BCL2 expression, which had greater prognostic ability for overall survival than molecular or clinical factors (COO, MYC/BCL2 expression, IPI) alone.19 Additionally, the studies by Schmitz et al.8 and Chapuy et al.9 helped elucidate some of the reported clinical and genetic heterogeneity in transcriptionally defined COO subsets of front-line DLBCL.8,9 Using a set of common genetic alterations, both studies identified distinct molecular subtypes and evaluated their clinical prognostic outcome. Both studies identified a number of common mutational profiles, including two distinct subsets of ABC (one enriched for mutations in MYD88 and CD79B, and another for BCL6 and NOTCH mutations) and a GCB subset enriched for BCL2 translocations and mutations in CREBBP and EZH2. Importantly, these clusters had distinct prognostic profiles, many reflecting the established prognostic impact of the dominant mutations in each group (e.g. worse prognosis for the BCL2 and MYD88 subsets).9 Here, we perform an integrated analysis to evaluate if somatic mutations in DLBCL provide clinical prognostic value over established clinical and biological risk factors, including COO and IPI. Using data from the phase III GOYA study, the largest (n=1,418) randomized clinical trial in patients with previously untreated DLBCL to date, we analyzed the mutational profile of DLBCL using a wellestablished, highly validated targeted next-generation sequencing (NGS) platform, and evaluated the prognostic impact of somatic mutations and their relationship with COO. A previous exploratory analysis in the GOYA study showed that patients with GCB DLBCL achieved a better outcome in terms of progression-free survival (PFS) than those with the ABC subtype, irrespective of treatment.3

Methods Patient treatment and assessments The GOYA study design has been described previously.3 Patients included in the study had previously untreated, histologihaematologica | 2020; 105(9)

cally documented, CD20+ DLBCL; details of the inclusion criteria are available in the Online Supplementary Methods. The study was conducted in accordance with the European Clinical Trial Directive (for European centers), the Declaration of Helsinki, and the International Conference on Harmonisation Guidelines for Good Clinical Practice. The protocol was approved by the ethics committees of participating centers and registered at clinicaltrials.gov identifier: NCT01287741. All patients provided written informed consent. Staging investigations included computed tomography (CT) scanning and bone marrow biopsy. Tumor response and progression were assessed by the investigator using regular clinical and laboratory examinations and CT scans. Response was evaluated according to the Revised Response Criteria for Malignant Lymphoma20 4-8 weeks after last study treatment, or at early discontinuation.

Cell-of-origin analysis Cell-of-origin classification was based on gene expression profiling using the NanoString Lymphoma Subtyping Research-UseOnly assay (NanoString Technologies Inc., Seattle, WA, USA). COO data were available in 933 patients. Reasons for non-availability were: restricted Chinese export license (n=252), CD20+ DLBCL not confirmed by central pathology (n=102) and missing/inadequate tissue (n=131).

Immunohistochemical analyses Pre-treatment tumor samples were analyzed by a central laboratory using the Ventana BCL2 (124) and MYC (Y69) investigational use only immunohistochemical assays. The pre-specified scoring algorithm incorporated percentage of tumor cells stained and their intensity: BCL2 immunohistochemistry-positive was defined as moderate/strong cytoplasmic staining in ≥50% of tumor cells and MYC immunohistochemistry-positive was defined as nuclear staining at any intensity in ≥40% of tumor cells.

Targeted next-generation sequencing Genomic DNA was extracted from diagnostic formalin-fixed, paraffin-embedded tissue sections containing ≥20% tumor cells. Samples were submitted to a central laboratory for NGS-based genomic profiling and processed as previously described.21,22 Adaptor-ligated DNA underwent hybrid capture for all coding exons of 465 cancer-related genes [FoundationOne HemeTM platform, Foundation Medicine Incorporated (FMI), MA, USA] (Online Supplementary Methods). NGS data were available for 499 of the 1,418 patients included in the intent-to-treat (ITT) population of the GOYA study; both NGS and COO were available in 482 patients. Information about known drug targets and ongoing clinical trials targeting individual mutations was queried on March 23, 2018, through an FMI internal database populated using data from clinicaltrials.gov and other publicly available sources.

Validation of mutational models We sought to confirm the prognostic value of the mutational genomic risk model generated by Reddy et al.,19 Chapuy et al.,9 and Schmitz et al.,8 as described in the Online Supplementary Methods.

Statistical analysis Only genetic alterations with known somatic and functional status were included in the statistical analysis.21 Univariate and multivariate Cox regression analyses were used to evaluate the prognostic effect of a genetic alteration if there were ≥10 progression events in mutated patients or ≥40 patients in total with the mutation. Multivariate Cox regression analysis was performed to control for COO, IPI, treatment arm, number of planned 2299

C.R. Bolen et al. A

Figure 1. Frequently observed gene alterations in patients with diffuse large B-cell lymphoma (DLBCL) in the GOYA trial (clinicaltrials.gov identifier: NCT01287741). (A) Most frequently (≥2% of cases) observed gene alterations: single nucleotide variant (SNV), amplifications and deletions. (B) Genes with significant differences in mutation rates* between the activated B-cell-like (ABC) and germinal center B-cell-like (GCB) DLBCL subtypes. (C) Frequency of BCL2 and CDKN2A alterations in the ABC and GCB DLBCL subtypes. *False discovery rate (FDR) <0.05. CNA: copy number abnormality; trans: translocation.

chemotherapy cycles, and geographic region. Multiple testing adjustment was performed by estimating false discovery rates (FDR) using the Benjamini-Hochberg procedure (significance <5% FDR).

Results Baseline disease characteristics were similar between patients with NGS available and the overall GOYA ITT population, except for race (Online Supplementary Table S1) and geographic region (data not shown) due to lack of access to samples from China.

Genomic alterations detectable by targeted next-generation sequencing Of 465 sequenced genes, 59 (13%) were identified as functionally altered (i.e. having mutations that significant2300

ly alter the function of a gene in a manner that has been previously reported to drive cancer progression) in at least 10 of 499 patient samples (≥2% prevalence), and 334 additional genes with alterations were identified in ≥1 patient; 3% of patients had no identified mutation. The median number of gene alterations per patient was 6 (range 0-17). The median number of single nucleotide variants (SNV) and copy number abnormalities (CNA) per patient were 4 (range, 0-16) and 0 (range, 0-10), respectively. Ninetyseven percent of cases harbored ≥1 alteration and 93% of cases harbored multiple (≥2) alterations. The most frequently (≥2% cases) observed gene alterations (SNV, amplifications and deletions) are shown in Figure 1A. SNV were the most common mutation type, while CNA were specific to a few genes, including CDKN2A/B and REL. Of the 31 analyzed gene rearrangements, BCL2, MYC and BCL6 were the most frequently rearranged; for these genes, the most frequently observed haematologica | 2020; 105(9)

Somatic mutations in DLBCL

translocation partner was the immunoglobulin heavy chain locus, found in 92 of 92 (100%), 29 of 32 (90.6%), and 57 of 100 (57.0%) cases where the rearrangement partner could be determined, respectively (Online Supplementary Table S2).

Frequencies of genomic alterations among cell-of-origin subsets Of the patients for whom both COO and NGS were available (n=482), 272 (56%), 78 (16%), and 132 (27%) were classified as GCB, unclassified, and ABC DLBCL, respectively (Online Supplementary Table S1). This was similar to findings for the overall COO population [n=933; GCB, n=540 (58%); unclassified, n=150 (16%); ABC, n=243 (26%)]. Within the GCB subtype, the most prevalent mutated genes were BCL2 [88 of 272 (32%)], MLL2 (KMT2D) [82 of 272 (30%)] and CREBBP [60/272 (22%)]; loss of CDKN2A [64 of 132 (49%)] and CDKN2B [40 of 132 (30%)] and mutations of MYD88 [45 of 132 (34%)] were most frequently observed in the ABC subtype (Table 1 and Online Supplementary Table S3). Fifteen genes were found to be significantly differentially mutated between the GCB and ABC subtypes at FDR <0.05 (Figure 1B). Alterations of BCL2, CREBBP, TNFRSF14, EZH2, REL, BCL7A and SGK1 were more frequently observed in GCB DLBCL whereas BCOR, ETV6, PRDM1, PIM1, CD79b, CDKN2B, MYD88 and CDKN2A were more frequently mutated in ABC DLBCL (Figure 1B). In the case of BCL2

and CDKN2A, specific types of alterations displayed different frequencies between the GCB and ABC subtypes (Figure 1C). While BCL2 translocations and SNV were more frequently found in the GCB subtype, high-level BCL2 amplifications (≥6 copies) were enriched within the ABC subtype [ABC, 9 of 132 (6.8%); GCB, 4 of 272 (1.5%); Fisher’s exact test P=0.012]. An analysis of lowlevel BCL2 amplifications (≥1 copy above median ploidy and ≥3 copies) confirmed the enrichment in ABC DLBCL samples [ABC, 83 of 132 (62.9%); GCB, 45 of 272 (16.5%); Fisher’s exact test P<0.001]. The enrichment of CDKN2A alterations within the ABC subtype was pronounced only

B Table 1. Prevalence of most frequent* gene mutations according to diffuse large B-cell lymphoma cell-of-origin (COO) subtype.

BCL2 KMT2D CREBBP TP53 BCL6 B2M TNFRSF14 EZH2 TNFAIP3 REL BCL7A CDKN2A MYD88 CD58 TMEM30A CD70 PIM1 CDKN2B NOTCH2 CD79B PRDM1 ETV6

GCB, n=272 (%)

Unclassified, n=78 (%)

ABC, n=132 (%)

32.4 30.1 22.1 19.5 18.8 17.6 17.3 16.2 15.4 13.2 10.7 10.3 8.8 8.5 8.1 7.7 7.0 5.1 4.0 2.2 1.5 0.7

5.1 21.8 7.7 17.9 35.9 12.8 1.3 6.4 11.5 5.1 2.6 21.8 15.4 10.3 11.5 17.9 5.1 11.5 10.3 9.0 3.8 5.1

4.5 28.8 3.8 15.2 22.0 12.9 0.0 0.8 9.1 0.8 2.3 48.5 34.1 6.8 8.3 6.1 24.2 30.3 6.8 25.0 19.7 10.6

Listed in order of frequency in the germinal center B-cell-like (GCB) subgroup. *Gene mutations occurring in ≥10% of patients in any COO subgroup. n=number; ABC: activated B-cell-like.

haematologica | 2020; 105(9)

Figure 2. Association between BCL2 gene alterations and progression-free survival (PFS) in diffuse large B-cell lymphoma (DLBCL). (A) All BCL2 alterations. (B) BCL2 single nucleotide variant. (C) BCL2 translocations. CI: confidence interval; FDR: false discovery rate; HR: hazard ratio; MUT: mutant; WT: wild-type.

2301

C.R. Bolen et al.

Figure 3. BCL2 alterations according to cell-of-origin (COO) subtype. *Germinal center B-cell-like (GCB), 31%; unclassified, 5.1%; activated B-cell-like (ABC), 0.8%. â&#x20AC; GCB, 11%; unclassified, 0%; ABC, 4.5%. â&#x20AC;ĄGCB, 1.5%; unclassified, 2.6%; ABC, 6.8%. amp: amplification; NA: not available; SNV: single nucleotide variant; trans: translocation.

for CDKN2A deletions; SNV occurred to the same degree in all COO subtypes.

Correlation of individual alterations with clinical outcomes Alterations of 23 genes (fulfilling the predefined criteria based on their prevalence in the analyzed cohort) were evaluated for association with PFS on univariate and multivariate analyses. Prognostic trends were observed among a number of previously studied biomarkers, including BCL2, CREBBP, REL, TP53 and CDKN2A (all P<0.05, unadjusted). However, alterations of BCL2 (including translocations, SNV and high-level amplifications) were the most strongly associated with PFS [hazard ratio (HR): 2.6; 95% confidence interval (CI): 1.6-4.2; FDR, 0.0037] independent of COO, IPI, treatment arm, number of planned chemotherapy cycles, and geographic region (Table 2). None of the 23 biomarkers showed significant differences in prognostic impact between treatment arms. The BCL2 prognostic effect was observed for both BCL2 SNV (HR: 2.6; 95%CI: 1.5-4.7; FDR, 0.022) and translocations (HR: 2.5; 95%CI: 1.4-4.2; FDR, 0.0028) (Table 2 and Figure 2) individually. The prognostic role of high-level BCL2 amplification was not tested separately due to the low prevalence of this alteration in the current study. No association was found between survival and low-level BCL2 amplifications (HR: 1.2; 95%CI: 0.8-1.9; FDR, 0.58). BCL2 alterations were detected in 20% (102 of 499) of patients, with 92 of 102 patients having a BCL2 translocation, 90% (83 of 92) of whom were GCB patients, with only one translocated ABC patient. Of 39 patients with BCL2 SNV, 80% (31 of 39) and 15% (6 of 39) were in the GCB and ABC subgroups, respectively. The majority of patients with BCL2 SNV harbored BCL2 translocations (74%, 29 of 39) (Figure 3), but BCL2 SNV were still associated with worse prognosis among patients without a BCL2 translocation (HR: 2.8; 95%CI: 1.0-7.9; P=0.047). BCL2 mutations were also significantly correlated with 2302

BCL2 gene and protein expression levels (Online Supplementary Figure S1). Alterations of CREBBP (HR: 2.1; 95%CI: 1.3-3.4; FDR, 0.054) and TP53 (HR: 1.6; 95%CI: 1.1-2.5; FDR, 0.22) were also associated with PFS on multivariate analysis, but did not fulfill the predefined criteria for significance (FDR <0.05). Alterations of CREBBP were detected in 15% (73 of 499) of patients; 82% (60 of 73), 8% (6 of 73), and 7% (5 of 73) of whom belonged to the GCB, unclassified and ABC subtypes, respectively. Four of the 73 patients harbored two different CREBBP mutations. In the majority of cases, CREBBP alterations were SNV (97%, 71 of 73), with only two cases of CREBBP deletion. Alterations of TP53 were found in 18% (92 of 499) of patients, of whom 58% (53 of 92), 15% (14 of 92), and 22% (20 of 92) had the GCB, unclassified, and ABC DLBCL subtype, respectively. Overall, 105 TP53 alterations were observed in 92 patients, with 13 of 92 patients harboring two simultaneous TP53 mutations. SNV were the most frequently observed TP53 alterations (98%, 103 of 105), while TP53 deletions and rearrangements were observed in two cases, and one case, respectively. CDKN2A alterations were associated with shorter PFS on univariate analysis (HR: 1.7; 95%CI: 1.2-2.5; FDR, 0.13). This effect was driven by CDKN2A deletions (HR: 1.6; 95%CI: 1.1-2.4; FDR, 0.058). No significant association with PFS was observed on multivariate analysis for all CDKN2A alterations or for CDKN2A deletions only (Table 2). CDKN2A alterations were observed in 23% (113 of 499) of DLBCL patients. Of all cases with any CDKN2A alteration, 25% (28 of 113), 15% (17 of 113), and 57% (64 of 113) belonged to the GCB, unclassified, and ABC subtypes, respectively. The majority of the CDKN2A alterations were homozygous gene deletions, which were enriched within the ABC subtype. Patients with CDKN2A deletions had adverse clinical disease characteristics (IPI, extranodal sites, age, and serum lactate dehydrogenase) compared with patients without a CDKN2A deletion, haematologica | 2020; 105(9)

Somatic mutations in DLBCL

Table 2. Results from prognostic evaluation of prioritized candidate genes.

Gene BCL2 BCL2 translocation BCL2 SNV CREBBP REL CD274 TP53 TP53 SNV TNFRSF14 KMT2D CD58 MYC MYC translocation ARID1A CDKN2A CDKN2A deletion CDKN2B BCL7A TNFAIP3 MYD88 B2M EZH2 BCL6 PIM1 CD79B CD70 CARD11 TMEM30A

Univariate HR (95%CI)*

FDR

Multivariate HR (95%CI)†

FDR

1.7 (1.1-2.5) 1.6 (1.0-2.4) 2.2 (1.3-3.8) 1.4 (0.9-2.2) 1.3 (0.8-2.3) 1.6 (0.9-3.2) 1.6 (1.0-2.4) 1.5 (1.0-2.3) 1.2 (0.7-2.1) 1.2 (0.8-1.7) 1.2 (0.7-2.1) 1.6 (0.9-2.8) 1.8 (0.9-3.2) 1.2 (0.6-2.2) 1.7 (1.2-2.5) 1.6 (1.1-2.4) 1.5 (1.0-2.4) 1.1 (0.6-2.1) 0.9 (0.5-1.5) 1.2 (0.8-1.9) 0.8 (0.5-1.4) 0.5 (0.3-1.2) 1.0 (0.7-1.6) 0.8 (0.5-1.4) 0.9 (0.5-1.6) 1.0 (0.5-1.9) 0.5 (0.2-1.1) 0.6 (0.3-1.3)

0.012 0.036 0.0025 0.14 0.32 0.13 0.034 0.044 0.49 0.46 0.59 0.15 0.064 0.66 0.0056 0.014 0.077 0.81 0.63 0.44 0.52 0.12 0.86 0.48 0.77 0.93 0.076 0.19

0.14 0.096 0.041 0.37 0.67 0.37 0.26 0.35 0.74 0.74 0.79 0.37 0.096 0.79 0.13 0.058 0.35 0.88 0.79 0.74 0.74 0.37 0.9 0.74 0.88 0.93 0.35 0.43

2.6 (1.6-4.2) 2.5 (1.4-4.2) 2.6 (1.5-4.7) 2.1 (1.3-3.4) 1.9 (1.0-3.4) 1.7 (0.9-3.3) 1.6 (1.1-2.5) 1.6 (1.0-2.5) 1.4 (0.8-2.7) 1.3 (0.9-1.9) 1.3 (0.7-2.4) 1.2 (0.6-2.2) 1.4 (0.7-2.5) 1.2 (0.6-2.4) 1.2 (0.8-1.9) 1.1 (0.7-1.7) 1.1 (0.7-1.7) 1.1 (0.6-2.3) 1.0 (0.6-1.8) 0.9 (0.5-1.4) 0.9 (0.5-1.5) 0.8 (0.4-1.7) 0.8 (0.5-1.2) 0.7 (0.4-1.2) 0.7 (0.4-1.3) 0.7 (0.4-1.4) 0.6 (0.3-1.4) 0.6 (0.3-1.4)

0.00016 0.00095 0.0014 0.0047 0.043 0.13 0.029 0.034 0.26 0.23 0.38 0.60 0.30 0.55 0.46 0.85 0.82 0.68 0.85 0.52 0.63 0.50 0.27 0.21 0.28 0.38 0.22 0.25

0.0037 0.0028 0.022 0.054 0.25 0.54 0.22 0.18 0.54 0.54 0.62 0.72 0.30 0.70 0.70 0.99 0.85 0.75 0.85 0.70 0.72 0.70 0.54 0.54 0.54 0.62 0.54 0.54

Listed in order of multivariate hazard ratio (HR). Significant alterations on multivariate analysis [false discovery rate (FDR) <0.05] shown in bold. *Adjusted for treatment only. † Adjusted for treatment arm, International Prognostic Index, cell-of-origin, number of planned chemotherapy cycles, and geographic region. CI: confidence interval.

both in the total FMI evaluable patients and among the ABC subtype (Online Supplementary Table S4). In a survival analysis according to COO subtype, BCL2 translocations (HR: 2.3; 95%CI: 1.3-4.2; P=0.0049; FDR, 0.017) were significantly associated with shorter PFS independent of clinical factors in the GCB subtype, while none of the identified genetic alterations were significantly prognostic within the ABC subtype (Online Supplementary Table S5).

Correlation of combined genomic risk model with clinical outcomes We evaluated the performance of a combined genomic risk model for predicting clinical outcomes using a single comprehensive NGS assay. When applying a modified mutational model generated by Reddy et al.,19 the risk scores ranged from -3 to 7, with most patients centered at 0 (Figure 4A). Low-risk was defined by a score <0 (n=112), low-intermediate-risk with a score 0 (n=215), high-intermediate-risk patients had a score >0 and <3 (n=107), and high-risk had a score ≥3 (n=29). This genomic scoring system provided clear separation between the low/low-intermediate and high/high-intermediate groups (Figure 4B). haematologica | 2020; 105(9)

Using a simple dichotomization of the score into low- and high-risk subgroups, the overall univariate HR for the prognostic score was 0.61 (95%CI: 0.42-0.88; P=0.0087). The risk groups were highly correlated with COO subtypes, and after correcting for COO, the model was no longer significant in the entire cohort (HR: 0.77; 95%CI: 0.49-1.2; P=0.27). When tested within COO subtypes, no significant prognostic signal was found, although there was a trend for added prognostic information among the GCB subset (HR: 0.5; 95%CI: 0.24-1.04; P=0.06) but not the ABC subset (HR: 1.2; 95%CI: 0.66-2.32; P=0.5).

Validation of new molecular classifications Although there is no publicly available tool for classifying samples into molecular subtypes as defined by Schmitz et al.8 and Chapuy et al.,9 we sought to validate these classifications using an approximation of their clusters. For Schmitz et al.,8 we approximated the EZB, BN2, N1 and MCD clusters using each cluster’s founder alterations (EZH2 or BCL2; BCL6 or NOTCH2; NOTCH1; and MYD88, L265P or CD79B, respectively; see Methods). Prevalence of these four clusters was consistent with those reported by Schmitz et al.8 (Figure 5A); however, we 2303

C.R. Bolen et al.

observed no difference in prognosis among any of the four mutational subgroups (log-rank P=0.94), although the mutational subsets did perform worse than the unclassified â&#x20AC;&#x153;other GCBâ&#x20AC;? subset (pooled mutational clusters vs. other GCB P=0.021; EZB vs. other GCB P=0.023 ) (Figure 5B). To recreate the Chapuy classifications, we applied the non-negative matrix factorization (NMF) clustering algorithm to the set of mutations overlapping with those reported by Chapuy et al.9 This resulted in five clusters (plus an unmutated cluster: C0) sharing very similar mutational profiles and distribution of COO subsets with the clusters of Chapuy et al.9 (Figure 5C and Online Supplementary Figure S2), with the notable exception that CDKN2A/2B (9p21) deletions significantly co-occurred with MYD88 and CD79B alterations, rather than with TP53 alterations as observed in Chapuy et al.9 We observed similar prognostic trends among these subsets, with our clusters G2, G3 and G5 (equivalent to Chapuy C2, C3 and C5) showing significantly worse prognosis when compared with clusters G0, G1 and G4 (Chapuy C0, C1 and C4, respectively) (HR: 1.8; 95%CI: 1.2-2.6; P=0.0033) (Figure 5D) .

Discussion In this study, we analyzed the mutational profile and prognostic impact of genomic alterations in newly diagnosed DLBCL patients who were uniformly treated with anti-CD20-based immunochemotherapy [obinutuzumab or rituximab plus cyclophosphamide, doxorubicin, vincristine and prednisone (G-/R-CHOP)] in the phase III GOYA trial. Using a well-established, highly validated targeted NGS platform, we analyzed SNV and CNA in 465 cancer-related genes and 31 select gene rearrangements in 499 patients. This is the largest prospectively collected dataset in DLBCL so far. These data serve as a valuable resource for understanding the clinical relevance of mutations as measured by this platform. Alteration of the BCL2 gene was the only genetic abnormality significantly associated with shorter PFS independent of molecular or clinical factors (treatment arm, COO, IPI, number of planned chemotherapy cycles, and geographic region). This effect was observed for both BCL2 translocations and SNV. The co-occurrence of BCL2 SNV with BCL2 translocations, possibly as a consequence of aberrant somatic hypermutation,23 may partially explain the negative prognostic

Figure 4. (A) Distribution of risk scores using the applied Reddy et al.19 prognostic model, and (B) progression-free survival (PFS) by risk group (n=443). int: intermediate.

2304

haematologica | 2020; 105(9)

Somatic mutations in DLBCL

impact of BCL2 SNV, although the negative prognostic effect of BCL2 SNV among patients without BCL2 translocations may point to an independent biological role for these alterations. BCL2 translocations were significantly enriched within the GCB subtype and were associated with shorter PFS within this subtype. BCL2 translocations were associated with high levels of BCL2 mRNA and protein expression, both of which have been shown to be associated with an adverse prognosis in DLBCL, independent of COO and IPI, including in the GOYA study.24 Our data suggest that pharmacological inhibition of the BCL2 protein could be a promising treatment strategy in a subset of DLBCL patients. Venetoclax, a highly specific BCL2 inhibitor,25 is currently being tested in clinical trials in patients with newly diagnosed DLBCL; however, the subpopulation of DLBCL patients who could benefit from venetoclax needs to be defined. Given the molecular uniqueness and prognostic value of the particular COO subtypes, we aimed to analyze the prognostic impact of genetic alterations within these subtypes. The only genetic alteration significantly associated

with shorter PFS within the GCB subtype was BCL2 translocation. None of the tested genetic alterations were significantly associated with outcome within the prognostically-inferior ABC subtype, supporting the strong prognostic significance of COO assessed by gene expression profiling. In this study, we observed prognostic trends in several genes, including TP53, CREBBP and CDKN2A, but none met our thresholds for significance. There are several potential explanations for this observation. First, in the current study we used robust statistical methods with strict pre-defined criteria for significance to test the association of particular gene alterations with clinical outcomes. Second, only truncating/frameshift mutations and previously reported loss-of-function mutations were included in this study. Alteration of several genes, such as CREBBP and TP53, were associated with shorter PFS in our study, in the absence of multiple testing correction. When validating the genomic risk model from Reddy et al.,19 although the model was prognostic in our population when stratified into high- and low-risk groups (HR: 0.61;

4-way P=0.94 EZB-like vs. other GCB

P=0.023

G2/3/5 vs. G0/1/4:

P=0.0033

Figure 5. Diffuse large B-cell lymphoma (DLBCL) mutational subset validation. (A) Prevalence and (B) association of Schmitz et al.8 classifications with progressionfree survival (PFS). Schmitz clusters were approximated using the seed mutations: EZB - EZH2 or BCL2; BN2 - BCL6 or NOTCH2; N1 - NOTCH1; MCD - MYD88, L265P or CD79B; Multi: multiple seed mutations from more than one cluster. (C) Chapuy et al.9 clusters were approximated by application of non-negative matrix factorization (NMF) to the GOYA Foundation Medicine Incorporated (FMI) dataset and selecting five clusters (G1-G5). Mutations with significant enrichment in one or more clusters are shown. (D) Association between NMF clusters and PFS. ABC: activated B-cell-like; alt: alteration; CNA: copy number abnormality; COO: cell-of-origin; GCB: germinal center B-cell-like; HR: hazard ratio; SNV: single nucleotide variant.

haematologica | 2020; 105(9)

2305

C.R. Bolen et al.

