Haematologica, Volume 105, Issue 12 by Haematologica

haematologica Journal of the Ferrata Storti Foundation

Editor-in-Chief Jacob M. Rowe (Haifa)

Deputy Editor Carlo Balduini (Pavia), Jerry Radich (Seattle)

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), David C. Rees (London), Francesco Rodeghiero (Vicenza), Davide Rossi (Bellinzona), Gilles Salles (New York), Aaron Schimmer (Toronto), Richard F Schlenk (Heidelberg), Sonali Smith (Chicago)

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 2020 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 12: December 2020 About the Cover 2699

100-year-old Haematologica images: the contribution of Camillo Golgi to the first issue Paolo Mazzarello https://doi.org/10.3324/haematol.2020.272948

Editorials 2700

Hematopoietic stem and progenitor cells take the route through the bone marrow endothelium Lydia Kalafati and Triantafyllos Chavakis https://doi.org/10.3324/haematol.2020.262113

2702

T(o) be, or (not) to B, or both? Somatically mutated clonal T cells in common variable immunodeficiency and related immunodeficiencies Fumihiro Ishida and Hideyuki Nakazawa https://doi.org/10.3324/haematol.2020.261982

2704

Hypoxia-inducible erythropoietin expression: details matter Thomas Kietzmann https://doi.org/10.3324/haematol.2020.261966

2706

COVID-19: risk of infection is high, independently of ABO blood group Willy Albert Flegel https://doi.org/10.3324/haematol.2020.266593

2708

Expanding dasatinib beyond KIT in acute myeloid leukemia John S. Welch https://doi.org/10.3324/haematol.2020.262147

2710

Convalescent plasma for administration of passive antibodies against viral agents Giovanni Di Minno et al. https://doi.org/10.3324/haematol.2020.267427

Centenary Review Article 2716

History of hematopoietic cell transplantation: challenges and progress Noa Granot and Rainer Storb https://doi.org/10.3324/haematol.2019.245688

Review Articles 2730

Why is it critical to achieve a deep molecular response in chronic myeloid leukemia? Susan Branford https://doi.org/10.3324/haematol.2019.240739

2738

Discontinuation of tyrosine kinase inhibitors in chronic myeloid leukemia: when and for whom? Ehab Atallah and Charles A Schiffer https://doi.org/10.3324/haematol.2019.242891

Articles Hematopoiesis

2746

Hematopoietic stem and progenitor cells use podosomes to transcellularly cross the bone marrow endothelium Timo Rademakers et al. https://doi.org/10.3324/haematol.2018.196329

Haematologica 2020; vol. 105 no. 12 - December 2020 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

Immunodeficiency

2757

Somatic mutations and T-cell clonality in patients with immunodeficiency Paula Savola et al. https://doi.org/10.3324/haematol.2019.220889

Red Cell Biology & its Disorders

2769

No evidence of hemoglobin damage by SARS-CoV-2 infection Anthony W. DeMartino et al. https://doi.org/10.3324/haematol.2020.264267

2774

Distal and proximal hypoxia response elements co-operate to regulate organ-specific erythropoietin gene expression Ilaria M. C. Orlando et al. https://doi.org/10.3324/haematol.2019.236406

Myelodysplastic Syndromes

2785

Eltrombopag monotherapy can improve hematopoiesis in patients with low to intermediate risk-1 myelodysplastic syndrome Alana Vicente et al. https://doi.org/10.3324/haematol.2020.249995

Acute Myeloid Leukemia

2795

Dasatinib response in acute myeloid leukemia is correlated with FLT3/ITD, PTPN11 mutations and a unique gene expression signature Sigal Tavor et al. https://doi.org/10.3324/haematol.2019.240705

Non-Hodgkin Lymphoma

2805

Treatment of patients with MYC rearrangement positive large B-cell lymphoma with R-CHOP plus lenalidomide: results of a multicenter phase II HOVON trial Martine E.D. Chamuleau et al. https://doi.org/10.3324/haematol.2019.238162

Plasma Cell Disorders

2813

Targeting CD47/TNFAIP8 by miR-155 overcomes drug resistance and inhibits tumor growth through induction of phagocytosis and apoptosis in multiple myeloma Nasrin Rastgoo et al. https://doi.org/10.3324/haematol.2019.227579

Hemostasis

2824

Functional plasminogen activator inhibitor 1 is retained on the activated platelet membrane following platelet activation Gael B. Morrow et al. https://doi.org/10.3324/haematol.2019.230367

Transfusion Medicine

2834

Mortality reduction in 46 patients with severe COVID-19 treated with hyperimmune plasma. A proof-of-concept, single-arm, multicenter trial Cesare Perotti et al. https://doi.org/10.3324/haematol.2020.261784

Letters to the Editor 2841

ABO blood groups are not associated with the risk of acquiring SARS-CoV-2 infection in young adults Laurys Boudin et al. https://doi.org/10.3324/haematol.2020.265066

2844

Perforin gene variant A91V in young patients with severe COVID-19 Oscar Cabrera-Marante et al. https://doi.org/10.3324/haematol.2020.260307

2847

Complement C5 inhibition in patients with COVID-19 - a promising target? Regis Peffault de Latour et al. https://doi.org/10.3324/haematol.2020.260117

2851

The proteome of neutrophils in sickle cell disease reveals an unexpected activation of the interferon alpha signaling pathway Patricia Hermand et al. https://doi.org/10.3324/haematol.2019.238295

Haematologica 2020; vol. 105 no. 12 - December 2020 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

2855

Plasticity in growth behavior of patientsâ&#x20AC;&#x2122; acute myeloid leukemia stem cells growing in mice Sarah Ebinger et al. https://doi.org/10.3324/haematol.2019.226282

2861

Donor cell leukemia: is reappearance of gene mutations in donor cells more than an incidental phenomenon? Tal Shahar Gabay et al. https://doi.org/10.3324/haematol.2019.242347

2864

Chronic lymphocytic leukemias with trisomy 12 show a distinct DNA methylation profile linked to altered chromatin activation Maria Tsagiopoulou et al. https://doi.org/10.3324/haematol.2019.240721

2868

Lenalidomide, dexamethasone and alemtuzumab or ofatumumab in high-risk chronic lymphocytic leukemia: final results of the NCRI CLL210 trial Andrew R. Pettitt et al. https://doi.org/10.3324/haematol.2019.230805

2872

Clinical characteristics and outcome of multiple myeloma patients with concomitant COVID-19 at Comprehensive Cancer Centers in Germany Monika Engelhardt, et al. https://doi.org/10.3324/haematol.2020.262758

2879

Ixazomib-thalidomide-low dose dexamethasone induction followed by maintenance therapy with ixazomib or placebo in newly diagnosed multiple myeloma patients not eligible for autologous stem cell transplantation; results from the randomized phase II HOVON-126/NMSG 21.13 trial Sonja Zweegman et al. https://doi.org/10.3324/haematol.2019.240374

Case Reports 2883

Apparent recessive inheritance of sideroblastic anemia type 2 due to uniparental isodisomy at the SLC25A38 locus Immacolata Andolfo et al. https://doi.org/10.3324/haematol.2020.258533

2887

Eculizumab for complement mediated thrombotic microangiopathy in sickle cell disease Satheesh Chonat et al. https://doi.org/10.3324/haematol.2020.262006

Comment 2892

SARS-CoV-2 severity in African Americans â&#x20AC;&#x201C; a role for Duffy null? Robert P. Hebbel and Gregory M. Vercellotti https://doi.org/10.3324/haematol.2020.269415

Haematologica 2020; vol. 105 no. 12 - December 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 contribution of Camillo Golgi to the first issue Paolo Mazzarello Department of Brain and Behavioral Sciences and the University Museum System, University of Pavia, Italy E-mail: PAOLO MAZZARELLO - paolo.mazzarello@unipv.it doi:10.3324/haematol.2020.272948

amillo Golgi, despite having been a pupil of Giulio Bizzozero, one of the founders of hematology, did not have a primary interest in the study of blood.1-3 However, in his scientific work he dealt with important hematologic problems. In 1873, Golgi described the alterations of the bone marrow in smallpox. In 1880, he treated cases of anemia with peritoneal transfusions and, from 1885, he studied the alterations of the blood in the course of malarial infection.3 The foundation of Haematologica gave him the opportunity to publish a couple of works that were, instead, of a purely hematologic nature.4,5 The first of these two papers inaugurated the new journal in 1920 and was preceded, on 12th June 1919, by a lecture

given by Golgi to the Medical-Surgical Society of Pavia.4 This study explored a new coloring method based on gold chloride. In the erythrocytes, Golgi observed a "circumscribed rounded area, with clear boundaries and with different shades of color, from red to more or less intense brown" which had "a finely dotted appearance" or, sometimes, "with a hint of streak and a very tenuously fibrillar constitution". This suggested the existence of a nucleus, although an atypical one. Golgi immediately distanced himself from some researchers who had actually supported this thesis by noting that the reaction to gold chloride was negative when tested in erythrocytes of fish, birds, reptiles, amphibians, and mammalian embryonic red blood cells, all elements with nuclei. Finally, the method used did not stain the nuclear substance of the white blood cells. The possibility of the presence of a nucleus in human erythrocytes had been advanced by Angelo Petrone in 1897, but two years later this was refuted in Golgi's laboratory by his pupil Adelchi Negri. The idea continued to be a subject for discussion by some hematologists and was revived on the basis of new observations made by Petrone. Golgi's skepticism appeared very timely, even if he was unable to put forward alternative hypotheses on the nature of what he had observed. These features were later considered to be artifacts related to the staining methods rather than morphological structures that actually exist in red blood cells.3 Golgi's article also described the centrosome in white blood cells, while in a subsequent note, again published in Haematologica, he discussed the problem of the possible existence of centrosomes in erythrocytes.5 This research reveals much of Golgi's tenacious and passionately devoted approach to his laboratory research. He was seventy-seven years old at the time, had won the Nobel Prize for Medicine fourteen years earlier, had been a Senator of the Kingdom of Italy for twenty years, had been awarded an honorary degree by the University of Cambridge in 1898, and was twice Rector of the University of Pavia. Yet he still devoted himself with passion to research in fields new to him and he took part in the meetings of the Medical-Surgical Society of Pavia, reporting his studies on minute morphological peculiarities of red blood cells of secondary interest. Evidently, continuing to work in the lab was really the way in which Golgi still felt mentally young and alive.

References Figure 1. On the pseudo-nucleus of erythrocytes. Hand-drawn color drawing illustrating the first article by Camillo Golgi published in Haematologica.4 Several red blood cells (RBC), under the action of a special gold chloride dye used by Golgi, show circumscribed thickenings that resemble a nucleus. Fig 1: the two white blood cells have a discolored nucleus. Fig. 2: RBC are seen at the beginning of the reaction with a discolored circular area in the center while in Fig 3 the intensely colored erythrocytes are seen 15-20 days after the start of the reaction. Fig 4: fetal RBC. The erythrocyte at bottom left is nucleated, but the nucleus remains discolored. Fig 5: the RBC have a finely fibrillar appearance in the center.

haematologica | 2020; 105(12)

1. Mazzarello P, Calligaro AL, Calligaro A. Giulio Bizzozero: a pioneer of cell biology. Nat Rev Mol Cell Biol. 2001;2(10):776-781. 2. Mazzarello P. Golgi. Oxford & New York: Oxford University Press; 2010. 3. Mazzarello P. One hundred years of Haematologica. Haematologica. 2020;105(1):12-21. 4. Golgi C. [Sulla struttura dei globuli rossi dell’uomo e di altri mammiferi]. Haematologica. 1920;1(1):1-16. 5. Golgi C. [Il centrosoma dei globuli rossi del sangue circolante dell’uomo e di altri animali]. Haematologica. 1920;1:333-359.

2699

EDITORIALS Hematopoietic stem and progenitor cells take the route through the bone marrow endothelium Lydia Kalafati1,2 and Triantafyllos Chavakis1 1

Institute for Clinical Chemistry and Laboratory Medicine, University Hospital and Faculty of Medicine, Technische UniversitĂ¤t Dresden, Dresden and 2National Center for Tumor Diseases, Partner Site Dresden, Dresden and German Cancer Research Center Heidelberg, Heidelberg, Germany E-mail: TRIANTAFYLLOS CHAVAKIS - triantafyllos.chavakis@uniklinikum-dresden.de doi:10.3324/haematol.2020.262113

ematopoietic stem cell transplantation (HSCT) is an established therapeutic approach for various hematologic diseases, particularly hematologic malignancies. A prerequisite for successful HSCT is the recruitment of infused hematopoietic stem and progenitor cells (HSPC) to the bone marrow (BM). A critical component in this process is the transmigration of HSPC through the BM endothelium. But how do HSPC actually cross the BM endothelium? Given that HSCT has been performed in the clinics for decades now, surprisingly little is known about the mechanisms underlying the transmigration of HSPC through the BM endothelium. A study by Rademakers et al.1 in this issue of Haematologica provides major novel insights into this important question. In particular, the authors show that HSPC engage preferentially in the transcellular route to migrate through the BM endothelium by using podosome-like structures. Hematopoiesis, i.e., the generation of blood cells, takes place in the BM. Hematopoietic stem cells (HSC) lie at the top of the hematopoietic hierarchy and are characterized by the capacity of self-renewal and multi-lineage differentiation giving rise to all mature blood cells.2 HSPC play also an important role in demand-adapted hematopoiesis, as they respond rapidly to inflammatory or infectious challenges by proliferative expan-

sion and enhanced differentiation to myeloid cells.3 The function of HSPC in the BM is supported by the special environment of the hematopoietic niche that consists of the extracellular matrix and several cells, such as endothelial cells and mesenchymal stromal cells.4 The niche facilitates HSPC functions, including their maintenance and their differentiation via different mechanisms.4 While vascular endothelial cells in the BM contribute to the HSC niche, they also represent the main barrier that HSC need to cross in order to arrive in the BM and repopulate the BM niche in the context of HSCT, which is an established treatment for several hematologic disorders, especially hematologic malignancies.5,6 HSC engraftment in the BM in HSCT is regulated by various interactions with different components of the BM niche,4 however, less is known regarding the mechanisms that facilitate the initial BM homing of HSC upon transplantation.7 The BM vascular endothelium expresses various adhesion molecules that facilitate adhesion of circulating cells.7 As the shear stress in sinusoidal vessels is significantly lower than in arterioles, BM sinusoids are a preferential site for HSPC homing.8 The adhesive interactions of HSPC with the BM endothelium share major similarities with the leukocyte adhesion cascade, which mediates leukocyte recruitment in inflammation.9

Figure 1 Hematopoietic stem and progenitor cells bone marror homing in the context of hematopoietic stem cell transplantation. Upon hematopoietic stem cell transplantation (HSCT), the hematopoietic stem and progenitor cells (HSPC) extravasate through the bone marrow (BM) vascular endothelium enter and repopulate the BM niche. The BM endothelium expresses several adhesion molecules, such as selectins and VCAM-1, which facilitate rolling and firm adhesion of HSPC on the endothelial cells. Subsequently, the HSPC preferentially migrate through the endothelial cell body, designated as transcellular transmigration. VE-cadherin regulates vascular endothelial integrity by stabilizing the endothelial cell-cell junctions. Following inhibition of VE-cadherin, vascular permeability increases and HSPC may also migrate through the endothelial cell junctions in a paracellular manner.

2700

haematologica | 2020; 105(12)

Editorials

Specifically, the rolling of HSPC and their subsequent firm adhesion onto the vascular endothelial cells of the BM are mediated by endothelium-expressed selectins and vascular cell adhesion molecule-1 (VCAM-1), respectively.10,11 VCAM-1 on the BM endothelium interacts primarily with integrin α4b1, also named very late antigen-4 (VLA-4) on HSPC.11 Furthermore, the interaction between integrin α4b7 on HSPC and mucosal addressin cell adhesion molecule 1 (MAdCAM1) on the BM endothelium facilitates HSPC recruitment in the context of transplantation.12 After their adhesion on the endothelium, HSPC extravasate into the BM niche. In the leukocyte adhesion cascade, immune cells may take two distinct routes to migrate across the endothelial barrier. The paracellular route is through the cell-cell junctions, while the transcellular route occurs through the endothelial cell body, with paracellular transendothelial migration being the predominant pathway at least for innate immune cells.13 However, lymphocytes may preferentially engage in the transcellular transmigration route.14 The choice between transcellular and paracellular transmigration may be context-dependent; different immune cells may have a preference for one route over the other, however, this choice may also depend on the type of endothelium in different vascular beds and tissues.15 The major regulator of endothelial cell-cell junctions and vascular integrity is vascular endothelial (VE)-cadherin;16 VE-cadherin is, therefore also a modulator of the paracellular transmigration route. Antibodies blocking VE-cadherin can promote increased vessel permeability and enhanced leukocyte diapedesis.17,18 The exact role of VE-cadherin in HSPC transendothelial migration, however remains elusive. An older in vitro study by van Buul et al.19 showed that inhibition of VE-cadherin in human BM endothelial cells not only increased their permeability but also promoted the transendothelial migration of human CD34+ hematopoietic progenitor cells.19 The same group followed up on this previous finding in their article by Rademakers and collegues,1 published in this issue of Haematologica. In this article, the authors provide important novel insights into the role of VE-cadherin for the vascular integrity in the BM and on the mechanisms that HSPC employ to cross the BM endothelial barrier. They show that VE-cadherin is a major regulator of vascular endothelial permeability in the BM.1 Antibodymediated inhibition of VE-cadherin led to increased permeability in both sinusoids and arterioles in the BM under steady-state conditions and upon low-dose irraditation. Conversely, VE-cadherin-α-catenin chimera knock-in animals, in which endothelial junctions are stabilized by the expression of the VE-cadherin-alpha-catenin fusion protein,20 displayed reduced BM vascular permeability under normal conditions and upon irradiation. Additionally, BM homing of HSPC transferred to low-dose irradiated mice was enhanced in the presence of blocking antibody against VE-cadherin.1 Surprisingly, homing of HSPC to the BM was not altered in VE-cadherin-α-catenin chimera knock-in mice. This finding could be attributed to the fact that HSPC engage preferentially the transcellular route to cross the BM endothelial barrier. Hence, transendothelial migration of HSPC is regulated in a different way than vascular permeability in the BM and only the latter is strictly dependent on VE-cadherin.1 Further mechanistic experihaematologica | 2020; 105(12)

ments suggested that podosome structures on HSPC mediate their transendothelial migration. Despite the enhanced HSPC recruitment to the BM upon antibody-mediated blockade of VE-cadherin long-term engraftment of HSPC in the BM was not improved. Hence, blocking VE-cadherin does not necessarily promote hematopoietic regeneration in the context of transplantation. Taken together, this article provides an important novel insight into the mechanisms underlying HSPC homing to the BM in the context of transplantation. The findings may be of importance with regards to developing strategies to enhance HSPC homing and increase the efficacy of therapeutic HSC transplantation.

References 1. Rademakers T, Goedhart M, Hoogenboezem M, et al, Hematopoietic stem and progenitor cells use podosomes to transcellularly cross the bone marrow endothelium. Haematologica. 105(12):2746-2756. 2. Jacobsen SEW, Nerlov C. Haematopoiesis in the era of advanced single-cell technologies. Nat Cell Biol, 2019;21(1):2-8. 3. Chavakis T, Mitroulis I, Hajishengallis G. Hematopoietic progenitor cells as integrative hubs for adaptation to and fine-tuning of inflammation. Nat Immunol. 2019;20(7):802-811. 4. Wei Q, Frenette PS. Niches for hematopoietic stem cells and their progeny. Immunity. 2018;48(4):632-648. 5. To LB, Haylock DN, Simmons PJ, Juttner CA. The biology and clinical uses of blood stem cells. Blood. 1997;89(7):2233-2258. 6. Copelan EA. Hematopoietic stem-cell transplantation. N Engl J Med. 2006;354(17):1813-1826. 7. Lapidot T, Dar A, Kollet O. How do stem cells find their way home? Blood. 2005;106(6):1901-1910. 8. Perlin JR, Sporrij A, Zon LI. Blood on the tracks: hematopoietic stem cell-endothelial cell interactions in homing and engraftment. J Mol Med (Berl). 2017;95(8):809-819. 9. Nourshargh S, Alon R. Leukocyte migration into inflamed tissues. Immunity. 2014;41(5):694-707. 10. Frenette PS, Subbarao S, Mazo IB, von Andrian UH, Wagner DD. Endothelial selectins and vascular cell adhesion molecule-1 promote hematopoietic progenitor homing to bone marrow. Proc Natl Acad Sci U S A. 1998;95(24):14423-14428. 11. Papayannopoulou T, Craddock C, Nakamoto B, Priestley GV, Wolf NS. The VLA4/VCAM-1 adhesion pathway defines contrasting mechanisms of lodgement of transplanted murine hemopoietic progenitors between bone marrow and spleen. Proc Natl Acad Sci U S A. 1995;92 (21):9647-9651. 12. Katayama, Y, Hidalgo A, Peired A, Frenette PS. Integrin alpha4beta7 and its counterreceptor MAdCAM-1 contribute to hematopoietic progenitor recruitment into bone marrow following transplantation. Blood. 2004,104(7):2020-6. 13. Vestweber D, Wessel F, Nottebaum AF. Similarities and differences in the regulation of leukocyte extravasation and vascular permeability. Semin Immunopathol. 2014;36(2):177-192. 14. Carman CV. Mechanisms for transcellular diapedesis: probing and pathfinding by 'invadosome-like protrusions'. J Cell Sci. 2009;122(Pt 17):3025-3035. 15. Martinelli R, Zeiger AS, Whitfield M, et al. Probing the biomechanical contribution of the endothelium to lymphocyte migration: diapedesis by the path of least resistance. J Cell Sci. 2014;127(Pt 17):3720-3734. 16. Lagendijk AK, Hogan BM. VE-cadherin in vascular development: a coordinator of cell signaling and tissue morphogenesis. Curr Top Dev Biol. 2015;112:325-352. 17. Corada M, Mariotti M, Thurston G, et al. Vascular endothelial-cadherin is an important determinant of microvascular integrity in vivo. Proc Natl Acad Sci U S A. 1999;96(17):9815-9820. 18. Gotsch U, Borges E, Bosse R, et al. VE-cadherin antibody accelerates neutrophil recruitment in vivo. J Cell Sci. 1997;110( Pt 5):583-588. 19. van Buul JD, Voermans C, van den Berg V, et al. Migration of human hematopoietic progenitor cells across bone marrow endothelium is regulated by vascular endothelial cadherin. J Immunol. 2002;168(2):588-596. 20. Schulte D, Küppers V, Dartsch N, et al. Stabilizing the VE-cadherincatenin complex blocks leukocyte extravasation and vascular permeability. EMBO J. 2011;30(20):4157-4170.

2701

Editorials

T(o) be, or (not) to B, or both? Somatically mutated clonal T cells in common variable immunodeficiency and related immunodeficiencies Fumihiro Ishida1 and Hideyuki Nakazawa2 1

Academic Assembly School of Medicine and Health Sciences, Institute of Health Science and School of Medicine and Department of Biomedical Laboratory Sciences, Shinshu University and 2Division of Hematology, Department of Internal Medicine, Shinshu University School of Medicine, Matsumoto, Japan E-mail: FUMIHIRO ISHIDA- fumishi@shinshu-u.ac.jp doi:10.3324/haematol.2020.261982

n common variable immunodeficiency (CVID), autoimmune diseases and lymphoproliferative disorders (LPD) often develop in addition to recurrent infections due to hypogammaglobulinemia and the decreased number of antigen-specific memory B cells.1,2 While various T-cell abnormalities, such as CD8+ T-cell expansion and suppressed regulatory T cells have been observed in CVID,3 their pathophysiological backgrounds are unknown. In this issue of Haematologica, Savola et al.4 investigated the somatic mutations of T cells from patients with congenital immunodeficiency, including CVID, using deep amplicon sequencing with 2,355 gene panels and a T-cell receptor (TCR) b gene analysis to seek possible relation-

ships between genetic alterations and T-cell abnormalities in these diseases. They found that 6 of 8 patients with CIVD harbored somatically mutated T cells and, in total, 59% of patients with congenital immunodeficiency were positive for somatic mutations in CD4+ or CD8+ T cells, which would be expected to have deleterious effects on the cellular functions of T cells (Figure 1). Clonal hematopoiesis-related gene mutations, including DNMT3A, were found in CD3+ T cells from 24% of the patients. T-cell somatic mutations were also identified, albeit less frequently, in age-matched heathy controls. Patients with immunodeficiency had more convergent, namely restricted, TCR b chain CDR3 sequences, although these were not specific to previously known

Figure 1. Germline mutations, abnormalities of B cells and somatically mutated clonal T cells identified in common variable immunodeficiency and other immunodeficiencies by Savola et al.4 Savola et al.4 demonstrated somatic mutation of selected genes in the T cells of patients with CVID and related immunodeficiencies. Somatically mutated clonal T cells would lead to T-cell abnormalities and might contribute to the recurrence of infection due to hypogammaglobulinemia and B-cell lymphoproliferative disorders in patients with these immunodeficiencies.

2702

haematologica | 2020; 105(12)

Editorials

antigens. In CD8+ T cells, the somatically mutated gene burden was correlated with the T-cell clone size. The germline mutations of the genes associated with CVID are heterogenous, and only 30-50% of patients with CVID were positive for germline mutations, such as NFKB1 in B-cell-signaling pathways,5,6 while genetic abnormalities are still unknown in a significant proportion of CVID patients. In this study, Savola et al.4 identified germline TAC1 mutations in CVID patients, and STAT3 and ADA2 mutations in other immunodeficient patients. What role the somatically mutated T cells play in CVID and other immunodeficient states associated with these genetic backgrounds remains unclear. Furthermore, it is not known how or to what extent these identified mutations affect T-cell function. Further steps are needed to clarify the pathophysiology of a population of somatically mutated clonal T cells in whole T-cell networks in the settings of autoimmune diseases, LPD, and hypogammaglobulinemia in CVID or related immunodeficiency. The results presented in this paper provide new insights into the T-cell abnormalities of CVID and immunodeficiency, suggesting that clonal T cells with somatic mutations may contribute to the development of B-cell LPD, and that they may be attributed to B-cell abnormalities, such as decreased numbers of isotypeswitched memory B cells, leading to hypogammaglobulinemia in CVID. To date, B-cell dysfunction and reduced concentrations of immunoglobulins have been considered fundamental characteristics of CVID.1,7 Emerging evidence on T-cell abnormalities,3,8,9 including clonal T cells with somatic mutations, in addition to frequent complication of autoimmune diseases and LPD, together with a variety of germline gene mutations, underlies the heterogeneity and complex nature of CVID and related immunodeficiencies. Beyond this work, recent evidence shows that somatic mutations of non-neoplastic cells are indeed relevant in various diseases.10-12 Clonal hematopoiesis of myeloid lineage cells in association with bone marrow failure, the development of hematologic neoplasms and arteriosclerosis is one of the issues in the field of hematology.13-15 Now congenital immunodeficiency has been added to the list. Approaches such as a single cell analysis would provide further insight into the inherent genetic instability of T cells in association with the mechanism of TCR rearrangement, clonal hematopoiesis, or any other novel system pertinent to somatic mutations in T cells.

haematologica | 2020; 105(12)

Acknowledgments FI is supported by Kaken 20K08709 from Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

References 1. Bonilla FA, Barlan I, Chapel H, et al. International consensus document (ICON): common variable immunodeficiency disorders. J Allergy Clin Immunol Pract. 2016;4(1):38-59. 2. Yakaboski E, Fuleihan RL, Sullivan KE, Cunningham-Rundles C, Feuille E. Lymphoproliferative disease in CVID: a report of types and frequencies from a US patient registry. J Clin Immunol. 2020;40(3):524-530. 3. Wong GK, Huissoon AP. T-cell abnormalities in common variable immunodeficiency: the hidden defect. J Clin Pathol. 2016;69(8):672676. 4. Savola P, Martelius T, Kankainen M, et al. Somatic mutations and Tcell clonality in patients with immunodeficiency. Haematologica. 2020;105(12):2757-2768. 5. Abolhassani H, Hammarstrom L, Cunningham-Rundles C. Current genetic landscape in common variable immune deficiency. Blood. 2020;135(9):656-667. 6. Seidel MG, Kindle G, Gathmann B, et al. The European Society for Immunodeficiencies (ESID) registry working definitions for the clinical diagnosis of inborn errors of immunity. J Allergy Clin Immunol Pract. 2019;7(6):1763-1770. 7. Ameratunga R, Lehnert K, Woon ST, et al. Review: diagnosing common variable immunodeficiency disorder in the era of genome sequencing. Clin Rev Allergy Immunol. 2018;54(2):261-268. 8. Ramesh M, Hamm D, Simchoni N, Cunningham-Rundles C. Clonal and constricted T cell repertoire in common variable immune deficiency. Clin Immunol. 2017;178:1-9. 9. Le Saos-Patrinos C, Loizon S, Blanco P, Viallard JF, Duluc D. Functions of Tfh cells in common variable immunodeficiency. Front Immunol. 2020;11:6. 10. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371(26):2488-2498. 11. Vijg J, Dong X. Pathogenic mechanisms of somatic mutation and genome mosaicism in aging. Cell. 2020;182(1):12-23. 12. Savola P, Kelkka T, Rajala HL, et al. Somatic mutations in clonally expanded cytotoxic T lymphocytes in patients with newly diagnosed rheumatoid arthritis. Nat Commun. 2017;8:15869. 13. Jaiswal S, Natarajan P, Silver AJ, et al. Clonal hematopoiesis and risk of atherosclerotic cardiovascular disease. N Engl J Med. 2017;377 (2):111-121. 14. Yoshizato T, Dumitriu B, Hosokawa K, et al. Somatic mutations and clonal hematopoiesis in aplastic anemia. N Engl J Med. 2015;373(1):35-47. 15. Shen W, Clemente MJ, Hosono N, et al. Deep sequencing reveals stepwise mutation acquisition in paroxysmal nocturnal hemoglobinuria. J Clin Invest. 2014;124(10):4529-4538.

2703

Editorials

Hypoxia-inducible erythropoietin expression: details matter Thomas Kietzmann Faculty of Biochemistry and Molecular Medicine, Biocenter Oulu, University of Oulu, Oulu, Finland E-mail: THOMAS KIETZMANN - thomas.kietzmann@oulu.fi doi:10.3324/haematol.2020.261966

ecombinant human erythropoietin (rhEPO) and its derivatives are some of the most important treatment options against anemia associated with chronic kidney disease, chemotherapy, and low-risk myelodysplastic syndrome.1 On the other hand, it has also gained fame as a doping agent in endurance sports. All this became possible by translating basic research findings that started more than 100 years ago with the recognition of an increase in red blood cell number in animals living at high altitude (i.e., at a decreased pO2) and that went, via purification, partial sequencing, cloning, and basic understanding of EPO gene regulation, into clinics (for excellent reviews of EPO history see2-5). With its cloning and the fast-improving molecular biological tools, findings from hypoxia-inducible EPO gene regulation also became a paradigm to understand basic principles of how cells sense and adapt to oxygen availability. The discovery of those principles were acknowledged by part of the 2019 Nobel prize in Physiology or Medicine.6 The major production sites of EPO are peritubular fibroblasts in the kidney as well as hepatocytes and hepatic stellate cells (about 90% of circulating EPO stems from kidney and 10% from liver). But EPO expression can also be found in brain, testis, uterus, and osteoblasts.7-10 Apart from its major function as a driver of erythropoiesis, EPO was also found to exert organ- and tissuerestricted protective functions in the brain, cardiovascular system, adipose tissue, and bones.11,12 Experiments with transgenic mice revealed that critical parts of the EPO gene that are located away from the proper EPO coding exons in the distal 5’- and 3’-flanking

regions are responsible for tissue-specific and hypoxiadriven EPO gene expression. Up to then, it had been known that an array consisting of the kidney-inducible element (KIE; between -14 kb to -9.5 kb 5’ from the promoter) and a negative regulatory element (NRE, between -6kb to -0.4kb) in the 5’-part seem to be of special importance for EPO expression in kidney (Figure 1). By contrast, in liver, hypoxia-inducible EPO expression appeared to depend largely on an enhancer in the so-called liverinducible element (LIE) that was found 3’-distal from the EPO polyadenylation side (Figure 1). That area contained, among other functional sites, an HRE (hypoxia response element) that served as a binding site for a hypoxiainducible factor (HIF).13-16 From the three HIFs known, HIF-2α has been defined as a major part of the hypoxiainducible EPO gene activating complex.17,18 While the role of the HRE in the 3’-enhancer is well established and found to be necessary and sufficient to confer liver-specific EPO gene expression, the DNA element responsible for kidney-specific EPO gene expression is far less well characterized and nothing was known about the presence and function of HREs in EPO-producing none-kidney tissues such as neurons of the brain. This problem has been tackled by Orlando and co-workers, and the outcome of their studies is described in an article in the current issue of Haematologica.19 In their studies, the authors built on findings where they discovered a distal 5’-HRE within a DNaseI hypersensitive site -9.2 kb upstream of the EPO transcriptional start site which was supposed to contribute to oxygenregulated EPO expression in the kidney.20 As the EPOproducing cells in the kidney derive from neural crest and

Figure 1. Hypoxia-regulated human erythropoietin (EPO) gene expression in neuronal and liver cells. Hypoxia leads to stabilization of hypoxia-inducible factor (HIF), and mainly HIF-2 containing complexes are able to induce EPO gene expression. In neuronal cells, HIF-containing complexes bind most strongly to the newly identified hypoxia response element (HRE) in the promoter and to a lesser extent to the 3'-HRE. Although HIF-complexes appear not to bind the 5'-HRE, the 5´-HRE co-operates in the hypoxic induction process. In liver cells, HIF-complexes bind mainly to the 3´-HRE, weakly to the pHRE, and not to the 5´- HRE. Lighter colors indicate less HIF binding; green arrows indicate co-operation between the HREs. KIE: kidney-inducible element; NRE: negative regulatory element; LIE: liver-inducible element; HRE: hypoxia response element; NRLE: negative regulatory liver element; Prom: promoter; pHRE: promoter HRE functioning as one entity but consisting of a tandem dimeric repeat with two HRE sequences.

2704

haematologica | 2020; 105(12)

Editorials

neuroepithelial cells,21 the authors went on to compare the function of this part containing the 5’ HRE and the established 3’-HRE in human neuroblastoma cells and liver hepatocellular carcinoma cells; both cell types are able to express EPO in response to hypoxia.22 To do this, they used gene editing by CRISPR/Cas9 to specifically mutate the -9.2 kb 5’-HRE and the +3.0 kb 3’HRE in the EPO gene locus, and combined this with subsequent HIF-DNA interaction studies, RNA and protein expression measurements and reporter gene assays. Intriguingly, when analyzing EPO mRNA and secreted EPO protein levels upon exposure of the engineered cells to hypoxia (0.2% O2) for 24 hours, the authors found that the neuroblastoma cells with mutations in both the 5’HRE and the 3’-HRE did not increase EPO mRNA and protein levels to the same extent as wild-type cells. By contrast, only mutation of the 3’-HRE reduced the hypoxic EPO induction in hepatoma cells. When the authors expanded these experiments by transiently transfecting cells with minimal EPO promoter-driven reporter genes, enhanced by various DNA fragments containing the 5’- and/or 3’-HRE, they found that EPO promoter-driven reporter gene activity in neuroblastoma cells was best when a 100 bp 5’-region containing the HRE was combined with a 126 bp HRE containing 3’fragment. Further experiments with different EPO promoter HRE constructs indicated that the most efficient hypoxiadependent EPO induction requires co-operation between the EPO 5’- and 3’-HRE specifically in neuronal cells, and that additional distal and proximal 5’-flanking elements further contribute to tissue-specific and conditional EPO regulation. In addition, the minimal EPO gene promoter alone was able to promote hypoxia-dependent reporter gene activation in neuroblastoma cells and hepatoma cells indicating that this part may contain previously unidentified HRE. Indeed, the authors found a tandem dimeric repeat with two HRE sequences that had not previously been reported. Importantly, mutation of each single HRE reduced reporter gene activity, but simultaneous mutation of both HREs did not have an additional effect, suggesting that both HREs behave as an entity. Next, the authors were able to link the observed findings to HIF binding. Chromatin immunoprecipitation assays in neuroblastoma cells revealed that none of the HIF subunits bound to the 5’-HRE, whereas a significant hypoxic increase in HIF-2α/HIF-b binding to the 3’-HRE could be detected. By contrast, in hepatocellular carcinoma cells, HIF-2α/HIF-b binding could be detected with the promoter region and the 3’-HRE. Importantly, nonEPO producing cells did not show any HIF binding to the examined regions of the EPO gene. Further experiments with the CRISPR/Cas9 engineered cells showed that mutation of the 3’-HRE also decreased HIF binding to the promoter region; vice versa, mutation of the 5’-HRE impaired HIF binding to the 3’-HRE. Taken together, the study shows that hypoxic EPO expression in hepatic cells appears only to depend on HIF interaction with the 3'-HRE and not on the 5'-HRE or the HIF binding EPO promoter. By contrast, neuronal EPO expression seems to require co-operation with intact 5'haematologica | 2020; 105(12)

and 3’-HRE, with HIF interacting strongly with the new HRE in the EPO promoter, and to a lesser extent with the 3' HRE. Another important aspect of this study is that hypoxiadependent gene expression is not a uniform process and needs to be seen and analyzed in a cell type-specific context. This is because some cells’ HIF may not bind to all HREs, although they may confer oxygen sensitivity. This may be of importance in gene-wide association studies where different SNPs or disease-associated polymorphisms may represent HREs which may, depending on the context, gain or lack their function. Overall, the investigation by Orlando et al.19 presents novel and interesting findings highlighting that a complex interplay between various HREs, HIFs and other factors at the EPO locus contribute to its hypoxia-dependent and organ-specific expression. Acknowledgments TK was supported by the Academy of Finland SA296027, the Jane and Aatos Erkko Foundation, the Finnish Cancer Foundation, the Sigrid Juselius Foundation, the University of Oulu, and Biocenter Oulu.

References 1. Rizzo JD, Brouwers M, Hurley P, et al. American Society of Hematology/American Society of Clinical Oncology clinical practice guideline update on the use of epoetin and darbepoetin in adult patients with cancer. Blood. 2010;116(20):4045-4059. 2. Fandrey J. Oxygen-dependent and tissue-specific regulation of erythropoietin gene expression. Am J Physiol Regul Integr Comp Physiol. 2004;286(6):R977-988. 3. Wenger RH, Kurtz A. Erythropoietin. Compr Physiol. 2011;1(4):1759-1794. 4. Bunn HF. Erythropoietin. Cold Spring Harb Perspect Med. 2013;3(3):a011619. 5. Jelkmann W. Erythropoietin. Front Horm Res. 2016;47:115-127. 6. Kietzmann T. The air that we breeze: from 'Noble' discoveries of a general oxygen-sensing principle to its clinical use. Acta Physiol (Oxf). 2020;228(2):e13416. 7. Marti HH. Erythropoietin and the hypoxic brain. J Exp Biol. 2004;207(Pt 18):3233-3242. 8. Magnanti M, Gandini O, Giuliani L, et al. Erythropoietin expression in primary rat Sertoli and peritubular myoid cells. Blood. 2001;98(9):2872-2874. 9. Yasuda Y, Masuda S, Chikuma M, Inoue K, Nagao M, Sasaki R. Estrogen-dependent production of erythropoietin in uterus and its implication in uterine angiogenesis. J Biol Chem. 1998;273(39):25381-25387. 10. Rankin EB, Wu C, Khatri R, et al. The HIF signaling pathway in osteoblasts directly modulates erythropoiesis through the production of EPO. Cell. 2012;149(1):63-74. 11. Suresh S, Rajvanshi PK, Noguchi CT. The many facets of erythropoietin physiologic and metabolic response. Front Physiol. 2020;10:1534. 12. Peng B, Kong G, Yang C, Ming Y. Erythropoietin and its derivatives: from tissue protection to immune regulation. Cell Death Dis. 2020;11(2):79. 13. Semenza GL, Traystman MD, Gearhart JD, Antonarakis SE. Polycythemia in transgenic mice expressing the human erythropoietin gene. Proc Natl Acad Sci U S A. 1989;86(7):2301-2305. 14. Semenza GL, Dureza RC, Traystman MD, Gearhart JD, Antonarakis SE. Human erythropoietin gene expression in transgenic mice: multiple transcription initiation sites and cis-acting regulatory elements. Mol Cell Biol. 1990;10(3):930-938. 15. Madan A, Lin C, Hatch SL 2nd, Curtin PT. Regulated basal, inducible, and tissue-specific human erythropoietin gene expression in transgenic mice requires multiple cis DNA sequences. Blood. 1995;85(10):2735-2741. 16. Madan A, Curtin PT. A 24-base-pair sequence 3' to the human erythropoietin gene contains a hypoxia-responsive transcriptional

2705

Editorials enhancer. Proc Natl Acad Sci U S A. 1993;90(9):3928-3932. 17. Imeri F, Nolan KA, Bapst AM, et al. Generation of renal Epo-producing cell lines by conditional gene tagging reveals rapid HIF-2 driven Epo kinetics, cell autonomous feedback regulation, and a telocyte phenotype. Kidney Int. 2019;95(2):375-387. 18. Kapitsinou PP, Liu Q, Unger TL, et al. Hepatic HIF-2 regulates erythropoietic responses to hypoxia in renal anemia. Blood. 2010;116(16):3039-3048. 19. Orlando IMC, Lafleur VN, Storti F, et al. Distal and proximal hypoxia response elements co-operate to regulate organ-specific erythropoi-

etin gene expression. Haematologica. 2020;105(12):2774-2784. 20. Storti F, Santambrogio S, Crowther LM, et al. A novel distal upstream hypoxia response element regulating oxygen-dependent erythropoietin gene expression. Haematologica. 2014;99(4):e45-e48. 21. Hirano I, Suzuki N. The neural crest as the first production site of the erythroid growth factor erythropoietin. Front Cell Dev Biol. 2019;7:105. 22. Stolze I, Berchner-Pfannschmidt U, Freitag P, et al. Hypoxiainducible erythropoietin gene expression in human neuroblastoma cells. Blood. 2002;100(7):2623-2628.