95%CI: 0.42-0.88; P<0.01), when corrected for COO, the model was no longer significant (HR: 0.77; 95%CI: 0.491.2; P=0.27), indicating that it provided little additional benefit over the most commonly used gene expression profiling and fluorescence in situ hybridization assays, and that COO evaluation in combination with BCL2 and MYC translocation status may be a simpler approach with similar overall prognostic relevance, although other genomic features such as TP53 or CREBBP may provide additional information that is worth considering. However, it should be noted that we were unable to apply the Reddy et al.19 model in its entirety due to some differences in gene availability on the FMI platform, and for the fact that Reddy et al.19 evaluated the model in terms of overall survival, whereas our study evaluated it in terms of PFS. The current study also demonstrated the molecular heterogeneity of DLBCL, with the majority of the observed genetic alterations shared by COO subtypes; however, the frequency of mutations in 15 genes was enriched between GCB and ABC subtypes. In addition, approximating the molecular clusters described by Schmitz et al.8 and Chapuy et al.9 revealed a consistent set of molecular subgroups, with some specific to either GCB (EZB-like, G3), ABC (MCD- or N1-like, G5), or Unclassified (BN2like) COO subtypes, and others appearing to be independent of the tumor COO. Among the clusters defined by NMF, we observed a significantly worse prognosis for clusters G2, G3 and G5, consistent with Chapuyâ&#x20AC;&#x2122;s C2, C3 and C5 clusters.9 This is most likely driven by the enrichment of individual prognostic alterations among these subgroups (BCL2 and CREBBP in G3; TP53 and REL in G2), or by enrichment for the ABC subset (G5). By contrast, our approximation of the Schmitz clusters identified four sets of clusters with approximately equivalent prognosis, suggesting that the founder alterations used to define these clusters are not sufficient to identify patients with worse prognosis. Although we cannot directly recapitulate the clusters defined by Schmitz et al.8 and Chapuy et al.,9 both due to limitations of the FMI panel and because algorithms for classifying DLBCL samples are not publicly available, our results here show that we can successfully capture the molecular heterogeneity of DLBCL using this targeted mutational panel. Since 2011, several studies have characterized the landscape of somatic mutations in DLBCL by whole exome NGS technologies5-7,26 or the FMI targeted exome-sequencing platform,4 and have identified recurrent genetic alterations. Our study identified a relatively lower number of genetic alterations compared with whole-exome studies, but it was relatively consistent with the frequencies of mutations identified by Intlekofer et al.4 This is most likely because both our study and the study by Intlekofer et al. focused on mutations with known or likely somatic and functional status. FMI may also lack some alterations of potential relevance in DLBCL, including alterations in the human leukocyte antigen genes, potentially limiting the scope of this analy-

References 1. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profil-

2306

sis. In contrast, the relatively low prevalence of MYC translocations in this dataset may be reflective of an accrual bias during patient recruitment. Patients with these alterations, particularly in combination with BCL2 translocations (double-hit lymphoma) have been well characterized as having particularly aggressive disease and are generally more difficult to recruit for clinical trials. These patients may also benefit from more aggressive chemotherapy than G-/R-CHOP, which could also explain why these patients were not enrolled in GOYA. Our data show that DLBCL contains mutations in a variety of potentially targetable pathways. In total, a majority (59%) of patients harbor â&#x2030;Ľ1 alteration in genes that would be eligible for potential targeted therapies approved in other indications (e.g. venetoclax for BCL2 translocations/amplifications, everolimus for PTEN loss, and ruxolitinib and tofacitinib for JAK2 mutations) and over 70% of patients would potentially qualify to be enrolled in ongoing clinical trials based on genomic information, according to the FMI clinical trial database. Genes enriched between GCB and ABC subtypes also included previously reported driver mutations and gene alterations that can be targeted by novel therapies, such as the gain of function mutation of EZH2 in the GCB DLBCL subtype,27 and the BCL2 translocations and amplifications.28 These mutations, along with COO subtype information, would be useful for the design of clinical trials involving combinations of novel targeted therapies. In conclusion, using the largest prospective dataset in previously untreated DLBCL to date, we demonstrated the molecular heterogeneity of DLBCL, with potential treatment targets harbored by the distinct COO subtypes. Only alterations in BCL2 were significantly associated with clinical outcome independent of COO and clinical factors, thereby demonstrating the strong prognostic value of COO for clinical outcome in DLBCL. Data sharing Qualified researchers may request access to individual patient level data through the clinical study data request platform. Further details on Roche's criteria for eligible studies are available here (https://vivli.org/members/ourmembers/). For further details on Roche's Global Policy on the Sharing of Clinical Information and how to request access to related clinical study documents, see here(https://www.roche.com/research_and_development/ who_we_are_how_we_work/clinical_trials/our_commitment_to_ data_sharing.htm). Acknowledgments The authors would like to thank the GOYA study team investigators, coordinators, nurses and patients. Funding GOYA was supported by F. Hoffmann-La Roche Ltd, with scientific support from the Fondazione Italiana Linfomi. Editorial support was provided by Louise Profit, PhD (Gardiner-Caldwell Communications Ltd, Macclesfield, UK), and was funded by F. Hoffmann-La Roche Ltd.

ing. Nature. 2000;403(6769):503-511. 2. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med. 2002;

346(25):1937-1947. 3. Vitolo U, Trneny M, Belada D, et al. Obinutuzumab or rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone in previously

haematologica | 2020; 105(9)

Somatic mutations in DLBCL

10.

11.

12.

untreated diffuse large B-cell lymphoma. J Clin Oncol. 2017:35(31):3529-3537. Intlekofer AM, Joffe E, Batlevi CL, et al. Integrated DNA/RNA targeted genomic profiling of diffuse large B-cell lymphoma using a clinical assay. Blood Cancer J. 2018; 8(6):60. Morin RD, Mendez-Lago M, Mungall AJ, et al. Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature. 2011;476(7360):298-303. Pasqualucci L, Trifonov V, Fabbri G, et al. Analysis of the coding genome of diffuse large B-cell lymphoma. Nat Genet. 2011; 43(9):830-837. Zhang J, Grubor V, Love CL, et al. Genetic heterogeneity of diffuse large B-cell lymphoma. Proc Natl Acad Sci U S A. 2013; 110(4):1398-1403. Schmitz R, Wright GW, Huang DW, et al. Genetics and pathogenesis of diffuse largeB-cell lymphoma. N Engl J Med. 2018; 378(15):1396-1407. Chapuy B, Stewart C, Dunford AJ, et al. Molecular subtypes of diffuse large-B-cell lymphoma are associated with distinct pathogenic mechanisms and outcomes. Nat Med. 2018;24(5):679-690. Lunning MA, Green MR. Mutation of chromatin modifiers; an emerging hallmark of germinal center B-cell lymphomas. Blood Cancer J. 2015;5:e361. Jardin F, Jais JP, Molina TJ, et al. Diffuse large B-cell lymphomas with CDKN2A deletion have a distinct gene expression signature and a poor prognosis under RCHOP treatment: a GELA study. Blood. 2010;116(7):1092-1104. Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor kappaB

haematologica | 2020; 105(9)

13.

14.

15. 16.

17.

18.

19. 20.

activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med. 2001;194(12):1861-1874. Davis RE, Ngo VN, Lenz G, et al. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature. 2010; 463(7277):88-92. Young RM, Shaffer AL, Phelan JD, Staudt LM. B-cell receptor signaling in diffuse large B-cell lymphoma. Semin Hematol. 2015;52(2):77-85. Lenz G, Wright G, Dave SS, et al. Stromal gene signatures in large-B-cell lymphomas. N Eng J Med. 2008;359(22):2313-2323. Scott DW, Wright GW, Williams PM, et al. Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffinembedded tissue. Blood. 2014;123(8):12141217. Xu-Monette ZY, Wu L, Visco C, et al. Mutational profile and prognostic significance of TP53 in diffuse large B-cell lymphoma patients treated with R-CHOP: report from an International DLBCL Rituximab-CHOP Consortium Program Study. Blood. 2012;120(19):3986-3996. Fernandez-Rodriguez C, Bellosillo B, Garcia-Garcia M, et al. MYD88 (L265P) mutation is an independent prognostic factor for outcome in patients with diffuse large B-cell lymphoma. Leukemia. 2014; 28(10):2104-2106. Reddy A, Zhang J, Davis NS, et al. Genetic and functional drivers of diffuse large B cell lymphoma. Cell. 2017;171(2):481-494.e15. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579586.

21. Frampton GM, Fichtenholtz A, Otto GA, et al. Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing. Nat Biotechnol. 2013;31(11):1023-1031. 22. He J, Abdel-Wahab O, Nahas MK, et al. Integrated genomic DNA/RNA profiling of hematologic malignancies in the clinical setting. Blood. 2016;127(24):3004-3014. 23. Schuetz JM, Johnson NA, Morin RD, et al. BCL2 mutations in diffuse large B-cell lymphoma. Leukemia. 2012;26(6):13831390. 24. Sehn LH, Oestergaard MZ, TrnÄ&#x203A;nĂ˝ M, et al. Prognostic impact of BCL2 and MYC expression and translocation in untreated DLBCL: results from the phase III GOYA study. Hematol Oncol. 2017;35(S2):131133. 25. Souers AJ, Leverson JD, Boghaert ER, et al. ABT-199, a potent and selective BCL-2 inhibitor, achieves antitumor activity while sparing platelets. Nat Med. 2013;19(2):202208. 26. Lohr JG, Stojanov P, Lawrence MS, et al. Discovery and prioritization of somatic mutations in diffuse large B-cell lymphoma (DLBCL) by whole-exome sequencing. Proc Natl Acad Sci U S A. 2012; 109(10):3879-3884. 27. McCabe MT, Ott HM, Ganji G, et al. EZH2 inhibition as a therapeutic strategy for lymphoma with EZH2-activating mutations. Nature. 2012;492(7427):108-112. 28. Roberts AW, Huang D. Targeting BCL2 with BH3 mimetics: basic science and clinical application of venetoclax in chronic lymphocytic leukemia and related B cell malignancies. Clin Pharmacol Ther. 2017; 101(1):89-98.

2307

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2308-2315

Non-Hodgkin Lymphoma

DA-EPOCH-R combined with high-dose methotrexate in patients with newly diagnosed stage II-IV CD5-positive diffuse large B-cell lymphoma: a single-arm, open-label, phase II study Kana Miyazaki,1 Naoko Asano,2 Tomomi Yamada,3 Kohta Miyawaki,4 Rika Sakai,5 Tadahiko Igarashi,6 Momoko Nishikori,7 Kinya Ohata,8 Kazutaka Sunami,9 Isao Yoshida,10 Go Yamamoto,11 Naoki Takahashi,12 Masataka Okamoto,13 Hiroki Yano,14 Yuki Nishimura,15 Satoshi Tamaru,15 Masakatsu Nishikawa,15 Koji Izutsu,11,16 Tomohiro Kinoshita,17 Junji Suzumiya,18 Koichi Ohshima,19 Koji Kato,4 Naoyuki Katayama1 and Motoko Yamaguchi1

Department of Hematology and Oncology, Mie University Graduate School of Medicine, Mie; 2Department of Molecular Diagnostics, Nagano Prefectural Shinshu Medical Center, Suzaka; 3Department of Medical Innovation, Osaka University Hospital, Suita; 4Department of Medicine and Biosystemic Science, Kyushu University Graduate School of Medical Sciences, Fukuoka; 5Department of Medical Oncology, Kanagawa Cancer Center, Yokohama; 6Hematology/Oncology Division, Gunma Cancer Center, Ohta; 7 Department of Hematology and Oncology, Kyoto University Graduate School of Medicine, Kyoto; 8Department of Hematology, Kanazawa University, Kanazawa; 9Department of Hematology, National Hospital Organization Okayama Medical Center, Okayama; 10 Department of Hematologic Oncology, National Hospital Organization Shikoku Cancer Center, Matsuyama; 11Department of Hematology, Toranomon Hospital, Tokyo; 12 Department of Hematology and Oncology, International Medical Center, Saitama Medical University, Hidaka; 13Department of Hematology, Fujita Health University School of Medicine, Toyoake; 14Department of Hematology, Kainan Hospital Aichi Prefectural Welfare Federation of Agricultural Cooperatives, Yatomi; 15Clinical Research Support Center, Mie University Hospital, Tsu; 16Department of Hematology, National Cancer Center Hospital, Tokyo; 17Department of Hematology and Cell Therapy, Aichi Cancer Center Hospital, Nagoya; 18Department of Oncology and Hematology, Shimane University Hospital, Izumo and 19Department of Pathology, Kurume University School of Medicine, Kurume, Japan 1

ABSTRACT

Correspondence: KANA MIYAZAKI kmiyazaki@clin.medic.mie-u.ac.jp Received: July 11, 2019. Accepted: October 21, 2019. Pre-published: October 24, 2019. doi:10.3324/haematol.2019.231076 ÂŠ2020 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.

2308

D5-positive diffuse large B-cell lymphoma (CD5+ DLBCL) is characterized by poor prognosis and a high frequency of central nervous system relapse after standard immunochemotherapy. We conducted a phase II study to investigate the efficacy and safety of dose-adjusted (DA)EPOCH-R (etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab) combined with high-dose methotrexate (HD-MTX) in newly diagnosed patients with CD5+ DLBCL. Previously untreated patients with stage II to IV CD5+ DLBCL according to the 2008 World Health Organization classification were eligible. Four cycles of DA-EPOCH-R followed by two cycles of HD-MTX and four additional cycles of DAEPOCH-R (DA-EPOCH-R/HD-MTX) were planned as the protocol treatment. The primary end point was 2-year progression-free survival (PFS). Between September 25, 2012, and November 11, 2015, we enrolled 47 evaluable patients. Forty-five (96%) patients completed the protocol treatment. There were no deviations or violations in the DA-EPOCH-R dose levels. The complete response rate was 91%, and the overall response rate was 94%. At a median follow up of 3.1 years (range, 2.0-4.9 years), the 2year PFS was 79% [95% confidence interval (CI): 64-88]. The 2-year overall survival was 89% (95%CI: 76-95). Toxicity included grade 4 neutropenia in 46 (98%) patients, grade 4 thrombocytopenia 12 (26%) patients, and febrile neutropenia in 31 (66%) patients. No treatment-related death was noted during the study. DA-EPOCH-R/HD-MTX might be a first-line therapy option for stage II-IV CD5+ DLBCL and warrants further investigation. (Trial registered at: UMIN-CTR: UMIN000008507.) haematologica | 2020; 105(9)

DA-EPOCH-R/HD-MTX for CD5+ DLBCL

Introduction

Methods

Diffuse large B-cell lymphoma (DLBCL) includes a heterogeneous group of lymphomas with various clinicopathological and genetic features.1,2 The current standard first-line therapy of DLBCL is rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). Gene expression profiling (GEP) has identified activated B-cell-like (ABC), germinal center B-cell-like (GCB), and unclassified DLBCL.3 ABC DLBCL exhibits worse prognosis than GCB DLBCL after R-CHOP.4 CD5-positive (CD5+) DLBCL is an immunohistochemical subgroup of DLBCL, not otherwise specified (NOS) in the 2008 World Health Organization (WHO) classification and accounts for 5-10% of DLBCL.1 It is associated with various clinical characteristics, such as elderly onset, advanced stage at diagnosis, elevated lactate dehydrogenase (LDH) level, and frequent involvement of extranodal sites.5-10 Ninety percent of CD5+ DLBCL cases are positive for BCL2,6,9 which has been reported as a risk factor for short survival in DLBCL. The patients with CD5+ DLBCL show significantly worse survival than those with CD5â&#x20AC;&#x201C; negative (CD5â&#x20AC;&#x201C;) DLBCL after R-CHOP,7,9-14 and is also characterized by a high incidence of central nervous system (CNS) relapse (8-13%), mainly brain parenchymal relapse.8,10 CNS relapse strongly affects the prognosis of CD5+ DLBL.8 Up to 80% of CD5+ DLBCL are classified as ABC DLBCL by GEP,9,15,16 and CD5 expression is significantly associated with the poor prognosis of patients with ABC DLBCL.9,15 Dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (DAEPOCH-R) is an infusional regimen that incorporates a dynamic dose adjustment and has shown excellent efficacy in BCL2+ DLBCL.17,18 High-dose methotrexate (HDMTX) is commonly used in the treatment of primary DLBCL of the CNS.19 To explore a more effective first-line chemotherapy for CD5+ DLBCL, we have been investigating DA-EPOCH-R combined with high-dose methotrexate (HD-MTX) therapy (DA-EPOCH-R/HD-MTX) since 2009.20 In 2009, when we were planning this study, eight cycles of R-CHOP was used as the standard treatment for the patients with advanced DLBCL.21 In previous phase II trials using DA-EPOCH-R, patients received two cycles beyond complete response (CR) or stable changes after six cycles of DA-EPOCH-R.17,18 Considering the aggressiveness of CD5+ DLBCL, we selected eight cycles of DAEPOCH-R based on the maximum number of cycles in original studies of DA-EPOCH-R.17,18,22 Because most CNS relapse events in patients with CD5+ DLBCL during R-CHOP have been documented within 6 months after diagnosis,8 we added two cycles of HD-MTX between the fourth and the fifth cycles of DA-EPOCH-R. In our retrospective analysis of five patients with untreated stage IV CD5+ DLBCL who received DA-EPOCH-R/HD-MTX, the overall response rate (ORR) was 100%, and all toxicities were manageable.20 Of note, all of the patients belonged to the high-risk group of the International Prognostic Index (IPI),23 and three patients had CNS involvement at diagnosis. To further evaluate the efficacy of DA-EPOCH-R/HDMTX for newly diagnosed CD5+ DLBCL, we have been conducting a prospective, multicenter, single-arm, openlabel, phase II trial in 31 hospitals in Japan since 2015 (registered at: UMIN000008507). Here, we report the primary analysis results of that trial.

Study design and participants

haematologica | 2020; 105(9)

In this multicenter phase II trial, eligible patients were 20-75 years of age and had newly diagnosed CD5+ CD20+ DLBCL with Ann Arbor stage II-IV disease (see Online Supplementary Methods). The diagnosis of CD5+ DLBCL was made according to the 2008 WHO criteria1 as DLBCL, NOS with CD5 and CD20 expression in tumor cells based on immunohistochemistry and/or flow cytometry. A complete list of inclusion and exclusion criteria is shown in Online Supplementary Table S1. The study was approved by the protocol review committee of the study as well as the institutional review board of each institution in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients before enrollment.

Treatment The protocol treatment comprised four cycles of DA-EPOCH-R followed by two cycles of HD-MTX and four additional cycles of DA-EPOCH-R. DA-EPOCH-R was administered according to original reports.17,18,22 The treatment details are described in the Online Supplementary Methods. The fourth and fifth cycles of DAEPOCH-R were administered at the same dose level. MTX was administered intravenously at a dose of 3.5 g/m2 on a 14-day cycle with leucovorin rescue. The patients did not receive intrathecal chemotherapy. Acyclovir and trimethoprim-sulfamethoxazole were administered prophylactically. Radiation therapy was not permitted during the protocol treatment.

Diffuse large B-cell lymphoma subtyping Following patient enrollment, formalin-fixed paraffin-embedded (FFPE) sections were histologically reviewed according to both the 2008 WHO criteria and the 2017 WHO criteria by the central pathology review board because the WHO classification was updated during the study period.1,2 DLBCL cell-of-origin (COO) subtypes were determined based on an immunohistochemical method developed by Hans et al.24 and gene-expression-based Lymph2Cx assay.25 The details of immunohistochemistry and the methods of Lymph2Cx assay are given in the Online Supplementary Methods.

End points The primary end point was 2-year progression-free survival (PFS) (see the Online Supplementary Methods). Secondary end points were CR rate, ORR, 2-year overall survival (OS), 2-year CNS relapse rate, and toxicity. As a preplanned exploratory analysis, we assessed survival in each subgroup of COO, IPI, and CNS-IPI categories.26 Treatment responses were assessed at each participating institute using positron emission tomography-computed tomography within 6-8 weeks after the beginning of the eighth cycle of DA-EPOCH-R according to the revised International Working Group Criteria.27 Toxicity was graded according to the Common Terminology Criteria for Adverse Events version 4.0.

Statistical analysis Because this was the first clinical trial for untreated CD5+ DLBCL, we obtained historical controls of 2-year PFS (51%) from the data set of our previous retrospective study of CD5+ DLBCL.8 (See the Online Supplementary Methods and Online Supplementary Figure S1 for details) We anticipated that 41 evaluable patients were needed to achieve 80% power to detect a 20% difference in 2-year PFS from the historical control (51%) with a one-sided type I error of 0.05. The planned sample size was 45 patients with the expectation that 10% would be ineligible. Survival estimates were calculated using the Kaplan-Meier method, and 95% confidence 2309

K. Miyazaki et al.

Figure 1. Trial profile.

intervals (CI) were estimated using Greenwoodâ&#x20AC;&#x2122;s formula. All analyses were performed using SAS v9.4 (Cary, NC, USA) statistical package.

Results

tions in determining the dose levels of DA-EPOCH-R. All patients were administered with DA-EPOCH-R/HDMTX entirely in the hospital. Of the 47 patients evaluable for response, 43 patients achieved a CR (91%; 95%CI: 8098%), one patient achieved a partial response, and two patients experienced progressive disease. The ORR was 94% (95%CI: 82-99%).

Patients' characteristics A total of 47 patients from 24 participating institutions were enrolled in the study between July 2012 and November 2015 (Figure 1). The central pathology review confirmed the diagnosis of all patients as CD5+ DLBCL based on the 2008 WHO criteria. The review also revealed that two patients had high-grade B-cell lymphoma, NOS based on the 2017 WHO classification. The baseline clinical features and disease characteristics are listed in Table 1. Thirty-six percent of the patients were over 65 years of age. Skin/subcutaneous tissue was the most frequent site of extranodal involvement (15%). One patient had primary testicular DLBCL. For one case, the quality of the RNA extracted from the FFPE tissue sample was inadequate, but the remaining 46 cases were analyzed using GEP. Thirty-nine (85%) of 46 cases were classified as ABC DLBCL, four (9%) were classified as GCB DLBCL, and three (7%) were unclassified. The morphological and immunophenotypical features are summarized in Table 2. Based on Hansâ&#x20AC;&#x2122; criteria, 72% of the 46 patients examined were classified as non-GCB DLBCL.

Treatment and response Forty-five (96%) patients completed the protocol treatment. In the remaining two patients, treatment was discontinued because of tumor lysis syndrome (TLS) with hyperkalemia after the first administration of rituximab and grade 3 stomatitis during the sixth cycle of DAEPOCH-R. The dose-adjustment map of DA-EPOCH-R is shown in Online Supplementary Figure S2. The range of the dose level of DA-EPOCH-R was 1-4; median dose level of DA-EPOCH-R was two. The maximum dose level of DAEPOCH-R was 1 in 26.1%, 2 in 32.6%, 3 in 26.1%, 4 in 15.2%, and >4 in 0%. There were no deviations or viola2310

Survival With a median follow up of 3.1 years (range, 2.0-4.9), the 2-year PFS was 79% (90%CI: 67-87%; 95%CI: 6488%) (Figure 2A). This finding compared favorably with that of the historical control (51%). The 2-year OS was 89% (90%CI: 79-95%; 95%CI: 76-95%) (Figure 2B). One patient with CR died in a traffic accident 0.8 years after enrollment. In a preplanned exploratory subgroup analysis, there was no significant difference in PFS (Online Supplementary Figure S3A) or OS (Online Supplementary Figure S3B) according to COO categories. The 2-year PFS and OS for CD5+ ABC DLBCL (n=39) were 77% and 87%, respectively. There was no significant difference in PFS according to gender and each risk factor for IPI (data not shown).

Central nervous system relapse The 2-year CNS relapse rate was 9% (n=4; 90%CI: 419%, 95%CI: 3-21%) (Figure 3A). During the follow-up period, no patient experienced CNS relapse more than 2 years from enrollment. There was no significant difference in PFS among the CNS-IPI categories (P=0.58) (Figure 3B). Ten patients in the CNS IPI high-risk group were classified as ABC (n=8), GCB (n=1), and unclassified DLBCL (n=1). The clinical characteristics of four patients who experienced CNS relapse are summarized in Online Supplementary Table S2. Among them, one patient had primary testicular DLBCL. Two patients were diagnosed as having high-grade B-cell lymphoma, NOS with MYC rearrangement based on the 2017 WHO criteria. Neither of these patients had the BCL2 translocation, and both were positive for CD10 and BCL2 by immunohistochemistry. For the remaining one patient, the protocol treathaematologica | 2020; 105(9)

DA-EPOCH-R/HD-MTX for CD5+ DLBCL

Table 1. Baseline clinical features and disease characteristics.

Characteristic Age, years Median Range >60 years Sex Male Female Stage II III-IV ECOG PS 0 or 1 >1 Serum LDH level Normal Elevated Extranodal sites 0 or 1 >1 IPI risk category Low/Low-intermediate High-intermediate/High CNS-IPI risk category Low Intermediate High Cell-of-origin* ABC GCB Unclassified

62 37-74 28 (60%) 18 (38%) 29 (62%) 20 (43%) 27 (57%) 45 (96%) 2 (4%) 16 (34%) 31 (66%) 31 (66%) 16 (34%) 25 (53%) 22 (47%) 14 (30%) 23 (49%) 10 (21%) 39 (85%) 4 (9%) 3 (7%)

ABC: activated B-cell-like; CNS: central nervous system; EBER: Epstein-Barr virus-encoded RNA-1. GCB: germinal center-B-cell-like; IPI: International Prognostic Index; PS: performance status. *Examined by Lymph2Cx using formalin-fixed and paraffin-embedded tissues (n=46).

ment was discontinued after the first drug administration due to grade 4 TLS. CNS relapses occurred in two patients (one patient had leptomeningeal disease with tumor of ethmoid sinus, and the other patient had brain parenchymal disease alone) before HD-MTX of the protocol treatment.

Toxicity Toxicity was assessed in all 445 cycles of DA-EPOCHR/HD-MTX with 357 cycles of DA-EPOCH-R in all 47 patients. Table 3 lists adverse events in all 445 cycles of DA-EPOCH-R/HD-MTX, and Online Supplementary Table S3 lists adverse events in 357 cycles of DA-EPOCH-R. There were no treatment-related deaths. The most common grade 3 non-hematologic toxicity was elevated ALT (28%). Four patients experienced grade 3 infection, including catheter-related infection, cellulitis, endocarditis, and urinary tract infection. Grade 3 peripheral motor neuropathy and peripheral sensory neuropathy occurred in 4% of patients, separately. There was no grade 3 cardiac toxicity. Grade 4 neutropenia and thrombocytopenia were present in 98% and 26% of patients, respectively. The targeted ANC (<0.5x109/L) occurred in 77% of haematologica | 2020; 105(9)

Table 2. Morphological and immunophenotypical features.

All patients (n=47)

All patients (n=47) Morphology Common variant Giant cell rich variant Polymorphic variant Immunoblastic variant Immunohistochemistry CD5* CD20 CyclinD1 CD10† BCL6† MUM1† MYC† BCL2† MYC positive/BCL2 positive† COO subtype by Hans' criteria GCB† Non-GCB† EBER in situ†

31 (66%) 3 (6%) 12 (26%) 1 (2%) 47 (100%) 47 (100%) 0 (0%) 13 (28%) 39 (85%) 44 (96%) 26 (57%) 45 (98%) 26 (57%) 13 (28%) 33 (72%) 0 (0%)

COO: cell-of-origin; EBER: Epstein-Barr virus-encoded RNA-1; GCB: germinal center-Bcell-like. *Assessed by flow cytometry in 4 patients. †Examined in 46 patients.

cycles. Thrombocytopenia of <25x109/L occurred in 8% of cycles. Febrile neutropenia (FN) occurred in 66% of patients and 23% per cycle of DA-EPOCH-R (Table 3 and Online Supplementary Table S3). All patients who experienced FN recovered from the toxicity soon after initial antibiotics therapy. Possible second malignancies were documented in three patients (age, 64-74 years; acute myeloid leukemia, n=1; glioblastoma, n=1; and colon cancer in adenoma, n=1). In two of these patients, the second malignancy was diagnosed after salvage chemotherapy for relapsed CD5+ DLBCL.

Discussion To the best of our knowledge, this is the first clinical trial for CD5+ DLBCL. At the beginning of this trial in 2015, we calculated the threshold 2-year PFS using our previous retrospective study including 337 patients with CD5+ DLBCL. In 149 patients with newly diagnosed stage II-IV CD5+ DLBCL who received rituximab-containing chemotherapy (R-chemo) and had available follow-up data, the 2-year PFS was 51%. Recently, the results of several retrospective studies of CD5+ DLBCL in the rituximab-era (R-era) were reported. In both cohorts of 102 patients with all stage CD5+ DLBCL who received R-chemo at nine institutes in the United States10 and 31 patients with all stage CD5+ DLBCL who received RCHOP in other Western countries,9 the 2-year OS was less than 50%. Therefore, we consider that our threshold setting was reasonable and that the 2-year PFS (79%) by the protocol treatment in our present study was superior to the historical control. Few reports have described the use of DA-EPOCH-R for CD5+ DLBCL. For all DLBCL subtypes, there was no sig2311

K. Miyazaki et al.

nificant difference in survival between patients who received R-CHOP and patients who received DA-EPOCHR in a large randomized controlled trial.28 In that trial, the prognosis stratified by CD5 was not shown. In the largest study including 130 DLBCL patients who were examined for CD5 expression and received DA-EPOCH-R, 16 (12.3%) patients had CD5+ DLBCL.29 At a median follow up of 28.5 months, 37.5% of patients in the CD5+ DLBCL group had died, which showed a significantly worse survival than those in the CD5- DLBCL. The 2-year OS was 89% (90%CI: 79-95%; 95%CI: 76-95%) for DA-EPOCHR/HD-MTX in our present trial. In a large retrospective study including 102 patients with CD5+ DLBCL, 6 of 7 patients with untreated CD5+ DLBCL who received DA-EPOCH-R achieved a CR and

were disease-free with a median follow up of 30 months.10 In another study, CD5 expression in DLBCL was identified as a risk factor for short OS in a cohort of patients who received R-EPOCH,29 which indicated that R-EPOCH alone may be insufficient to cure this disease. In the present study of stage II-IV CD5+ DLBCL, HD-MTX therapy was added between the fourth and fifth cycles of DAEPOCH-R. In our trial, the completion rate of the protocol treatment was high. Moreover, there were no deviations or violations in the dose-adjustment of DA-EPOCH-R. All patients received DA-EPOCH-R/HD-MTX only in the hospital. These features may have contributed to the excellent efficacy of DA-EPOCH-R/HD-MTX in our trial. However, a considerable number of female patients (62%), patients with PS <2 (96%), and those with

Figure 2. Survival curves for patients with CD5-positive diffuse large B-cell lymphoma. (A) Progression-free survival and (B) overall survival for all patients (n=47).

Figure 3. Central nervous system (CNS) relapse in CD5-positive diffuse large B-cell lymphoma patients (n=47). (A) Probability of CNS relapse. (B) Probability of CNS relapse according to the CNS-International Prognostic Index.

2312

haematologica | 2020; 105(9)

DA-EPOCH-R/HD-MTX for CD5+ DLBCL

low/low-intermediate IPI risk (53%) were enrolled in the present study. In the subgroup analysis of PFS, there was no significant difference between those with and those without risk factors for gender, PS, and IPI group, although the sample size was small. Considering these results, further studies are needed to establish DA-EPOCH-R/HDMTX as a standard therapy for newly diagnosed stage IIIV CD5+ DLBCL. Nevertheless, we consider that this regimen might be regarded as a treatment option, because CD5 expression can be easily examined in clinical practice. The incidence of CNS relapse for all patients with DLBCL treated with R-CHOP has been reported to be 1.96.4%,30-32 whereas the incidence of CNS relapse for CD5+ DLBCL in the R-era were reported as 13% in our retrospective study,8 8% in a study in Western countries,9 and 33% in ten patients treated with DA-EPOCH-R in a retrospective study.29 Although our present trial does not have a sufficient statistical power to assess the reduction in CNS relapse, the incidence of CNS relapse in the present study (9%; 95%CI: 3-21%) seems to be lower than that for DA-EPOCH-R alone. To further reduce the incidence of CNS relapse, development of systemic chemotherapy in combination with one or more newer agents that cross

the blood-brain barrier, including ibrutinib,33,34 lenalidomide35 and tumor necrosis factor-a coupled with NGR,36 is warranted. Combining a high CNS-IPI score and ABC/classified COO were reported to identify a patient subgroup at high risk for developing CNS relapse.37 Of note, in the present trial, 90% of patients in the CNS-IPI high-risk group were ABC/unclassified DLBCL. However, in four patients with CNS relapse, one patient had primary testicular lymphoma, which should be treated with a specific treatment strategy including local radiotherapy.38 Moreover, 2 of the 4 patients with CNS relapse had highgrade B-cell lymphoma, NOS with MYC translocation, which is known to result in a relatively high incidence of CNS relapse.39 These results suggest that CD5+ DLBCL diagnosed according to the 2008 WHO criteria was a heterogeneous lymphoma and that DA-EPOCH-R/HD-MTX might be beneficial for patients with CD5+ DLBCL in the WHO 2017 era. As expected, the incidence of grade 3 peripheral motor neuropathy, peripheral sensory neuropathy and cardiac events was low (4%, 4%, and 0%, respectively) and was comparable to that in previous studies including DAEPOCH-R.17,18 Two-thirds (66%) of the patients in this study experienced FN, and the incidence was higher than

Table 3. Hematologic and non-hematologic toxicities. Hematologic adverse event Neutropenia Leukopenia Thrombocytopenia Anemia Febrile neutropenia Non-hematologic adverse event Blood bilirubin increased AST increased ALT increased Hyperglycemia Hyponatremia Hyperkalemia Hypokalemia Hypocalcaemia Cardiac disorders Constipation Ileus Nausea Vomiting Infection Allergic reaction Tumor lysis syndrome Peripheral motor neuropathy Peripheral sensory neuropathy Pneumonitis Others

Grade 1-2

Grade 3

Grade 4

Grade 5

0 1 (2%) 13 (28%) 11 (23%) -

1 (2%) 0 22 (47%) 26 (55%) 31 (66%)*

46 (98%) 46 (98%) 12 (26%) 10 (21%) 0

0 0 0 0 0

9 (19%) 36 (77%) 30 (64%) 40 (85%) 36 (77%) 20 (43%) 38 (81%) 44 (94%) 31 (66%) 1 (2%) 33 (70%) 9 (19%) 11 (23%) 0 11 (23%) 29 (62%) 2 (4%) -

2 (4%) 5 (11%) 13 (28%) 1 (2%) 7 (15%) 0 8 (17%) 1 (2%) 0 3 (6%) 2 (4%) 1 (2%) 1 (2%) 4 (9%)‡ 0 3 (6%) 2 (4%) 2 (4%) 1 (2%) 10 (21%)§

0 0 0 0 0 1 (2%)† 0 1 (2%) 0 0 0 0 0 0 0 1 (2%) 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

AST: aspartate aminotransferase; ALT: alanine aminotransferase. *Observed in 23% of all cycles of DA-EPOCH-R (Online Supplementary Table S3). †One patient experienced tumor lysis syndrome with hyperkalemia after the first administration of rituximab. ‡Catheter related infection (n=1), cellulitis (n=1), infective endocarditis (n=1), and urinary tract infection (n=1). §Oral mucositis (n=3), hypertension (n=2), pulmonary embolism (n=2), deep vein thrombosis (n=1), Palmar-Planter erythrodysesthesia syndrome (n=1), and anorexia (n=1).

haematologica | 2020; 105(9)

2313

K. Miyazaki et al. that in previous reports (36-37%).17,18,28 However, the incidence of FN per cycle of DA-EPOCH in our study was 23%, which was comparable to that in previous reports for DA-EPOCH-R (19%).17,40 The long period (7 months) of this protocol treatment may have resulted in the high proportion of patients who experienced FN. All patients with FN promptly recovered after initial antibiotic therapy and had no serious events. Hence, the toxicity could be considered manageable. Although secondary malignancy was observed in three patients, all of them were older, and two of them were treated with salvage chemotherapy after relapse. A longer follow up is needed to adequately evaluate second malignancies. Our study has several limitations. First, it is a non-randomized phase II study with a small number of patients. Second, 96% of the patients had good PS (< 2). Third, the follow-up period for the evaluation of secondary malignancy was short. Fourth, considering the rarity of the disease and the feasibility of the clinical trial, this study was not designed to assess 2-year CNS relapse rate as a primary end point. Nevertheless, our results showed promis-

References 11. 1. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. 4th ed. Lyon: IARC Press, 2008. 2. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Revised 4th ed. Lyon: IARC Press, 2017. 3. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-Bcell lymphoma. N Engl J Med. 2002; 346(25):1937-1947. 4. Lenz G, Wright G, Dave SS, et al. Stromal gene signatures in large-B-cell lymphomas. N Engl J Med. 2008;359(22):2313-2323. 5. Yamaguchi M, Seto M, Okamoto M, et al. De novo CD5+ diffuse large B-cell lymphoma: a clinicopathologic study of 109 patients. Blood. 2002;99(3):815-821. 6. Yamaguchi M, Nakamura N, Suzuki R, et al. De novo CD5+ diffuse large B-cell lymphoma: results of a detailed clinicopathological review in 120 patients. Haematologica. 2008;93(8):1195-1202. 7. Ennishi D, Takeuchi K, Yokoyama M, et al. CD5 expression is potentially predictive of poor outcome among biomarkers in patients with diffuse large B-cell lymphoma receiving rituximab plus CHOP therapy. Ann Oncol. 2008;19(11):1921-1926. 8. Miyazaki K, Yamaguchi M, Suzuki R, et al. CD5-positive diffuse large B-cell lymphoma: a retrospective study in 337 patients treated by chemotherapy with or without rituximab. Ann Oncol. 2011;22(7):1601-1607. 9. Xu-Monette ZY, Tu M, Jabbar KJ, et al. Clinical and biological significance of de novo CD5+ diffuse large B-cell lymphoma in Western countries. Oncotarget. 2015; 6(8):5615-5633. 10. Alinari L, Gru A, Quinion C, et al. De novo CD5+ diffuse large B-cell lymphoma: Adverse outcomes with and without stem cell transplantation in a large, multicenter,

2314

12.

13.

14.

15.

16.

17.

18.

19. 20.

ing efficacy and manageable toxicity of DA-EPOCHR/HD-MTX for untreated stage II-IV CD5+ DLBCL, and this trial is an important initial step in developing a more effective treatment strategy. To assess long-term efficacy and toxicity including second malignancy, a 5-year follow up is scheduled in November 2020. Acknowledgments We would like to thank all of the patients, hemato-oncologists, and pathologists at the participating institutes for invaluable contributions to this multicenter study. Funding This study was supported by grants-in-aid from the Japan Agency for Medical Research and Development, AMED (the Practical Research for Innovative Cancer Control; JP15Ack0106157, JP16ck0106157, JP17ck0106157, JP18ck0106439), the Ministry of Labour, Health, and Welfare of Japan (201438142A), the director of Mie University Hospital (2012, 2013), and the National Cancer Center Research and Development Fund (26-A-4, 29-A-3).

rituximab treated cohort. Am J Hematol. 2016;91(4):395-399. Hyo R, Tomita N, Takeuchi K, et al. The therapeutic effect of rituximab on CD5-positive and CD5-negative diffuse large B-cell lymphoma. Hematol Oncol. 2010;28(1):2732. Johnson NA, Boyle M, Bashashati A, et al. Diffuse large B-cell lymphoma: reduced CD20 expression is associated with an inferior survival. Blood. 2009;113(16):37733780. Salles G, de Jong D, Xie W, et al. Prognostic significance of immunohistochemical biomarkers in diffuse large B-cell lymphoma: a study from the Lunenburg Lymphoma Biomarker Consortium. Blood. 2011; 117(26):7070-7078. Chuang WY, Chang H, Shih LY, et al. CD5 positivity is an independent adverse prognostic factor in elderly patients with diffuse large B cell lymphoma. Virchows Arch. 2015;467(5):571-582. Suguro M, Tagawa H, Kagami Y, et al. Expression profiling analysis of the CD5+ diffuse large B-cell lymphoma subgroup: development of a CD5 signature. Cancer Sci. 2006;97(9):868-874. Miyazaki K, Yamaguchi M, Imai H, et al. Gene expression profiling of diffuse large BCell lymphomas supervised by CD5 expression. Int J Hematol. 2015;102(2):188-194. Wilson WH, Dunleavy K, Pittaluga S, et al. Phase II study of dose-adjusted EPOCH and rituximab in untreated diffuse large B-cell lymphoma with analysis of germinal center and post-germinal center biomarkers. J Clin Oncol. 2008;26(16):2717-2724. Wilson WH, Jung SH, Porcu P, et al. A Cancer and Leukemia Group B multi-center study of DA-EPOCH-rituximab in untreated diffuse large B-cell lymphoma with analysis of outcome by molecular subtype. Haematologica. 2012;97(5):758-765. Grommes C, DeAngelis LM. Primary CNS Lymphoma. J Clin Oncol. 2017;35(21):24102418. Miyazaki K, Yamaguchi M, Umino A, et al. DA-EPOCH-R Combined With High-dose

Methotrexate For Newly Diagnosed CD5positive Diffuse Large B-cell Lymphoma. Hematol Oncol. 2013;31:227. 21. Feugier P, Van Hoof A, Sebban C, et al. Longterm results of the R-CHOP study in the treatment of elderly patients with diffuse large B-cell lymphoma: a study by the Groupe d'Etude des Lymphomes de l'Adulte. J Clin Oncol. 2005;23(18):41174126. 22. Wilson WH, Grossbard ML, Pittaluga S, et al. Dose-adjusted EPOCH chemotherapy for untreated large B-cell lymphomas: a pharmacodynamic approach with high efficacy. Blood. 2002;99(8):2685-2693. 23. A predictive model for aggressive nonHodgkin's lymphoma. N Engl J Med. 1993; 329(14):987-994. 24. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275-282. 25 . Scott DW, Wright GW, Williams PM, et al. Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffinembedded tissue. Blood. 2014;123(8):12141217. 26. Schmitz N, Zeynalova S, Nickelsen M, et al. CNS International Prognostic Index: a risk model for CNS relapse in patients with diffuse large B-cell lymphoma treated with RCHOP. J Clin Oncol. 2016;34(26):3150-3156. 27. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579-586. 28. Bartlett NL, Wilson WH, Jung SH, et al. Dose-Adjusted EPOCH-R Compared With R-CHOP as Frontline Therapy for Diffuse Large B-Cell Lymphoma: Clinical Outcomes of the Phase III Intergroup Trial Alliance/CALGB 50303. J Clin Oncol. 2019; 37(21):1790-1799. 29. Thakral B, Medeiros LJ, Desai P, et al. Prognostic impact of CD5 expression in diffuse large B-cell lymphoma in patients treated with rituximab-EPOCH. Eur J Haematol. 2017;98(4):415-421.

haematologica | 2020; 105(9)

DA-EPOCH-R/HD-MTX for CD5+ DLBCL

30. Boehme V, Schmitz N, Zeynalova S, Loeffler M, Pfreundschuh M. CNS events in elderly patients with aggressive lymphoma treated with modern chemotherapy (CHOP-14) with or without rituximab: an analysis of patients treated in the RICOVER-60 trial of the German High-Grade Non-Hodgkin Lymphoma Study Group (DSHNHL). Blood. 2009;113(17):3896-3902. 31. Villa D, Connors JM, Shenkier TN, et al. Incidence and risk factors for central nervous system relapse in patients with diffuse large B-cell lymphoma: the impact of the addition of rituximab to CHOP chemotherapy. Ann Oncol. 2010;21(5):1046-1052. 32. Gleeson M, Counsell N, Cunningham D, et al. Central nervous system relapse of diffuse large B-cell lymphoma in the rituximab era: results of the UK NCRI R-CHOP-14 versus 21 trial. Ann Oncol. 2017;28(10):2511-2516. 33. Grommes C, Pastore A, Palaskas N, et al. Ibrutinib Unmasks Critical Role of Bruton Tyrosine Kinase in Primary CNS Lymphoma.

haematologica | 2020; 105(9)

Cancer Discov. 2017;7(9):1018-1029. 34. Grommes C, Tang SS, Wolfe J, et al. Phase 1b trial of an ibrutinib-based combination therapy in recurrent/refractory CNS lymphoma. Blood. 2019;133(5):436-445. 35. Ghesquieres H, Chevrier M, Laadhari M, et al. Lenalidomide in combination with intravenous rituximab (REVRI) in relapsed/refractory primary CNS lymphoma or primary intraocular lymphoma: a multicenter prospective 'proof of concept' phase II study of the French Oculo-Cerebral lymphoma (LOC) Network and the Lymphoma Study Association (LYSA). Ann Oncol. 2019; 30(4):621-628. 36. Ferreri AJM, Calimeri T, Conte GM, et al. RCHOP preceded by blood-brain barrier permeabilization with engineered tumor necrosis factor-alpha in primary CNS lymphoma. Blood. 2019;134(3):252-262. 37. Klanova M, Sehn LH, Bence-Bruckler I, et al. Integration of cell of origin into the clinical CNS International Prognostic Index

improves CNS relapse prediction in DLBCL. Blood. 2019;133(9):919-926. 38. Vitolo U, Chiappella A, Ferreri AJ, et al. Firstline treatment for primary testicular diffuse large B-cell lymphoma with rituximabCHOP, CNS prophylaxis, and contralateral testis irradiation: final results of an international phase II trial. J Clin Oncol. 2011;29(20):2766-2772. 39. Barrans S, Crouch S, Smith A, et al. Rearrangement of MYC is associated with poor prognosis in patients with diffuse large B-cell lymphoma treated in the era of rituximab. J Clin Oncol. 2010;28(20):33603365. 40. Dunleavy K, Fanale MA, Abramson JS, et al. Dose-adjusted EPOCH-R (etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab) in untreated aggressive diffuse large B-cell lymphoma with MYC rearrangement: a prospective, multicentre, single-arm phase 2 study. Lancet Haematol. 2018;5(12):e609-e617.

2315

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2316-2326

Plasma Cell Disorders

RAL GTPases mediate multiple myeloma cell survival and are activated independently of oncogenic RAS Marcel Seibold,1 Thorsten Stühmer,2 Nadine Kremer,2 Anja Mottok,3 Claus-Jürgen Scholz,4 Andreas Schlosser,5 Ellen Leich,6 Ulrike Holzgrabe,7 Daniela Brünnert,2 Santiago Barrio,8 K. Martin Kortüm,1 Antonio G. Solimando,1 Manik Chatterjee,2 Hermann Einsele,1 Andreas Rosenwald,6 Ralf C. Bargou2 and Torsten Steinbrunn1

Department of Medicine II, University Hospital of Würzburg, Würzburg, Germany; Comprehensive Cancer Center Mainfranken, Chair of Translational Oncology, University Hospital of Würzburg, Würzburg, Germany; 3Institute of Human Genetics, University of Ulm, Ulm, Germany; 4Core Unit Systems Medicine, University of Würzburg, Würzburg, Germany; 5Rudolf Virchow Center for Experimental Biomedicine, University of Würzburg, Würzburg, Germany; 6Institute of Pathology, University of Würzburg, Würzburg, Germany; 7 Institute of Pharmacy and Food Chemistry, University of Würzburg, Würzburg, Germany and 8Hematology Department, Hospital 12 de Octubre, Complutense University, Madrid, Spain 1 2

ABSTRACT

Correspondence: TORSTEN STEINBRUNN steinbrunn_t@ukw.de Received: March 26, 2019. Accepted: October 10, 2019. Pre-published: October 10, 2019.

ncogenic RAS provides crucial survival signaling for up to half of multiple myeloma (MM) cases, but has so far remained a clinically undruggable target. RAS-like protein (RAL) is a member of the RAS superfamily of small GTPases and is considered to be a potential mediator of oncogenic RAS signaling. In primary MM, we found RAL to be overexpressed in the vast majority of samples when compared with pre-malignant monoclonal gammopathy of undetermined significance or normal plasma cells. We analyzed the functional effects of RAL abrogation in myeloma cell lines and found that RAL is a critical mediator of survival. RNAi-mediated knockdown of RAL resulted in rapid induction of tumor cell death, an effect which was independent from signaling via mitogen-activated protein kinase, but appears to be partially dependent on Akt activity. Notably, RAL activation was not correlated with the presence of activating RAS mutations and remained unaffected by knockdown of oncogenic RAS. Furthermore, transcriptome analysis yielded distinct RNA expression signatures after knockdown of either RAS or RAL. Combining RAL depletion with clinically relevant anti-myeloma agents led to enhanced rates of cell death. Our data demonstrate that RAL promotes MM cell survival independently of oncogenic RAS and, thus, this pathway represents a potential therapeutic target in its own right.

Introduction doi:10.3324/haematol.2019.223024 ©2020 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.