COVID-19: risk of infection is high, independently of ABO blood group Willy Albert Flegel1,2 1

Department of Transfusion Medicine, NIH Clinical Center, National Institutes of Health, Bethesda, MD, USA and 2Huazhong University of Science and Technology, Wuhan, Hubei, China E-mail: WILLY ALBERT FLEGEL - waf@nih.gov doi:10.3324/haematol.2020.266593

hen a French aircraft carrier set sail on 22 January, 2020 for a mission of several months, its 1,769 crewmembers were unaware of a stowaway in the form of a novel virus. The SARS-CoV-2 virus, assumed in early 2020 to be a recent arrival in Europe, was already on board. Upon the ship’s return to Toulon, the main naval base of France on the Mediterranean Sea, most of the crew were confined to their barracks and 1,688 sailors participated in health monitoring. In this issue of Haematologica, Boudin and colleagues report data from this unique epidemiological setting,1 which could have hardly been better designed, if it had been set up for the purpose of studying a SARS-CoV-2 outbreak among young professionals. After 1 month at sea, the first case of COVID-19 was recognized. Another month went by before an epidemic broke out, which forced the ship’s early return to base within 2 weeks. Several viral strains were detected by nucleotide sequencing.1 This observation could imply the embarkation of multiple sailors who were independently infected, an unlikely scenario in Europe so early in the pandemic. Possibly, only one crewmember was the source, and the initial strain evolved within the 12 weeks’ voyage while spreading among the crew. Due to its exponential rate of spread, the SARS-CoV-2 virus rapidly infected at least 1,279 sailors, 76% of the participants of the study, whose median age was 28 years. Only 13% were female, without difference in the infection rate between males and females.1,2 This rate seemed strikingly high among young, healthy individuals,1 although it may not differ so much from that of other SARS-CoV-2 outbreaks, but rather reflected an exceptionally thorough follow-up and documentation. Only 14% of the infected participants remained asymptomatic.1 The median age of the 19 patients requiring only oxygen therapy was 45 years; the five patients admitted to intensive care units were older than 50 years. All infected sailors recovered eventually. These relatively benign clinical courses may not be representative of COVID-19 among the general population or cruise ship passengers, with a decidedly older age profile and related comorbidities.3 2706

A PubMed search for “ABO in COVID-19” yielded more than 50 publications including reviews and metaanalyses,4,5 documenting this possible correlation as a topic of intense research in the past 9 months.3,6 The study by Boudin et al.1 provides data leading to an important clarification: the rate of infection among young adults is independent of ABO blood group. This study can be considered the definitive conclusion on this aspect, as the quality of the epidemiological data was optimal. Studies in smaller cohorts7-10 and less well-defined epidemiological settings7-9,11 should be considered with caution, even if there are many. They are more likely to be affected by unknown cofounders. Particular precaution should be applied when COVID-19 was associated with ABO along with other blood group systems.10 Better data on ABO blood group and SARS-CoV-2 infection may not be accrued soon, and any future study would have to measure up to the quality of the study by Boudin et al.1 Can the ABO in COVID-19 topic be considered settled? An early study did not claim an influence of ABO on the SARS-CoV-2 infection rate.12 Rather the clinical course and disease outcome in patients, once infected, may differ depending on the ABO blood group.3,6,7,9,12,13 The lack of convincing evidence for an association between ABO and outcome in some,10,14 even many, studies cannot be construed as convincing evidence for lack of such an association. The largest and most comprehensive data set so far was from patients with respiratory failure.15 This genome-wide association study15 reported a small association signal coinciding with the chromosomal position of the ABO blood group system. Outcome was better for patients with blood group O than for those with blood group A. The study design was criticized for using blood donors as the majority of controls.14 Using flawed control cohorts is a notorious cause of erroneous conclusions,16 and blood donors are generally selected in favor of blood group O.17 However, it remains to be explored whether the odds ratio introduced by this well-founded bias of Spanish6 and Italian donor recruitment, could entirely explain the odds ratio of excess death associated with blood group A.15 Even haematologica | 2020; 105(12)

Editorials

a modest influence of ABO blood group on outcome should not be neglected, as happened once the SARS-CoV-1 epidemic abated,18 but should be investigated and resolved. Minor effects are important in precision medicine enabling comprehensive treatment of patients with COVID-19. Less than 20% of patients with COVID-19 received blood transfusions.19-21 The details of ABO matching can easily be reported,22 as they are routinely known. When properly documented, the combined data from small prospective observational studies can amount to impressive case series.23 ABO data may offer surprising insights, as cellular blood components contain residual plasma, and often a lot. For logistical reasons, components that are not ABO-identical can be transfused in a “major ABO-compatible” way (for instance, O red cells to an A recipient), without explicitly informing the hematologists. In such cases, anti-A or anti-B from the donor will bind to cell surfaces of the recipient’s tissues and form immune complexes with A or B antigens that are soluble in the recipient’s plasma.24 This pathophysiology applies particularly to the transfusion of platelet components.18 Convalescent plasma use in randomized clinical trials,25-27 in other clinical studies,28-33 and outside of them, is monitored for safety,27,34,35 best including ABO matching. Evidence is lacking to direct our practice on ABO matching of convalescent plasma. Hence, policies reflect the understanding and application of the basic principles of ABO compatibility, which have not been corroborated for COVID-19. “Minor ABO-compatible” plasma (for instance, AB plasma to an O recipient) transfers soluble A and B antigens. When bound by the recipient’s anti-A or anti-B, immune complexes are formed.24 These are a known trigger of the innate immune system,36 which receives another boost from the complement in the transfused plasma.37 Coagulation factors differ based on the patient’s ABO blood group and are, of course, also transfused by plasma.3,38 The interaction of complement and coagulation39 is not well understood in critically ill patients with COVID-19, whose potential harm from convalescent plasma should be considered and limited.40 Convalescent plasma, containing high-titer anti-SARS-CoV-2 and neutralizing antibody,41 can be tested for isoagglutinin titers, too. Convalescent plasma with high-titer anti-A or anti-B should be transfused to ABO-identical recipients,18 and a low-titer product is best for immune globulin manufacturing. Any indication to transfuse blood components containing plasma, particularly if not ABO identical, should be carefully considered and the exposure evaluated in studies. The ABO blood group system may have some influence on disease progression, once an individual is infected by SARS-CoV-2 and falls ill. The study by Boudin et al. in its unique epidemiological setting offered convincing evidence that becoming infected with the virus was not influenced by the ABO blood group in young professionals.1 This difference is not surprising, as the mechanisms involved likely differ between infection and disease progression. The sailors’ experience in spring 2020 should serve as a reminder: the risk of acquiring a SARS-CoV-2 infection is exceptionally high among young adults exposed to the virus in certain circumstances and no ABO blood group type can protect an individual from becoming infected. haematologica | 2020; 105(12)

Funding NIH Clinical Center, Intramural Research Program, project ID ZIA CL002128-01. The views expressed do not necessarily represent the view of the National Institutes of Health, the Department of Health and Human Services, or the U.S. Federal Government.

References 1. Boudin L, Janvier F, Bylicki O, Dutasta F. ABO blood groups are not associated with risk of acquiring the SARS-CoV-2 infection in young adults. Haematologica 2020;105(12):2841-2843. 2. Takahashi T, Ellingson MK, Wong P, et al. Sex differences in immune responses that underlie COVID-19 disease outcomes. Nature. 2020 Aug 26. [Epub ahead of print] 3. Yamamoto F, Yamamoto M, Muñiz-Diaz E. Blood group ABO polymorphism inhibits SARS-CoV-2 infection and affects COVID-19 progression. Vox Sang. 2020 Sep 23. [Epub ahead of print] 4. Wu BB, Gu DZ, Yu JN, Yang J, Shen WQ. Association between ABO blood groups and COVID-19 infection, severity and demise: A systematic review and meta-analysis. Infect Genet Evol. 2020;84:104485. 5. Pourali F, Afshari M, Alizadeh-Navaei R, Javidnia J, Moosazadeh M, Hessami A. Relationship between blood group and risk of infection and death in COVID-19: a live meta-analysis. New Microbes New Infect. 2020;37:100743. 6. Muñiz-Diaz E, Llopis J, Parra R, et al. Relationship between the ABO blood group and COVID-19 susceptibility, severity and mortality in two cohorts of patients. Blood Transfusion 2020: in press. 7. Wu Y, Feng Z, Li P, Yu Q. Relationship between ABO blood group distribution and clinical characteristics in patients with COVID-19. Clin Chim Acta. 2020;509:220-223. 8. Fan Q, Zhang W, Li B, Li DJ, Zhang J, Zhao F. Association between ABO blood group system and COVID-19 susceptibility in Wuhan. Front Cell Infect Microbiol. 2020;10:404. 9. Zalba Marcos S, Antelo ML, Galbete A, Etayo M, Ongay E, Garcia-Erce JA. Infection and thrombosis associated with COVID-19: possible role of the ABO blood group. Med Clin (Barc). 2020;S0025-7753(20):3044330447.2712 10. Latz CA, DeCarlo C, Boitano L, et al. Blood type and outcomes in patients with COVID-19. Ann Hematol. 2020;99:2113-2118. 11. Zhao J, Yang Y, Huang H, et al. Relationship between the ABO blood group and the COVID-19 susceptibility. Clin Infect Dis. 2020 Aug 4. [Epub ahead of print] 12. Leaf RK, Al-Samkari H, Brenner SK, Gupta S, Leaf DE. ABO phenotype and death in critically ill patients with COVID-19. Br J Haematol. 2020 Jul 1. [Epub ahead of print] 13. Sardu C, Marfella R, Maggi P, et al. Implications of AB0 blood group in hypertensive patients with covid-19. BMC Cardiovasc Disord. 2020;20:373. 14. Dzik S, Eliason K, Morris EB, Kaufman RM, North CM. COVID-19 and ABO blood groups. Transfusion. 2020 Jun 19. [Epub ahead of print] 15. Ellinghaus D, Degenhardt F, Bujanda L, et al. Genomewide association study of severe Covid-19 with respiratory failure. Genomewide association study of severe Covid-19 with respiratory failure. N Engl J Med. 2020 June 17. [Epub ahead of print] 16. Thomson G, Bodmer WF. Population stratification as an explanation of IQ and ABO association. Nature. 1975;254(5498):363-364. 17. Hirani R, Wong J, Diaz P, et al. A national review of the clinical use of group O D- red blood cell units. Transfusion. 2017;57(5):1254-1261. 18. Flegel WA. CoVID-19 insights from transfusion medicine. Br J Haematol. 2020;190(5):715-717. 19. Cai X, Ren M, Chen F, Li L, Lei H, Wang X. Blood transfusion during the COVID-19 outbreak. Blood Transfus. 2020;18(2):79-82. 20. Doyle AJ, Danaee A, Furtado CI, et al. Blood component use in critical care in patients with COVID-19 infection: a single centre experience. Br J Haematol. 2020 Jul 8. [Epub ahead of print] 21. Barriteau CM, Bochey P, Lindhom PF, Hartman K, Sumugod R, Ramsey G: Blood transfusion utilization in hospitalized COVID-19 patients. Transfusion 2020 June 24. [Epub ahead of print] 22. Zaidi FZ, Zaidi ARZ, Abdullah SM, Zaidi SZA. COVID-19 and the ABO blood group connection. Transfus Apher Sci. 2020 Jun 3. [Epub ahead of print] 23. Jiang SQ, Huang QF, Xie WM, Lv C, Quan XQ. The association between severe COVID-19 and low platelet count: evidence from 31 observational studies involving 7613 participants. Br J Haematol. 2020;190(1):229-e33. 24. Flegel WA. Pathogenesis and mechanisms of antibody-mediated hemolysis. Transfusion. 2015;55 Suppl 2(0):S47-S58.

2707

Editorials 25. Li L, Zhang W, Hu Y, et al. Effect of convalescent plasma therapy on time to clinical improvement in patients with severe and life-threatening COVID-19: a randomized clinical trial. JAMA. 2020;324(5):460470. 26. Liu Z. Errors in trial of effect of convalescent plasma therapy on time to clinical improvement in patients with severe and life-threatening COVID-19. JAMA. 2020;324(5):518-519. 27. Casadevall A, Joyner MJ, Pirofski LA. A randomized trial of convalescent plasma for COVID-19-potentially hopeful signals. JAMA. 2020;324(5):455-457. 28. Zhang B, Liu S, Tan T, et al. Treatment with convalescent plasma for critically Ill patients with severe acute respiratory syndrome coronavirus 2 infection. Chest. 2020;158(1):e9-e13. 29. Duan K, Liu B, Li C, et al. Effectiveness of convalescent plasma therapy in severe COVID-19 patients. Proc Natl Acad Sci U S A. 2020;117(17):9490-9496. 30. Ahn JY, Sohn Y, Lee SH, et al. Use of convalescent plasma therapy in two COVID-19 patients with acute respiratory distress syndrome in Korea. J Korean Med Sci. 2020;35(14):e149. 31. Ye M, Fu D, Ren Y, et al. Treatment with convalescent plasma for COVID-19 patients in Wuhan, China. J Med Virol. 2020 Apr 15. [Epub ahead of print] 32. Shen C, Wang Z, Zhao F, et al. Treatment of 5 critically Ill patients with COVID-19 with convalescent plasma. JAMA. 2020;323(16):1582-1589.

33. Zeng QL, Yu ZJ, Gou JJ, et al. Effect of convalescent plasma therapy on viral shedding and survival in patients with coronavirus disease 2019. J Infect Dis. 2020;222(1):38-43. 34. Murphy M, Estcourt L, Grant-Casey J, Dzik S. International survey of trials of convalescent plasma to treat COVID-19 infection. Transfus Med Rev. 2020;34(3):151-157. 35. Joyner MJ, Bruno KA, Klassen SA, et al. Safety update: COVID-19 convalescent plasma in 20,000 hospitalized patients. Mayo Clin Proc. 2020;95(9):1888-1897. 36. Birra D, Benucci M, Landolfi L, et al. COVID 19: a clue from innate immunity. Immunol Res. 2020;68(3):161-168. 37. Gralinski LE, Sheahan TP, Morrison TE, et al. Complement activation contributes to severe acute respiratory syndrome coronavirus pathogenesis. mBio. 2018;9(5): e01753-18. 38. Al-Samkari H, Karp Leaf RS, Dzik WH, et al. COVID-19 and coagulation: bleeding and thrombotic manifestations of SARS-CoV-2 infection. Blood. 2020;136(4):489-500. 39. Dzik S. Complement and coagulation: cross talk through time. Transfus Med Rev. 2019;33(4):199-206. 40. Dzik S. COVID-19 convalescent plasma: now is the time for better science. Transfus Med Rev. 2020;34(3):141-144. 41. Li L, Tong X, Chen H, et al. Characteristics and serological patterns of COVID-19 convalescent plasma donors: optimal donors and timing of donation. Transfusion. 2020 Jul 6. [Epub ahead of print]

Expanding dasatinib beyond KIT in acute myeloid leukemia John S. Welch Washington University School of Medicine, St. Louis, MO, USA E-mail: JOHN S. WELCH - jwelch@wustl.edu doi:10.3324/haematol.2020.262147

he goal of personalized medicine is to match patient-specific factors with relevant therapeutic options. The therapeutic conundrum in acute myeloid leukemia (AML) remains the heterogeneity of the disease and the paucity of treatment options for which there are highly predictive biomarkers. AML is broadly heterogeneous across all measured axes, including morphological presentation, cytogenetics, point mutations, expression signatures, epigenetic signatures, and chromatin signatures.1,2 Furthermore, within patients, subclonal architecture suggests ongoing acquisition of new variants and expression signatures,3 providing for intrapatient leukemic heterogeneity as well as interpatient heterogeneity. Just as heterogeneity can be seen across a range of measurements, response diversity has been mapped to clinical and molecular diversity, providing prognostic opportunities, but not yet personalized opportunities.2 AML results in unrestrained growth of the leukemia cells in vivo, but this has not translated into easy in vitro growth sufficient for effective cell manipulation in laboratory settings. Primary human AML cells grow poorly in liquid tissue culture media or methylcellulose. Only a fraction of samples will effectively engraft into immunodeficient mice, and among these, it is often only a subclone that engrafts.4 Recent progress has made short-term ex vivo culture possible, and improvements in the immunodeficient hosts have improved xenograft potential.5 Ex vivo analysis of chemotherapy has been championed by several groups, including the large-scale BEAT AML project.6-9 Studies have increasingly suggested ex vivo corre2708

lations with clinical response and the feasibility of scaling up to achieve sufficient throughput to identify useful functional biomarkers of sensitivity and resistance to chemotherapy. In this issue, Tavor et al. present a focused analysis that leverages careful sample selection with ex vivo drug sensitivity.10 They applied a 384-well approach to interrogates cell viability in liquid culture after 48 hours assessed across 46 drugs, each at 12 concentrations, which provided a broad area under the curve (AUC) measurement. In this assay, the authors used cytokine combinations (colonystimulating factor, interleukin 3, interleukin 6, thrombopoietin) in liquid culture and avoided stromal cell co-culture to facilitate viability read-out using a streamlined ATP-dependent assay (Cell Titer Glo). This approach provides an efficient read-out of early chemotoxicity, but does not provide an effective measure of differentiation or the cell toxicity that occurs after several days of exposure or cell divisions. In evaluating outcomes using this design, it is worth noting that the strong cytokine stimuli in the tissue culture may bias cell survival and chemosensitivity to cells that are capable of utilizing those signaling pathways or dependent on their stimuli for survival, and the small cell numbers evaluated in 384-well formats focus outcomes on phenotypes in the bulk cell population. Tavor et al. found that relapse samples were less chemosensitive than the paired diagnostic sample, across diverse classes of chemotherapy. Indeed, there were statistical differences between the sensitivity of diagnostic and relapse pairs to some common salvage agents, including haematologica | 2020; 105(12)

Editorials

etoposide and fludarabine, while cytarabine and venetoclax were associated with strong trends. This phenomenon of reduced sensitivity at relapse, compared with diagnosis, has played out across diverse chemotherapy regimens and newer venetoclax-based approaches in vivo.11,12 This provides a possible clinical validation of the platform and insight into relapsed AML as a broadly chemoresistant disease and not simply a cytarabine-resistant problem. Although the sample size was relatively small (29 patients), sensitivity and resistance could be clustered by functional groups, with tyrosine kinase inhibitors emerging with shared phenotypes. Drug targets included MEK, PI3K, PIM, JAK, mTORC1/2, FLT3, and BET inhibitors. To understand the mechanism of activity and to expand potential biomarker targets, Tavor et al. integrated mutation and transcriptome data with the functional chemosensitivity results. This identified increased frequency of FLT3-ITD variants in the chemosensitive cases. In particular, dasatinib emerged as a compound for which there were wide differences in chemosensitivity between these groups of patients. The authors further leveraged available results from the BEAT-AML study with their transcriptional data to identify overlapping signatures associated with dasatinib sensitivity, including dasatinib targets (CSF1R, FGR, HCK, and LYN). Using further model building and validation between the two cohorts, a model with an AUC 0.78 could be generated, providing an intriguing new biomarker for dasatinib. Finally, response to dasatinib could be observed in xenograft models of FLT3-ITD AML. Dasatinib has been studied in combination with multichemotherapy regimens in both acute lymphoblastic leukemia and AML,13-15 and thus is a natural compound to consider for quick integration into biomarker-driven AML therapy. In AML, the interest in dasatinib has focused on its use for the core-binding factor leukemias (CBF; t(8;21)

and inv(16)), which are frequently associated with mutations in KIT, and where the presence of the secondary KIT mutation is associated with inferior outcomes.16 Dasatinib has been explored in single-arm studies by the Cancer and Leukemia Group B (CALGB) and the Acute Myeloid Leukemia Study Group (AMLSG) to determine whether KIT inhibition could mitigate the negative effects of these variants. In both studies, responses and survival were favorable compared with those of historical controls. In the CALGB study, the adverse impact of KIT mutations appeared to have been effaced by dasatinib treatment.13 Analysis in the AMLSG study included paired evaluation of nine patients with KIT mutations. Five of these patients lost the variant at relapse, suggesting a potential selective disadvantage in the presence of dasatinib.15 In contrast to the molecular focus of dasatinib as an antiKIT agent in the CBF leukemia studies, Tavor et al. identified broader kinase mutations and kinase activity that correlate with dasatinib sensitivity, suggesting that additional biomarkers may be relevant for predicting sensitivity to dasatinib in non-CBF AML.10 In AML, biomarker-driven personalized medicine is challenged by the therapeutic timeline. Treatment needs to begin quickly once the diagnosis has been made. However, the delays in therapy initiation appear to affect predominantly younger patients,17 and the RADIFY study suggested that a delay of 8 days in adding a tyrosine kinase inhibitor to induction chemotherapy did not compromise efficacy in the younger age group.18 This provides a reasonable window for either ex vivo drug screening to be completed or mutation and transcriptomic analysis to be finalized while chemotherapy options are fine-tuned. AML is diversely heterogeneous between patients, but outcomes will be patient-specific. The application of ex vivo drug screening and the integration of results with

Figure 1. Integration of ex vivo drug sensitivity data and molecular signatures described in Tavor et al.10 AUC: area under the curve.

haematologica | 2020; 105(12)

2709

Editorials

expression and mutation signatures may provide the means of ultimately matching patients with treatment and matching treatment with response mechanisms. Given that relapsed disease appears to be chemoresistant across multiple classes of therapy, integration of personalized treatment is likely to be most effective when applied as early as possible.