2316

Mutated RAS is one the most frequent oncogenic drivers in human cancers, yet it has so far confounded efforts to render it a clinically exploitable drug target.1–4 Consequently, the identification and targeting of RAS effector pathways has been pursued to establish therapeutic approaches that counter RAS-driven tumors.5–7 Multiple myeloma (MM) harbors oncogenic NRAS or KRAS mutations in up to half of the cases and we have shown that RNA-mediated knockdown of oncogenic RAS induces apoptosis in MM cell lines.8–10 In vitro, the so-called classical RAS-associated pathways which signal via RAF/MAPK and PI3K/Akt, respectively, have been studied at different levels in MM cells and have been shown to be crucial for MM cell survival.11–19 In addition, we have demonstrated that although inhibition of one or both of these pathways can strongly affect MM cell growth and survival in vitro,8,14 their constitutive activation appears not to be directly correlated with the presence of oncogenic RAS.8,11 However, early clinical trials including MM patients treated with pharmacological inhibitors of either one of these pathways have shown only limited efficacy,20–22 haematologica | 2020; 105(9)

RAL GTPases mediate multiple myeloma cell survival

whereas combined blockade in patients with solid tumors resulted in high levels of toxicity.23–26 The identification of alternative RAS-driven pathways to target MM cells is therefore highly warranted. Here, we investigated the functional role of RAS-like protein (RAL) in MM, which has sometimes been branded “the third pathway” in the context of RAS-dependent oncogenic signaling.27–29 RAL belongs to the RAS superfamily of small GTPases that – like RAS itself – are characterized by cycling between a GTP-bound active and a GDP-bound inactive state. The two isoforms of RAL: RALA and RALB have both been shown to be involved in malignant transformation, tumor cell survival, and tumor cell growth and metastasis, although their functional role(s) may depend to some extent on the tumor entity and/or model tested.30–32 In our study, we sought to analyze the functional importance of RAL in MM as bona fide downstream effector of oncogenic RAS by using RNAi-mediated knockdown approaches. We found that RAL is important for MM cell survival, but that its constitutive activation is not directly linked to oncogenic RAS. Furthermore, knockdown of RAL entails very different transcriptomic changes than RAS depletion. Therefore, we infer that the RAL pathway constitutes a potential clinical target in its own right.

Methods Culture of MM cell lines and preparation of primary MM cells Cell culture conditions of human myeloma cell line (HMCL) and isolation of CD138-positive primary MM cells were previously described.33 Bone marrow aspirates of MM patients were obtained after informed consent according to the Declaration of Helsinki, and with permission of the Ethics Committee of the University of Würzburg (reference no. 76/13). See the Online Supplementary Materials and Methods for details.

Immunohistochemical stainings of bone marrow biopsies To evaluate protein expression of the RAL isoforms in plasma cells we performed immunohistochemical analysis in formalinfixed, paraffin-embedded bone marrow biopsies from 26 patients with MM as previously described.8,14 For comparison, we analyzed patients with monoclonal gammopathy of undetermined significance (MGUS) (n=10) and bone marrow trephines containing reactive, polyclonal plasma cells (n=5). Slides were evaluated by experienced hematopathologists. See the Online Supplementary Materials and Methods for details.

Construction of shRNA expression vectors Construction of pSUPER-based small hairpin RNA (shRNA) expression vectors was performed as previously described.35 See the Online Supplementary Materials and Methods for sequences.36,37

Transfection of MM cells by electroporation Transient transfection of HMCL was previously described in detail.34 HMCL were electroporated with pSUPER-based shRNA expression vectors. ShRNA expression plasmid concentrations in the final electroporation mix were 20 μg/mL (15 μg/mL for transfections with subsequent drug treatment). Strongly transfected cells were purified by microbead selection for co-expressed CD4 or, in the case of AMO-1, by fluorescence-activated cell sorting for co-expressed enhanced green-fluorescent protein (EGFP).

RALA activity assay INA-6 and MM.1S cells were transfected with shRNA expression plasmids and harvested two days after electroporation. The activation status of RALA was measured using the RAL Activation Assay from Cell Biolabs (no. STA-408, San Diego, CA, USA) according to the manufacturer's instructions. Subsequent Western blotting was performed to analyze RAL-GTP levels and total RAL protein loads. Antibodies against RALA were diluted 1:500 or 1:1,000.

Western analysis Western blotting of cell lysates was performed according to standard protocols as previously described.12,34 See the Online Supplementary Materials and Methods for details.

RNA sequencing analysis For transcriptome analyses, MM.1S cells were transfected with pSUPER-based shRNA expression vectors against either KRAS or RALA. Control cells were transfected with empty pSUPER plasmids. RNA sequencing data are deposited in Gene Expression Omnibus in entry GSE126794. See the Online Supplementary Materials and Methods for details.

Mass spectrometry-based interactome analysis To identify RAL interaction partners we performed quantitative mass spectrometric analysis of MM.1S cells with stable expression of HA-tagged RALA protein. Detailed description of sample preparation and analysis is provided in the Online Materials and Methods and by Cox et al.38,39

Statistical analysis Statistical significance (P<0.05) was determined by a two-tailed Student’s t-test. Three independent experiments were performed.

Results

Cell death assay

RAL expression in multiple myeloma cells

Fractions of unaffected and (pre-)apoptotic cells were measured by flow cytometry after staining with propidium iodide (PI) and annexin V labeled with either PromoFluor 647, allophycocyanin (APC) or fluorescein isothiocyanate (FITC) as previously described.34 Cell death measurements were conducted at days 3 and 4 after transfection.

RAL protein expression in a panel of MM cell lines (n=7) and primary MM samples (n=10) was analyzed for each isoform using Western blotting. Both proteins were detected at fairly equal levels in all (RALA), and in 6 of 7 (RALB) HMCL, respectively (Figure 1A, top). Cell line U266 was notable for its complete lack of RALB expression. Interestingly, all cell lines showed constitutive RALA activation through the presence of GTP-bound RALA as detected by a pulldown assay (Figure 1A, bottom). In CD138-positive primary MM cells isolated from bone marrow aspirates of MM patients, both RAL proteins were always present. Accounting for differences in the amounts of sample loading, RALA and RALB expression

Cell metabolism, proliferation and cell cycle assays Alamar Blue and bromodeoxyuridine (BrdU)/PI assays were performed to analyze cell metabolism, proliferation and cell cycle distribution after RAL knockdown or pharmacological inhibition with RBC8. See the Online Supplementary Materials and Methods for details. haematologica | 2020; 105(9)

2317

M. Seibold et al.

Figure 1. Expression of the two RAL isoforms in multiple myeloma cells. Representative Western analyses showing expression levels of RALA, RALA-GTP and RALB as well as expression and phosphorylation levels of MAPK and Akt signaling in (A) human myeloma cell line (HMCL) (n=7) and (B) primary MM samples (n=10). Îątubulin served as loading control. (C) In situ expression of RALA and RALB in bone marrow trephines of multiple myeloma (MM) patients (n=26). CD138 staining as well as RALA and RALB staining shown for 3 different patients (I, II, III) at 200x (I) and 400x (II, III) magnification. Scale bars: 50 Îźm. Samples I, II, and III correspond to sample numbers 22, 25, and 26 in the Online Supplementary Table S2, respectively.

2318

haematologica | 2020; 105(9)

RAL GTPases mediate multiple myeloma cell survival

Figure 2. Effects of abrogation of RAL signaling with different small hairpin RNA expression vectors on multiple myeloma cell survival and signaling. RALA and RALB knockdown was achieved with two different small hairpin RNA (shRNA) expression vector constructs for each RAL isoform in L-363 cells (A) and MM.1S cells (B). Upon RALA knockdown, MM cell survival was significantly reduced 3 days and 4 days after electroporation. Similarly, RALB knockdown also reduced cell survival, but to a lesser extent than RALA depletion. Shown are mean values and standard deviation (SD) from three independent experiments. Percentages were calculated relative to the respective empty vector control. Cell viability was monitored by annexin V/PI staining. Exemplarily, Western blots of L-363 (C) and MM.1S (D) cell lysates show that RAS-dependent signaling in form of ERK1/2 phosphorylation and PI3K-dependent signaling illustrated by Akt and GSK-3β phosphorylation are not influenced by RALA or RALB knockdown 2 days after electroporation. (E) L-363 and (F) MM.1S cells were transfected with shRNA targeting RALA or RALB and purified cells were harvested after 2 and 3 days. Onset of apoptosis after RALA knockdown as indicated by cleavage of PARP 1 and caspase-3 was accompanied by reduction of phosphorylated Akt after 3 days, whereas after 2 days, signaling remained still unchanged. α-tubulin and β-actin served as loading control.

haematologica | 2020; 105(9)

2319

M. Seibold et al. A

Figure 3. Distinct regulation of the RAL and RAS pathways. (A) NRAS (G12D)-mutated INA-6 cells were transfected with a samll hairpin RNA (shRNA) expression vector against mutated NRAS, KRAS (G12A)-mutated MM.1S cells were transfected with an shRNA expression vector against mutated KRAS. As shown by Western analyses, RALA activation was not changed by depletion of oncogenic RAS (and its cognate wild-type form) in either cell line. RALA activation was measured by RALA-GTP pulldown with RALPB1 protein-binding domain agarose beads 48 hours (h) after transfection. RALA and a-tubulin total load samples were taken before the pulldown procedure. (B) To analyze RAL- versus RAS-dependent gene expression, MM.1S cells were transfected with shRNA expression vectors against KRAS or RALA and successful knockdown was confirmed by Western blotting. RNA was isolated 48 h after electroporation and analyzed with RNA-Seq. Three independent experiments were performed. (C) Of 1,473 genes that were expressed differentially after KRAS knockdown, 656 were up- and 817 downregulated. After RALA knockdown, 771 genes showed an altered expression, whereof 336 were up- and 435 downregulated. Of the 235 genes in the overlap, 135 were up- and 100 downregulated under both conditions. The diagram shows all genes with altered expression with a false discovery rate (FDR) <0.05. In total, 28,440 genes were analyzed. (D) Ontology mapping of differential gene expression highlighting the most distinct functional gene groups with relevance for MM growth and survival after RAL versus RAS knockdown was performed using the Molecular Signatures Database Hallmark Gene Set Collection.40 Adjusted P-value <0.05.

again appeared quite similar between primary samples, with a few notable digressions to the upside (RALA in samples 1 and 5, RALB in sample 5; Figure 1B). Expression of RAL isoforms in primary plasma cells was also analyzed in situ by immunohistochemical staining of bone marrow biopsies from MM patients (n=26) and compared with sections from non-MM (n=5) and MGUS patients (n=10). Co-staining was performed with the plasma cell marker CD138 (Figure 1C). Normal plasma cells showed no detectable expression of RALA except for one sample, which displayed weak staining in 10% of the cells. All normal plasma cell samples were negative for 2320

RALB staining. In 5 of 10 samples with pre-malignant MGUS plasma cells RALA was not detectable. The remaining 5 samples showed slightly elevated RALA expression levels. RALB expression did not reach the detection level in any MGUS sample (Online Supplementary Table S1 and Online Supplementary Figure S1). In contrast, 20 of 26 primary samples from MM patients showed medium to strong RALA expression in 80% to 100% of the cells. Two samples showed RALA expression in 25% and 50% of the cells, respectively. In 14 of 26 samples, RALB stained weakly in at least 5 % of MM cells (Online Supplementary Table S2). haematologica | 2020; 105(9)

RAL GTPases mediate multiple myeloma cell survival

Figure 4. Pharmacological RAL inhibition in multiple myeloma cell lines. (A) AMO-1, INA-6, L-363 and MM.1S cells were subjected to increasing concentrations of the RAL inhibitor RBC8. Cell survival was measured by annexin V/PI staining after 72 hours (h) of treatment. RBC8 treatment reduced cell survival of AMO-1 and INA-6 at concentrations higher 10 μM. In contrast, L-363 and MM.1S cells showed no sensitivity towards RBC8 treatment with concentrations up to 20 μM. (B) Effect of RBC8 treatment on RAL activation status was tested in INA-6 and MM.1S cells. INA-6 cells were treated with 10 μM and 20 μM of RBC8 for 3 h, MM.1S cell were treated with 20 μM of RBC8. RAL activation assays were performed subsequently. RALA total load served as loading control in addition to a-tubulin. RAL-GTP levels were not influenced by treatment with RBC8 in MM.1S. In INA-6 20 µM of RBC8 reduced the amount of RAL-GTP compared to DMSO-treated cells, while RAL-GTP levels of cells treated with 10 μM of RBC remained unchanged. (C) Combined blockade of RAL and PI3K/Akt or MEK/MAPK signaling. MM cell lines (n=4) were treated for 72 h with 10 μM of RBC8, 1 μM of PD0325901, 10 µM (AMO-1, INA-6, L-363) or 2.5 µM (MM.1S) of BYL-719, 10 μM (AMO-1, L-363) or 5 μM (INA-6, MM.1S) of Akti-1,2 and the combination of RBC8 with one of the other drugs. In AMO-1 and INA-6 combination of RAL-blockade by RBC8 with blockade of MEK, PI3K or Akt1,2 by PD0325901, BYL-719 or Akti-1,2, respectively, led to a significant reduction in cell survival. MM.1S and L-363 showed no stronger decrease in cell survival after combination of RAL blockade with MEK, PI3K or Akt. Cell viability was measured with annexin V/PI staining. Bar charts show mean values and standard deviation (SD) from three independent experiments. Percentages were calculated relative to DMSO-treated control.

haematologica | 2020; 105(9)

2321

M. Seibold et al.

Figure 5. Combined RAL-knockdown and blockade of PI3K/Akt or MEK/MAPK signaling. L-363 cells were transfected with shRNA-expression vectors against RALA or RALB, purified next day by selection for strongly transfected cells, and then treated with PD0325901 (1 µM), MK2206 (1 µM) or BYL-719 (2.5 μM). Cell survival was measured by annexin V/PI staining after 2 days (= day 3 post-transfection). Combination of RALA knockdown with PD0325901, BYL-719 or MK-2206 treatment did not further enhance the already strong apoptosis-induction resulting from RALA depletion alone. RALB knockdown in combination with PD0325901 showed only slight additional apoptosis-induction, whereas in combination with BYL-719 or MK-2206 treatment, the rate of cell death was strongly enhanced and matched that achieved by RALA knockdown.

RNAi-mediated RAL knockdown induces cell death in MM cells To assess whether RAL proteins contribute to MM cell survival we used an RNAi-mediated knockdown approach in HMCL with subsequent cell death assays and Western analysis. Two different target sequences against each of the respective isoforms, RALA and RALB, were cloned into pSUPER-type shRNA expression vectors. Cell survival was quantified 3 and 4 days after transfection by flow cytometry and assessment of annexin V-FITC-negative/ PI-negative events. In 4 of 7 cell lines tested, and with both shRNA-constructs, RALA depletion yielded stronger cell death effects than knockdown of RALB (Figure 2A-B and Online Supplementary Figure S2). Specifically, for cell line L363, viability decreased to below 40% (day 3 post-transfection) and to below 30% (day 4 post-transfection) of control cells after RALA knockdown (Figure 2A). Knockdown of RALB, too, led to significantly decreased viability, albeit to a lesser extent (63-71% at day 3 post-transfection, 39-59% at day 4 post-transfection (Figure 2A). Similarly, in MM.1S cells, knockdown of RALA led to significantly reduced cell survival to 57-64% at day 3 and to 32-42% at day 4 post-transfection. Knockdown of RALB significantly induced apoptosis leading to cell survival rates of 69-87% at 3 days and 52-79% at 4 days after electroporation (Figure 2B). RAL knockdown also led to cell death in other MM cell lines tested (INA-6, KMS-11, KMS-12-BM, and U-266), whereas AMO-1 cells remained largely unaffected by RAL depletion (Online Supplementary Figure S2 and Online Supplementary Table S3). Of note, concomitant knockdown of both RALA and RALB (tested in cell lines MM.1S, L-363, and INA-6) resulted in rapid and near complete cell death, precluding further functional analyses (data not shown). Effects on cell metabolism and cell cycle distribution 2322

were less pronounced than the induction of apoptosis described above. The Alamar Blue mitochondrial activity assay showed a significant decrease to 64% in L-363 cells after RALA knockdown, but only minor effects were found for MM.1S cells (Online Supplementary Figure S3A). Likewise, RALA knockdown in L-363 cells led to a significant increase of the G2/M-phase from 16% to 27% after 2 days at the expense of the S-phase (decreased from 36% to 21%). After 3 days, similar effects were observed for both RALA and RALB knockdown (G2/M-phase: 20% > 24% or 25%, respectively; S-phase: 34% > 24% or 22%, respectively). For MM.1S cells, the most notable change occurred after 3 days, at which time point the share of cells in S-phase had decreased from 20 % to 14% after RALA knockdown, and to 10% after RALB knockdown (Online Supplementary Figure S3B).

Targeting of RAL does not affect activity of the MEK/MAPK pathway but RALA appears to sustain AKT activity To investigate whether cell death induction after RAL knockdown is linked to down-regulation of the classical RAS downstream apoptosis and proliferation pathways, we analyzed the phosphorylation levels of ERK1/2 (MEK/MAPK pathway) and of Akt and GSK-3 (PI3K/Akt pathway) in L-363 and MM.1S cells after knockdown of either RALA or RALB by Western blotting. Cells were harvested at day 2 after transfection, i.e. before the onset of significant amounts of cell death, and at day 3, at which time-point care was taken to perform sample collection such that equivalent numbers of trypan-blue negative cells were collected for control and RAL-knockdown samples. RALA or RALB depletion had no discernible effect on the phosphorylation levels of any of the above-mentioned signaling intermediates at day 2 (Figures 2C-D), whereas RALA knockdown specifically led to lower levels of phoshaematologica | 2020; 105(9)

RAL GTPases mediate multiple myeloma cell survival

pho-Akt (of both, the Thr308 and Ser473 phosphorylation sites) at day 3 (Figures 2E-F). This effect was quite pronounced in both of these cell lines which display relatively high constitutive levels of activated Akt.

Constitutive RAL activation in MM cells remains unaffected by knockdown of oncogenic RAS To address our hypothesis that RAL activation is dependent on oncogenic RAS, we analyzed the change of the levels of activated GTP-bound RAL after knockdown of oncogenic KRAS or NRAS in HMCL harboring the respective mutated RAS isoform. Effective silencing of oncogenic RAS was verified by Western analysis 48 hours (h) after electroporation with the respective shRNA expression vectors. RAL-GTP levels were measured by performing RALA pulldown using RALBP1 protein-binding domain agarose beads. Notably, the RAL activation status was not altered by knockdown of oncogenic KRAS in MM.1S cells or oncogenic NRAS in INA-6 cells (Figure 3A). Expression levels of total load of RALA proteins remained also unchanged.

RNAi-mediated knockdown of KRAS or RALA in MM.1S cells entails distinct transcriptomic effects Because mutated RAS did not appear to be directly linked to RAL activation, we next analyzed in more detail the influence of both signaling hubs on the transcriptome of KRAS-mutated MM.1S cells using an RNA sequencing technique for 28,440 gene transcripts. Cells were transfected with shRNA expression vectors against either KRAS or RALA, or with the pSU empty-vector. Strongly transfected cells were purified via CD4Δ microbead selection and harvested at day 2 post-transfection for preparation of samples for transcriptomic analysis and Western blotting to confirm successful target knockdown (Figure 3B). As displayed in the Venn diagram (Figure 3C), KRAS knockdown led to changes in gene transcription in about double the number of genes (n=1,473) than RALA knockdown (n=771). Taken together, the number of transcripts that is altered in a mutually exclusive fashion (n=1,744) far outweighs the number affected by both, KRAS- or RALAknockdown (n=235). Using the Molecular Signatures Database Hallmark Gene Set Collection,40 ontology gene mapping was performed for the classification of differential gene expression after RAL versus RAS knockdown, highlighting the most distinct functional gene groups with relevance for MM biology (Figure 3D).

Effects of the small molecule compound RBC8 on survival and RAL activation of MM cells The small molecule inhibitor RBC8 has recently been described as selective allosteric inhibitor of RALA and RALB, which stabilizes RAL in its inactive GDP-bound state.41 We treated MM cell lines (n=4) and primary MM cells (n=6) for 3 days with different concentrations of RBC8 and measured cell survival by flow cytometry using annexin V/PI staining. INA-6 cells were most sensitive to the drug with EC50/90 values of 12,5 and 17,5 μM, respectively. MM.1S cells, on the other hand, were unaffected by RCB-8 even at the highest concentration tested (20 μM) (Figure 4A). In accordance with these results, analysis of RALA activation by pulldown of RALA-GTP in the sensitive INA-6 cells showed a strong reduction after treatment with 20 μM of RBC8 for 3 h, whereas at 10 μM no marked effects were seen. Conversely, the levels of activated haematologica | 2020; 105(9)

RALA remained unaltered after 3 h-treatment with 20 μM of RBC8 in MM.1S cells (Figure 4B). Primary MM samples remained largely unaffected by RBC8 (20 μM) (Online Supplementary Figure S5A). Data from combined RAL (RBC8) and MEK/MAPK or PI3K/Akt inhibition showed increased anti-myeloma effects in RBC8-sensitive cell lines, but no combination advantage in RBC8-insensitive cells (Figure 4C). These drug combinations showed at best mild effects on cell survival in the primary MM samples tested (Online Supplementary Figure S5B). We also performed Alamar Blue assays to test possible effects of RBC8 treatment on cell metabolism and proliferation. We found only minor impacts at concentrations up to 20 µM, most pronounced in INA-6 and AMO1 cells (Online Supplementary Figure S3C). Although higher concentrations (up to 40 μM) of RBC8 enhanced these effects, these concentrations may also exert unspecific cytotoxicity.42 Cell cycle distribution after RBC8 treatment was analyzed by BrdU/PI staining and revealed that in INA-6 cells which are most sensitive to treatment with RBC8, the G2/ M-phase significantly increased from 19% to 30 % at the expense of the S-phase, which decreased from 38% to 20% (Online Supplementary Figure S3D). No relevant changes were observed in AMO-1, L-363, or MM.1S cells. Finally, no effects on the constitutive levels of phospho-ERK and phospho-Akt were observed after RBC8 treatment of HCML (Online Supplementary Figure S4).

Combination of RALB depletion with PI3K/Akt inhibitors leads to enhanced MM cell death Because RAL blockade had either no or differential effects on the activity of the MEK/MAPK or PI3K/Akt pathways (see above and Figures 2C-F), we tested the potential usefulness of pharmacologically targeting these pathways in combination with RAL knockdown. After knockdown of either RALA or RALB, L-363 cells were treated with pharmacological inhibitors of MEK1/2 (PD0325901), Akt (MK2206) or PIK3CA (BYL-719). While combinations of either of these compounds with RALA abrogation did not significantly enhance apoptosis induction in excess of the rather strong effects of RALA knockdown alone, combined depletion of RALB and Akt or PI3K inhibition, respectively, led to significantly higher rates of cell death (Figure 5). Combination experiments in MM cell lines using the pharmacological RAL inhibitor RBC8 showed statistically significant (and functionally relevant) synergistic effects for the combination with PI3K/Akt primarily in the aforementioned cell line INA-6 (Figure 4C).

Mass spectrometric analysis identifies the exocyst complex as a predominant RAL interaction partner To analyze potential downstream signaling partners of RAL in MM cells, we performed quantitative mass spectrometry of MM.1S cells with stable expression of HAtagged RALA protein. A total of 48 proteins were identified as specific partners of RALA, including six members of the exocyst complex (EXOC-1, 2, 3, 4, 7, 8) (Online Supplementary Figure S6). Except for EXOC-7, all of these exocyst components are listed as highly confident interaction partners in the HitPredict database for protein-protein interactions.43 Moreover, EXOC-2/Sec5 and EXOC8/Exo84 have previously been described to play a role in RAL-mediated tumor cell proliferation.44 2323

M. Seibold et al.

Discussion In this study, we demonstrate the functional importance of the small GTPase RAL for survival of MM cells. Both RAL isoforms were strongly expressed in the majority of HMCL and primary MM cells when directly compared to normal plasma cells or pre-malignant MGUS cells. Moreover, GTP-bound and thus activated RAL was present in all MM cell lines analyzed, pointing to a potential functional role of RAL in transition to and/ or maintenance of the malignant tumor clone. To test this hypothesis we performed isoform-specific RNAi-mediated RAL knockdown and found that abrogation of RAL led to fast and strong cell death induction in the majority of MM cell lines. These experiments thus identified the RAL GTPases as potent pro-survival mediators in MM. Because activation of RAL has been described as a predominantly RAS-dependent oncogenic survival pathway in various cancer entities,29,45–48 we also sought to test the putative functional link between oncogenic RAS and RAL activation in MM using pulldown assays and RNA sequencing. We found that in the MM cells tested, RAL activation could not be ascribed to the presence of oncogenic RAS (as defined by harboring activating point mutations in NRAS or KRAS). Neither did shRNA-mediated knockdown of oncogenic RAS alter the RAL activation status. As opposed to the well-defined activating mutations of NRAS and KRAS, data available from the CoMMpass trial cohort (these data were generated as part of the Multiple Myeloma Research Foundation Personalized Medicine Initiatives [https://research.themmrf.org and www.themmrf.org]) and other large next generation sequencing studies49–51 revealed no oncogenic bona fide mutations affecting RAL in MM. Activation of RAL by RAS-independent mechanisms has also been shown by other groups in solid tumors such as melanoma,52 bladder carcinoma53 and malignant peripheral nerve sheath tumors54 in which deregulated RAL guanine exchange factors, direct RAL phosphorylation by protein kinase C, or involvement of secondary GTPases, respectively, may lead to increased RAL-dependent tumor cell proliferation. De Gorter et al. showed that RAL could be activated by chemokines originating from the bone marrow microenvironment. In their study, treatment of MM cells with stromal cell-derived factor-1 resulted in increased levels of GTP-bound RAL and led to enhanced cell migration.55 These effects occurred independently of RAS, which is in line with our observation that no direct link between oncogenic RAS and activated RAL could be established. Additionally, in our transcriptome analysis we observed distinct changes of gene expression after RAL versus RAS knockdown, underpinning the notion that RAL functions as a survival pathway in its own right and warrants further validation for potential therapeutic intervention. Due to the high affinity of the guanine nucleotides at their binding sites, small GTPases such as RAS and RAL are hard to target pharmacologically, however. Whereas to date, no clinically suitable RAS inhibitors are available,1,4,56 a small molecule RAL inhibitor has recently been described,41 showing in vitro effects in adipose tissue57 and in chronic myelogenous leukemia.58 This allosteric compound has been developed to stabilize RAL in its inactive GDP-bound state and thus prevent its activation.41 In our hands, in the most sensitive MM cell line INA-6, RAL acti2324

vation was indeed abrogated and apoptosis induced at drug concentrations starting from 10 μM, whereas survival of primary MM cells and of other HMCL was less affected even at 20 μM, warranting development of more potent second generation RAL inhibitors. To this end, Walsh et al.42 have recently observed in a murine platelet RAL knockout model that RBC8 does indeed exert specific as well as unspecific effects within similar concentration ranges, which may explain its inconsistency when tested across different (cell line) models. In our mass-spectrometric analysis which we performed to define RAL interaction partners serving as potential downstream mediators of the RAL pathway, we identified six members of the exocyst complex among the highest scoring hits. They included the complex members EXOC2/Sec5 and EXOC-8/Exo84 which are known to contribute to RAL-induced proliferation in tumor cells.44 Interestingly, RALBP1 which is another well-defined binding partner to RAL, did not appear to play a predominant role in our MM cell line analysis. In MM, the interconnection with signals from a per se altered bone marrow microenvironment59–61 may bypass otherwise important signaling hubs such as RAS. We have previously made this observation for the PI3K/Akt pathway in MM, which can be constitutively activated independently of oncogenic RAS, possibly by involvement of upstream receptor tyrosine kinases.12,13,51 Whereas we found no indication for RAL involvement in RAS/MAPK signaling, we did find distinctly lower levels of activated Akt after extended knockdown of RALA. In keeping with this observation, while the already strong apoptotic effects of RALA depletion could not further be enhanced by simultaneous pharmacological PI3K/Akt blockade, such treatment considerably enhanced the cytotoxic effects of RALB knockdown. These observations suggest that both RAL isoforms may at least in parts play differential roles in cellular signaling, and point specifically to a role for RALA as one of the potential mediators for high intrinsic levels of active Akt in a subgroup of MM cells. Given the heterogeneity of oncogenic pathways in MM, synergistically acting combination therapies seem to be the most promising targeted treatment strategies. To this extent, our data demonstrate that RAL abrogation may be effective in combination with inhibitors of the PI3K/Akt pathway. This is particularly important because in early clinical trials, PI3K inhibitors displayed limited effectivity and will therefore most likely play a role as combination partners in tumor therapy.18,62 Taken together, our data indicate that RAL is a promising molecular target for MM therapy that is functionally independent of oncogenic RAS. However, because the one existing pharmacological inhibitor targeting RAL in our hands does not perfectly mimmick the strong effects of RAL knockdown, development of more potent secondgeneration inhibitors for MM treatment is mandatory for clinical translation. Funding This work was supported by grants from the Deutsche Forschungsgemeinschaft (KFO 216) and the Interdisziplinäres Zentrum für Klinische Forschung of the Universitätsklinikum Würzburg (B-188). TSte was supported by a fellowship of the Else Kröner Fresenius-Stiftung (2010_Kolleg.52). EL and RB were supported by a grant of the Deutsche Krebshilfe (70112693). haematologica | 2020; 105(9)

RAL GTPases mediate multiple myeloma cell survival

References 1. Cox AD, Fesik SW, Kimmelman AC, Luo J, Der CJ. Drugging the undruggable RAS: Mission Possible? Nat Rev Drug Discov. 2014;13(11):828-851. 2. Stephen AG, Esposito D, Bagni RK, McCormick F. Dragging Ras back in the ring. Cancer Cell. 2014;25(3):272-281. 3. Athuluri-Divakar SK, Vasquez-Del Carpio R, Dutta K, et al. A small molecule RASmimetic disrupts RAS association with effector proteins to block signaling. Cell. 2016;165(3):643-655. 4. McCormick F. K-Ras protein as a drug target. J Mol Med. 2016;94(3):253-258. 5. Beeram M, Patnaik A, Rowinsky EK. Raf: A strategic target for therapeutic development against cancer. J Clin Oncol. 2005; 23(27):6771-6790. 6. Baines AT, Xu D, Der CJ. Inhibition of Ras for cancer treatment: the search continues. Future Med Chem. 2011;3(14):1787-1808. 7. Mandal R, Becker S, Strebhardt K. Stamping out RAF and MEK1/2 to inhibit the ERK1/2 pathway: an emerging threat to anticancer therapy. Oncogene. 2016; 35(20):2547-2561. 8. Steinbrunn T, Stühmer T, Gattenlöhner S, et al. Mutated RAS and constitutively activated Akt delineate distinct oncogenic pathways, which independently contribute to multiple myeloma cell survival. Blood. 2011;117(6):1998-2004. 9. Lionetti M, Barbieri M, Todoerti K, et al. Molecular spectrum of BRAF, NRAS and KRAS gene mutations in plasma cell dyscrasias: implication for MEK-ERK pathway activation. Oncotarget. 2015; 6(27):24205-24217. 10. Walker BA, Boyle EM, Wardell CP, et al. Mutational spectrum, copy number changes, and outcome: results of a sequencing study of patients with newly diagnosed myeloma. J Clin Oncol. 2015;33(33):39113920. 11. Steinbrunn T, Stühmer T, Sayehli C, et al. Combined targeting of MEK/MAPK and PI3K/Akt signalling in multiple myeloma. Br J Haematol. 2012;159(4):430-440. 12. Hofmann C, Stühmer T, Schmiedl N, et al. PI3K-dependent multiple myeloma cell survival is mediated by the PIK3CA isoform. Br J Haematol. 2014;166(4):529-539. 13. Munugalavadla V, Mariathasan S, Slaga D, et al. The PI3K inhibitor GDC-0941 combines with existing clinical regimens for superior activity in multiple myeloma. Oncogene. 2014;33(3):316-325. 14. Müller E, Bauer S, Stühmer T, et al. Pan-Raf co-operates with PI3K-dependent signalling and critically contributes to myeloma cell survival independently of mutated RAS. Leukemia. 2017;31(4):922-933. 15. Xu J, Pfarr N, Endris V, et al. Molecular signaling in multiple myeloma: association of RAS/RAF mutations and MEK/ERK pathway activation. Oncogenesis. 2017; 6(5):e337. 16. McMillin DW, Ooi M, Delmore J, et al. Antimyeloma activity of the orally bioavailable dual phosphatidylinositol 3kinase/mammalian target of rapamycin inhibitor NVP-BEZ235. Cancer Res. 2009; 69(14):5835-5842. 17. Mayer IA, Arteaga CL. The PI3K/AKT pathway as a target for cancer treatment. Annu Rev Med. 2016;67(1):11-28. 18. Ramakrishnan V, Kumar S. PI3K/AKT/mTOR pathway in multiple myeloma: from basic biology to clinical

haematologica | 2020; 105(9)

19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

29. 30.

31.

32. 33.

34.

35.

promise. Leuk Lymphoma. 2018; 59(11):2524-2334. Breitkreutz I, Podar K, Figueroa-Vazquez V, et al. The orally available multikinase inhibitor regorafenib (BAY 73-4506) in multiple myeloma. Ann Hematol. 2018; 97(5):839-849. Chang-Yew Leow C, Gerondakis S, Spencer A. MEK inhibitors as a chemotherapeutic intervention in multiple myeloma. Blood Cancer J. 2013;3(3):e105. Holkova B, Zingone A, Kmieciak M, et al. A phase II trial of AZD6244 (Selumetinib, ARRY-142886), an oral MEK1/2 inhibitor, in relapsed/refractory multiple myeloma. Clin Cancer Res. 2016;22(5):1067-1075. Ramakrishnan V, D’Souza A. Signaling pathways and emerging therapies in multiple myeloma. Curr Hematol Malig Rep. 2016;11(2):156-164. Bedard PL, Tabernero J, Janku F, et al. A phase Ib dose-escalation study of the oral Pan-PI3K inhibitor buparlisib (BKM120) in combination with the oral MEK1/2 inhibitor trametinib (GSK1120212) in patients with selected advanced solid tumors. Clin Cancer Res. 2015;21(4):730738. Tolcher AW, Patnaik A, Papadopoulos KP, et al. Phase I study of the MEK inhibitor trametinib in combination with the AKT inhibitor afuresertib in patients with solid tumors and multiple myeloma. Cancer Chemother Pharmacol. 2015;75(1):183189. Grilley-Olson JE, Bedard PL, Fasolo A, et al. A phase Ib dose-escalation study of the MEK inhibitor trametinib in combination with the PI3K/mTOR inhibitor GSK2126458 in patients with advanced solid tumors. Invest New Drugs. 2016; 34(6):740-749. Wainberg ZA, Alsina M, Soares HP, et al. A multi-arm phase I study of the PI3K/mTOR inhibitors PF-04691502 and gedatolisib (PF05212384) plus irinotecan or the MEK inhibitor PD-0325901 in advanced cancer. Target Oncol. 2017;12(6):775-785. Rodriguez-Viciana P, McCormick F. RalGDS comes of age. Cancer Cell. 2005; 7(3):205-206. Thomas JC, Cooper JM, Clayton NS, et al. Inhibition of Ral GTPases using a stapled peptide approach. J Biol Chem. 2016; 291(35):18310-18325. Yan C, Theodorescu D. RAL GTPases: biology and potential as therapeutic targets in cancer. Pharmacol Rev. 2018;70(1):1-11. Rangarajan A, Hong SJ, Gifford A, Weinberg RA. Species- and cell type-specific requirements for cellular transformation. Cancer Cell. 2004;6(2):171-183. Lim K-H, O’Hayer K, Adam SJ, et al. Divergent roles for RalA and RalB in malignant growth of human pancreatic carcinoma cells. Curr Biol. 2006;16(24): 2385-2394. Martin TD, Der CJ. Differential involvement of RalA and RalB in colorectal cancer. Small GTPases. 2012;3(2):126-130. Stühmer T, Arts J, Chatterjee M, et al. Preclinical anti-myeloma activity of the novel HDAC-inhibitor JNJ-26481585. Br J Haematol. 2010;149(4):529-536. Steinbrunn T, Chatterjee M, Bargou RC, Stühmer T. Efficient transient transfection of human multiple myeloma cells by eectroporation - an appraisal. PLoS One. 2014; 9(6):e97443. Brummelkamp TR, Bernards R, Agami R. A system for stable expression of short inter-

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52. 53.

54.

fering RNAs in mammalian cells. Science. 2002;296(5567):550-553. Lim K-H, Baines AT, Fiordalisi JJ, et al. Activation of RalA is critical for Rasinduced tumorigenesis of human cells. Cancer Cell. 2005;7(6):533-545. Oxford G, Owens CR, Titus BJ, et al. RalA and RalB: rntagonistic Relatives in cancer cell migration. Cancer Res. 2005; 65(16):7111-7120. Cox J, Hein MY, Luber CA, et al. Accurate Proteome-wide Label-free Quantification by Delayed Normalization and Maximal Peptide Ratio Extraction, Termed MaxLFQ. Mol Cell Proteomics. 2014;13(9):25132526. Cox J, Mann M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat Biotechnol. 2008;26(12):1367-1372. Liberzon A, Birger C, Thorvaldsdóttir H, et al. The molecular signatures database hallmark gene set collection. Cell Syst. 2015; 1(6):417-425. Yan C, Liu D, Li L, et al. Discovery and characterization of small molecules that target the GTPase Ral. Nature. 2014;515(7527):443-447. Walsh TG, Wersäll A, Poole AW. Characterisation of the Ral GTPase inhibitor RBC8 in human and mouse platelets. Cell Signal. 2019;59:34-40. López Y, Nakai K, Patil A. HitPredict version 4: comprehensive reliability scoring of physical protein–protein interactions from more than 100 species. Database (Oxford). 2015;2015. Shin H, Kaplan REW, Duong T, Fakieh R, Reiner DJ. Ral signals through a MAP4 kinase-p38 MAP kinase cascade in C. elegans cell fate patterning. Cell Rep. 2018; 24(10):2669-2681. González-García A, Pritchard CA, Paterson HF, et al. RalGDS is required for tumor formation in a model of skin carcinogenesis. Cancer Cell. 2005;7(3):219-226. Mishra PJ, Ha L, Rieker J, et al. Dissection of RAS downstream pathways in melanomagenesis: a role for Ral in transformation. Oncogene. 2010;29(16):2449-2456. Yin J, Pollock C, Tracy K, et al. Activation of the RalGEF/Ral pathway promotes prostate cancer metastasis to bone. Mol Cell Biol. 2007;27(21):7538-7550. Guin S, Theodorescu D. The RAS-RAL axis in cancer: evidence for mutation-specific selectivity in non-small cell lung cancer. Acta Pharmacol Sin. 2015;36(3):291-297. Chapman MA, Lawrence MS, Keats JJ, et al. Initial genome sequencing and analysis of multiple myeloma. Nature. 2011; 471(7339):467-472. Lohr JG, Stojanov P, Carter SL, et al. Widespread genetic heterogeneity in multiple myeloma: implications for targeted therapy. Cancer Cell. 2014;25(1):91-101. Leich E, Weißbach S, Klein H-U, et al. Multiple myeloma is affected by multiple and heterogeneous somatic mutations in adhesion- and receptor tyrosine kinase signaling molecules. Blood Cancer J. 2013; 3(2):e102-e102. Zipfel PA, Brady DC, Kashatus DF, et al. Ral activation promotes melanomagenesis. Oncogene. 2010;29(34):4859-4864. Wang H, Owens C, Chandra N, et al. Phosphorylation of RalB is important for bladder cancer cell growth and metastasis. Cancer Res. 2010;70(21):8760-8769. Bodempudi V, Yamoutpoor F, Pan W, et al.

2325

M. Seibold et al. Ral overactivation in malignant peripheral nerve sheath tumors. Mol Cell Biol. 2009; 29(14):3964-3974. 55. De Gorter DJJ, Reijmers RM, Beuling EA, et al. The small GTPase Ral mediates SDF-1 induced migration of B cells and multiple myeloma cells. Blood. 2008;111(7):33643372. 56. Wilson CY, Tolias P. Recent advances in cancer drug discovery targeting RAS. Drug Discov Today. 2016;21(12):1915-1919. 57. Skorobogatko Y, Dragan M, Cordon C, et al. RalA controls glucose homeostasis by

2326

regulating glucose uptake in brown fat. Proc Natl Acad Sci. 2018;115(30):78197824. 58. Gu C, Feng M, Yin Z, Luo X, Yang J. RalA, a GTPase targeted by miR 181a, promotes transformation and progression by activating the Ras related signaling pathway in chronic myelogenous leukemia. Oncotarget. 2016;7(15):20561-20573. 59. Schinke C, Qu P, Mehdi SJ, et al. The pattern of mesenchymal stem cell expression is an independent marker of outcome in multiple myeloma. Clin Cancer Res. 2018;

24(12):2913-2919. 60. Garayoa M, Garcia JL, Santamaria C, et al. Mesenchymal stem cells from multiple myeloma patients display distinct genomic profile as compared with those from normal donors. Leukemia. 2009;23(8):15151527. 61. Bianchi G, Munshi NC. Pathogenesis beyond the cancer clone(s) in multiple myeloma. Blood. 2015;125(20):3049-3058. 62. Markman B, J. Tao J, Scaltriti M. PI3K pathway inhibitors: better not left alone. Curr Pharm Des. 2013;19(5):895-906.

haematologica | 2020; 105(9)

ARTICLE

Hemostasis

Thrombin generation in cardiovascular disease and mortality – results from the Gutenberg Health Study Pauline C.S. van Paridon,1,2* Marina Panova-Noeva,2* Rene van Oerle,1 Andreas Schulz,3 Iris M. Hermanns,3,4 Jürgen H. Prochaska,2,3,5 Natalie Arnold,3 Harald Binder,6,7 Irene Schmidtmann,7 Manfred E. Beutel,8 Norbert Pfeiffer,9 Thomas Münzel,5,10 Karl J. Lackner,10,11 Hugo ten Cate,1,2 Philipp S. Wild2,3,10# and Henri M.H. Spronk1#

Laboratory for Clinical Thrombosis and Hemostasis, Department of Internal Medicine, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University Medical Center, Maastricht, the Netherlands; 2Center for Thrombosis and Hemostasis (CTH), University Medical Center of the Johannes Gutenberg-University Mainz, Germany; 3 Preventive Cardiology and Preventive Medicine, Center for Cardiology, University Medical Center of the Johannes Gutenberg-University Mainz, Germany; 4University of Applied Sciences, Hochschule Fresenius, Idstein, Germany; 5Cardiology I, Center for Cardiology, University Medical Center of the Johannes Gutenberg-University Mainz, Germany; 6Institute of Medical Biometry and Statistics, Faculty of Medicine and Medical Center, University of Freiburg, Freiburg, Germany; 7Institute of Medical Biostatistics, Epidemiology and Informatics, University Medical Center of the Johannes GutenbergUniversity Mainz, Germany; 8Department of Psychosomatic Medicine and Psychotherapy, University Medical Center of the Johannes Gutenberg-University Mainz, Germany; 9Department of Ophthalmology, University Medical Center of the Johannes Gutenberg-University Mainz, Germany; 10DZHK (German Center for Cardiovascular Research), Partner Site RhineMain, Mainz, Germany and 11Institute for Clinical Chemistry and Laboratory Medicine, University Medical Center of the Johannes Gutenberg-University Mainz, Germany 1

* #

Ferrata Storti Foundation

Haematologica 2020 Volume 105(9):2327-2334

PCSvP and MP-N contributed equally as co-first authors.

PSW and HMHS contributed equally as co-senior authors.

ABSTRACT

hrombin generation may be a potential tool to improve risk stratification for cardiovascular diseases. The aim of this study was to explore the relation between thrombin generation and cardiovascular risk factors, cardiovascular diseases, and total mortality. For this study, 5,000 subjects from the population-based Gutenberg Health Study were analyzed in a highly standardized setting. Thrombin generation was assessed by the Calibrated Automated Thrombogram method at 1 and 5 pM tissue factor triggers in platelet-poor plasma. Lag time, endogenous thrombin potential, and peak height were derived from the thrombin generation curve. Sex-specific multivariable linear regression analysis adjusted for age, cardiovascular risk factors, cardiovascular diseases and therapy, was used to assess clinical determinants of thrombin generation. Cox regression models adjusted for age, sex, cardiovascular risk factors and vitamin K antagonists investigated the association between thrombin generation parameters and total mortality. Lag time was positively associated with obesity and dyslipidaemia for both sexes (P<0.0001). Obesity was also positively associated with endogenous thrombin potential in both sexes (P<0.0001) and peak height in males (1 pM tissue factor, P=0.0048) and females (P<0.0001). Cox regression models showed an increased mortality in individuals with lag time (1 pM tissue factor, hazard ratio=1.46, 95% confidence interval: 1.07-2.00; P=0.018) and endogenous thrombin potential (5 pM tissue factor, hazard ratio=1.50, 95% confidence interval: 1.06-2.13; P=0.023) above the 95th percentile of the reference group, independently of the cardiovascular risk profile. haematologica | 2020; 105(9)

Correspondence: HENRI M.H. SPRONK henri.spronk@maastrichtuniversity.nl Received: March 13, 2019. Accepted: December 3, 2019. Pre-published: December 5, 2019. doi:10.3324/haematol.2019.221655 ©2020 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.

2327

P.C.S. van Paridon et al.

This large-scale study demonstrates that traditional cardiovascular risk factors, particularly obesity, are relevant determinants of thrombin generation. Lag time and endogenous thrombin potential were found to be potentially relevant predictors of increased total mortality, observations which deserve further investigation.

Introduction Thrombin generation (TG) is one of the key enzymatic processes that direct the activity of the hemostatic system and coagulation cascade up to and including the formation of a fibrin clot.1 Physiologically, thrombin formation is essential to maintain hemostasis and bleeding tendencies are associated with reduced thrombin (and hence fibrin) formation.2 An enhanced plasma potential to generate thrombin has been linked to an increased risk of venous thromboembolism, while the associations with arterial vascular disease are still inconsistent.3-8 The TG assay is an important method addressing the overall potential of a plasma sample to form thrombin. More than 95% of TG occurs after initial formation of fibrin, so routine diagnostic coagulation tests, such as prothrombin time and activated partial thromboplastin time fail to reproduce this overall potential. Hence, there is a strong research interest in TG as a promising diagnostic tool for hypo- and particularly hyper-coagulability phenotypes.9 In a study of healthy individuals, fibrinogen, factor XII, free tissue factor pathway inhibitor (TFPI) and antithrombin have been identified as major determinants of TG parameters.10 Relation to demographic characteristics, such as age and sex, has been previously addressed but the studies have been small and results are not entirely consistent.11,12 As TG analysis is a promising tool to estimate a subject’s risk for thrombosis or more broadly cardiovascular diseases (CVD), it is of eminent importance to fully understand the nature and direction of effects of cardiovascular risk factors (CVRF). Hence, we undertook the present investigation in the first 5,000 participants of the population-based Gutenberg Health Study. The primary aim of this study was to investigate CVRF and CVD as major clinical determinants of increased TG in a large population-based sample. Additional aims were to obtain age- and sex-related reference values for TG parameters in a representative subsample of adults who were healthy from a cardiovascular point of view. Finally, having prospective data on total mortality allowed us to investigate the relation between TG parameters and all-cause mortality.

Methods Research design The Gutenberg Health Study, a population-based, prospective, observational, single-center cohort study in the Rhine-Main region in Western Mid-Germany, was designed to improve the individual risk prediction of CVD. At baseline examination, the study included a total of 15,010 individuals. A detailed description of the research design is provided in Online Supplementary Material, Part A. Further details of the study protocol and purpose are discussed elsewhere.13 The study was designed in accordance with the tenets of the revised Helsinki protocol, and the protocol and sampling design 2328

were approved by the local ethics committee. The sampling design was additionally approved by local and state data safety commissioners.

Study sample and reference sample The study sample consisted of the first 5,000 subjects enrolled into the Gutenberg Health Study between April 2007 and October 2008. After excluding subjects without biomaterial available or without complete TG assessment (one or several TG parameters missing), 4,843 individuals were successfully included in the overall study sample for the present analysis. The reference group was defined as subjects apparently healthy from a cardiovascular point of view, without a history of CVD (myocardial infarction, congestive heart failure, coronary artery disease, venous thromboembolism, atrial fibrillation or peripheral artery disease), presence of CVRF (obesity, dyslipidemia, arterial hypertension, diabetes mellitus) or use of antithrombotic agents, oral contraceptives or hormonal replacement therapy. In addition, individuals with a self-reported history of inherited coagulation abnormalities were excluded from the reference sample. A detailed definition of traditional CVRF and categorization of medications are provided in Online Supplementary Material, Part A.

Clinical assessment and laboratory measurements Clinical examination and determination of CVRF were performed as published elsewhere.14,15 Standard laboratory measurements were carried out at the Institute of Clinical Chemistry and Laboratory Medicine, University Medical Center Mainz, Germany. Details on venous blood sampling and plasma preparation are provided in Online Supplementary Material, Part A. TG was assessed according to the recommendations16 for the Calibrated Automated Thrombogram (CAT) assay (Thrombinoscope BV, Maastricht, the Netherlands) in plateletpoor plasma. The TG was triggered by 1 pM tissue factor (TF) with 4 μM phospholipids at 20:20:60 mol% phosphatidylserine/phosphatidylethanolamine/phosphatidylcholine, or 5 pM TF with 4 μM phospholipids. Trigger reagents were selected for commercial availability, e.g. PPP Reagent and PPP Low Reagent. The CAT method employs a low affinity fluorogenic substrate for thrombin (Z-Gly-Gly-Arg-AMC) in order to monitor thrombin activity continuously in clotting plasma. TG measurements were calibrated against the fluorescence curve obtained in a sample from the same plasma (80 μL), supplemented with a fixed amount of thrombin-a 2-macroglobulin complex (20 μL of Thrombin Calibrator; Thrombinoscope BV, Maastricht, the Netherlands) and 20 μL of the fluorogenic substrate and calcium chloride mixture.16 TG parameters were derived from the TG curve and include lag time (time to minimum thrombin formed, in min), peak height (the maximum amount of thrombin formed, in nM) and endogenous thrombin potential (ETP or area under the curve, in nM.min). All samples were tested as one batch using one batch of reagents within a period of 6 months. Two technicians performed the analyses on three validated systems and normal pooled plasma was included in each assay run for in-house quality control according to our ISO9001 certification (Coagulation Profile BV, Maastricht, the Netherlands). haematologica | 2020; 105(9)

Clinical determinants of thrombin generation

Data management and statistical analysis are described in Online Supplementary Material, Part A.