9. 10.

11.

References 1. Cancer Genome Atlas Research Network. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):2059-2074. 2. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209-2221. 3. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264-278. 4. Klco JM, Spencer DH, Miller CA, et al. Functional heterogeneity of genetically defined subclones in acute myeloid leukemia. Cancer Cell. 2014;25(3):379-392. 5. Krevvata M, Shan X, Zhou C, et al. Cytokines increase engraftment of human acute myeloid leukemia cells in immunocompromised mice but not engraftment of human myelodysplastic syndrome cells. Haematologica. 2018;103(6):959-971. 6. Tyner JW, Tognon CE, Bottomly D, et al. Functional genomic landscape of acute myeloid leukaemia. Nature. 2018;562(7728):526-531. 7. Klco JM, Spencer DH, Lamprecht TL, et al. Genomic impact of transient low-dose decitabine treatment on primary AML cells. Blood. 2013;121(9):1633-1643. 8. Kurtz SE, Eide CA, Kaempf A, et al. Molecularly targeted drug combi-

12.

13. 14.

15. 16. 17. 18.

nations demonstrate selective effectiveness for myeloid- and lymphoid-derived hematologic malignancies. Proc Natl Acad Sci U S A. 2017;114(36):E7554-E7563. Malani D, Murumagi A, Yadav B, et al. Enhanced sensitivity to glucocorticoids in cytarabine-resistant AML. Leukemia. 2017;31(5):11871195. Tavor S, Shalit T, Ilani NC, et al. Dasatinib response in acute myeloid leukemia is correlated with FLT3/ITD, PTPN11 mutations and a unique gene expression signature. Haematologica. 2020;105(12):27952804. Estey EH. Treatment of relapsed and refractory acute myelogenous leukemia. Leukemia. 2000;14(3):476-479. DiNardo CD, Rausch CR, Bent on C, et al. Clinical experience with the BCL2-inhibitor venetoclax in combination therapy for relapsed and refractory acute myeloid leukemia and related myeloid malignancies. Am J Hematol. 2018;93(3):401-407. Marcucci G, Geyer S, Laumann K, et al. Combination of dasatinib with chemotherapy in previously untreated core binding factor acute myeloid leukemia: CALGB 10801. Blood Adv. 2020;4(4):696-705. Ravandi F, O'Brien S, Thomas D, et al. First report of phase 2 study of dasatinib with hyper-CVAD for the frontline treatment of patients with Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukemia. Blood. 2010;116(12):2070-2077. Paschka P, Schlenk RF, Weber D, et al. Adding dasatinib to intensive treatment in core-binding factor acute myeloid leukemia-results of the AMLSG 11-08 trial. Leukemia. 2018;32(7):1621-1630. Ayatollahi H, Shajiei A, Sadeghian MH, et al. Prognostic importance of C-KIT mutations in core binding factor acute myeloid leukemia: a systematic review. Hematol Oncol Stem Cell Ther. 2017;10(1):1-7. Sekeres MA, Elson P, Kalaycio ME, et al. Time from diagnosis to treatment initiation predicts survival in younger, but not older, acute myeloid leukemia patients. Blood. 2009;113(1):28-36. Stone RM, Mandrekar SJ, Sanford BL, et al. Midostaurin plus Chemotherapy for acute myeloid leukemia with a FLT3 mutation. N Engl J Med. 2017;377(5):454-464.

Convalescent plasma for administration of passive antibodies against viral agents Giovanni Di Minno,1 Pier Mannuccio Mannucci,2 James W. Ironside,3 Carlo Federico Perno,4 Lutz GĂźrtler5 and Louis Aledort6 1

Dipartimento di Medicina Clinica e Chirurgia, Centro Hub per le Emocoagulopatie, Napoli, Italy; 2Fondazione IRCCS Caâ&#x20AC;&#x2122; Granda Ospedale Maggiore Policlinico, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Milan, Italy; 3Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, UK; 4Departmental Medicine Laboratory, ASST Niguarda Hospital, Milan, Italy; 5Max von Pettenkofer Institute, Ludwig Maximilians University of Munich, Munich, Germany and 6Mary Weinfeld Professor of Clinical Research in Hemophilia at the Icahn School of Medicine at Mount Sinai, New York City, NY, USA E-mail: GIOVANNI DI MINNO - diminno@unina.it doi:10.3324/haematol.2020.267427

dministration of passive antibodies through transfusion of plasma from donors recovering from a viral infection has long been employed to treat individuals infected with the same pathogen.1 However, in studies with convalescent plasma (CP), differences and inherent limitations (e.g., sensitivity/specificity of tests to quantify neutralizing antibodies; sample size; scheduling of treatment [early/late CP administration vs. degree of disease severity], the presence of confounders [concomitant treatments]), and restricted generalizability of data argued for large-scale, randomized, controlled trials.1,2 The results of a multicenter proof-of-concept, observational Italian study in 46 patients with moderate or severe acute respiratory distress syndrome due to infection with the novel coronavirus, SAR-CoV-2, who needed mechanical ventilation and/or continuous positive airway pressure are reported in this 2710

issue of the Journal.3 The interval between symptom onset and study inclusion was highly variable (2-29 days). The 7day mortality rate was 6% in patients given CP compared with an expected 15% according to Italian statistics and 30% in a small concurrent cohort not treated with CP. Weaning from continuous positive airway pressure was achieved in 26 of 30 patients, and three of the seven intubated patients were extubated. Whether those who received CP earlier improved more or faster than patients who received plasma later in the course of the disease is not clarified, nor are the reasons for administering one, two or three CP bags provided. In this larger than previous uncontrolled reports, five serious adverse events (including 1 transfusion-related acute lung injury [TRALI]) occurred in four patients. Although TRALI may be triggered by transfused antibodies,4 CP was safe in this study as it was in haematologica | 2020; 105(12)

Editorials

Table 1. Summary on the use of convalescent plasma# in moderately to severely ill patients infected with major emerging viruses.

Virus, Ref(s)

Design/Patients/Treatments/Objectives

Results/Interpretation/Limitations

SARS-CoV-2 JAMA. 2020;323(16):1582-1589.

Case series. 5 critically ill patients (36-65 years old, 2 women), all with confirmed SARS-CoV-2 infection + ARDS, rapidly progressing severe pneumonia and high viral load despite treatment; all receiving mechanical ventilation; PaO2/FiO2 <300. 10-22 days after admission (mean, 18.2), patients received 400 mL CP with a SARS-CoV-2–specific IgG binding titer ≥1:1000 and a neutralization titer ≥40.

SARS-CoV-2 Proc Natl Acad Sci U S A. 2020; 117(17):9490-9496.

Prospective observational. 10 severely ill patients (34-78 years old) with confirmed COVID-19 infection. Maximal supportive care + 200 mL CP (neutralizing antibody titers >1:640) given to the patients 16.5 days (median) after onset of illness. Primary endpoint: CP safety. Secondary endpoints: improvement of symptoms and laboratory parameters within 3 days after administration of CP. Case reports. 4 critically ill patients (31-73 years old) with confirmed SARS-CoV-2 infection and respiratory failure requiring mechanical ventilation (ECMO in 2 cases) were given 200-2400 mL of CP 11-18 days (mean: 15.25 days) after admission.

Following CP transfusion, body temperature normalized within 3 days (4/5 patients). Within 12 days, the SOFA score decreased; PaO2/FiO2 increased (range, 284-366 after vs. 172-276 before); viral loads became negative, and SARS-CoV-2-specific and neutralizing antibody titers increased (range, 1:80-1:320 on day 7 vs. 1:40-1:60 before). ARDS resolved in 4 patients 12 days after transfusion, 3 patients stopped mechanical ventilation within 2 weeks of treatment. 3/5 patients were discharged from the hospital, 2 in a stable condition 37 days after CP. By day 3 after CP transfusion, improved clinical symptoms and laboratory values, increases in neutralizing antibody titers, patients’ oxygen saturation and lymphocyte count; decreases in CRP, SARS-Cov-2 viral load, and radiological lung lesions (varying degrees of absorption of lung lesions within 7 days). No severe adverse events following CP administration.

SARS-CoV-2 Chest. 2020;158(1):e9-e13.

SARS-CoV-2 JAMA 2020;324(5):460-470.

SARS-CoV-2 Haematologica 2020;105(12):2834-2840.

Influenza A or B § Lancet Respir Med. 2019;7(11):941-950.

All recovered from the infection. Resolution or partial absorption of lung lesions (all cases); reduced viral load (2 cases), 3/4 discharged between days 18-43. Recovery/discharge within 1 to 4 weeks after starting CP transfusion; one patient discharged on supplemental oxygen, another required continued critical care for multi-organ failure. Open-label, multicenter, randomized trial. The original sample At the time of the termination, 103/200 patients (58.3% males, size was 100 for each group to provide 80% power with a median age 70 years) had been enrolled. Of these, 98.1% two-sided significance level of α=0.05. (101/103) completed the trial. Within 28 days, clinical Patients (males/females, ≥18 years of age) with confirmed improvement was detected in 51.9% (27/52) in the CP group COVID-19 and severe ARDS and/or hypoxemia or life-threatening vs. 43.1% (22/51) in the control group (HR, 1.40 [95% CI: conditions (shock, organ failure, or requiring mechanical 0.79-2.49]; P=0.26). Among those with severe disease, ventilation) were stratified by age and disease severity. improvement was found in 91.3% (21/23) of the CP group vs. Intervention: 4-13 mL/kg volume of ABO-compatible 68.2% (15/22) in the control group (HR, 2.15 [95% CI: CP with IgG titer ≥1:640 + standard treatment (vs. standard 1.07-4.32]; P=0.03). Among those with life-threatening disease, treatment alone). Primary outcome: time to improvement within improvement was found in 20.7% (6/29) of the CP group vs. 28 days, secondary outcomes: 28-day mortality, time to discharge, 24.1% (7/29) in the control group (HR, 0.88 [95% CI: 0.30-2.63]; negativity of viral PCR within 72 h of treatment. The study was P=0.83). At day 28, there was no difference in mortality (15.7% terminated early due to the containment of the SARS-CoV-2 vs. 24.0%; OR, 0.65 [95% CI: 0.29-1.46]; P=0.30) or time to epidemic in China. discharge (51.0% vs. 36.0%; HR, 1.61 [95% CI: 0.88-2.93]; P=0.12). CP treatment was associated with a higher conversion rate to negative viral PCR at 72 h (87.2% in the CP group vs. 37.5% in the control group (OR, 11.39 [95% CI: 3.9133.18]; P<0.001). Two adverse events occurred in two patients in the CP group. Multicenter, one-arm, interventional study. 46 patients, mean Patients had been symptomatic for a mean of 14 days 62 years old (SD 11), 28 males (61%), confirmed SARS-CoV-2 (SD, 7) and had had ARDS for a mean of 6 days (SD 3) prior infection + moderate-to-severe ARDS, elevated CRP and need to receiving CP. Three patients (6.5%) died within 7 days. for mechanical ventilation and/or CPAP. 1-3 units (250-300 mL each) Among survivors, PaO2/FiO2 increased by 112 units of CP (neutralizing antibody titers: 1:80-1:320) +usual treatment. (95% CI: 82-142); severity of radiological signs decreased Primary outcome: 7-day hospital mortality. Secondary outcomes: in 23% (95% CI: 5-42%); CRP, ferritin and lactate dehydrogenase PaO2/FiO2 , laboratory/radiologic changes, weaning from mechanical levels decreased by 60%, 36% and 20%, respectively. ventilation, CP safety. Weaning from CPAP was achieved in 26/30 patients and 3/7 intubated patients could be extubated. Five serious adverse events occurred in four patients (1 TRALI), of which two were possibly treatment related. Randomized, double-blind phase III prospective trial. Patients 92/138 randomized to the high-titer, 48/138 to the low-titer group. of all ages with severe influenza A infection; onset of illness At baseline, 60 (43%) participants were in intensive care; 55/78 within 6 days of randomization. Randomization based on disease (71%) of participants were not in intensive care requiring severity and age (< vs. >18 years): either two units (or pediatric oxygen. Early termination. No superiority of high-titer over equivalent) of high titer (≥1:80) or low titer (≤1:10) low-titer plasma (OR on day 7: 1.22; 95% CI: 0.65-2.29, P=0.54). anti-influenza virus hemagglutinin antibodies CP; 28 days of 34% of participants (47/138) experienced 88 serious adverse follow-up. Objectives: efficacy of high-titer vs. low-titer CP. events, including ARDS. Ten patients died (6 [7%] in the high-titer group, 4 [9%] in the low-titer group, P=0.73), worsening of ARDS was the most common cause of death.

continued on the nex page

haematologica | 2020; 105(12)

2711

Editorials continued from the previous page

Virus, Ref(s)

Design/Patients/Treatments/Objectives

Results/Interpretation/Limitations

Influenza A or B § Lancet Respir Med. 2019;7(11):951-963.

Randomized, double-blind, placebo-controlled trial. Adults (≥18 years of age), confirmed severe influenza A or B infection needing hospital treatment, symptoms starting within 7 days before randomization, assigned to standard care (which included antiviral therapy) + either 500 mL infusion of high-titer H-Ig (0.25 g/kg bodyweight, 24·75 g maximum) or saline placebo. Primary outcome: clinical status at day 7.

Influenza A or B § Lancet Respir Med. 2017;5(6):500-511.

Randomized, open-label, multicenter, phase II retrospective trial. Hospitalized patients with severe (hypoxia/tachypnea) influenza A or B assigned on day 0 to standard care ± 2 units (225–350 mL/unit or 8 mL/kg pediatric equivalent) of compatible anti-influenza plasma hemagglutination inhibition antibody titers ≥1:80 CP, follow-up: 28 days. Objectives: time to normalization of respiratory status by day 28 of hospitalization.

Influenza A (H1N1) °^ Clin Infect Dis 2011;52:447-456.

Prospective cohort study. Within 7 days of symptom onset, 93 patients ≥18 years old with severe H1N1 2009 infection needing intensive care were given the possibility of a 500 mL CP infusion over a 4 h period (neutralizing antibody titer ≥1:160).

Influenza A (H1N1) °^ Chest. 2013:144(2):464-473.

Multicenter, prospective, double-blind, randomized controlled trial. Patients with severe H1N1 infection on standard antiviral treatment requiring intensive care and ventilatory support randomized to receive one dose of 0.4 g/kg of H-Ig (17 patients) or 0.4 g/kg normal intravenous immunoglobulins (18 matched patients) over a period of 4 h.

Ebola virus disease^^ N Engl J Med. 2016;374(1):33-42, N Engl J Med. 2017;376(13):1297.

Non-randomized, comparative study. Patients of various ages, confirmed EVD, two consecutive transfusions of 200 to 250 mL of ABO-compatible CP with varying levels of neutralizing antibodies. Transfusions initiated up to 2 days after diagnosis. Controls: 418 patients who had been treated at the same center during the previous 5 months. Primary outcome: risk of death from 3 to 16 days after diagnosis, adjustments for age; patients who died before day 3 excluded.

Ebola virus disease^^ & J Infect. 2017;74(3):302-309.

Case series. Patients treated vs those who did not receive 1 unit (450 mL) of ABO-compatible convalescent whole blood (CWB) within the first 24 h of admission over a period of 1-4 h.

SARS-CoV-1 *§§ Clin Microbiol Infect. 2004;10(7):676-678.

Retrospective. 19 patients (38.7±12.39 years old) given 200-400 mL CP (coronavirus titer range 1:160-1: 2560) compared to 21 patients (47.9±19.60 years old) given methylprednisolone pulses.

SARS-CoV-1 *§§ Eur J Clin Microbiol Infect Dis. 2005;24(1):44-46.

Retrospective. Patients (n=33/80) given CP (median volume 279 mL, coronavirus titer range, 1:160-1:2,560) within day 14 after the onset of symptoms vs. patients given CP more than 14 days after hospital admission.