Results Study sample and reference subsample characteristics While there was a balanced sex ratio in the overall study sample, there was a slight preponderance of women (55.3%) in the reference subsample. The median and interquartile range (IQR) of age of the study sample was 56 years (IQR, 46-66) in males and 55 years (IQR, 45-64) in females. The reference sample included 1,210 subjects, of whom 541 (44.7%) were male and 669 (55.3%) female. The median age of the reference sample was 47 years (IQR, 42-55) in males and 48 years (IQR, 41-54) in females. In the study population, hypertension was the most prevalent CVRF, being present in 56.6% of the male population and 46.1% of the female population, followed by dyslipidemia. Antithrombotic agents were taken by 15.9% of males and 9.4% of females. Among females, 6.4% were taking oral contraceptives and 12.3% hormone replacement therapy. Detailed characteristics of the study population and reference sample are presented in Online Supplementary Table S1.

Thrombin generation reference values and parameters in the overall study sample The results of the TG parameters in males and females from the study sample and reference subsample (reference values) are shown in Table 1. Females presented with a shorter lag time at 1 pM TF and 5 pM TF (P<0.0001 for both), higher ETP at 1 pM TF (P<0.0001) and higher peak height at 1 pM TF (P=0.014) and at 5 pM TF (P<0.0001) compared to males from the reference subsample. In the study sample, females presented with a shorter lag time at 1 pM TF (P<0.0001), as well as higher ETP and peak height at both 1 pM TF and 5 pM TF (P<0.0001 for both), compared to males.

Clinical determinants of thrombin generation in the overall study sample As shown in Table 2, age was associated with longer lag time, both in males (Table 2A) and females (Table 2B), at 1 pM TF (males: P=0.014; females: P<0.0001) and at 5 pM TF (Online Supplementary Table S2A, B). In males, age was pos-

itively associated with ETP at 1 pM TF (Table 2A) and peak height at 1 pM TF (Table 2A) and at 5 pM TF (Online Supplementary Table S2A). In contrast, in females, age was associated with lower ETP at 1 pM TF (P=0.015) and lower peak height at 5 pM TF (Online Supplementary Table S2B). Of the various CVRF considered, obesity showed a positive association with lag time (males: P<0.0001; females: P<0.0001), ETP (males: P<0.0001; females: P<0.0001 ) and peak height (males: P=0.0048; females: P<0.0001) at 1 pM TF. Dyslipidemia was positively associated with lag time in both males (P<0.0001) and females (P<0.0001) and with ETP in males only (P=0.0057) at 1 pM TF. Similar findings for both obesity and dyslipidemia were observed at 5 pM TF as shown in Online Supplementary Table S2A, B. No associations were found for TG and history of CVD.

Therapeutic agents and thrombin generation parameters in the overall study sample Females using oral contraceptives or hormone replacement therapy presented with shorter lag time, higher ETP and peak height at both 1 pM TF (Table 2B) and 5 pM TF (Online Supplementary Table S2B). Use of vitamin K antagonists reduced TG as shown by prolonged lag time, and reduced ETP and peak height (Table 2A, B). These effects were detectable at 1 pM TF in both males (P<0.0001; ETP: P<0.0001; peak height: P<0.0001) and females (P<0.0001; ETP: P<0.0001; peak height: P<0.0001;). Similar results were obtained at 5 pM TF (Online Supplementary Table S2A, B). Results from multiple linear regression analysis for TG parameters at 1 pM TF and at 5 pM TF, demonstrated associations between other medications, in addition to vitamin K antagonists, and TG as shown in Table 3A, B. The lag time at 1 pM TF was positively associated with intake of cardiac drugs (P<0.0001), diuretics (P=0.00043), anti-gout preparations (P=0.00038) and immunosuppressants (P=0.00021), and inversely associated with hormone-containing drugs (i.e., hormone replacement therapy and oral contraceptives, P<0.0001). Differently, ETP was inversely associated with cardiac drugs, ATC code C01 (P<0.0001) at 5 pM TF and positively with hormonecontaining drugs (P<0.0001). Peak height showed a positive association with hormone-containing drugs (P<0.0001). The results at 5 pM TF were comparable to the results at 1 pM TF.

Table 1. Parameters of thrombin generation in the reference subsample and the study sample.

Male (n=541) Lag time (1 pM TF), min ETP (1 pM TF), nM.min Peak height (1 pM TF), nM Lag time (5 pM TF), min ETP (5 pM TF), nM.min Peak height (5 pM TF), nM

5.07 1047 108 2.67 1322 236

Reference subsample (N=1,210) Female P-value (n=669)

(4.67;5.67) (216) (51) (2.33;3.00) (196) (52.2)

4.67 1099 115 2.39 1318 259

(4.33;5.33) (203) (48.7) (2.33;2.67) (212) (53.3)

<0.0001 <0.0001 0.014 <0.0001 0.71 <0.0001

Male (n=2,471) 5.33 1068 113 2.67* 1352 238

(4.74;6.07) (267) (51.9) (2.40;3.00) (267) (59.3)

Study sample (N=4,843) Females Females without OC with OC (n=2,218) (n=151) 5 1115 117 2.67* 1370 257

(4.40;5.67) 260 51 (2.33;3.00) (266) (61)

4 (3.40;4.33) 1491 (308) 201 (63) 2.06 (2;2.33) 1661 (350) 365 (71)

P-value

<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

The values presented are thrombin generation parameters at 1 pM and 5 pM tissue factor in the reference subsample and study sample. Medians (interquartile range) of lag time and means (standard deviation) of endogenous thrombin potential and peak height are presented. *Due to equal times, the median values in males and females are the same; however, the distribution of the lag time values is different in males and females. TF: tissue factor; ETP: endogenous thrombin potential.

haematologica | 2020; 105(9)

2329

P.C.S. van Paridon et al.

Kaplan-Meier survival curves and Cox-regression models During the follow-up period until April 2017, with a median follow-up of 9.21 (8.83-9.65) years, a total of 308 deaths were registered. As presented in Figure 1, a longer lag time above the 90th percentile of the reference, at both 1 pM TF (Figure 1A) and 5 pM TF (Figure 1B), was signif-

icantly associated with worse survival (P<0.0001). In addition, higher ETP above the 90th percentile (P=0.034) and 97.5th percentile (P=0.00097) of the reference sample (Figure 1C), measured at 5 pM TF was associated with worse survival. No such associations were observed for ETP at 1 pM TF.

Table 2A. Multivariable linear regression in the overall study sample for parameters of thrombin generation in males at 1 pM tissue factor.

Variable Age (10 years) Diabetes Obesity Smoking Hypertension Dyslipidemia FH of MI/stroke History of MI History of stroke History of CAD History of AF History of PAD History of VTE History of CHF History of cancer VKA use

Males (N)

242 628 515 1,395 917 853 116 57 167 93 107 71 39 194 391

0.0098 -0.0318 0.0392 0.0151 0.0135 0.0428 -0.00624 0.0263 0.0352 -0.0306 0.0418 0.0367 -0.0342 -0.00240 -0.000508 0.646

Log (lag time [min]) 95% CI P-value (0.00199;0.0175) (-0.0593;-0.00441) (0.0214;0.0570) (-0.00328;0.0334) (-0.00263;0.0297) (0.0275;0.0580) (-0.0217;0.00919) (-0.0193;0.0719) (-0.0169;0.0874) (-0.0674;0.00617) (-0.00183;0.0854) (-0.00267;0.0761) (-0.0794;0.0111) (-0.0652;0.0604) (-0.0284;0.0273) (0.0589;0.702)

0.014 0.023 <0.0001 0.11 0.10 <0.0001 0.43 0.26 0.19 0.10 0.061 0.068 0.14 0.94 0.11 <0.0001

β 28.9 -6.00 54.7 18.2 -6.70 30.6 17.1 -57.5 -39.7 25.7 -19.5 -53.0 -4.91 -15.1 0.764 -716

ETP (nM.min) 95% CI P-value (17.9;39.9) (-45.0;33.0) (29.5;80.0) (-7.82;44.3) (29.7;16.3) (8.96;52.3) (-4.84;39.0) (-122;7.32) (-114;34.4) (-26.6;77.9) (-81.4;42.4) (-109;2.95) (-69.2;59.4) (-104;74.0) (-38.8;40.3) (-796;-637)

<0.0001 0.76 <0.0001 0.17 0.57 0.0057 0.13 0.082 0.29 0.34 0.54 0.064 0.88 0.74 0.97 <0.0001

β 7.60 5.16 7.83 -1.64 -3.22 1.39 -0.283 -8.36 -2.67 1.21 -6.31 -10.6 -1.44 5.46 -2.11 -66.3

Peak height (nM) 95% CI P-value (5.23;9.97) <0.0001 (-3.21;13.5) 0.23 (2.40;13.3) 0.0048 (7.24;3.96) 0.57 (-8.15;1.72) 0.20 (-3.27;6.04) 0.56 (-4.99;4.42) 0.91 (-22.3;5.56) 0.24 (-18.6;13.2) 0.74 (-10.0;12.4) 0.83 (-19.6;6.99) 0.35 (-22.6;1.40) 0.083 (-15.3;12.4) 0.84 (-13.7;24.6) 0.58 (-10.6;6.39) 0.63 (-83.5;-49.2) <0.0001

Multivariable linear regression models were calculated in the overall study sample for each parameter of thrombin generation as a dependent variable separately. The analysis was adjusted for age, vitamin K antagonist use, presence of cardiovascular risk factors and cardiovascular diseases. ETP: endogenous thrombin potential; 95% CI: 95% confidence interval; FH: family history; MI: myocardial infarction; CAD: coronary artery disease; AF: atrial fibrillation; PAD: peripheral artery disease; VTE: venous thromboembolism; CHF: congestive heart failure; VKA: vitamin K antagonist.

Table 2B. Multivariable linear regression in the overall study sample for parameters of thrombin generation in females at 1 pM tissue factor.

Variable

Females (N)

Age (10 years) Diabetes 123 Obesity 542 Smoking 415 Hypertension 1,092 Dyslipidemia 517 FH of MI/stroke 924 History of MI 35 History of stroke 35 History of CAD 50 History of AF 35 History of PAD 93 History of VTE 124 History of CHF 36 History of cancer 236 HRT 223 Oral contraceptives 151 VKA use 292

β 0.0410 -0.00149 0.0587 0.0222 0.0214 0.0429 -0.0089 -0.0273 -0.0826 0.0336 -0.0320 0.0390 -0.0348 -0.0588 0.00886 -0.0457 -0.145 0.816

Log (lag time [min]) 95% CI P-value (0.0332;0.0488) (-0.0342;0.0312) (0.0411;0.0763) (0.00372;0.0408) (0.00582;0.0370) (0.0254;0.0604) (-0.0233;0.00546) (-0.099;0.0445) (-0.149;-0.0159) (-0.0178;0.0850) (-0.106;0.0417) (0.00303;0.0749) (-0.0687;-0.000971) (-0.13;0.0122) (-0.015;0.0327) (-0.0669;-0.0246) (-0.175;-0.116) (0.743;0.889)

<0.0001 0.93 <0.0001 0.019 0.0072 <0.0001 0.22 0.46 0.015 0.20 0.39 0.034 0.044 0.10 0.47 <0.0001 <0.0001 <0.0001

ETP (nM.min) 95% CI P-value

-15.5 -34.2 110 -17.3 -14.7 24.5 5.08 -19.2 105 -46.4 104 -25.1 44.4 -3.77 -33.6 60.1 379 -751

(-28.0;-2.99) 0.015 (86.8;18.3) 0.20 (81.9;139) <0.0001 (-47.1;12.4) 0.25 (-39.8;10.4) 0.25 (-3.63;52.6) 0.088 (-18.0;28.1) 0.67 (-135;96.2) 0.74 (-2.54;212) 0.056 (-129;36.1) 0.27 (-14.8;222) 0.086 (-82.9;32.6) 0.39 (-10.0;98.8) 0.11 (-118;110) 0.95 (-71.9;4.75) 0.086 (26.2;94.1) 0.00053 (331;427) <0.0001 (-869;-633) <0.0001

0.657 2.24 11.7 -7.23 -4.76 -2.02 1.50 8.04 21.2 3.34 20.3 -3.96 1.70 -9.50 -3.68 14.3 91.1 -92.1

Peak height (nM) 95% CI P-value (-1.90;3.22) (-8.54;13.0) (5.86;17.5) (-13.3;-1.12) (-9.90;0.390) (-7.80;3.75) (-3.24;6.23) (-15.6;31.7) (-0.799;43.2) (-13.6;20.3) (-3.95;44.6) (-15.8;7.89) (-9.47;12.9) (32.9;13.9) (-11.5;4.19) (7.28;21.2) (81.3;101) (-116;-67.9)

0.62 0.68 <0.0001 0.020 0.070 0.49 0.54 0.51 0.059 0.70 0.10 0.51 0.77 0.43 0.36 <0.0001 <0.0001 <0.0001

Multivariable linear regression models were calculated in the overall study sample for each parameter of thrombin generation as a dependent variable separately. The analysis was adjusted for age, use of vitamin K antagonists, oral contraceptives, and hormone replacement therapy, cardiovascular risk factors and cardiovascular diseases. ETP: endogenous thrombin potential; 95% CI: 95% confidence interval; FH: family history; MI: myocardial infarction; CAD: coronary artery disease; AF: atrial fibrillation; PAD: peripheral artery disease; VTE: venous thromboembolism; CHF: congestive heart failure; HRT: hormone replacement therapy; VKA: vitamin K antagonist.

2330

haematologica | 2020; 105(9)

Clinical determinants of thrombin generation

Considering the strong positive association between TG and the use of oral contraceptives or hormone replacement therapy, individuals taking these medications were excluded from Cox regression analysis. In the first model, adjustments were made for age, sex, and vitamin K antagonist use, whereas in the second and third models, additional adjustments were made for CVRF and CVD, respectively. As demonstrated by the second model in Table 4A, B, Cox regression analysis confirmed an increased mortality for individuals with a lag time at 1 pM TF above the 95th percentile of the reference [hazard ratio (HR) = 1.55, 95% confidence interval (95% CI): 1.14-2.11; P=0.0058] and with ETP at 5 pM TF above the 95th percentile of the reference (HR=1.53, 95% CI: 1.09-2.15; P=0.015), independently of the presence of CVRF. After additional adjustment for CVD, lag time at 1 pM TF (HR=1.46, 95% CI: 1.07-2.00; P=0.018) and ETP at 5 pM TF (HR=1.50, 95% CI: 1.06-2.13; P=0.023) remained associated with mortality.

Discussion The formation of thrombin is one of the key processes underlying thrombotic diseases and its role in CVD due to atherosclerosis attracted new interest with recent data showing superior efficacy of a combined strategy of aspirin and low-dose direct oral anticoagulation in reducing atherothrombotic events.17 Hence, limiting TG by inhibiting factor Xa provides an interesting approach to lower cardiovascular risk. In this study we explore the clinical determinants of TG measured in plasma, in a large population-derived study. Our data provide important

insights into the effects of CVRF in males and females. This study is the first to demonstrate the positive association of TG parameters, ETP as a global measure of procoagulant and anticoagulant action in plasma and lag time, with total mortality, independent of age, sex and CVRF. The presented reference values may be generalized to other laboratories. However, the reference ranges should be used with caution as the (pre-)analytical conditions of the assay may influence the reference ranges and standardization between laboratories is needed, as well as confirmation of the observed data. The reference values of the TG parameters as well as the mean and median values of the TG parameters in the overall study sample showed sex-specific differences with females having shorter lag times and higher ETP and peak height, compared to males. The sex differences in TG can be partly explained by the strong influence of female endogenous sex hormones on the coagulation cascade, as higher levels of fibrinogen and lower levels of protein S, antithrombin and protein C were observed in females, compared to males, irrespective of hormonal treatment.18 Following a sex-stratified, fully adjusted, large, multivariable model analysis we show that age, obesity and dyslipidemia are the most important clinical factors linked with higher TG potential. Furthermore, this study demonstrates the effect of different groups of medication on TG, with hormone-containing drugs being positively associated and anticoagulant and antiarrhythmic drugs being inversely associated with TG potential. Few studies have described the effect of age on TG parameters.10,11,19 Collectively, these studies suggest that TG potential enhances with increasing age, indicated by shorter lag time and higher ETP and peak height.

Table 3A. Multiple linear regression analysis of the effects of drugs on thrombin generation parameters in the study sample at 1 pM tissue factor.

Drug Sex hormones and modulators of the genital system Antithrombotic agents Cardiac therapy Immunosuppressants Anti-gout preparations

Log (lag time [min]) 95% CI

P-value

-0.0795

(-0.0979; -0.0611)

<0.000001

151

0.0862 0.112 0.114 0.0552

(0.0661; 0.106) (0.0763; 0.147) (0.0537; 0.174) (0.0248; 0.0856)

<0.000001 <0.000001 0.00021 0.00038

-92.8 -83.3 -28.6 0.270

ETP (nM.min) 95% CI P-value (124; 178)

Peak height (nM) 95% CI P-value

<0.000001

35.1

(-122; -63.4) <0.000001 (-135; -31.5) 0.0016 (-116; 59.0) 0.52 (-44.1; 44.6) 0.99

-5.01 -3.94 -1.37 3.76

(29.7; 40.4) <0.000001

(-11.0; 0.952) (-14.4; 6.54) (-19.1; 16.3) (-5.20; 12.7)

0.10 0.46 0.88 0.41

The analysis was adjusted for age, sex, cardiovascular risk factors and cardiovascular diseases. Bonferroni corrected P-value (0.00079) is used. For categorization of medication groups see Online Supplementary Material Part A. ETP: endogenous thrombin potential; 95% CI: 95% confidence interval.

Table 3B. Multiple linear regression analysis of the effects of drugs on thrombin generation parameters in the study sample at 5 pM tissue factor.

Drug Sex hormones and modulators of the genital system Antithrombotic agents Cardiac therapy Diuretics

Log (lag time [min]) 95% CI P-value

-0.0558

(-0.0706; -0.0411)

<0.000001

120

0.0656 0.0916 0.0354

(0.0495; 0.0817) (0.0633; 0.120) (0.0157; 0.0522)

<0.000001 <0.000001 0.00043

-122 -122 -37.0

ETP (nM.min) 95% CI P-value (93.6; 147)

<0.000001

(-151; -93.4) <0.000001 (-173; -71.1) 0.0000028 (-72.5; -1.45) 0.041

β 44.9

Peak height (nM) 95% CI P-value (38.7; 51.1) <0.000001

-19.4 (-26.3; -12.6) <0.000001 -16.0 (-28.0; -3.89) 0.0096 -4.55 (-12.9; 3.84) 0.29

haematologica | 2020; 105(9)

2331

P.C.S. van Paridon et al. Table 4A. Prognostic value of markers of thrombin generation, measured at 1 pM tissue factor, for mortality.

TG parameter Lag time below 5% of reference Lag time above 95% of reference ETP below 5% of reference ETP above 95% of reference Peak height below 5% of reference Peak height above 95% of reference

First model* 95% CI

P-value

1.22 1.54 1.08 1.41 1.14 1.28

(0.58; 2.60) (1.14; 2.08) (0.75; 1.56) (0.98; 2.03) (0.78; 1.67) (0.82; 2.01)

0.60 0.0055 0.66 0.061 0.51 0.27

1.11 1.55 1.11 1.44 1.18 1.29

Second model** 95% CI P-value (0.52; 2.37) (1.14; 2.11) (0.77; 1.60) (0.99; 2.07) (0.81; 1.74) (0.83; 2.03)

0.78 0.0058 0.58 0.051 0.39 0.26

HR 1.14 1.46 1.05 1.41 1.12 1.39

Third model*** 95% CI P-value (0.53; 2.42) (1.07; 2.00) (0.72; 1.52) (0.97; 2.04) (0.76; 1.65) (0.89; 2.18)

0.74 0.018 0.80 0.071 0.56 0.15

Cox regression model performed in the overall study sample with mortality as outcome and markers of thrombin generation as predictors. Models were *adjusted for age, sex and vitamin K antagonist (VKA) use. **adjusted for age, sex, VKA use and presence of cardiovascular risk factors (CVRF). ***adjusted for age, sex, VKA, use, presence of CVRF, and cardiovascular disease. TG: thrombin generation; HR: hazard ratio; 95% CI: 95% confidence interval; ETP: endogenous thrombin potential.

Table 4B. Prognostic value of markers of thrombin generation, measured at 5 pM tissue factor, for mortality.

TG parameter Lag time below 5% of reference Lag time above 95% of reference ETP below 5% of reference ETP above 95% of reference Peak height below 5% of reference Peak height above 95% of reference

First model* 95% CI

P-value

1.08 1.29 1.35 1.55 1.07 1.18

(0.40; 2.90) (0.92; 1.79) (0.94; 1.94) (1.11; 2.17) (0.72; 1.59) (0.78; 1.78)

0.89 0.14 0.11 0.0099 0.76 0.44

0.98 1.25 1.36 1.53 1.09 1.21

Second model** 95% CI P-value (0.36; 2.63) (0.89; 1.75) (0.94; 1.97) (1.09; 2.15) (0.73; 1.63) (0.80; 1.83)

0.96 0.20 0.10 0.015 0.66 0.37

Third model*** 95% CI P-value

1.08 1.17 1.26 1.50 1.02 1.23

(0.40; 2.92) (0.83; 1.65) (0.87; 1.83) (1.06; 2.13) (0.68; 1.52) (0.81; 1.86)

0.88 0.36 0.23 0.023 0.92 0.34

However, these studies had rather small sample sizes and included a homogenous population of healthy volunteers. In the present analysis, age was positively associated with lag time in both males and females. In males, ETP and peak height increased with age, whereas in females the amount of TG showed a rather negative trend with less strong associations compared to those in males. Other positive determinants of lag time observed in this study were obesity and dyslipidemia, which may partly be explained by increased levels of TFPI, a lipoproteinassociated coagulation inhibitor. It has been suggested that free TFPI is a major determinant of lag time.10 Elevated TFPI levels have been reported in individuals with impaired glucose tolerance and type 2 diabetes mellitus20 and it has been suggested that these TFPI levels were elevated due to related obesity.21 Smid and colleagues found that a prolongation in lag time in patients with previous myocardial infarction may be due to release of TFPI.7 In addition to lag time, both ETP and peak height showed positive associations with obesity and dyslipidemia. Total body fat percentage and body mass index have been positively associated with lag time, ETP and peak height in females, independently of age, prior CVD, glucose metabolism and smoking status, though no associations were observed in males.22 The present study demonstrates strong relations of obesity with a longer lag time and higher ETP and peak height in both males and females, independently of potential confounders. The association of ETP and peak height with obesity may be attributed to a low-grade inflammation observed in obese individuals.23 The results on associations between therapy and TG parameters showed that use of vitamin K antagonists was positively associated with lag time and negatively associ2332

ated with ETP and peak height, as expected from previous studies.24-26 Aspirin showed no effect on TG (data not shown), in line with recent findings from the COMPASS trial, in which treatment with a combination of aspirin and rivaroxaban, a direct factor Xa inhibitor, showed a superior effect on prevention of the manifestation of atherothrombosis in atherosclerotic disease, as compared to treatment with aspirin alone.17 Furthermore, intake of oral contraceptives or hormone replacement therapy was associated with a shorter lag time and higher ETP and peak height, in line with previous reports.10,27-29 The influence of estrogen-containing medication on the TG potential has been linked through increased levels of the coagulation factors II, VII, VIII, and X and fibrinogen, decreased levels of the natural anticoagulants, antithrombin and protein S, and acquired resistance to activated protein C.27,28 Hitherto, only a limited number of studies have explored the association between TG and mortality. The PROSPER study, including only elderly individuals, showed positive associations of vascular mortality with lag time and peak height and total mortality with lag time.5 However, after adjustment for interleukin-6 and Creactive protein levels, the associations were no longer significant, indicating that inflammation may be contributing to higher TG in these individuals. In another smaller study, higher ETP and peak height (at 5 pM TF), independently of age, sex and CVRF, were associated with increased risk of cardiovascular death in patients with acute coronary syndrome.30 In the present large, adult, population-based study, we demonstrate a positive association between lag time at 1 pM TF and total mortality, which remained significant after adjusting for traditional CVRF and CVD. Furthermore, this study highlights the haematologica | 2020; 105(9)

Clinical determinants of thrombin generation

Figure 1. Survival over 10 years for markers of thrombin generation above and below reference limits. Kaplan-Meier survival curves of the overall study sample demonstrating the 10-year survival of individuals with the thrombin generation parameters lag time (left panels), endogenous thrombin potential (ETP) (middle panels), and peak height (right panels) within the range of the reference group (green line), individuals above the 90th percentile of the reference group (blue line), and individuals above the 97.5th percentile of the reference group (red line), at 1 (upper panels) and 5 pM (lower panels) tissue factor (TF). For the lag times at both 1 and 5 pM TF, P<0.001 for the difference between the reference and the 90th percentile, as well as for the reference and the 97.5th percentile. For the ETP at 5 pM TF, P=0.034 for the difference between the reference and the 90th percentile and P=0.00097 for the difference between the reference and the 97.5th percentile.