156 received 500 mL of H-Ig, 152, placebo (224/308 influenza A serotypes and 84/308 influenza B serotypes). Clear rise in hemagglutination inhibition titers against influenza A; smaller rise in influenza B titers in the treated group. In subgroup analyses, the OR was 0.94 (0.55-1.59) in patients with influenza A and 3.19 (1.21-8.42) in those with influenza B (interaction P=0.023). Through 28 days of follow-up, 47/156 (30%) of patients in the H-Ig group and 45/152 (30%) in the placebo group died or experienced a serious adverse event (HR 1.06, 95% CI: 0.70-1.60; P=0.79). Respiratory status normalized in 28/42 (67%) of the participants in the plasma group vs. 24/45 (53%) in the standard care alone group (HR: 1.71, 95% CI: 0.96-3·06, P=0.069). One patient in the plasma group, and five (10%) in the standard care group died (HR 0·19 [95% CI: 0·02-1.65], P=0·093). No difference between groups in days of hospital stay (median, 6 days vs. 11 days, P=0.13) or in mechanical ventilation (median, 0 days vs. 3 days, P=0.14). Serious adverse events including ARDS lower in the plasma group than in the standard care group (9/46 [20%] vs. 20/52 [38%], P=0.041. 20/93 patients (21.5%) received CP. Mortality was lower in the treatment group than in the matched control group (20.0% vs. 54.8%; P=0.01, OR, 0.20; 95% CI: .06-.69; P=0.011). CP-treated individuals showed significantly lower viral loads, and post-treatment levels of interleukin-6, interleukin-10 and tumor necrosis factor-α than those who refused CP treatment (P<0.05). Patients receiving H-Ig had lower post-treatment (days 5 and 7) viral loads than controls (P=0.04 and P=0.02, respectively). Initial serum cytokine level, significantly higher in the H-Ig group, fell to control levels 3 days after treatment. In patients receiving treatment within 5 days of symptom onset, H-Ig was the only factor related to mortality (0% vs. 40%, respectively, OR, 0.14; 95% CI: 0.02-0.92; P=0.04). No treatment-related adverse events. 84 patients treated with plasma included in the analysis. From day 3 to day 16 after diagnosis, the risk of death was 31% in the CP group vs. 38% in the control group (risk difference, -7%; 95% CI: -18 to 4). No serious adverse reactions were associated with the use of CP. CP treatment was not associated with a significant improvement in survival, in patients with confirmed EVD. However: (i) three-fourths of CP donors had low titers of neutralizing antibodies (1:10-1:40), only a minority (5%) had higher titers (1:160), and (ii) patients receiving plasma with high doses anti-Ebola virus IgG antibodies had larger decreases in viral loads. Compared with 25 non-treated patients, improved survival was documented in 44 subjects who received CWB (deaths, 44% vs. 27.9%, respectively; odds ratio: 2.3, 95% CI: 0.8-6.5). There were significant difference between viral load on admission and after 24 h of treatment with CWB (P<0.01). No adverse events. Discharge at the end of a 3-week hospitalization in 74% of subjects receiving CP and in 19% of those on methylprednisolone. Mortality: 0/19 (CP group) vs. 5/21 (steroid group). Unknown titer or type of antibodies affecting outcomes. Anti-SARS-CoV-1 antibodies contained within the CP not standardized. At completion of a 3-week hospitalization, compared to the overall SARS-related mortality for admitted patients (17%, n = 299), those receiving plasma earlier had a lower mortality rate (12.5%). No adverse events reported. No correlation between clinical response and antibody titers or transfused volumes. continued on the nex page

2712

haematologica | 2020; 105(12)

Editorials #Donor selection is critical for the safety of plasma and plasma products. For SARS-CoV-2 studies; the FDA established that plasma should be collected from patients who had recovered from COVID-19 who had a negative SARS-CoV-2 polymerase chain reaction test and had been symptom-free for at least 14 days. Donors should meet the eligibility criteria for standard blood donors as set out in federal regulation 21 CFR 630.10 and 21 CFR 630.15, and standard testing should follow regulation 21 CFR 610.40. § A meta-analysis of studies published between 1918 and 1925 on the so-called Spanish influenza pneumonia (Ann Intern Med. 2006;145(8):599-609), revealed lower case-fatality rates in patients receiving CP than in controls (16% vs. 37%, respectively); maximal benefit being found in those receiving early treatment compared to patients who were transfused later (19% vs. 59%, respectively). §§Case reports and case series concerning the Middle East respiratory syndrome coronavirus (MERS-CoV) responsible for the 2015 MERS did not show clinical improvement with the administration of CP, and raised the issue of the quality of plasma and the role of neutralizing antibody titers in studies of CP (Sci Rep. 2016;6:25359, Emerg Infect Dis. 2015;21(12):2186-2189). ° Data from this prospective, non-randomized trial are in keeping with information of a contemporary retrospective observational study in the setting of H1N1 (Hong Kong Med J. 2010;16(6):447-454). Comparable support stems from case reports and case series of the 2006 avian influenza A/H5N1 outbreak and the 2015 outbreak influenza A (H7N9) ( Hong Kong Med J. 2006;12:489; N Engl J Med. 2007;357:1450-1451; Int J Infect Dis. 2015;41:3-5; PLoS One. 2008;3:e2985). The latter studies also call for selecting convalescent donors based on the viral sequence of individual patients, a key issue for infections, such as influenza, caused by multiple strains of viruses. ^ The overall evidence that the administration of CP, serum, or H-Ig might be beneficial for the treatment of severe acute respiratory infections due to influenza or SARS-CoV-1 has been systematically reviewed and meta-analyzed (J Infect Dis. 2015;211(1):80-90). As a whole, a 75% reduction in the odds of mortality was found among patients with severe acute respiratory infections of viral etiology who were treated with convalescent plasma (or serum), together with a clinically relevant impact of the treatment on reducing the viral load. No evidence of serious adverse events or complications due to therapy were found and there was some evidence of a reduced use of critical care resources and length of hospital stay. Maximal reduction in mortality was achieved when CP was administered early after symptom onset. ^^The efficacy of CP in Ebola virus disease has been reported in two case series (Clin Infect Dis. 2015;61(4):496-502, N Engl J Med. 2016;374(7):636-646). & In addition to Ebola virus disease, CP has been evaluated in the treatment of other viral hemorrhagic fevers e.g. Bolivian (Am J Med. 1966;40(2):217-230), Argentine (Presse Med. 1986;15(45):2239-2242.), and Lassa fever (Trans R Soc Trop Med Hyg. 1984;78(3):319-324). Inherent limitations and confounders compel caution in the analysis of these small studies. *Data from case reports (Transfus Apher Sci. 2003;29(1):101; Hong Kong Med J. 2003;9(3):199-201; Zhonghua Yi Xue Za Zhi. 2003;83(12):1018-1022) and case series (J Antimicrob Chemother. 2005;56(5):919-922) are in keeping with information from retrospective, non-randomized studies, in this setting. SARS-CoV-2: severe acute respiratory syndrome of the coronavirus 2019, which causes COVID-19; SARS-CoV-1: 2003 severe acute respiratory syndrome-associated coronavirus 1 outbreak in Hong Kong; EVD: Ebola virus disease; ARDS: acute respiratory distress syndrome; PaO2/FiO2: ratio of partial arterial oxygen pressure to fraction of inspired oxygten; TRALI: transfusion-related acute lung injury (whose features are similar to those of SARS-CoV2 infection); SOFA: Sequential Organ Failure Assessment; ECMO: extracorporeal membrane oxygenation support; CPAP: continuous positive airway pressure; H-Ig: hyperimmune immunoglobulin; CP: convalescent plasma; CRP: C-reactive protein; OR: odds ratio; HR: hazard ratio; CI: confidence interval; SD: standard deviation; PCR: polymerase chain reaction.

Table 2. Major emerging viruses potentially transmitted by blood and plasma products.§

Major pathogens Enveloped viruses West Nile virus Dengue virus Chikungunya virus Hepatitis B virus escape variants

Description

Epidemiology

Detection method

Inactivation/elimination method

Flavivirus causing encephalitis Flavivirus causing myalgia, arthralgia, hemorrhagic fever Alphavirus causing arthralgia Causes hepatitis, hepatocarcinoma

South Europe, Africa, Americas South Asia, South America, Africa

NAT, antibody IgM* NAT/antigen antibody.

Solvent/detergent, heat Solvent/detergent, heat

South Asia, Africa Asia, Americas, Europe

NAT/antibody. NAT/sequence HBsAg.

Solvent/detergent, heat Heat, vaccination.

NAT/antibody NAT

Prolonged heat Heat

NAT/PCR

Heat

No screening test; diagnosis by histopathology and western blot of brain or tonsils after onset of symptoms

Resists conventional forms of inactivation

Non-enveloped viruses Parvoviruses Cause myalgia, anemia, fetal malformation Worldwide Enterovirus Picornaviruses causing pharyngitis, myalgia, Worldwide neuritis, encephalitis Circoviruses No known disease Unknown Major emerging prions Variant Creutzfeldt-Jakob Prion infectivity in brain, lymphoid tissues, disease prion† and blood

Europe, North America, Japan, Taiwan, Saudi Arabia.

§ The term “Emerging” denotes those pathogens whose incidence in humans has increased or may increase in the future, based on their appearance in a new host population or changes in epidemiology in an existing population. This definition includes causative agent(s) leading to a new infection, a re-emerging infection, or an infection with new attributes (e.g., drug resistance or virulence) making them pathogenic. Like SARS-CoV-2, most of these previously unknown entities are capable of quickly overcoming the original geographical boundaries and spreading among the human population. In the globalized world of the third millennium, millions of people and huge amounts of merchandise are transported every day by air from one part of the planet to another.This facilitates the 'traffic' of microbes and the diseases they cause from ecological niches in which they were confined, and their rapid spread to every corner of the earth. Unpredictable global and national events (wars, world tourism, migration) can further increase the spread of blood-borne diseases. With regards to the roles of zoonotic transmission in the introduction of an unknown pathogen into a new human host population, recombination and re-assortment (adaptation) to produce ‘new’ mutations that allow pathogens to acquire new biological characteristics to adapt to new ecologies and to infect new hosts (dissemination), leading to short, explosive outbreaks (e.g. Ebola virus) or slower, silent spreading (e.g. HIV-1) have been documented. Factors that influence the course of the SARS-CoV-2 disease are largely unknown. Suggested reading: Nature. 2004;430(6996):242-249; Nat Med. 2004;10 (12 Suppl):S70-76; Nature. 2007;447(7142):279-283; Nature. 2008;451(7181):990-993; Blood Transfus. 2009;7(3):167-171; Semin Thromb Hemost. 2013;39(7):779-793; Haematologica 2013;98(10):1495-1498. *Strategy currently only used in North America and Canada, †Only four cases reported. IgM: immunoglobulin M; NAT: nucleic acid testing; PCR: polymerase chain reaction.

5,000 patients in another study.5 Also considering a risk/benefit analysis performed to improve the treatment of severe acute respiratory distress syndrome caused by SARS-CoV-2 infection,1 the Food and Drug Administration (FDA) issued guidance on CP collection and distribution in the USA and recommended conducting clinical trials with CP in the setting of SARS-CoV-2 infection.6 It has been proposed that, in such trials, one CP unit is used for post-expohaematologica | 2020; 105(12)

sure prophylaxis and one to two units for treatment of SARS-CoV-2 infection.1 For patients who fail to meet the criteria for enrollment in clinical trials, the FDA has approved protocols for emergency use and expanded access.4 In parallel, the plasma industry joined forces (the CoVIg-19 ALLIANCE) to increase plasma collection and produce safe and effective CP and hyperimmune immunoglobulins (H-Ig).7,8 Beside the USA, other countries9 2713

Editorials

are collecting CP to be used in SARS-CoV-2 infections, and many studies are ongoing.10 Parallel to the submission of the Italian study, an openlabel, multicenter, randomized trial from China appeared, in which patients with SARS-CoV-2 and severe acute respiratory distress syndrome were randomized to 4-13 mL/kg of CP plus standard treatment vs. standard treatment alone (Table 1). The calculated sample size was 100 patients for each group. Due to the containment of the SARS-CoV-2 epidemic in China, the study was terminated when 103 of the 200 planned patients had been enrolled. At termination of the trial, improvement was found in 21/23 patients in the CP group vs. 15/22 in the control group (P=0.03) among those with severe disease, and in 6/29 of the CP group vs. 7/29 in the control group (P=0.83) among those with lifethreatening disease. There was no between-group difference in mortality (P=0.30) and two adverse events were detected in two patients in the CP group. While antibody administration by means of CP is indeed a reliable strategy for conferring immediate immunity against viral agents to individuals with SARS-CoV-2 infection, there is uncertainty about whether CP or H-Ig is the more effective product to be administered.10,11 While CP is characterized by donor-dependent variability in antibody specificity and titers, H-Ig contains standardized antibody concentrations. On the other hand, while the IgM fraction, a key weapon against some viruses, is removed from plasma during H-Ig fractionation, CP also provides coagulation factors (to fight hemorrhagic fevers, such as Ebola).2 Although specific antibodies hamper viral replication, the SARS-CoV spike (S) protein is the main antigenic component responsible for biological effects, e.g., host immune responses, neutralizing-antibody formation, T-cell responses and ultimately protective immunity.12 On the whole, the proportion of anti-S protein antibodies, relationships between IgG/IgA/IgM, standardization of antibody titers and optimal dosing and scheduling of CP administration are still major unknowns from studies conducted so far in the frame of the SARS-CoV-2 pandemic. This scenario of growing interest from clinicians, patients, policy-makers, health systems and pharmaceutical industries provides an unprecedented opportunity to exert a major imprint on the practice of medicine.2 A concerted effort is warranted to achieve globally uniform, high-quality standards for CP or H-Ig preparations. In high-income countries, the industrial production of plasma-derived products has never been safer than nowadays both because of the guidelines produced by the FDA and European Medicines Agency on donor selection and screening and because of the availability of viral inactivation methods. Plasma is collected at plasmapheresis centers using technologies regularly inspected by governing bodies before approval for commercial use. Plasma is screened after each donation and re-screened in mini-pools for human immunodeficiency virus-1, hepatitis A, B and C viruses, and parvovirus B19, and Plasma Master Files are subject to yearly approval by regulatory agencies.13 Once collected, plasma from single donors may be administered directly to patients or pooled to manufacture plasma-derived products such as H-Ig, coagulation factors and others. The resulting products may be treated with solvent/detergent and/or super-heated (at 80° C for 3 days), pasteurized or nano-filtered. The 2714

aforementioned approaches are highly effective in minimizing pathogen transmission, as demonstrated by the fact that no blood-borne pathogen transmission has been reported since 1987 for commercially prepared plasma products received by patients with hemophilia, the epitome of multi-transfused patients.13 In theory, however, risks remain pertaining to emerging and re-emerging pathogens (prions, non-lipid enveloped viruses) (Table 2), for which diagnosis and inactivation methods are still a challenge.14 The reasons for this caveat concerning risks include the lack of reliable screening tests for some pathogens (e.g. prions), no screening for unknown pathogens, and relative poor sensitivity/specificity of the available assay methods.15 Furthermore, some viral mutants may escape screening,16 which may also not pick up potential plasma contamination from infectious but not yet seropositive donors. In addition, there may be low-level chronic carriers among donors who remain undetected and yet contribute to infect the plasma pool.17,18 Finally, determining the prevalence of emerging pathogens may be difficult when there is a long latency between infection and symptom onset.19 On this background and with these knowledge gaps, the adaptation of screening methods is a constant challenge,13 and public health organizations and plasma pharmaceutical industries have combined efforts to tackle the risks. In the framework of its global perspective, the World Health Organization tries to minimize pathogen transmission through early information and public health vigilance on the emergence of regional pathogens capable of causing transfusion-transmitted infections (e.g. Zika virus in Brazil), even before local authorities manage to develop means to prevent blood-borne transmission.20 Because ‘zero risk’ in terms of product safety is unlikely, governing bodies provide guidance to identify factors relevant for pathogen transmission. As an example, the presence of blood-borne hepatitis E virus may pose significant threats to some people (e.g., the elderly, immunocompromised individuals) despite being of low risk to other potential recipients. Thus, in addition to the circumstances under which blood products are collected and manufactured, the nature of the pathogen (e.g., its physical characteristics, level of virulence, prevalence) and the patients’ characteristics (age, immune status, geographical location, lifestyle, treatment urgency) should be considered when choosing the individual treatment (and assessing an acceptable level of risk). Alongside this scenario of basically satisfactory blood product safety in high-income countries, it should be considered that in most low/middle-income countries procedures for blood collection are seldom standardized, and donor selection, screening and viral inactivation often fail to meet the criteria validated and implemented by regulatory agencies in high-income countries. If insufficient antiSARS-CoV-2 CP is available from high-income countries to meet global needs, the use of plasma from low/middleincome countries may become necessary but may also raise some issues, because the type and prevalence of infectious agents likely differ in different populations.13 To sum up, if worldwide uniform advancements in blood-banking quality are encouraged in low/middleincome countries, there is now a global opportunity to perform clinical studies on the efficacy of CP or H-Ig in viral infections and address uncertainties on the occurrence of haematologica | 2020; 105(12)

Editorials

serious adverse events related to the administration of these products.10 Removing regulatory barriers that limit the use of pathogen-reduction technology for CP collections would be a major help in this respect.2 The process of obtaining informed consent requires communication of risks and benefits of treatments to patients. SARS-CoV-2 is an easily inactivated enveloped virus,13 and strict regulations for plasma product manufacturing minimize the risk of known and unknown pathogens. Apart from emergency situations, the extent to which people should be further informed on specific risks associated with any particular product will depend on a variety of factors including availability of alternative treatments, and the patientsâ&#x20AC;&#x2122; characteristics (e.g., age, physical/mental condition, education/level of understanding, language barriers/religious beliefs). A good understanding by health care professionals of the sources and modes of production of plasma derivatives and of pathogen-reduction/inactivation techniques might be an additional benefit of studies involving CP. Addendum Parallel to the submission of this Editorial, a systematic review of completed (as of June 4, 2020) and ongoing (n=98) studies on the efficacy and safety of convalescent plasma or hyperimmune immunoglobulins to reduce mortality in patients with SARSCoV-2 infection appeared (Cochrane Database Syst Rev. 2020;7(7):CD013600. Its provisional conclusions support the clinical relevance of the concepts summarized in the present report.

References 1. Bloch EM, Shoham S, Casadevall A, et al. Deployment of convalescent plasma for the prevention and treatment of COVID-19. J Clin Invest. 2020;130(6):2757-2765. 2. Roback JD, Guarner J. Convalescent plasma to treat COVID-19: possibilities and challenges. JAMA. 2020; 323(16):1561-1562. 3. Perotti C, Baldanti F, Bruno R et al. Mortality reduction in 46 severe Covid-19 patients treated with hyperimmune plasma. A proof of concept single arm multicenter trial. Haematologica. 2020;105(12): 2834-2840. 4. Semple JW, Rebetz J, Kapur R. Transfusion-associated circulatory over-

haematologica | 2020; 105(12)

5. 6.

7. 8. 9.

10. 11. 12. 13.

14. 15. 16.

17. 18. 19. 20.

load and transfusion-related acute lung injury. Blood. 2019;133 (17):1840-1853. Joyner M, Wright RS, Fairweather D, et al. Early safety indicators of COVID-19 convalescent plasma in 5,000 patients. medRxiv. 2020 May 14. [Epub ahead of print] Food and Drug Administration. Recommendations for Investigational COVID-19 Convalescent Plasma. Available from: https://www.fda.gov/vaccines-blood-biologics/investigational-newdrug-ind-or-device-exemption-ide-process-cber/recommendationsinvestigational-covid-19-convalescent-plasma#Collection%20 of%20COVID-19 [Accessed May 9, 2020] https://clinicaltrials.gov/ct2/show/NCT04488081?term=i+spy+covid &draw=2&rank=1[Accessed August 11, 2020] 7 bis. https://www.covidrdalliance.com/pdf/ISPY_PressRelease_ August_3_Final.pdf Sanquin. Sanquin starts collecting plasma from cured corona patients. Available from: https://www.sanquin.org/news/2020/mar/sanquinstarts-collecting-plasma-from-cured-corona-patients. [Accessed May 9, 2020] Piechotta V, Chai KL, Valk SJ, et al. Convalescent plasma or hyperimmune immunoglobulin for people with COVID-19: a living systematic review. Cochrane Database Syst Rev. 2020;7:CD013600. [Editorial: no authors listed]. The resurgence of convalescent plasma therapy. Lancet Haematol. 2020;7(5):e353. Du L, He Y, Zhou Y, Liu S, Zheng BJ, Jiang S. The spike protein of SARS-CoV--a target for vaccine and therapeutic development. Nat Rev Microbiol. 2009;7(3):226-236. Di Minno G, Navarro D, Perno CF, et al. Pathogen reduction/inactivation of products for the treatment of bleeding disorders: what are the processes and what should we say to patients? Ann Hematol. 2017;96(8):1253-1270. Srivastava A, Brewer AK, Mauser-Bunschoten EP, et al. Guidelines for the management of hemophilia. Haemophilia. 2013;19(1):e1-47. Di Minno G, Canaro M, Ironside JW, et al. Pathogen safety of longterm treatments for bleeding disorders: still relevant to current practice. Haematologica. 2013;98(10):1495-1498. Salpini R, Piermatteo L, Battisti A, et al. A hyper-glycosylation of HBV surface antigen correlates with HBsAg-negativity at immunosuppression-driven HBV reactivation in vivo and hinders HBsAg recognition in vitro. Viruses. 2020;12(2):251. Schreiber GB, Busch MP, Kleinman SH, Korelitz JJ. The risk of transfusion-transmitted viral infections. The Retrovirus Epidemiology Donor Study. N Engl J Med. 1996;334(26):1685-1690. Di Minno G, Perno CF, Tiede A, et al. Current concepts in the prevention of pathogen transmission via blood/plasma-derived products for bleeding disorders. Blood Rev. 2016;30(1):35-48. Ludlam CA, Powderly WG, Bozzette S, et al. Clinical perspectives of emerging pathogens in bleeding disorders. Lancet. 2006;367(9506):252261. Barjas-Castro ML, Angerami RN, Cunha MS, et al. Probable transfusion-transmitted Zika virus in Brazil. Transfusion. 2016;56(7):16841688.

2715

CENTENARY REVIEW ARTICLE Ferrata Storti Foundation

History of hematopoietic cell transplantation: challenges and progress Noa Granot and Rainer Storb

Fred Hutchinson Cancer Research Center and the University of Washington, Seattle, WA, USA

ABSTRACT

Haematologica 2020 Volume 105(12):2716-2729

fter more than 60 years of research in allogeneic hematopoietic cell transplantation (HCT), this therapy has advanced from one that was declared dead in the 1960s to a standard treatment of otherwise fatal malignant and non-malignant blood diseases. To date, close to 1.5 million hematopoietic cell transplants have been performed in more than 1,500 transplantation centers worldwide. This review will highlight the enormous efforts by numerous investigators throughout the world who have brought the experimental field of HCT to clinical reality, examine ongoing challenges, and provide insights for the future.

The beginnings

Correspondence: RAINER STORB rstorb@fredhutch.org Received: July 30, 2020. Accepted: September 25, 2020. Pre-published: October 9, 2020. doi:10.3324/haematol.2019.245688 ©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.

2716

The late 1940s saw major research efforts directed at repairing or preventing radiation damage to organs in response to observations made on survivors of the atomic bomb explosions in Japan. A pivotal report by Jacobson et al. in 1949 demonstrated protection of mice from lethal total body irradiation (TBI) damage to the bone marrow when shielding their spleens or femora with lead.1,2 Two years later, Lorenz et al. saw similar protection when mice were given an intravenous infusion of syngeneic marrow following TBI.3 Of note, Jacobson4 and others attributed the radiation protection to some humoral factor present in spleen or bone marrow, the “humoral hypothesis”, a supposition that was controversial and not shared by others who thought the ‘rescue’ of the irradiated mice had cellular origins. It was not until the mid-1950s that several laboratories unequivocally documented, with the help of blood genetic markers, that the radioprotection was due to repopulation of the irradiated marrow spaces by transplanted donor cells, thereby validating the “cellular hypothesis”.5,6 Much of that early work has been comprehensively described in the 1967 book “Radiation Chimeras” written by van Bekkum and de Vries.7 The unequivocal validation of the cellular hypothesis was greeted enthusiastically by immunologists, radiation biologists, and clinicians because of its implications for cell biology and because it promised clinical translation for treating patients with life-threatening blood disorders. Investigators thought that high doses of chemoradiation therapy could be used both to destroy diseased marrow and suppress the host immune system, thereby preventing rejection of an infused marrow graft from a healthy donor. Already one year after the studies in rodents were published, Thomas et al. reported in 1957 in The New England Journal of Medicine that marrow could be infused into irradiated leukemia patients and then engraft, even though, in the end, the patients were not cured of their leukemia.8 In 1965, Mathé et al. described a patient with acute leukemia who was given TBI followed by a marrow infusion from each of six relatives.9 The marrow of one of the relatives engrafted. While the patient eventually succumbed to an immunologic complication, initially called secondary disease and now known as graft-versus-host disease (GvHD), his leukemia remained in remission. With this observation, Mathé et al. corroborated a previous report by Barnes and Loutit from 1956 in mice showing that GvHD could lead to eradication of leukemic cells.10 In order to describe this phenomenon, Mathé coined the term “graft-versus-leukemia” effect. Unfortunately, all human allogeneic marrow grafts in these early years failed, as meticulously documented in a paper from 1970 by Bortin.11 Of the 200 patients reported between 1957 and 1967, 73 were transplanted for aplastic anemia, 115 for advanced and refractory hematologic malignancies, and 12 for immunodeficiency diseases. In the end, all 200 patients died, 125 with graft failure, 47 with GvHD, haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation

and others with infections or recurrence of their underlying malignancies. These early human transplants were performed before a full understanding of conditioning regimens and GvHD prevention was achieved, and before the discovery of the importance of histocompatibility matching for the outcome of marrow transplantation. The early transplants were based on observations in inbred mice, for which major histocompatibility complex (MHC) matching was not an absolute requirement. As a result of the complete failure of translating findings from mice to humans, most investigators abandoned the idea that allogeneic HCT could ever become a valuable asset in clinical medicine, and prominent immunologists doubted that the immunological barrier from one human to another could ever be crossed.

Back to the laboratory Discouraged by the disastrous clinical results, most investigators left the field, declaring it a dead end. However, a few small laboratories in Europe and the United States persisted in systematic efforts to understand and overcome the perceived “insurmountable” obstacles encountered in early human marrow transplantation. Much of the work was carried out in large animals, including monkeys and dogs. An important paper published in 1968 showed that canine littermates matched for the MHC antigens by in vitro tissue typing had far better HCT outcomes than MHC-mismatched recipients.12 In vitro typing for the MHC, called human leukocyte antigen (HLA) region in humans, H2 in mice, and DLA in dogs, was very primitive at the time and, moreover, the complexity of the MHC was not yet understood. Typing consisted of serologic measurements using multi-specific antibodies collected from parous women or transfusion recipients in trypan blue exclusion or leuko-agglutination assays,13-16 combined with testing of donor and recipient lymphocytes for reactivity in a mixed leukocyte culture.17 While trail-blazing at the time, these primitive testing techniques are now history, and typing is accomplished by genetic sequencing of up to 14 HLA-alleles. The second, completely unexpected observation made in the original canine experiment was that GvHD, either in acute, subacute or chronic form, developed in MHCmatched littermates, even though significantly later than in mismatched littermates.18 This finding was surprising since it had not been encountered in mice that were mismatched with their donors for non-H2 antigens. It pointed out the need for investigating methods to prevent and control GvHD even in well-matched human donor-recipient combinations. Consequently, studies of numerous immunosuppressive agents were conducted in a canine model that eventually led to identifying the antimetabolite methotrexate as the best drug for GvHD prevention.19 By balancing the drug’s toxicities against its efficacy, a regimen of intermittent methotrexate was established, with administration of the drug 1, 3, 6, and 11 days after transplantation and then weekly for at least 3 months. This regimen entered the clinic in 1969 and was used until the early 1980s. Other research efforts focused on effective and tolerable conditioning regimens. In the beginning, single-dose TBI up to 10 Gray (Gy) was utilized. However, extensive studhaematologica | 2020; 105(12)

ies in canines revealed that delivering TBI in multiple fractions of 2 Gy each reduced damage to slow-responding tissues, such as liver, lung and others, while barely diminishing radiation effects on marrow and lymphoid tissues. Based on these studies, fractionating TBI has remained the standard.20 Additionally, effective drug-based conditioning regimens were developed. Among those, cyclophosphamide has become the standard for patients with aplastic anemia since the drug had outstanding immunosuppressive qualities. However, cyclophosphamide spared stem cells and was not myeloablative,21 and so was not deemed suitable for conditioning patients with leukemia. In contrast, another alkylating agent, busulfan, proved to be highly myeloablative but lacked the immunosuppressive qualities of cyclophosphamide or TBI. Prospective, randomized trials showed that busulfan was better tolerated than TBI and had equivalent efficacy to TBI for conditioning patients with myeloid malignancies;22 however, in order to ensure hematopoietic engraftment, busulfan needed to be combined with immunosuppressive drugs such as cyclophosphamide23 or fludarabine. Other studies addressed transfusion-induced sensitization to minor histocompatibility antigens, which often resulted in marrow graft rejection among HLA-identical recipients with aplastic anemia. Rejection rates in early transplants for aplastic anemia ranged from 36% to 60%.24,25 In order to surmount this major problem, methods were identified that minimized the risk of sensitization from transfusions. Also, a more immunosuppressive conditioning regimen was developed that combined antithymocyte globulin (ATG) with cyclophosphamide.26,27 This regimen has become standard practice for aplastic anemia patients with HLA-identical sibling donors. Other studies showed that, unlike in solid organ transplantation, post-HCT immunosuppression was not required for the remainder of the patients’ life but could often be discontinued after 6 months. After 6 months, donor-derived regulatory T cells (then called “suppressor T cells”) were found in the peripheral blood, which were thought to enable and maintain a state of graft-versus-host tolerance; of note, these cells were absent in patients with chronic GvHD.28 It was also shown that successful grafts could be accomplished using hematopoietic cells derived from the peripheral blood in mice, dogs and baboons.29 In later years, it was found that large numbers of these cells could be “mobilized” from the marrow into the peripheral blood (peripheral blood stem cells or PBSC) with granulocytecolony stimulating factor (G-CSF).30,31 Finally, since dogs share spontaneous blood disorders with humans, preclinical exploration of treating such diseases by allogeneic HCT was possible. For example, dogs with severe combined immunodeficiency (SCID) were cured by marrow transplants, as were dogs with severe hemolytic anemia due to pyruvate kinase deficiency. The latter dogs had massive iron deposits in their inner organs from the severe hemolysis.32 Long-term follow-up of transplanted dogs showed impressive resolution of iron deposits in the liver over time. This finding encouraged the first successful transplantation for multiply transfused human patients with thalassemia major.33 Dogs with spontaneous non-Hodgkin lymphoma (NHL) served to establish the value of autologous HCT in the treatment of this disease. Moreover, comparisons with results of allogeneic HCT confirmed the presence of graft-versus-lymphoma effects in dogs. 2717

N. Granot and R. Storb

Return to the clinic: 1968 to 1980 By 1968, the progress made in preclinical transplantation and the advances in the understanding of HLA set the stage for clinical trials to resume. In 1968/1969, three publications reported the first successful marrow grafts for patients with primary immune deficiency disorders.34-36 However, during the subsequent years, most clinical transplants were performed in patients with advanced hematologic malignancies and severe aplastic anemia.37,38 These early trials posed serious challenges not only in the field of transplantation biology but also in aspects of supportive care, especially infections and transfusion support. Therefore, these trials stimulated incredible progress in infectious disease and transfusion research. Even though donors and recipients in nearly all early trials were HLA-identical siblings, and despite prophylaxis with methotrexate, GvHD occurred in almost half of the patients. Major advances in GvHD prevention and, consequently, improvement in overall patient survival were accomplished when, based on preclinical canine studies, methotrexate was combined with calcineurin inhibitors such as cyclosporine or tacrolimus.39-41 These synergistic drug combinations have remained among the most widely used methods for GvHD prevention to date. The first grading system for acute GvHD was described in 1974,42 and the first effective treatment of acute GvHD with anti-thymocyte globulin (ATG) was reported in the same year.43 In those early years, ATG was not commercially available and the drug was produced in our laboratory by immunizing rabbits with human thymocytes. Early results in patients with aplastic anemia conditioned with cyclophosphamide showed 45% long-term survival.44,45 One reason for the disappointing findings was fatal GvHD. However, as predicted from canine studies, the most serious fatal complication was graft rejection due to sensitization through transfusions to minor histocompatibility antigens for which donors and recipients were disparate. Preclinical studies transformed how clinical transfusions and transplants were conducted and paved the way for better transplantation outcomes. Changing transfusion support to leukocyte-depleted, in vitro irradiated platelet and red blood cell products reduced the risk of sensitization to minor antigens and, with it, the risk of graft rejection not only for patients with aplastic anemia but also those with hemoglobinopathies. In addition, the newly developed cyclophosphamide/ATG regimen more effectively suppressed recipient immunity thereby enabling almost uniform marrow engraftment. The cumulative effects of these changes have resulted in survivals for patients with aplastic anemia given HLA-identical sibling marrow grafts ranging from 64% to 100%.39,46-55 All early transplantations for acute leukemia were performed in patients who were in refractory relapse. As a result, in addition to fatalities from GvHD, many patients died from post-transplantation relapse. A decision in the mid-1970s to transplant patients earlier in the course of their disease, while the leukemia burden was low, reduced the relapse risk and led to a significant improvement in survival among patients with acute leukemias.37,38 Two pivotal publications from 1979 and 1981 in The New England Journal of Medicine described powerful graft-versus-leukemia effects associated with acute and chronic GvHD.56,57 This work provided the rationale for the subsequent introduction of donor lymphocyte infusions in the 1990s to prevent or combat relapse after HCT.58,59 2718

Some transplant centers focused on removing T cells from the marrow in order to reduce the risk of GvHD. However, initial studies showed unacceptably high incidences of mortality from graft rejection, disease relapse and infections.60 When T-cell depletion was combined with high-intensity conditioning regimens before and careful monitoring after transplantation for recurrence of acute leukemia and prompt treatment by donor lymphocyte infusions, outcomes were improved. This approach has remained an acceptable procedure in patients with acute leukemia.61 In the late 1980s, G-CSF-mobilized PBSC were introduced for allogeneic transplants. Randomized, prospective trials showed marrow and PBSC to be equivalent as far as engraftment and overall survival were concerned (Table 1). However, PBSC caused more chronic GvHD than marrow; because of this, marrow has remained the preferred source of stem cells for patients with non-malignant diseases such as aplastic anemia or hemoglobinopathies. However, PBSC continue to be the predominant graft source for patients with hematologic malignancies, in part due to donor preference. One limitation in early allogeneic HCT was that only approximately 35% of patients had HLA-identical siblings who could serve as marrow donors. In order to get around that limitation, and assisted by an increasing understanding of the genetics of the HLA region, along with improved HLA-typing techniques,62-64 registries were established in the 1980s that collected HLA data from unrelated volunteer donors, first in the UK with the Anthony Nolan Foundation, in the United States with the National Marrow Donor Registry, and then other national registries (Table 1). Early canine studies had already indicated the feasibility of â&#x20AC;&#x153;matchedâ&#x20AC;?, unrelated HCT,65,66 and the first successful human transplant from an HLAmatched unrelated donor was carried out in 1979 for a patient with acute lymphoblastic leukemia.67 Currently, HLA data from more than 36 million unrelated volunteers are accessible in the various national registries. For Caucasian patients, the likelihood of finding an HLAmatched unrelated donor is approximately 80%; however, this percentage declines dramatically for patients from ethnic groups.68-71 In order to provide potentially curative HCT for these otherwise unserved patients, transplant methods have been developed that use grafts either from unrelated umbilical cord blood (UCB) or from HLA-haploidentical relatives (see below). This way donors can be found for 95% of transplant candidates regardless of age and ethnic background.

Moving forward: the 1990s Over the past 25 years, more and more transplant centers have been established worldwide. In order to collect and analyze outcome data from the ever-increasing numbers of transplants, data registries have been set up, such as the European Bone Marrow Transplant Registry (EBMT) and the Center for International Bone Marrow Transplant Research (CIBMTR). To date, information on close to 1.5 million HCT has been collected.72,73 Large and mostly retrospective data analyses have generated information aimed at providing recommendations for the best HCT approaches for the various diseases and donor-recipient combinations. haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation

The 1990s saw many changes in the way transplantations have been carried out. Major advances in infectious disease prevention and treatment were made, including using acyclovir to prevent herpes simplex and varicella zoster virus reactivation, monitoring for cytomegalovirus (CMV) reactivation and, once reactivation occurred, preventing CMV disease with ganciclovir or foscarnet, preventing Pneumocystis Jirovecii infections with a synthetic antibacterial combination of sulfamethoxazole and trimethoprim, and introducing more effective anti-fungal agents and antibiotics.74-79 Conditioning regimens were intensified to the upper limit of tolerability in order to optimize tumor cell kill; however, investigators began realizing that these intensive, myeloablative regimens, including cyclophosphamide/TBI or busulfan/cyclophosphamide were too toxic for elderly patients or for those with comorbid conditions. This was especially unfortunate since most hematologic malignancies happen to occur in older patients. The problem was obvious when comparing the median ages of patients transplanted in those years at Fred Hutch (related grafts 40 years, and unrelated grafts 35 years) to the median patient age at diagnosis of the underlying hematologic malignancies (68 years).80 In short, most affected patients were excluded from transplantation. In order to address this serious problem, and extend allogeneic HCT to include older or medically infirm patients, less-intensive conditioning regimens were introduced. These regimens shifted the burden of tumor kill from high-dose chemo-irradiation therapy toward graftversus-tumor effects. The regimens were facilitated by the development of a then new immunosuppressive agent, mycophenolate mofetil (MMF), that blocked the de novo purine pathway needed for lymphocyte replication and worked in synergy with calcineurin inhibitors. This synergistic drug combination was not only effective in preventing GvHD but, importantly, also enhanced hematopoietic engraftment.81,82 MMF also was synergistic with another agent, sirolimus, which reduced the sensitivity of T cells to interleukin-2 through mTor inhibition.83-86 Combinations of these agents are now commonly used after allogeneic HCT, leading to relatively low and acceptable risk of non-relapse mortality (NRM).

The 21st century The early years of the 21st century saw tremendous growth in allogeneic HCT, in part because of non-myeloablative or reduced-intensity conditioning regimens, which enabled extending allogeneic HCT to include older patients, and, in part, the growth was due to increased use of grafts from alternative donors, including unrelated cord blood (UCB) and HLA-haploidentical relatives. Additionally, progress has been made in GvHD prevention among unrelated recipients. For example, a recently published randomized, prospective phase III trial compared a commonly used drug combination of MMF and cyclosporine to a triple-drug regimen of MMF, cyclosporine, and sirolimus.84 Patients on the triple drug arm had a significant reduction in overall acute GvHD, and acute grade III-IV GvHD was seen in only 2% of patients. This resulted in a significant improvement in overall survival. Another recent development has been the US Food and Drug Administration approval of ruxolitinib, haematologica | 2020; 105(12)

a JAK2 inhibitor for the treatment of steroid-refractory acute GvHD. The approval was prompted by the favorable outcome of the single-arm, phase II REACH 1 (Research Evaluation and Commercialization Hub) study.87 Extracorporeal photopheresis (ECP) has been used as an off-label second-line treatment for cutaneous manifestations of steroid-refractory acute and chronic GvHD since the early 2000s, with variable success.88 The introduction of the HCT comorbidity index (HCT-CI) in 2005 facilitated comparisons of results between centers worldwide, and has served as an important decision-making tool for choosing appropriate transplant regimens.89 Serum biomarkers derived from the gastrointestinal tract, specifically ST2 and REG-3Îą, have emerged as an additional method of predicting acute GvHD severity, as presented in a recent Myoblast Autologous Grafting in Ischemic Cardiomyopathy (MAGIC) trial.90 The application of this tool may enable tailored GvHD treatment and prediction of NRM. Results with minimal-intensity or reduced-intensity conditioning regimens have been summarized in numerous scientific publications. One publication from 2013 reported outcomes in nearly 1,100 elderly or medically infirm patients with advanced hematologic malignancies who were given HLA-matched related or unrelated grafts after minimal-intensity conditioning with fludarabine and low-dose (2-3 Gy) TBI.91 The oldest patients in that trial were 75 years of age. Half of the patients had serious comorbidities with HCT Comorbidity Index scores â&#x2030;Ľ3. These transplants were designed as an outpatient procedure. In fact, nearly half of the patients were never hospitalized while the remaining half had a median hospital stay of only 6 days.92 Living at home or in a private apartment and being able to move around freely were appreciated by patients and caregivers. At a median follow-up of 5 years, depending on the relapse risk of the underlying malignancies and on the comorbidity score, lasting remissions were seen in 45-75% of patients and 5-year survivals ranged from 25% to 60%. Overall 5-year NRM was 24%, for the most part associated with concurrent or preceding GvHD, and the overall relapse mortality was 34.5%. With the introduction of the triple-drug regimen of MMF, cyclosporine and sirolimus, the NRM has significantly declined among unrelated recipients. As a result, relapses have remained the major obstacle toward better outcome. Most relapses occurred within the first 2 years after HCT. When analyzing relapse risk per patient year, both disease and disease burden turned out to be major predictors of relapse.93 For example, the risk was 0.19 for multiple myeloma (MM) in remission and as high as 0.32 for patients not in remission. Comparable numbers for acute myeloid leukemia were 0.33 for patients in remission 1-3 and 0.65 for those with relapsed disease.91 These findings delineated the limitations of graft-versus-tumor effects and bore out the adverse impact of high tumor burden. They have encouraged future efforts to reduce relapse through both enhancing graft-versus-tumor effects and more effectively reducing the tumor burden before HCT. A more recent study94 reported remarkable improvements in allogeneic HCT outcomes among 1,720 patients with hematologic malignancies who received non-myeloablative conditioning over the past twenty years (Figure 1). These improvements were accomplished even though more recent patients were older (56% >60 years old in 2010-2017 vs. 27% in 1997-2003), had more comorbidities 2719

N. Granot and R. Storb

(45% with HCTCI scores equal to or >3 in 2010-2017 vs. 25% in 1997-2003), and had more frequently unrelated grafts (65% in 2010-2017 vs. 34% in 1997-2003). Explanations for the gradual improvements of outcome include use of ursodeoxycholic acid to prevent cholestasis and hyperbilirubinemia and reduce the incidence of liver GvHD,95 improved GvHD prevention, use of topically active oral glucocorticoids for gut GvHD,96 more judicious use of systemic glucocorticoid dosing for treatment of GvHD,97 and improved antibiotics and anti-fungal agents. As for the latter, a randomized, double-blinded trial showed that posaconazole and fluconazole were similarly effective in preventing overall fungal infections and reducing overall mortality among 600 patients with acute GvHD; however, posaconazole was superior to fluconazole in preventing invasive aspergillosis.98 Results from other transplant centers and from CIBMTR and EBMT analyses have shown similar outcomes with a variety of reduced-intensity or minimal-intensity conditioning regimens. Regimens included fludarabine and varying doses of melphalan with or without low-dose TBI, fludarabine and reduced doses of busulfan, reduced busulfan, cyclophosphamide and thiotepa and others. A 2016 review in the journal Haematologica summarized the findings with these regimens.99 Moreover, researchers at Johns Hopkins utilized the backbone of the

fludarabine/low-dose TBI regimen and added two small doses of pre-transplant cyclophosphamide, followed by two higher doses of cyclophosphamide on days 3 and 4 after HCT plus MMF and tacrolimus to enable engraftment of HLA-haploidentical related marrow or PBSC and minimize acute and chronic GvHD.100 This regimen has been surprisingly effective, although, owing in part to the reduction in acute and chronic GvHD, relapse has remained a prominent problem. The use of UCB as an innovative, alternative source of stem cells was introduced in the 1990s. Cord blood cells are immunologically naĂŻve and allow for greater HLA disparity with a given recipient. This feature enabled transplantation for patients who lacked HLA-matched donors. An EBMT report showed promising outcomes among 143 UCB transplantations performed in 45 centers.101 An observational study by Brunstein et al. in 2010 suggested that outcome after transplantation of two partially HLAmatched unrelated UCB units was comparable to those of HLA-matched related and unrelated HCT.102 A prospective, randomized trial comparing double-unit UCB to single-unit UCB transplantation among 224 patients with hematologic malignancies showed equivalent 1-year survivals of 65% versus 73%.103 Moreover, patients given single-unit UCB experienced more rapid platelet recovery and less grade 3-4 acute GvHD. Others have experiment-

Figure 1. Improved outcomes of allogeneic hematopoietic cell transplantation with non-myeloablative conditioning for older and medically infirm patients with hematologic malignancies over two decades at Fred Hutchinson Cancer Research Center (WA, USA).94 Overall survival (A), progression-free survival (PFS) (B), nonrelapse mortality (NRM) (C), and incidence of relapse (D) by time period of transplant: 1997-2003 (black line), 2004-2009 (blue line), and 2010-2017 (red line) (from Cooper et al.;94 with permission).