relation between higher ETP (above the 95th percentile of the reference group) at 5 pM TF, as a global measure of both procoagulant and anticoagulant forces in the plasma, and increased risk of death, independently of CVRF and CVD. These findings indicate that both lag time and ETP are potential biomarkers for increased mortality risk, beyond the traditional CVRF. As discussed for the previous published PROSPER study, the association between a prolonged lag time and total mortality is not only a surprising and counterintuitive observation, but also one that is difficult to explain. With a risk of being too speculative, potential mechanisms might include consumption of initiators of coagulation before the system overshoots to start actual thrombosis. In other words, a constant (weak) prothrombotic trigger activates the coagulation system which is subsequently downregulated by the natural anticoagulants antithrombin and TFPI until, at a certain moment, the prothrombotic trigger increases and the system becomes overactivated and anticoagulants can no longer prevent thrombosis. In such a scenario, consumption of factor VII or XII, for example, could lead to a prolonged lag time in a sensitive in vitro assay. Another potential contributor to the prolonged lag time could be altered TFPI levels between subjects. However, assessing the TFPI levels in the presented cohort is part of another study and beyond the scope of the current study. haematologica | 2020; 105(9)

Limitations of our study are that we measured TG in platelet-poor plasma after one-step centrifugation of whole blood (10 min at 2,000 x g) in contrast to recommendations (two-step centrifugation, 2000 x g for 5 min, 10,000 x g for 10 minutes). A previous small-scale analysis by Loeffen and colleagues16 showed that in order to eliminate residual platelets and microparticles, which may contribute to variability in TG results, double-centrifuged samples are preferable. We cannot, therefore, exclude that residual platelets and microparticles contributed to the observed associations between CVRF and TG parameters. Next, we had only cumulative mortality data available, so conclusions could not be made regarding associations between TG variables and specific causes of mortality. However, the standardized clinical investigation of the cardiovascular profile, standardized laboratory measurements of the large Gutenberg Health Study sample and availability of prospective mortality data are essential strenghts of our study, which delivers important evidence on the TG assay as a potential tool for improving risk stratification for CVD. In conclusion, this is the first, large, population-based study demonstrating an important relation between TG parameters, such as the time to minimum thrombin formed or the amount of thrombin formed, and total mortality. Further research is required on the underlying mechanism as well as to explore the potential role of the 2333

P.C.S. van Paridon et al.

parameters as independent biomarkers for increased mortality risk. The observed association of TG and traditional CVRF, particularly obesity, is an important finding in light of the growing “globesity’’ issue worldwide.31 Acknowledgments We are indebted to all study participants and all co-workers of the Gutenberg Health Study, who were involved in the planning and conduct of this study. Funding The Gutenberg Health Study is funded through the govern-

References 1. Spronk HM, Govers-Riemslag JW, ten Cate H. The blood coagulation system as a molecular machine. Bioessays. 2003;25(12): 1220-1228. 2. Dargaud Y, Beguin S, Lienhart A, et al. Evaluation of thrombin generating capacity in plasma from patients with haemophilia A and B. Thromb Haemost. 2005;93(3):475480. 3. Besser M, Baglin C, Luddington R, van Hylckama Vlieg A, Baglin T. High rate of unprovoked recurrent venous thrombosis is associated with high thrombin-generating potential in a prospective cohort study. J Thromb Haemost. 2008;6(10):1720-1725. 4. Loeffen R, van Oerle R, Leers MP, et al . Factor XIa and thrombin generation are elevated in patients with acute coronary syndrome and predict recurrent cardiovascular events. PLoS One. 2016;11(7): e0158355. 5. Loeffen R, Winckers K, Ford I, et al. Associations between thrombin generation and the risk of cardiovascular disease in elderly patients: results from the PROSPER study. J Gerontol A Biol Sci Med Sci. 2015;70(8):982-988. 6. Lutsey PL, Folsom AR, Heckbert SR, Cushman M. Peak thrombin generation and subsequent venous thromboembolism: the Longitudinal Investigation of Thromboembolism Etiology (LITE) study. J Thromb Haemost. 2009;7(10):1639-1648. 7. Smid M, Dielis AW, Spronk HM, et al. Thrombin generation in the Glasgow Myocardial Infarction Study. PLoS One. 2013;8(6):e66977. 8. ten Cate-Hoek AJ, Dielis AW, Spronk HM, et al. Thrombin generation in patients after acute deep-vein thrombosis. Thromb Haemost. 2008;100(2):240-245. 9. Hemker HC, Giesen P, AlDieri R, et al. The calibrated automated thrombogram (CAT): a universal routine test for hyper- and hypocoagulability. Pathophysiol Haemost Thromb. 2002;32(5-6):249-253. 10. Dielis AW, Castoldi E, Spronk HM, et al. Coagulation factors and the protein C system as determinants of thrombin generation in a normal population. J Thromb Haemost. 2008;6(1):125-131. 11. Haidl H, Cimenti C, Leschnik B, Zach D,

2334

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

ment of Rhineland-Palatinate (“Stiftung RheinlandPfalz für Innovation“, contract AZ 961–386261/733), the research programs “Wissen schafft Zukunft” and “Center for Translational Vascular Biology (CTVB)” of the Johannes GutenbergUniversity of Mainz, and its contract with Boehringer Ingelheim and PHILIPS Medical Systems, including unrestricted grants for the Gutenberg Health Study. This work was supported by the German Federal Ministry of Education and Research (BMBF 01EO1003) and the Center for Translational Vascular Biology (CTVB) of the University Medical Center Mainz (to PSW). HtC is a Fellow of the Gutenberg Research Foundation.

Muntean W. Age-dependency of thrombin generation measured by means of calibrated automated thrombography (CAT). Thromb Haemost. 2006;95(5):772-775. Chaireti R, Gustafsson KM, Bystrom B, Bremme K, Lindahl TL. Endogenous thrombin potential is higher during the luteal phase than during the follicular phase of a normal menstrual cycle. Hum Reprod. 2013;28(7):1846-1852. Wild PS, Zeller T, Beutel M, et al. [The Gutenberg Health Study]. Bundesgesundheitsblatt Gesundheitsforschu ng Gesundheitsschutz. 2012;55(6-7):824-829. Schnabel RB, Wilde S, Wild PS, Munzel T, Blankenberg S. Atrial fibrillation: its prevalence and risk factor profile in the German general population. Dtsch Arztebl Int. 2012;109(16):293-299. Wild PS, Sinning CR, Roth A, et al. Distribution and categorization of left ventricular measurements in the general population: results from the population-based Gutenberg Heart Study. Circ Cardiovasc Imaging. 2010;3(5):604-613. Loeffen R, Kleinegris MC, Loubele ST, et al. Preanalytic variables of thrombin generation: towards a standard procedure and validation of the method. J Thromb Haemost. 2012;10(12):2544-2554. Eikelboom JW, Connolly SJ, Bosch J, et al. Rivaroxaban with or without aspirin in stable cardiovascular disease. N Engl J Med. 2017;377(14):1319-1330. Lowe GD, Rumley A, Woodward M, et al. Epidemiology of coagulation factors, inhibitors and activation markers: the Third Glasgow MONICA Survey. I. Illustrative reference ranges by age, sex and hormone use. Br J Haematol. 1997;97(4):775-784. Spronk HM, Dielis AW, De Smedt E, et al. Assessment of thrombin generation II: validation of the Calibrated Automated Thrombogram in platelet-poor plasma in a clinical laboratory. Thromb Haemost. 2008;100(2):362-364. Leurs PB, Stolk RP, Hamulyak K, Van Oerle R, Grobbee DE, Wolffenbuttel BH. Tissue factor pathway inhibitor and other endothelium-dependent hemostatic factors in elderly individuals with normal or impaired glucose tolerance and type 2 diabetes. Diabetes Care. 2002;25(8):13401345. Vambergue A, Rugeri L, Gaveriaux V, et al.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Factor VII, tissue factor pathway inhibitor, and monocyte tissue factor in diabetes mellitus: influence of type of diabetes, obesity index, and age. Thromb Res. 2001;101(5): 367-375. Beijers HJ, Ferreira I, Spronk HM, et al. Body composition as determinant of thrombin generation in plasma: the Hoorn study. Arterioscler Thromb Vasc Biol. 2010;30(12): 2639-2647. Pruller F, Raggam RB, Posch V, et al. Trunk weighted obesity, cholesterol levels and low grade inflammation are main determinants for enhanced thrombin generation. Atherosclerosis. 2012;220(1):215-218. Brodin E, Seljeflot I, Arnesen H, Hurlen M, Appelbom H, Hansen JB. Endogenous thrombin potential (ETP) in plasma from patients with AMI during antithrombotic treatment. Thromb Res. 2009;123(4):573579. Wielders S, Mukherjee M, Michiels J, et al. The routine determination of the endogenous thrombin potential, first results in different forms of hyper- and hypocoagulability. Thromb Haemost. 1997;77(4):629-636. Gatt A, van Veen JJ, Bowyer A, et al. Wide variation in thrombin generation in patients with atrial fibrillation and therapeutic International Normalized Ratio is not due to inflammation. Br J Haematol. 2008;142(6): 946-952. Rotteveel RC, Roozendaal KJ, Eijsman L, Hemker HC. The influence of oral contraceptives on the time-integral of thrombin generation (thrombin potential). Thromb Haemost. 1993;70(6):959-962. Tchaikovski SN, van Vliet HA, Thomassen MC, et al. Effect of oral contraceptives on thrombin generation measured via calibrated automated thrombography. Thromb Haemost. 2007;98(6):1350-1356. Marchi R, Marcos L, Paradisi I. Comparison by sex between thrombin generation and fibrin network characteristics in a healthy population. Clin Chim Acta. 2015;441:8689. Attanasio M, Marcucci R, Gori AM, et al. Residual thrombin potential predicts cardiovascular death in acute coronary syndrome patients undergoing percutaneous coronary intervention. Thromb Res. 2016;147:52-57. James PT, Leach R, Kalamara E, Shayeghi M. The worldwide obesity epidemic. Obes Res. 2001;9(Suppl 4):228S-233S.

haematologica | 2020; 105(9)

ARTICLE

Coagulation & its Disorders

A mutated factor X activatable by thrombin corrects bleedings in vivo in a rabbit model of antibody-induced hemophilia A

Ferrata Storti Foundation

Toufik Abache,1 Alexandre Fontayne,1 Dominique Grenier,1 Emilie Jacque,1 Alain Longue,1 Anne-Sophie Dezetter,1 Béatrice Souilliart,2 Guillaume Chevreux,2 Damien Bataille,2 Sami Chtourou1 and Jean-Luc Plantier1 LFB Biotechnologies, Direction de l’Innovation Thérapeutique, Loos and 2LFB Biotechnologies, Direction Générale du Développement, Les Ulis, France

ABSTRACT

Haematologica 2020 Volume 105(9):2335-2343

endering coagulation factor X sensitive to thrombin was proposed as a strategy to bypass the need for factor VIII. In this study, this nonreplacement strategy was evaluated in vitro and in vivo for its ability to correct factor VIII but also factor IX, X and XI deficiencies. A novel modified factor X, named actiten, was generated and produced in the HEK293F cell line. The molecule possesses the required post-translational modifications, partially maintaining its ability to be activated by RVV-X, factor VIIa/tissue factor, and factor VIIIa/factor IXa and acquires the ability to be activated by thrombin. The potency of the molecule was evaluated in plasma samples with deficiencies of the respective factors and in plasma samples from patients with hemophilia A, some of which contained inhibitors. Actiten dose-dependently corrected all the deficient plasmas that were assayed. It was able to normalize the thrombin generation at 20 μg/mL although the lag time was increased. It was then assayed in a rabbit antibody-induced model of hemophilia A in which, in contrast to recombinant wild-type factor X, it normalized the bleeding time and the loss of hemoglobin. No sign of thrombogenicity was observed and the generation of activated factor X was controlled by the anticoagulation pathway in all the coagulation assays performed. These data indicate that actiten may be considered as a possible non-replacement factor to treat hemophilia, with the advantage of being a zymogen that corrects bleeding only when needed.

Correspondence: JEAN-LUC PLANTIER plantierj@lfb.fr

Introduction Hemophilia results from a default of coagulation factor IX or VIII (FIX or FVIII). It is treated by prophylactic or on-demand infusions of the missing or deficient factor.1 While offering a satisfying protection against bleeding, repeated infusions, required to maintain an active threshold of factor, are uncomfortable for patients being deleterious to the venous access, and bringing risks of infection and of developing inhibitors against the substitutive factor.2 These drawbacks justify a continuous search for improvement of hemophilia treatments, in particular prolonging the product’s circulating half-life.3,4 This property is sought in order to maintain a higher threshold of coagulation, aiming to increase the treatment efficiency and compliance.5 With regards to hemophilia B, the fusion of FIX to an IgG1 Fc fragment or to albumin allowed a significant increase in FIX half-life, a less frequent administration schedule and a higher product threshold.5-7 In contrast, there was a limited improvement for hemophilia A using therapeutics based on the FVIII backbone. Persistence in the circulation of these therapeutic compounds is driven by the halflife of the FVIII chaperone, von Willebrand factor (VWF), which is only 1.5-fold greater than that of FVIII. Thus, modifications to FVIII only moderately improve a patient’s exposure to the therapeutic protein.8 In recent years, a novel class of agents to treat hemophilia has emerged. These agents are based on non-replacement factors (NRF), i.e. they are independent of FVIII or FIX molecules. Some NRF diminish the level of anticoagulation, reinforcing haematologica | 2020; 105(9)

Received: February 18, 2019. Accepted: November 5, 2019. Pre-published: November 7, 2019. doi:10.3324/haematol.2019.219865 ©2020 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.

2335

T. Abache et al.

the potency of the traces of thrombin generated during the initiation of coagulation. These are a short interfering (si)RNA inhibiting the expression of antithrombin and several monoclonal antibodies targeting the tissue factor pathway inhibitor.9,10 Downmodulation of the anticoagulant system (activated protein C or protein S) also enters this category.11,12 Another NRF aims to substitute the function of FVIII. A bispecific antibody called emicizumab, which facilitates the interaction of endogenous FIX/FIXa with factor X (FX), demonstrated its potency in this setting.13-15 These NRF are pharmacological agents with a mechanism of action that is independent of the fate of FVIII or VWF, thus offering drugs with a longer half-life, for the patients’ comfort, while restoring a partial but clinically sufficient coagulation. A third, proposed NRF strategy is to redirect the activation of FX.16 FX is at the crossroads of the intrinsic and extrinsic coagulation pathways and is responsible for the activation of prothrombin to thrombin. Rendering FX activatable to thrombin allows the thrombin that appears during the initiation of coagulation to generate larger amounts of FXa. These supplemental amounts will be enough to bypass the need for FVIIIa, the natural amplifier of coagulation. Such a modified FX was demonstrated in vitro to correct FVIII-deficient plasma.16 In this study, a second generation of recombinant thrombin-activatable FX (actiten) was created, in which, notably, the activation peptide was preserved in order to maintain FX pharmacokinetics in vivo.17 The ability of this molecule to correct several coagulation factor deficiencies was assessed in vitro and in a rabbit antibodyinduced model of hemophilia A.

Staining Solution or transferred onto a polyvinylidene fluoride membrane and revealed by anti-Gla or anti-FX antibody.

Binding to phospholipids FX binding was evaluated by ELISA on phospholipid dry-coated plates. Details of the protocol are provided in the Online Supplement.

Activation of actiten All activation reactions of actiten were conducted in Hepes buffer and were stopped in EDTA-containing buffers before the amount of FXa generated was measured following the degradation of pNAPEP 1065. A standard curve of human FXa was used as the reference. The conditions for each activation are described in detail in the Online Supplement.

In vitro evaluation of the thrombotic potential of actiten FX-deficient plasma was re-calcified to a final concentration of 7.5 mM calcium. Plasma-derived FX (pdFX), pdFX + plasmaderived activated FX (pdFXa) or actiten was incubated at 37°C until clotting of the mixture. A similar experiment was conducted in FVIII-deficient plasma spiked with FVIII, pdFXa or actiten.

Thrombin generation assay The performance of the thrombin generation assay was based on the Calibrated Automated Thrombogram method developed by Hemker et al.18 The values of the assay are the means of duplicates of two to five independent experiments. Normal plasma values were from 13 independent assays. Values for actiten (20 μg/mL) are from at least three independent experiments.

In vivo evaluation of actiten Methods Material The materials used in this study are listed in the Online Supplement.

In vivo studies were conducted in accordance with procedures approved by the Institutional Animal Care and Use Committee (CEEA26: A16_004). The rabbit model was rendered hemophiliac following infusion of anti-FVIII and the bleeding assay is described in the Online Supplement.

Statistical analysis Recombinant protein preparation The generation and production of actiten and recombinant FX constructs are explained in detail in the Online Supplement.

Purification using an anti-Gla aptamer Re-calcified supernatant was loaded on an anti-Gla aptamer column containing the proprietary Mapt1.2CSO aptamer (Eurogentec, Liège, Belgium). The equilibration buffer was 50 mM Tris-HCl, 10 mM CaCl2, pH 7.5. After a wash with the equilibration buffer + 300 mM NaCl, the bound compounds were eluted with 50 mM Tris-HCl, 10 mM EDTA, pH 7.5. Fractions corresponding to the protein peak were pooled. Eventual traces of FXa were inhibited by 10 μM GGACK and the eluate was concentrated with a 50 kDa tangential flow filtration system then dialyzed in NaCl 0.9% and stored at -80°C. The absence of FXa was controlled by incubating the actiten preparation (100 nM) with FXa substrate (pNAPEP 1065) for 30 min at 37°C.

Antigen dosage, electrophoresis and immunoblotting An enzyme-linked immunosorbent assay (ELISA) for FX (Zymutest Factor X, Hyphen, France) was performed according to the manufacturer’s instructions. For sodium dodecylsulfate polyacrylamide gel electrophoresis, proteins were deposited on Gel NuPAGE Novex Bis-Tris 4-12% and migrated at 200 V. The gel was stained with PageBlue Protein 2336

In vitro data are presented as means ± standard deviation. In vivo data were analyzed using Prism 5 software (GraphPad Software Inc.) and are represented as means ± standard error of mean. In addition, all statistical comparisons were carried out using the nonparametric two-tailed Mann-Whitney test. Statistically significant differences are indicated in the figures: *P<0.05 and **P<0.01.

Results Expression of actiten, a thrombin-sensitive factor X A FX sensitive to thrombin was generated by inserting a 10-amino acid polypeptide between the carboxyl terminus of the activation peptide and the amino terminus of the heavy chain. The polypeptide was composed of six amino acids of fibrinopeptide A and four amino acids forming a thrombin cleavage site. The resulting molecule was called actiten (Figure 1). It was expressed in the HEK293F cell line and purified using a proprietary aptamer column recognizing the g-carboxylated (Gla) domain of the coagulation factor. The molecule was immunodetected by anti-Gla monoclonal and polyclonal anti-human FX antibodies (Figure 2). In native conditions, purified actiten appeared as a unique band at 64-66 kDa. Under reducing conditions, the anti-Gla antibody revealed the FX light haematologica | 2020; 105(9)

Mutated factor X to correct hemophilias

chain (20 kDa) and a minor fraction of the molecule that was not reduced. In reducing conditions, the polyclonal anti-FX mainly revealed the heavy chain at 50 kDa and barely the non-reduced FX fraction and the light chain. The molecular weight of actiten was 3-4 kDa greater than that of pdFX, probably because of its additional peptide. The molecule was analyzed by reverse phase high-performance liquid chromatography following activation with a FX activator from Russell’s viper venom (RVV-X) (Online Supplementary Figure S1). This analysis demonstrated the presence of the expected post-translational modifications. The light chain contained 11 Gla, in addition to Cterminal heterogeneity, also found in pdFX. The other post-translational modifications were identical to those in pdFX. The heavy chain was found complete with a minor population under a β form. O-glycosylation was also detected within the heavy chain.19 The activation peptide with the 10 amino-acid polypeptide added was completely liberated from the heavy chain by the RVV-X. It contained three O-glycosylations.19

The presence of the expected number of Gla allowed actiten to interact efficiently with phospholipids in a solid phase assay (Figure 3). The interaction was similar albeit slightly less efficient than that of pdFX. The half-life of actiten was then evaluated in rabbits in which pdFX had been previously studied. Both products had a similar halflife (at around 6 h) indicating that the modification of the FX did not affect its pharmacokinetics (Online Supplementary Figure S2). These data indicate that a mutated recombinant FX possessing an expected structure, pharmacokinetics and post-translational modifications was produced in HEK293F cells.

Evaluation of actiten activation Actiten was activated by RVV-X, the FVIIa/tissue factor (TF) complex and the FVIIIa/FIXa complex. The percentage of activation of actiten in comparison to that of pdFX was calculated from the initial velocity of FXa generation. Actiten maintained the ability to be activated by the natural FX effectors. However, whereas activation by RVV-X

Figure 1. Schematic representation of actiten. The scheme focuses on the factor X (FX) heavy chain: at the N-terminus of the heavy chain, the natural sequence of the FX activation peptide (AP) 1→52 ends at Arg234; in red a 10-amino acid peptide added between the AP and the catalytic domain (IVGGQ--) modifying the specificity of FX. The yellow arrow indicates the activation site. The light chain containing the g-carboxylation sites is represented in orange.

Figure 2. Visualization of purified actiten from HEK293F cells. Actiten was purified using an anti-Gla aptamer column and was separated by electrophoresis on a 4-12% sodium dodecylsulfate polyacrylamide gel. (A) PAGE Blue staining. Lane 1, molecular weight (values in kDa are on the left of the gel); lane 2, HEK293F supernatant; lane 3, flow-through; lane 4, washes; lane 5, purified actiten. (B) Detection of Gla using a monoclonal antibody (MoAb anti-Gla). Molecular weight values in kDa are on the left of the gel. Lane 1, non-reduced plasma-derived FX (pdFX); lane 2, non-reduced HEK293F supernatant; lane 3, non-reduced purified actiten; lane 4, dithiothreitol (DTT)-reduced pdFX; lane 5, DTT-reduced purified actiten. (C) Detection of FX by polyclonal antibodies (pAnti-FX). Lane 1, non-reduced pdFX; lane 2, non-reduced HEK293F supernatant; lane 3, non-reduced purified actiten; lane 4, DTT-reduced pdFX; lane 5, DTT-reduced purified actiten. Separated lanes from (B) and (C) are from the same gel but some bands not related to this article were removed. The signals were not adjusted.

haematologica | 2020; 105(9)

2337

T. Abache et al.

was moderately affected (63±11.97% that of pdFX), activation by FVIIa/TF was impaired, being 31±1.14%, and activation by the “tenase complex” was even more affected since actiten had only 20±1.23 % of the capacity of pdFX. The ability of actiten to be activated by thrombin was assessed in the absence of phospholipids (Figure 4). Phospholipids were omitted in this assay to eliminate the risk of auto-activation or degradation due to the appearance of FXa.20,21 Incubation of pdFX with or without thrombin or actiten alone for 48 h at 37°C did not generate any FXa. This result confirms that our preparation of actiten did not contain traces of FXa. In contrast, the incubation of actiten with thrombin led to the appearance of a regular amount of FXa, indicating that the modification of the activation peptide rendered actiten sensitive to thrombin. The kcat/Km was calculated and was found to be 1.25±0.1 x 103 M-1s-1, a value close to the modified FX (FXfpa) generated by Louvain-Quintard et al.16 To complete this evaluation, the molecule was incubated for 6 h in the presence of phospholipids and FXa, at 10% of the actiten concentration. There was no supplemental generation of FXa other than the 10% added, suggesting that actiten resisted FXa cleavage (data not shown).21 The presence of FXa and the stability of actiten were

then evaluated in re-calcified FX- or FVIII-deficient plasma samples (Figure 5A, B). As a control, pdFXa shortened the spontaneous clotting time of these plasma samples to 11±8.5 min and 6.7±2.9 min, respectively. The presence of pdFX in FX-deficient plasma or rFVIII in FVIII-deficient plasma allowed the plasma to clot in 60±35 min and 25±8.7 min, respectively. The time to clot in the presence of actiten was much longer in both plasma samples (96±47 min and >120 min, respectively). This result further confirms the absence of contaminating FXa and did not reveal any instability of actiten in different plasma samples.

In vitro evaluation of actiten potency Thrombin generation assays were performed to evaluate the ability of actiten to correct coagulation factor deficiencies. In all experiments, the deficient plasma was evaluated as a negative control and a pool of lyophilized normal plasma as a positive control. The thrombin generation parameters of normal plasma were chosen as a reference and compared with the parameters from deficient plasmas spiked with 20 μg/mL actiten (Table 1). In FVIII-deficient plasma, a dose-response relation was observed for actiten from 10 to 60 μg/mL before saturation at 120 μg/mL (Figure 6A). At 20 μg/mL of actiten, the

Figure 3. Binding of plasma-derived factor X and actiten to phospholipids. Plasma-derived factor X (pdFX) (●) and actiten (■) were incubated at increasing concentrations on coated phospholipids and detected by a polyclonal anti-FX coupled to horse radish peroxidase. (n=4). UDO: optical density unit.