2720

haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation

ed with ex vivo UCB expansion methods which may have led to higher engraftment rates.104,105 Additional potential advantages of UCB transplantation included the lack of risk to the donor, rapid availability, and ease of scheduling of transplantation. In 2001, the Blood and Marrow Transplant Clinical Trials Network (BMT CTN) was established through funding from the National Institutes of Health as a collaborative effort of the CIBMTR, the NMDP/Be The Match and the Emmes Company, together with 20 Core Transplant Centers. The BMT CTN has opened more than 30 multi-institutional clinical trials involving more than 100 transplant centers.

Current trends While early HCT involved grafts from HLA-identical sibling donors, from 2006 on unrelated donors became the most frequently used graft source in the USA with almost 4,500 transplants in 2018 alone (Figure 2A106). This increase could be attributed to: (i) the ever-larger number of unrelated, HLA-typed volunteers in the registries; (ii) advances in HLA-typing including the recognition of the importance of HLA-DPB1 expression for the development of GVHD;107 and (iii) increasing age of patients whose siblings were also older and often unfit to donate because of comorbidities. A rise in the use of UCB during the early 2000s has been offset and reversed by the remarkable increase in HLA-haploidentical, related grafts after the introduction of post-transplant cyclophosphamide for GvHD prevention. Post-HCT cyclophosphamide served to cause in vivo depletion of both donor-versus-host reactive T cells (GvHD prevention) and host-versus-donor T cells and natural killer (NK) cells (prevention of graft rejection).108 An alternative regimen for HLA-haploidentical grafts has been reported by Chinese investigators.109 They conditioned patients by busulfan, cyclophosphamide, ME-CCNU and ATG and combined cyclosporine, MMF and a short course of methotrexate for GvHD prevention and reported favorable outcomes. A very recent review summarized the various conditioning regimens used for HLA-haploidentical HCT.110 Retrospective comparisons of results from different centers have suggested comparable outcomes with HLA-haploidentical related versus UCB grafts. In contrast, two recent prospective, randomized trials indicated better outcomes with HLA-haploidentical grafts compared to UCB transplants.111,112 However, relapse has remained a major complication, as half of the patients relapsed two years after transplant in both studies, regardless of conditioning regimen intensity. Additional controlled comparisons of the two donor sources may be needed to validate the superiority of HLA-haploidentical over UCB grafts. Unfortunately, accrual to such trials may be challenging since a patient, when given the choice between the two modalities, might prefer the related donor over an UCB graft. Figure 2B shows trends in disease indications for HCT in North America for the past 18 years. Most notable has been a linear increase in acute myeloid leukemia, from 1,000 patients in 2000 to 3,500 in 2018. This increase was largely due to extending allogeneic HCT to include older patients in whom chemotherapy, as a rule, fails to maintain long-term remissions. For the same reason, increases, haematologica | 2020; 105(12)

although at lower levels, were also seen for patients with acute lymphoblastic leukemia and myelodysplastic syndromes (MDS). In contrast, allogeneic HCT for chronic myeloid leukemia, chronic lymphocytic leukemia (CLL), and MM have either remained at very low levels or declined. These trends were influenced by the introduction of alternative therapies for these diseases, including tyrosine kinase inhibitors, a BCL-2 antagonist, Bruton tyrosine kinase inhibitors, bi-specific or mono-specific monoclonal antibodies, proteasome inhibitors and chimeric antigen receptor (CAR) T cells, among others. However, these therapies including CART-T cells could also be used as a â&#x20AC;&#x2DC;bridgeâ&#x20AC;&#x2122; to allogeneic HCT, for example, in order to consolidate remissions in acute lymphoblastic leukemia (ALL) patients. Allogeneic HCT might also serve as a salvage treatment in NHL and CLL patients who relapsed after CAR-T cell treatment. A recent study reviewed our centerâ&#x20AC;&#x2122;s experience with allogeneic HCT after CAR-T cell therapy in 32 ALL, NHL and CLL patients and found no additional risk for infections and GvHD.113 In addition, recipients of allogeneic HCT after CAR-T cell therapy had longer event-free survival compared to patients given CAR-T cell therapy alone (P=0.014). A recently published Chinese study concurred with these results in ALL patients (1-year OS: 79.1% vs. 32.0%; leukemia-free survival: 76.9% vs. 11.6%; P<0.0001),114 while an earlier Memorial Sloan Kettering Cancer Center trial, also in ALL patients, showed no significant advantage with allogeneic HCT after CAR-T cell therapy compared to CAR-T cell therapy alone (P=0.89, P=0.64, respectively).115 In the aggregate, most reports recommended CAR-T cell therapy as a bridging treatment to allogeneic HCT in patients with high-risk B-cell malignancies. While allogeneic HCT has remained the only curative therapy for CLL, it is currently used mainly for patients in whom all other therapies failed and who are often in poor general condition. Given its curative potential, allogeneic HCT should be considered earlier in the disease course, for example, in patients with poor-risk CLL and no or few comorbidities, in whom the risk of HCT-associated NRM is very low. Because of novel alternative therapies, allogeneic HCT for NHL has declined from >1,000 cases in 2013 to approximately 600 in 2018. A recent retrospective CIBMTR analysis of allogeneic HCT showed that fludarabine/2 Gy TBI conditioning gave better results than a fludarabine/4 Gy TBI regimen, largely because of lower NRM and better overall 5-year survival (51% vs. 31%) with the former regimen while relapse rates at 5 years were comparable; this result was a testimony to powerful graft-versus-lymphoma effects.116 Allogeneic HCT can be considered as salvage therapy in select high-risk MM patients in the setting of a clinical trial, and has been reviewed in detail by a number of investigators.117,118 Allogeneic HCT has remained the therapy of choice for disorders such as congenital immunodeficiencies or autoimmune and immune dysregulation disorders. Given the rarity of these diseases, a better understanding of posttransplant complications and long-term outcome is only now emerging.119,120 Better timing of transplants, improved screening methods, lasting immune reconstitution post transplant, and reduced toxicity conditioning regimens have contributed to better outcomes for all immunodeficiencies. Allogeneic HCT has also remained standard of 2721

N. Granot and R. Storb

care for other non-malignant diseases such as severe aplastic anemia (AA), sickle cell disease, and thalassemia major. For AA, this is not only true for HLA-identical sibling marrow grafts, but increasingly also for unrelated HLAmatched marrow transplants, where survivals close to or of 100% have been reported.54 The most frequently used unrelated HCT regimens for AA patients, either fludarabine, ATG, cyclophosphamide and 2 Gy TBI or fludarabine combined with Campath antibody, appeared to produce similar results, including for older patients.121,122 Moreover, recent papers reported outstanding survival figures for AA patients given HLA-haploidentical grafts.123,124 Taken together, these findings suggested that upfront marrow transplantation should be considered for all patients with AA who have a suitable donor rather than waiting until failure of immunosuppressive therapy.

ATG has also been widely used as a form of in vivo T-cell depletion in conditioning regimens for patients with hematologic malignancies. Kumar et al. recently reported a systematic review of prospective, randomized trials comparing ATG to no ATG.125 They concluded that, while ATG reduced the incidence of acute and chronic GvHD, there was no statistically significant difference in NRM or overall survival. They suggested designing future studies with improved methodological quality to conclusively establish the role of ATG in allogeneic HCT. CMV positivity before HCT has remained an adverse risk factor despite monitoring for CMV reactivation and pre-emptive therapy in case of reactivation. Therefore, the results of a recent phase III double-blind trial by Marty et al. comparing prophylactic letermovir to placebo in CMVseropositive patients were encouraging.126 In that study,

Figure 2. Trends in allogeneic hematopoietic cell transplantation (HCT) over two decades in North America. Donor source (A), transplant indications (B). 2019 data from the Center for International Blood and Marrow Transplant Research (CIBMTR).106 ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; CLL: chronic lymphocytic leukemia; CML: chronic myelogenous leukemia; MDS: myelodysplastic syndrome; MM: multiple myeloma; NHL/HL: non-Hodgkin lymphoma/ Hodgkin disease; URDCB: unrelated donor cord blood; URD-PB: unrelated donor peripheral blood; URD-BM: unrelated donor bone marrow.

2722

haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation

patients randomized to letermovir showed a significant reduction in clinically significant CMV infection after HCT without encountering hematologic toxicities. A

promising option for management of resistant and refractory CMV infection includes maribavir, which is currently being investigated in a phase III trial.127 These newer treat-

Table 1. Timeline of list of notable events throughout the years.

Year

Event

1949-1955 1956 1957 1958 1958-1969 1958-1970 1968 1971 1972-1974 1973 1974 1974 1975 1979-1981 1980 1981 1981 1981 1981 1982 1983 1986 1986 1987-1993 1988 1991 1995 1995 1997 1998 1997-2001 1998 1983-2005 2001 2002 2005-2012 2008 2009 2012 2014 2015 2017 2018 2019-2020 2018-2020

Evidence of hematopoietic cell recovery after exposure to lethal radiation2-6 Bone marrow transplant induces graft-versus-host immune response (GvHD)10 First human bone marrow transplantation8,153 Major histocompatibility complex discovered: human leukocyte antigen (HLA)13,15,16,154,155 Major preparative regimens developed: total body irradiation (TBI), cyclophosphamide, busulfan156-160 Methotrexate for control GvHD in animal models19,161-163 First allogeneic transplants for primary immunodeficiencies34-36 First successful transplant for end-stage leukemia164 First allogeneic transplants for aplastic anemia, PNH and Fanconi anemia44,45 GvHD and graft-versus-leukemia effects are separate reactions9,165 Acute grading system and first effective treatment of acute GvHD42,43 The first bone marrow donor registry was established by the Anthony Nolan Foundation Transplantation earlier in the course of leukemia37 Establishing graft-versus-leukemia effect in human patients56,57 First successful unrelated HLA-matched transplant in acute leukemia patient67 Establishment of conditioning regimen for non-malignant diseases leading to successful full immune reconstitution166 First successful treatment of chronic GvHD with immunosuppression combination167 Introduction of the concept of fractionated total body irradiation20 Introduction of acyclovir for HSV and VZV prophylaxis79 First successful transplant for thalassemia major33 Busulfan-cyclophosphamide conditioning for acute myeloid leukemia23 Establishment of the National Marrow Donor Program in the USA More effective acute GvHD prophylaxis with a combination of methotrexate and cyclosporine or tacrolimus40 HLA class I and HLA class II structures are defined, and HLA-typing transitions from cellular to DNA based62-64 Standard treatment of chronic GvHD established; prednisone, cyclosporine168,169 Early treatment with ganciclovir after allogeneic HCT to prevent CMV disease77 Use of peripheral blood stem cells mobilized with granulocyte colony stimulating factor (G-CSF)30,31,170 Donor lymphocyte infusions for disease relapse58,59 Umbilical cord blood as an alternative source of hematopoietic cells101 Impact of matching for class II HLA-DRB1, HLA-DQB1 and class I HLA-C171 Less toxic conditioning regimens expand allogeneic transplant for older patients172-176 CMV monitoring assays177 HLA-haploidentical related grafts for severe combined immunodeficiency and leukemia patients178-181 BMT CTN established Dramatic reduction in liver GvHD with ursodeoxycholic acidl95 Novel antibacterial and antifungals improve transplantation outcomes74-76,78 Improved outcomes of HLA-haploidentical transplants with post-transplant cyclophosphamide100 More judicious dosing of systemic glucocorticoids for treatment of acute GvHD96,97 Same outcomes with PBSC versus bone marrow from unrelated donors, and less chronic GvHD with bone marrow182 Addition of sirolimus for control of GvHD83-85 Introduction of novel therapies149,150 Novel CMV prophylaxis with letermovir126 Ruxolitinib for treatment of steroid-refractory acute GvHD87 Improved outcomes of aplastic anemia patients with HLA-haploidentical transplants123,124 CAR-T cell therapy as a â&#x20AC;&#x2DC;bridgeâ&#x20AC;&#x2122; to allogeneic HCT113-115

BMT CTN: The Blood and Marrow Transplant Clinical Trials Network; CAR T-cell: chimeric antigen receptor T-cell; CMV: cytomegalovirus; G-CSF; granulocyte-colony stimulating factor; GvHD: graft-versus-host disease; HCT: hematopoietic cell transplantation; HSV: herpes simplex virus; PBSC: peripheral blood stem cell; PNH: paroxysmal nocturnal hemoglobinuria; VZV: varicella zoster virus.

haematologica | 2020; 105(12)

2723

N. Granot and R. Storb

ments have been recently reviewed by Einsele et al.128 Another approach is adoptive CMV-specific T-cell therapy.129 Adoptive T-cell therapy has been effective not only for CMV, but also for Epstein-Barr virus-associated lymphoproliferative syndrome.130

Future directions For allogeneic HCT to become an even more relevant treatment modality, advances must be made in mitigating three major interrelated problems: regimen-related toxicities, post-HCT relapse, and chronic GvHD. As for the first, younger patients have traditionally received systemic, high-intensity conditioning regimens for maximal tumor cell kill before HCT and reducing the risk of relapse after HCT. Regimen intensity seems to be especially important for patients with AML. A prospective, randomized trial in younger patients with AML in morphologic remission published in 1990131 showed that conditioning with cyclophosphamide and 1,575 cGy TBI resulted in a significantly lower relapse rate than conditioning with cyclophosphamide and 1,200 cGy TBI, even though this benefit was offset by an increase in NRM. The importance of regimen intensity for controlling relapse was underscored by a recent randomized BMT-CTNsponsored trial in patients with AML or MDS.132 That trial compared conditioning with various high-intensity regimens to that with a number of reduced-intensity or nonmyeloablative regimens. The less-intense regimens showed lower NRM, but higher relapse mortality compared to myeloablative regimens, resulting in a statistically significantly lower relapse-free survival (though not overall survival). A recent, retrospective analysis of outcomes for patients with AML showed a clear advantage of myeloablative regimens in patients without measurable residual disease over reduced-intensity or non-myeloablative regimens both with respect to relapse and survival.133 Counterintuitively, this benefit was not seen in patients with measurable residual disease. However, it has to be considered that high-intensity regimens not only target the malignant tumor cells but also affect every other cell in the body. As a result, transplanted patients experience numerous short- and long-term toxicities. These include mucositis, gastrointestinal damage, veno-occlusive disease of the liver, adverse effects on growth and development, sterility, endocrine imbalance, cataracts, subsequent neoplasms, and others. For example, a recent, retrospective analysis of 4,905 patients transplanted at one center found a cumulative incidence of subsequent malignancies of 22% at 30 years.134 The magnitude of the tumor incidence was associated with regimen intensity. This observation coupled with the remarkable graft-versus-tumor effects following the newer reduced or minimal-intensity regimens seen in older patients raises the possibility of using such regimens more broadly, including in younger patients. This would be especially desirable for children and young adults given their longer life expectancy after HCT. Minimizing the systemic regimen intensity would markedly reduce the risks of shortand long-term toxicities, including secondary cancers. For this to become reality, the problem of post-HCT relapse needs to be harnessed. This could be accomplished by: (i) increasing graft-versus-tumor effects after HCT; (ii) reducing the tumor burden before HCT through adding target2724

ed radioimmunotherapy (RIT) with few off-target effects to low-intensity conditioning regimens; and (iii) administering maintenance therapy after HCT. In order to design approaches to increase graft-versustumor effects, a better understanding of polymorphic minor histocompatibility antigens specific for hematopoietic cells and distinct from those expressed on other tissue cells will be required. Such understanding might result in the generation of vaccines or of activated T cells that would selectively target hematopoietic antigens and enable exclusive destruction of tumor cells, rather than causing general GvHD.135,136 Another promising approach has been reducing the pre-transplant tumor burden through targeted RIT. Most often RIT have included monoclonal antibodies to the hematopoietic cell surface antigen, CD45, or the B-cell antigen CD20, which are coupled to a radioactive isotope, and then added to a minimal-intensity conditioning regimen. Such targeted RIT, while adding intensity, do not add significant toxicity. First encouraging trials have used b-emitting radionuclides for this purpose including iodine-131, rhenium-188 or yttrium-90.137-140 However, while somewhat effective in patients transplanted for myeloid and lymphoid malignancies, these isotopes have the disadvantages of long half-lives, relatively low dose rates, relatively low energy, and long path lengths. More recent, extensive preclinical work has led to the introduction of an Îą-emitting radionuclide, astatine-211 (211At), that has several major advantages over the heretofore used b-emitters.141 First, it emits much higher energy in its decay, second, its half-life is a mere 7.2 hours, and finally, its path length is only 60 mm. The high energy of this radioisotope leads to complete destruction of targeted cells, without the possibility of DNA repair. The short pathlength results in few off-target effects. 211At decays as a pure alpha, and therefore no isolation of patients is required. Phase I-II, first-in-human clinical trials are ongoing in patients with advanced myeloid malignancies receiving HCT from HLA-matched and unrelated donors, and from HLA-haploidentical related donors. Also, other target antigens are being explored in patients with MM, e.g., CD 38 and B-cell maturation antigen.142,143 Furthermore, trials have begun using 211At-based RIT in order to reduce the intensity of conventional, systemic conditioning for patients with non-malignant blood disorders, which would result in fewer short- and long-term toxicities following HCT. Encouraging results with maintenance therapy after HCT have been reported in patients with FLT3-ITD AML. A randomized, prospective trial in 204 patients conditioned with busulfan/cyclophosphamide showed significantly less relapse with post-HCT sorafenib compared to controls (1-year relapse 7% vs. 24.5%) and improved leukemia-free and overall survival.144 Early results of the Routine Antenatal Diagnostic Imaging with Ultrasound (RADIUS) study showed that midostaurin reduced postHCT relapse in FLT3-mutated AML patients but the difference was as yet not statistically significant.145 Results with post-HCT azacytidine have been equivocal. Bortezimib maintenance after HCT was beneficial in patients with high-risk MM,146 while rituximab maintenance has been equivocal in patients with CLL147 or ineffective in patients with NHL.148 Apart from relapse, chronic GvHD has remained the most challenging complication of allogeneic HCT. While haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation

there are significant, beneficial graft-versus-tumor effects associated with chronic GvHD, these have been offset by morbidity and mortality from this immune complication. The conundrum of preventing chronic GvHD while not sacrificing graft-versus-tumor effects has, as yet, not been satisfactorily resolved. Many current approaches, for example, global in vivo T-cell depletion with ATG or in vitro depletion of naïve T cells from the graft, have used highintensity, myeloablative conditioning regimens to control post-HCT relapse, but this comes at the cost of regimenrelated sequelae. Emerging approaches in preventing GvHD have been achieved through understanding of immunologic pathways of chronic GvHD. These include therapies targeting alloreactive T cells, alloreactive and autoreactive B cells through direct depletion from stem cell grafts (e.g., post-transplantation cyclophosphamide, CD34 selection, IL-2 and IL-17 therapy), in vivo depletion (e.g., rituximab, ofatumumab, obinutuzumab), and signal inhibition (e.g., ITK, JAK 1/2 , ROCK-II, BTK, SYK inhibition); such studies were recently reviewed in depth by Cutler et al.149 and MacDonald et al.150 The multitude of approaches is an indication that no single method was found to be unequivocally effective. In addition, novel therapies focus on adoptive transfer and expansion of regulatory T cells (Tregs) to prevent and treat chronic GvHD through administration of low-dose IL-2 and T Tregs sparing therapy. Most recently, removal of naïve T cells from the graft has shown encouraging results for GvHD prevention in younger patients with high-risk leukemia.151 These patients were conditioned with a very intensive conditioning regimen consisting of fludarabine, thiotepa and 13.2 Gy TBI. This approach reduced rates of chronic GvHD (9% at 2 years) while preserving immune reconstitution, without increasing relapse or NRM, though observation periods are still short. Also, the intensity of the conditioning regimen, the TBI dose in particular, places patients at high risk for short- and longterm complications such as secondary cancer. Researchers are also looking for ways to avoid GvHD without compromising graft-versus-leukemia effects in