Figure 4. Activation of plasma-derived factor X and actiten by thrombin. Plasma-derived factor X (pdFX) (100 nM) with 10 nM thrombin (■) or without thrombin (●), and actiten (100 nM) with 10 nM thrombin (▼) or without thrombin (▲) were incubated without phospholipid (PL) for increasing periods of time. At different time points a sample was taken and the presence of activated factor X (FXa) was measured. The results are representative of three independent experiments. Signs for pdFX without thrombin (●) and actiten without thrombin (▲) are masked by (■)."

2338

haematologica | 2020; 105(9)

Mutated factor X to correct hemophilias

amount of actiten generated was greater than that in normal plasma (endogenous thrombin potential: 1819±202 vs. 1510±216 nM.min, respectively) although this was not a statistically significant difference. There was a slight increase in the lag time when actiten was used (10.2±1.1 min for actiten vs. 7.2±1.7 min for normal plasma; P<0.01). Actiten was then evaluated in hemophilia A plasma with or without a high titer of inhibitors (292 BU/mL) (Figure 6B, C). In this latter plasma, the addition of actiten, unlike that of FVIII, could restore thrombin generation in a dosedependent manner. At a dose of 20 μg/mL, a total amount of 689±24 nM.min of thrombin was generated, providing significant correction of the plasma. Actiten was next evaluated in FIX-deficient plasma, spiked or not with anti-FIX polyclonal antibodies, in which it behaved similarly as in the hemophilia A plasma or FVIII-deficient plasma (Figure 6D, E; Table 1). The potency of the molecule was then assessed in the absence of FX or FXI (Figure 6F, G). The behavior of the molecule was similar to that previously observed, with the amount

of thrombin generated (1010±73 nM.min and 1285±119 nM.min, respectively) being close to that in normal plasma (1510±216 nM.min) but again with an increased lag time compared to that of normal plasma (12.1±1.8 min and 18.6±8.7 min vs. 7.2±1.7 min; P<0.001 and P<0.001, respectively). It should be stressed that in all these thrombin generation assays, the thrombin generated was efficiently controlled by the anticoagulant system and that no runaway of coagulation was detected at up to 120 μg/mL actiten. These data indicate that even following clotting induction, actiten did not overreact or escape the anticoagulation pathway.

In vivo evaluation of actiten potency A rabbit model of hemophilia A was established to evaluate the potency of actiten in vivo. A cocktail of two antiFVIII monoclonal antibodies severely impaired thrombin generation in rabbit plasma (Figure 7A). An in vivo assay was designed and is schematically represented in Figure 7B. The anti-FVIII antibodies or NaCl were infused into rabbits and the bleeding times were recorded. A statistically significant difference was observed between rabbits administered NaCl or anti-FVIII (405±44 s vs. 2760.5±348 s; P<0.01), which validated the animal model (Figure 7C). The experiments were repeated but with the infusion of actiten, recombinant wild-type FX (recFX-WT) or NaCl following the infusion of anti-FVIII antibodies. RecFX-WT was produced from HEK293F cells and purified in the same way as actiten (Online Supplementary Figure S3). As a positive control, FVIIa (500 μg/kg) was shown to correct bleeding in this model (746±290 s; P<0.01) following its infusion. The bleeding times in anti-FVIII-treated rabbits (2861±2 s) were significantly longer than those in actitentreated rabbits (P<0.05); bleeding times of the latter were similar to those in untreated wild-type rabbits (387±111 s vs. 347±51 s, respectively; P=ns) (Figure 7D). In contrast, the infusion of recFX-WT did not diminish the bleeding time with regard to those of anti-FVIII-treated rabbits (2847±668 s vs. 2861±2 s; P=ns) confirming the absence of contaminating FXa in the preparation and the specificity of actiten (Online Supplementary Figure S3). The hemoglobin losses were also compared and identical efficacy profiles were found: rabbits treated with actiten lost the equivalent of 4±2 mg/dL hemoglobin; in contrast, the antiFVIII-treated rabbits lost the equivalent of 268±169 mg/dL (P<0.05) (Figure 7E). Recombinant FVIIa diminished the loss of hemoglobin (61±38 mg/mL) in contrast to recFXWT (420±222 mg/mL, P=ns). Markers of exaggerated coagulation (thrombin/antithrombin complexes, prothrombin fragment 1 and 2 and fibrinogen D-dimers) were also measured but none of them was significantly increased (Online Supplementary Table S1). These results demonstrate that actiten was able to restore coagulation in vivo, in a model of antibody-induced hemophilia A.

Discussion Figure 5. Evaluation of the stability of actiten in factor-deficient plasma samples. (A) Evaluation in factor X (FX)-deficient plasma. Plasma-derived FX (pdFX) (10 μg/mL), pdFX (10 μg/mL) plus plasma-derived activated factor X (pdFXa) (0.1 μg/mL) and actiten (20 μg/mL) were incubated in FX-deficient plasma recalcified by 7.5 mM CaCl2. Plasma clotting times were recorded manually (n=24). (B) Evaluation in factor VIII (FVIII)-deficient plasma. FVIII (1 U/mL), pdFXa (0.1 µg/mL) and actiten (20 μg/mL) were incubated in FVIII-deficient plasma re-calcified by 7.5 mM CaCl2. Plasma clotting times were recorded manually (n=3).

haematologica | 2020; 105(9)

Actiten as a NRF strategy has some specific features that are illustrated by the data presented here. First, actiten was able to normalize the amount of thrombin generated in vitro as well as in vivo during bleeding in a model of antibody-induced hemophilia A. In contrast, most of the other NRF strategies partially restore coagulation but some 2339

T. Abache et al.

Table 1. Thrombin generation parameters of normal and factor-deficient plasma samples, containing inhibitors or not, supplemented with actiten.

FIX-def FIX-def FIX-def FIX-def plasma plasma plasma + plasma + TF/PL TF/PL inhibit. inhibit. TF/PL (n=4) TF/PL (n=2) (n=4) (n=2) + buffer + buffer + actiten + buffer + actiten + buffer + actiten + buffer + actiten + buffer + actiten

Normal FVIII-def FVIII-def TG Parameter plasma plasma plasma TF/PL TF/PL TF/PL (n=5) (n=13) (n=5)

7.2 Lagtime ±1.7 (min) 1510 ETP (nM.min IIa) ±216 195 Peak ±87 (nM IIa) 12.1 ttPeak ±2.6 (min) 32.9 StartTail ±6.1 (min) 40.4 Velocity (nM IIa/min) ±25.4

8.7 ±1.4 543 ±77 19.6 ±4.0 25.7 ±2.1 67.8 ±6.8 1.1 ±0.25

10.2 ±1.1 1819 ±202 224 ±59 15.5 ±2.7 38.8 ±6.6 48.3 ±22.8

Hem. A TF/PL (n=3)

5.3 ±2.2 213 ±22 11.9 ±2.3 17.0 ±7.4 48 ±11 1.2 ±0.6

Hem. A Hem. A Hem. A TF/PL + inhibit. + inhibit. TF/PL (n=5) TF/PL (n=3) (n=3)

7.7 ±1.6 812 ±243 81.3 ±31.4 15.6 ±1.2 35.5 ±2.3 10.7 ±5.3

7.1 ±0.8 199 ±3.0 9.4 ±0.3 22.9 ±0.8 53,8 ±1.2 0.6 ±0.1

9.2 ±0.9 689 ±24 76 ±11 16.8 ±0.8 33.0 ±2.2 10.1 ±2.5

8.2 ±4.8 155 ±56 5.3 ±2.7 28.2 ±12.3 63.5 ±11 0.3 ±0.2

16.2 ±6.1 1436 ±101 168 ±34 23.1 ±6.6 40.9 ±6.9 25.2 ±8.7

8.6 ±4.1 186 ±263 9.5 ±10.2 35 ±26 28 ±39 0.8 ±1.0

12.8 ±3.1 1394 ±87 157 ±30 19.5 ±3.0 39.0 ±4.4 25.2 ±8.7

FX-def FX-def FXI-def FXI-def plasma plasma plasma plasma TF/PL TF/PL TF/PL TF/PL (n=3) (n=6) (n=2) (n=4) + buffer + actiten + buffer + actiten 31.3 ±6.3 0 ±0 0,3 ±0.1 42.6 ±4.0 18.7 ±32.3 0.1 ±0.01

12.1 ±1.8 1010 ±73 119 ±34 17.9 ±1.4 37.6 ±4.1 21.8 ±9.7

10.7 ±3.3 430 ±29 19.0 ±3.5 21.7 ±4.2 57 ±19.8 1.7 ±0.2

18.6 ±8.7 1285 ±119 146 ±52 25.6 ±10.3 44.6 ±12.2 23.3 ±13.4

Thrombin generation assays were performed in normal lyophilized pooled plasma (normal plasma), lyophilized pooled factor VIII-deficient plasma (FVIII-def.), lyophilized hemophilia A plasma (Hem. A), lyophilized hemophilia A plasma with 292 BU/mL inhibitors (Hem. A + inhibit.), lyophilized pooled factor IX-deficient plasma (FIX-def. plasma), lyophilized pooled FIX-deficient plasma spiked with 17 BU/mL inhibitors from a polyclonal anti-FIX antibody (FIX-def. plasma + inhibit.), lyophilized pooled factor X-deficient plasma (FX-def. plasma) and lyophilized pooled FXI-deficient plasma (FXI-def. plasma). The thrombin generation was initiated by 0.5 pM tissue factor and 4 μM phospholipids (TF/PL) in the presence of actiten (20 μg/mL) or buffer. N: number of duplicates performed.

bleeding continues to occur. Second, it is based on the use of a zymogen thereby being a unique strategy respecting the coagulation/anticoagulation balance. Actiten liberates a wild-type FXa, a natural target of the anticoagulation system for which antithrombin could easily be used as an antidote, if needed. Third, as the duration of the treatment is based on the half-life of FX, this strategy can offer a four-fold longer persistence of treatment efficacy than that using FVIII. Lastly, actiten was demonstrated to be active in several coagulation defects, offering a supplementary therapeutic option for rare coagulation defects such as hemophilia B with inhibitors (Figure 6E). In order to generate actiten, the cleavage site liberating the activation peptide was targeted to change the recognition of FX. The addition of the 10-amino acid peptide offers the possibility of two cleavage sites during actiten activation. As observed by mass spectroscopy following RVV-X cleavage, only the activation peptide containing the complete added peptide was detected, suggesting a preferential cleavage between LDRP and IVV from the heavy chain. Moreover, there was no loss in the activity of the FXa liberated from actiten with regard to pdFXa. These data confirm that the active species liberated corresponds to a wild-type FXa. The modification of activation was demonstrated since the presence of thrombin allowed the generation of FXa from actiten but not pdFX. Importantly, the recognition of the molecule by the FVIIa/TF complex was minimally affected and still allowed efficient participation of actiten in the initiation of the coagulation leading to the generation of primary traces of FXa and thrombin. The ability of the molecule to correct FX-deficient plasma confirmed that the activation of actiten by FVIIa/TF was sufficient to replace the endogenous FX. In contrast, the important loss of activation by the FVIIIa/FIXa complex is less crucial for the function of actiten in hemophilia since the molecule is designed to substitute for their absence. 2340

In vitro, actiten showed a velocity, a peak height and an endogenous thrombin potential sufficient to correct factordeficient plasma samples. However, an increased lag time to clotting was observed in all factor-deficient samples of plasmas. This delay likely results from: (i) the moderate loss of FVIIa/TF recognition to activate the molecule; (ii) activation by thrombin that might not be as efficient as natural initiation of coagulation; and/or (iii) the time required to generate the first traces of thrombin to amplify the coagulation.22 Nevertheless, this delay did not translate in vivo into a lack of efficiency since the presence of actiten allowed clotting in antibody-treated rabbits with the same kinetics as in wild-type animals. Factor V contained within platelets actively participates in the generation of thrombin through the prothrombinase complex.23-25 Thus, in vivo, there may be some elements favoring clotting, such as factor V and platelets, that are lacking from the in vitro assays. The presence of such elements might reinforce the clotting efficiency of actiten and eliminate differences in lag time, compared to that in normal plasma, observed in vitro. To evaluate actiten in vivo, a rabbit model was established since human FXa is equally efficient in rabbit and human plasma.26 The classic hemophilia A mouse model would have been preferred to evaluate the efficacy of actiten, but the compound is less efficient in mouse plasma, rendering this model poorly predictive (data not shown). Despite the fact that the pair of monoclonal antibodies severely impaired rabbit coagulation in vitro, a limitation of the model is that it was not possible to ascertain full FVIII inhibition in vivo. It has been reported that the half-life of FXa in blood is extremely short (<1.5 min).27 In our in vivo model, the rabbit nail cuticle was cut 35 min after infusion of the molecule. The occurrence of normal clotting at this time point demonstrated that the molecule was kept under a zymogen form in the circulation and that it was mobilized when needed. haematologica | 2020; 105(9)

Mutated factor X to correct hemophilias

Figure 6. Evaluation of actiten potency in several factor-deficient or hemophilia plasma samples. Each plasma sample was activated by 0.5 pM tissue factor and 4 µM phospholipids. For all panels, the following representations are used: normal plasma (●), deficient plasma (○). (A) Actiten potency was evaluated at different concentrations (10→120 μg/mL) in factor VIII (FVIII)-deficient plasma (blue ●). (B) Actiten potency was evaluated at 10 or 20 μg/mL (blue ●, ▲, respectively) in hemophilia A plasma. (C) Actiten potency was evaluated at different concentrations (10→30 μg/mL, blue ●, ▲, ■, respectively) in hemophilia A plasma with 292 BU/mL. (D) Actiten potency was evaluated at 10 and 20 μg/mL (blue ●, ▲, respectively) in factor IX (FIX)-deficient plasma. (E) Actiten potency was evaluated at 10 and 20 μg/mL (blue ●, ▲, respectively) in FIX-deficient plasma with spiked anti-FIX inhibitors (17 BU/mL). (F) Actiten potency was evaluated at 10 and 20 µg/mL (blue ●, ▲, respectively) in factor X (FX)-deficient plasma. (G) Actiten potency was evaluated at 10 and 20 μg/mL (blue ●, ▲, respectively) in factor XI (FXI)-deficient plasma. The data are derived from duplicate evaluations and are representative of 2-13 independent assays.

haematologica | 2020; 105(9)

2341

T. Abache et al. A

Figure 7. In vivo evaluation of actiten in a rabbit model of antibody-induced hemophilia A. (A) In vitro control of the efficiency of anti-factor VIII (FVIII) antibodies. Thrombin generation was induced by tissue factor/phospholipids in rabbit normal plasma (NP) in the presence or absence of anti-FVIII inhibitors (3 μg/mL each). (B) Schematic representation of the in vivo protocol. The rabbit’s paw is placed in phosphate-buffered saline at 37°C to measure time to clot and to collect blood loss. (C) Evaluation of rabbit bleeding following infusion of anti-FVIII antibodies. Bleeding times were recorded in rabbits infused with NaCl (n=8) or anti-FVIII antibodies (99 μg/kg each; n=4) following a nail cuticle cut. Data are presented as the means ± standard error of mean (SEM) and are representative of at least two separate experiments. (D) Evaluation of actiten potency in rabbits infused with anti-FVIII inhibitors. Bleeding times were recorded from rabbit controls (n=4), rabbits infused with anti-FVIII inhibitors and NaCl (n=7) or actiten (1.7 mg/kg, n=3) or recombinant wildtype factor X (recFX-WT) (1.7 mg/kg, n=4) or recombinant activated factor VII (recFVIIa) (500 μg/kg, n=6) following a nail cuticle cut. (E) Hemoglobin loss from rabbits infused with anti-FVIII inhibitors and NaCl (n=7) or actiten (1.7 mg/kg, n=3) or recFX-WT (1.7 mg/kg, n=4) or recFVIIa (500 μg/kg, n=4). Data are presented as the means ± SEM (statistical significance: ns: not significant; *P<0.05, **P<0.01). (D, E) Experiments were performed in two distinct rounds that are identified by the marker color. NP: normal plasma; MoAb: monoclonal antibodies; ND, not done.

The advantages of actiten include the interesting possibility that it could be used to correct various coagulation disorders. At a dose close to the physiological concentrations of FX, actiten normalizes clotting in FVIII-, FIX-, FX- and FXI-deficient plasma samples, independently of the presence of anti-FVIII or anti-FIX inhibitors. Thrombin generation assays performed in rabbit plasma were found to be remarkably predictive to determine the efficient dose of anti-FVIII antibodies, the effective concentrations of actiten and to help in the design of the in vivo assay. On the basis of this experience, corrections of the other coagulation defects observed in factor-deficient plasma samples would likely be confirmed in vivo in other preclinical models. Whereas the hemophilia A and B markets stimulate the research for innovative treatments, this is not true for FX and FXI deficiencies. Recently, a recombinant version of FX was offered to patients whereas FXI deficiency can only be treated by a few plasma-derived products.28,29 Given its limited activa2342

tion by the tenase complex, it could be expected that replacing FX by actiten might be less efficient than replacing it with natural FX, despite promising in vitro data. Nevertheless, if the in vivo potential of actiten turns out to be satisfactory, the compound could offer a supplementary treatment option for these rare diseases. Modifying the specificity of FX could induce a risk of thrombogenicity. However, no sign of such a drawback has been identified so far. In vitro incubation of the molecule in FVIII- and FX-deficient plasma samples without induction led to clotting at later times than when the samples were spiked with the missing factors. In all the thrombin generation assays performed, once the coagulation was initiated it was systematically controlled by the anticoagulation system since all peaks returned to baseline. In vivo, the rabbits showed no clinical signs of suffering and the markers of thrombosis measured remained at baseline levels. In addition, the infusion of actiten (at 0.33 haematologica | 2020; 105(9)

Mutated factor X to correct hemophilias

mg/kg) in wild-type rabbits did not lead to any sign of suffering suggesting that even in vivo in a blood normally prone to coagulate the molecule probably remains in the form of a zymogen (Online Supplementary Figure S2). Nevertheless, to confirm the safety of the molecule, more challenging in vivo models, such as the Wessler assay, are scheduled. Another primary concern regarding any novel biologic is its potential to elicit an immune response. This can be investigated during product development although preliminary evaluations using in silico, in vitro and in vivo models are recognized to be not fully predictive.30-32 Nevertheless, actiten was assessed by three independent in silico algorithms, and none identified any specific concern for its immunogenicity in comparison to that of pdFX (data not shown). Emicizumab is the most advanced NRF and has some

References 13. 1. Mannucci PM, Tuddenham EG. The hemophilias--from royal genes to gene therapy. N Engl J Med. 2001;344(23):1773-1779. 2. Berntorp E, Shapiro AD. Modern haemophilia care. Lancet. 2012;379(9824): 1447-1456. 3. Ragni MV. New-generation recombinant factor concentrates: bridge to gene therapy. Haemophilia. 2001;7(Suppl 1):28-35. 4. Schlesinger KW, Ragni MV. Safety of the new generation recombinant factor concentrates. Expert Opin Drug Saf. 2002;1(3):213223. 5. Lambert T, Benson G, Dolan G, et al. Practical aspects of extended half-life products for the treatment of haemophilia. Ther Adv Hematol. 2018;9(9):295-308. 6. Schulte S. Half-life extension through albumin fusion technologies. Thromb Res. 2009;124(Suppl 2):S6-S8. 7. Hoy SM. Eftrenonacog alfa: a review in haemophilia B. Drugs. 2017;77(11):12351246. 8. Mahlangu J, Young G, Hermans C, Blanchette V, Berntorp E, Santagostino E. Defining extended half-life rFVIII - a critical review of the evidence. Haemophilia. 2018;24(3):348-358. 9. Sehgal A, Barros S, Ivanciu L, et al. An RNAi therapeutic targeting antithrombin to rebalance the coagulation system and promote hemostasis in hemophilia. Nat Med. 2015;21(5):492-497. 10. Chowdary P, Lethagen S, Friedrich U, et al. Safety and pharmacokinetics of anti-TFPI antibody (concizumab) in healthy volunteers and patients with hemophilia: a randomized first human dose trial. J Thromb Haemost. 2015;13(5):743-754. 11. Prince R, Bologna L, Manetti M, et al. Targeting anticoagulant protein S to improve hemostasis in hemophilia. Blood. 2018;131(12):1360-1371. 12. Polderdijk SG, Adams TE, Ivanciu L, Camire RM, Baglin TP, Huntington JA. Design and characterization of an APC-specific serpin

haematologica | 2020; 105(9)

14.

15.

16.

17.

18.

19.

20.

21. 22.

crucial advantages for patients, in particular for those who have developed inhibitors.14 However, some patients treated with this compound still experienced bleeding whereas others developed anti-drug antibodies. Thus, there is still a need for alternative therapies with different mechanisms of action to improve patients’ care. The in vitro and in vivo data presented here validate the concept of re-directing the FX specificity to bypass the need for FVIII and FIX and may offer an additional opportunity to correct bleeding in several coagulation defects even in the presence of inhibitors. Acknowledgments The authors thank Bianca Boussier, Sylvie Le Ver, Bénedicte Fournes and Paul Martres for their participation in the in vivo experiments, Sophie Lecompte for helping to correct the manuscript and Céline Bourdon for her technical help.

for the treatment of hemophilia. Blood. 2017;129(1):105-113. Kitazawa T, Igawa T, Sampei Z, et al. A bispecific antibody to factors IXa and X restores factor VIII hemostatic activity in a hemophilia A model. Nat Med. 2012;18 (10):1570-1574. Oldenburg J, Mahlangu JN, Kim B, et al. Emicizumab prophylaxis in hemophilia A with inhibitors. N Engl J Med. 2017;377 (9):809-818. Mahlangu J, Oldenburg J, Paz-Priel I, et al. Emicizumab prophylaxis in patients who have hemophilia A without inhibitors. N Engl J Med. 2018;379(9):811-822. Louvain-Quintard VB, Bianchini EP, CalmelTareau C, Tagzirt M, Le Bonniec BF. Thrombin-activable factor X re-establishes an intrinsic amplification in tenase-deficient plasmas. J Biol Chem. 2005;280(50):4135241359. Gueguen P, Cherel G, Badirou I, Denis CV, Christophe OD. Two residues in the activation peptide domain contribute to the halflife of factor X in vivo. J Thromb Haemost. 2010;8(7):1651-1653. Hemker HC, Beguin S. Thrombin generation in plasma: its assessment via the endogenous thrombin potential. Thromb Haemost. 1995;74(1):134-138. Chevreux G, Tilly N, Faid V, Bihoreau N. Mass spectrometry based analysis of human plasma-derived factor X revealed novel posttranslational modifications. Protein Sci. 2015;24(10):1640-1648. Jesty J, Spencer AK, Nakashima Y, Nemerson Y, Konigsberg W. The activation of coagulation factor X. Identity of cleavage sites in the alternative activation pathways and characterization of the COOH-terminal peptide. J Biol Chem. 1975;250(12):4497-4504. Mertens K, Bertina RM. Pathways in the activation of human coagulation factor X. Biochem J. 1980;185(3):647-658. Dargaud Y, Wolberg AS, Luddington R, et al. Evaluation of a standardized protocol for thrombin generation measurement using the calibrated automated thrombogram: an international multicentre study. Thromb

Res. 2012;130(6):929-934. 23. Miletich JP, Majerus DW, Majerus PW. Patients with congenital factor V deficiency have decreased factor Xa binding sites on their platelets. J Clin Invest. 1978;62(4):824831. 24. Miletich JP, Jackson CM, Majerus PW. Properties of the factor Xa binding site on human platelets. J Biol Chem. 1978;253(19): 6908-6916. 25. Tracy PB, Giles AR, Mann KG, Eide LL, Hoogendoorn H, Rivard GE. Factor V (Quebec): a bleeding diathesis associated with a qualitative platelet factor V deficiency. J Clin Invest. 1984;74(4):1221-1228. 26. Grenier D, Samama MM, Chtourou S, Plantier J-L. Reverting Xarelto© (rivaroxaban) effect by activated factor X: assessment using thrombin generation assay. Blood. 2014;124(21):2868. 27. Ivanciu L, Toso R, Margaritis P, et al. A zymogen-like factor Xa variant corrects the coagulation defect in hemophilia. Nat Biotechnol. 2011;29(11):1028-1033. 28. Peyvandi F, Garagiola I, Biguzzi E. Advances in the treatment of bleeding disorders. J Thromb Haemost. 2016;14(11):2095-2106. 29. Bauduer F, de RE, Boyer-Neumann C, et al. Factor XI replacement for inherited factor XI deficiency in routine clinical practice: results of the HEMOLEVEN prospective 3-year postmarketing study. Haemophilia. 2015;21 (4):481-489. 30. Bryson CJ, Jones TD, Baker MP. Prediction of immunogenicity of therapeutic proteins: validity of computational tools. BioDrugs. 2010;24(1):1-8. 31. Mahlangu JN, Weldingh KN, Lentz SR, et al. Changes in the amino acid sequence of the recombinant human factor VIIa analog, vatreptacog alfa, are associated with clinical immunogenicity. J Thromb Haemost. 2015;13(11):1989-1998. 32. Lentz SR, Ehrenforth S, Karim FA, et al. Recombinant factor VIIa analog in the management of hemophilia with inhibitors: results from a multicenter, randomized, controlled trial of vatreptacog alfa. J Thromb Haemost. 2014;12(8):1244-1253.

2343