References 1. Jacobson LO, Simmons EL, Marks EK, Robson MJ, Bethard WF, Gaston EO. The role of the spleen in radiation injury and recovery. J Lab Clin Med. 1950;35(5):746770. 2. Jacobson LO, Marks EK, Robson MJ, Gaston EO, Zirkle RE. Effect of spleen protection on mortality following x-irradiation. J Lab Clin Med. 1949;34:1538-1543. 3. Lorenz E, Uphoff D, Reid TR, Shelton E. Modification of irradiation injury in mice and guinea pigs by bone marrow injections. J Natl Cancer Inst. 1951;12(1):197-201. 4. Jacobson LO. Evidence for a humoral factor (or factors) concerned in recovery from radiation injury: a review. Cancer Res. 1952;12 (5):315-325. 5. Barnes DWH, Loutit JF. What is the recovery factor in spleen? Nucleonics. 1954;12(5):6871. 6. Main JM, Prehn RT. Successful skin homografts after the administration of high dosage X radiation and homologous bone marrow. J Natl Cancer Inst. 1955;15(4):10231029.

haematologica | 2020; 105(12)

HLA-haploidentical transplants by co-infusion of donorderived Tregs and conventional T cells, and infusion of NK cells after transplantation. Another approach has been selective depletion of B cells and T cells by removal of CD45RA+ or α/b+ cells from the graft. In a recent multicenter clinical trial, 80 pediatric acute leukemia patients were transplanted with α/b+ T- and B-cell depleted HLA-haploidentical grafts, and no additional post-transplantation GvHD prophylaxis.152 This study resulted in 5-year probability of chronic GvHD-free, relapse-free survival of 71%. In conclusion, most current methods of preventing chronic GvHD have adversely impacted graft-versustumor effects thereby increasing the risk of relapse. In order to get around this problem, systemic, myeloablative conditioning regimens have been intensified for better tumor cell kill. However, this has increased the risk of short- and long-term toxicities. Also, high-intensity regimens cannot be tolerated in older patients. It remains to be seen whether in the future, high-dose systemic conditioning can be replaced by RIT that specifically destroy the malignant hematopoietic cells but spare normal tissues. In addition, vaccines to hematopoietic antigens or use of in vitro generated T cells that are cytotoxic for hematopoietic cells but not for target tissues involved in GvHD, might generate powerful graft-versus-tumor effects and reduce the risk of post-transplantation relapse. Acknowledgments The authors wish to thank Helen Crawford for her assistance with manuscript and figure preparation. Support funding This work was supported by NIH grants P01 CA078902, P30 CA015704 from the National Cancer Institute and P01 HL122173 from the National Heart, Lung, and Blood Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, which had no involvement in the study design; the collection, analysis and interpretation of data; the writing of the report; nor in the decision to submit the article for publication.

7. van Bekkum DW, de Vries MJ. Radiation chimaeras. London: Logos Press Limited; 1967. 8. Thomas ED, Lochte HL Jr, Lu WC, Ferrebee JW. Intravenous infusion of bone marrow in patients receiving radiation and chemotherapy. N Eng J Med. 1957;257(11):491-496. 9. Mathe G, Amiel JL, Schwarzenberg L, Cattan A, Schneider M. Adoptive immunotherapy of acute leukemia: experimental and clinical results. Cancer Res. 1965;25(9):1525-1531. 10. Barnes DWH, Corp MJ, Loutit JF, Neal FE. Treatment of murine leukaemia with x-rays and homologous bone marrow. Preliminary communication. Br Med J. 1956;2(4993): 626-627. 11. Bortin MM. A compendium of reported human bone marrow transplants. Transplantation. 1970;9(6):571-587. 12. Epstein RB, Storb R, Ragde H, Thomas ED. Cytotoxic typing antisera for marrow grafting in littermate dogs. Transplantation. 1968;6(1):45-58. 13. Payne R, Rolfs MR. Fetomaternal leukocyte incompatibility. J Clin Invest. 1958;37(12): 1756-1763. 14. Terasaki PI, McClelland JD. Microdroplet

assay of human serum cytotoxins. Nature. 1964;204:998-1000. 15. Dausset J. Iso-leuco-anticorps. Acta Haematol. 1958;20(1-4):156-166. 16. van Rood JJ, Eernisse JG, van Leeuwen A. Leukocyte antibodies in sera from pregnant women. Nature. 1958;181(4625):1735-1736. 17. Bach FH, Amos DB. HU-1: major histocompatibility locus in man. Science. 1967;156(3781):1506-1508. 18. Storb R, Rudolph RH, Thomas ED. Marrow grafts between canine siblings matched by serotyping and mixed leukocyte culture. J Clin Invest. 1971;50(6):1272-1275. 19. Storb R, Epstein RB, Graham TC, Thomas ED. Methotrexate regimens for control of graft-versus-host disease in dogs with allogeneic marrow grafts. Transplantation. 1970;9(3):240-246. 20. Deeg HJ, Storb R, Weiden PL, et al. Highdose total-body irradiation and autologous marrow reconstitution in dogs: dose-raterelated acute toxicity and fractionationdependent long-term survival. Radiat Res. 1981;88(2):385-391. 21. Storb R, Buckner CD, Dillingham LA, Thomas ED. Cyclophosphamide regimens in rhesus monkeys with and without mar-

2725

N. Granot and R. Storb row infusion. Cancer Res. 1970;30(8):21952203. 22. Clift RA, Buckner CD, Thomas ED, et al. Marrow transplantation for chronic myeloid leukemia: a randomized study comparing cyclophosphamide and total body irradiation with busulfan and cyclophosphamide. Blood. 1994;84(6):2036-2043. 23. Santos GW, Tutschka PJ, Brookmeyer R, et al. Marrow transplantation for acute nonlymphocytic leukemia after treatment with busulfan and cyclophosphamide. N Eng J Med. 1983;309(22):1347-1353. 24. Storb R, Prentice RL, Thomas ED. Marrow transplantation for treatment of aplastic anemia. An analysis of factors associated with graft rejection. N Eng J Med. 1977;296(2):6166. 25. van Bekkum DW, Bach F, Bergan JJ, et al. Bone marrow transplantation from histocompatible, allogeneic donors for aplastic anemia. A report from the ACS/NIH bone marrow transplant registry. JAMA. 1976;236(10):1131-1135. 26. Storb R, Etzioni R, Anasetti C, et al. Cyclophosphamide combined with antithymocyte globulin in preparation for allogeneic marrow transplants in patients with aplastic anemia. Blood. 1994;84(3):941-949. 27. Storb R, Floersheim GL, Weiden PL, et al. Effect of prior blood transfusions on marrow grafts: abrogation of sensitization by procarbazine and antithymocyte serum. J Immunol. 1974;112(4):1508-1516. 28. Tsoi MS, Storb R, Dobbs S, Thomas ED. Specific suppressor cells in graft-host tolerance of HLA-identical marrow transplantation. Nature. 1981;292(5821):355-357. 29. Storb R, Graham TC, Epstein RB, Sale GE, Thomas ED. Demonstration of hemopoietic stem cells in the peripheral blood of baboons by cross circulation. Blood. 1977;50(3):537542. 30. Bensinger WI, Weaver CH, Appelbaum FR, et al. Transplantation of allogeneic peripheral blood stem cells mobilized by recombinant human granulocyte colony-stimulating factor. Blood. 1995;85(6):1655-1658. 31. Schmitz N, Dreger P, Suttorp M, et al. Primary transplantation of allogeneic peripheral blood progenitor cells mobilized by filgrastim (granulocyte colony-stimulating factor). Blood. 1995;85(6):1666-1672. 32. Weiden PL, Hackman RC, Deeg HJ, Graham TC, Thomas ED, Storb R. Long-term survival and reversal of iron overload after marrow transplantation in dogs with congenital hemolytic anemia. Blood. 1981;57(1):66-70. 33. Thomas ED, Buckner CD, Sanders JE, et al. Marrow transplantation for thalassaemia. Lancet. 1982;2(8292):227-229. 34. de Koning J, van Bekkum DW, Dicke KA, Dooren LJ, Radl J, van Rood JJ. Transplantation of bone-marrow cells and fetal thymus in an infant with lymphopenic immunological deficiency. Lancet. 1969;1(7608):1223-1227. 35. Bach FH, Albertini RJ, Joo P, Anderson JL, Bortin MM. Bone-marrow transplantation in a patient with the Wiskott-Aldrich syndrome. Lancet. 1968;2(7583):1364-1366. 36. Gatti RA, Meuwissen HJ, Allen HD, Hong R, Good RA. Immunological reconstitution of sex-linked lymphopenic immunological deficiency. Lancet. 1968;2(7583):1366-1369. 37. Thomas ED, Storb R, Clift RA, et al. Bonemarrow transplantation (first of two parts). N Engl J Med. 1975;292(16):832-843. 38. Thomas ED, Storb R, Clift RA, et al. Bonemarrow transplantation (second of two parts). N Engl J Med. 1975;292(17):895-902.

2726

39. Storb R, Deeg HJ, Farewell V, et al. Marrow transplantation for severe aplastic anemia: methotrexate alone compared with a combination of methotrexate and cyclosporine for prevention of acute graft-versus-host disease. Blood. 1986;68(1):119-125. 40. Storb R, Deeg HJ, Whitehead J, et al. Methotrexate and cyclosporine compared with cyclosporine alone for prophylaxis of acute graft versus host disease after marrow transplantation for leukemia. N Eng J Med. 1986;314(12):729-735. 41. Storb R, Raff RF, Appelbaum FR, et al. FK506 and methotrexate prevent graft-versushost disease in dogs given 9.2 Gy total body irradiation and marrow grafts from unrelated dog leukocyte antigen-nonidentical donors. Transplantation. 1993;56(4):800807. 42. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-Amatched sibling donors. Transplantation. 1974;18(4):295-304. 43. Storb R, Gluckman E, Thomas ED, et al. Treatment of established human graft-versus-host disease by antithymocyte globulin. Blood. 1974;44(1):57-75. 44. Thomas ED, Buckner CD, Storb R, et al. Aplastic anaemia treated by marrow transplantation. Lancet. 1972;1(7745):284-289. 45. Storb R, Thomas ED, Buckner CD, et al. Allogeneic marrow grafting for treatment of aplastic anemia. Blood. 1974;43(2):157-180. 46. Champlin RE, Perez WS, Passweg JR, et al. Bone marrow transplantation for severe aplastic anemia: a randomized controlled study of conditioning regimens. Blood. 2007;109(10):4582-4585. 47. Locatelli F, Bruno B, Zecca M, et al. Cyclosporin A and short-term methotrexate versus cyclosporin A as graft versus host disease prophylaxis in patients with severe aplastic anemia given allogeneic bone marrow transplantation from an HLA-identical sibling: results of a GITMO/EBMT randomized trial. Blood. 2000;96(5):1690-1697. 48. Chu R, Brazauskas R, Kan F, et al. Comparison of outcomes after transplantation of G-CSF stimulated bone marrow grafts versus bone marrow or peripheral blood grafts from HLA-matched sibling donors for patients with severe aplastic anemia. Biol Blood Marrow Transplant. 2011;17(7):1018-1024. 49. Atta EH, de Sousa AM, Schirmer MR, Bouzas LF, Nucci M, Abdelhay E. Different outcomes between cyclophosphamide plus horse or rabbit antithymocyte globulin for HLA-identical sibling bone marrow transplant in severe aplastic anemia. Biol Blood Marrow Transplant. 2012;18(12):1876-1882. 50. Bacigalupo A, SociĂŠ G, Schrezenmeier H, et al. Bone marrow versus peripheral blood as the stem cell source for sibling transplants in acquired aplastic anemia: survival advantage for bone marrow in all age groups. Haematologica. 2012;97(8):1142-1148. 51. Kim H, Lee JH, Joo YD, et al. A randomized comparison of cyclophosphamide vs. reduced dose cyclophosphamide plus fludarabine for allogeneic hematopoietic cell transplantation in patients with aplastic anemia and hypoplastic myelodysplastic syndrome. Ann Hematol. 2012;91(9):14591469. 52. Rashidi A, Slade M, DiPersio JF, Westervelt P, Vij R, Romee R. Post-transplant high-dose cyclophosphamide after HLA-matched vs haploidentical hematopoietic cell transplantation for AML. Bone Marrow Transplant.

2016;51(12):1561-1564. 53. Sangiolo D, Storb R, Deeg HJ, et al. Outcome of allogeneic hematopoietic cell transplantation from HLA-identical siblings for severe aplastic anemia in patients over 40 years of age. Biol Blood Marrow Transplant. 2010;16(10):1411-1418. 54. Gallo S, Woolfrey AE, Burroughs LM, et al. Marrow grafts from HLA-identical siblings for severe aplastic anemia: does limiting the number of transplanted marrow cells reduce the risk of chronic GVHD? Bone Marrow Transplant. 2016;51(12):1573-1578. 55. Kahl C, Leisenring W, Deeg HJ, et al. Cyclophosphamide and antithymocyte globulin as a conditioning regimen for allogeneic marrow transplantation in patients with aplastic anaemia: a long-term followup. Br J Haematol. 2005;130(5):747-751. 56. Weiden PL, Flournoy N, Thomas ED, et al. Antileukemic effect of graft-versus-host disease in human recipients of allogeneic-marrow grafts. N Eng J Med. 1979;300(19):10681073. 57. Weiden PL, Sullivan KM, Flournoy N, Storb R, Thomas ED, and the Seattle Marrow Transplant Team. Antileukemic effect of chronic graft-versus-host disease. Contribution to improved survival after allogeneic marrow transplantation. N Eng J Med. 1981;304(25):1529-1533. 58. Kolb HJ, Schattenberg A, Goldman JM, et al. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. European Group for Blood and Marrow Transplantation Working Party Chronic Leukemia. Blood. 1995;86(5):20412050. 59. Slavin S, Naparstek E, Nagler A, et al. Allogeneic cell therapy with donor peripheral blood cells and recombinant human interleukin-2 to treat leukemia relapse after allogeneic bone marrow transplantation. Blood. 1996;87(6):2195-2204. 60. Marmont AM, Gale RP, Butturini A, et al. Tcell depletion in allogeneic bone marrow transplantation: progress and problems. Haematologica. 1989;74(3):235-248. 61. Keever CA, Small TN, Flomenberg N, et al. Immune reconstitution following bone marrow transplantation: comparison of recipients of T-cell depleted marrow with recipients of conventional marrow grafts. Blood. 1989;73(5):1340-1350. 62. Bjorkman PJ, Saper MA, Samraoui B, Bennett WS, Strominger JL, Wiley DC. Structure of the human class I histocompatibility antigen, HLA-A2. Nature. 1987;329(6139):506-512. 63. Brown JH, Jardetzky TS, Gorga JC, et al. Three-dimensional structure of the human class II histocompatibility antigen HLADR1. Nature. 1993;364(6432):33-39. 64. Levine JE, Yang SY. SSOP typing of the Tenth International Histocompatibility Workshop reference cell lines for HLA-C alleles. Tissue Antigens. 1994;44(3):174-183. 65. Storb R, Epstein RB, Bryant J, Ragde H, Thomas ED. Marrow grafts by combined marrow and leukocyte infusions in unrelated dogs selected by histocompatibility typing. Transplantation. 1968;6(4):587-593. 66. Storb R, Weiden PL, Graham TC, Lerner KG, Thomas ED. Marrow grafts between DLA-identical and homozygous unrelated dogs. Evidence for an additional locus involved in graft-versus-host disease. Transplantation. 1977;24(3):165-174. 67. Hansen JA, Clift RA, Thomas ED, Buckner CD, Storb R, Giblett ER. Transplantation of marrow from an unrelated donor to a

haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation patient with acute leukemia. N Eng J Med. 1980;303(10):565-567. 68. Gragert L, Eapen M, Williams E, et al. HLA match likelihoods for hematopoietic stemcell grafts in the US registry. N Eng J Med. 2014;371(4):339-348. 69. Petersdorf EW. The World Marrow Donor Association: 20 years of international collaboration for the support of unrelated donor and cord blood hematopoietic cell transplantation. Bone Marrow Transplant. 2010;45(5):807-810. 70. Oudshoorn M, van Leeuwen A, Zanden HG, van Rood JJ. Bone Marrow Donors Worldwide: a successful exercise in international cooperation. Bone Marrow Transplant. 1994;14(1):3-8. 71. Pidala J, Kim J, Schell M, et al. Race/ethnicity affects the probability of finding an HLA-A, -B, -C and -DRB1 allele-matched unrelated donor and likelihood of subsequent transplant utilization. Bone Marrow Transplant. 2013;48(3):346-350. 72. Gratwohl A, Pasquini MC, Aljurf M, et al. One million haemopoietic stem-cell transplants: a retrospective observational study. Lancet Haematol. 2015;2(3):e91-e100. 73. Niederwieser D, Baldomero H, Szer J, et al. Hematopoietic stem cell transplantation activity worldwide in 2012 and a SWOT analysis of the Worldwide Network for Blood and Marrow Transplantation Group including the global survey. Bone Marrow Transplant. 2016;51(6):778-785. 74. Ethier MC, Science M, Beyene J, Briel M, Lehrnbecher T, Sung L. Mould-active compared with fluconazole prophylaxis to prevent invasive fungal diseases in cancer patients receiving chemotherapy or haematopoietic stem-cell transplantation: a systematic review and meta-analysis of randomised controlled trials. Br J Cancer. 2012;106(10):1626-1637. 75. Gafter-Gvili A, Fraser A, Paul M, et al. Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy. Cochrane Database Syst Rev. 2012;1:CD004386. 76. Reuter S, Kern WV, Sigge A, et al. Impact of fluoroquinolone prophylaxis on reduced infection-related mortality among patients with neutropenia and hematologic malignancies. Clin Infect Dis. 2005;40(8):10871093. 77. Goodrich JM, Mori M, Gleaves CA, et al. Early treatment with ganciclovir to prevent cytomegalovirus disease after allogeneic bone marrow transplantation. N Eng J Med. 1991;325(23):1601-1607. 78. Robenshtok E, Gafter-Gvili A, Goldberg E, et al. Antifungal prophylaxis in cancer patients after chemotherapy or hematopoietic stem-cell transplantation: systematic review and meta-analysis (Review). J Clin Oncol. 2007;25(34):5471-5489. 79. Saral R, Burns WH, Laskin OL, Santos GW, Lietman PS. Acyclovir prophylaxis of herpes-simplex-virus infections. N Eng J Med. 1981;305(2):63-67. 80. Molina AJ, Storb RF. Hematopoietic stem cell transplantation in older adults. In: Rowe JM, Lazarus HM, Carella AM, editors. Handbook of Bone Marrow Transplantation. London, UK: Martin Dunitz Ltd.; 2000. p. 111-137. 81. Storb R, Yu C, Wagner JL, et al. Stable mixed hematopoietic chimerism in DLA-identical littermate dogs given sublethal total body irradiation before and pharmacological immunosuppression after marrow transplantation. Blood. 1997;89(8):3048-3054.

haematologica | 2020; 105(12)

82. Yu C, Seidel K, Nash RA, et al. Synergism between mycophenolate mofetil and cyclosporine in preventing graft-versus-host disease among lethally irradiated dogs given DLA-nonidentical unrelated marrow grafts. Blood. 1998;91(7):2581-2587. 83. Armand P, Kim HT, Sainvil MM, et al. The addition of sirolimus to the graft-versus-host disease prophylaxis regimen in reduced intensity allogeneic stem cell transplantation for lymphoma: a multicentre randomized trial. Br J Haematol. 2016;173(1):96-104. 84. Sandmaier BM, Kornblit B, Storer BE, et al. Addition of sirolimus to standard cyclosporine plus mycophenolate mofetilbased graft-versus-host disease prophylaxis for patients after unrelated non-myeloablative haemopoietic stem cell transplantation: a multicentre, randomised, phase 3 trial. Lancet Haematol. 2019;6(8):e409-e418. 85. Kornblit B, Maloney DG, Storer BE, et al. A randomized phase II trial of tacrolimus, mycophenolate mofetil and sirolimus after non-myeloablative unrelated donor transplantation. Haematologica. 2014;99(10): 1624-1631. 86. Hogan WJ, Little MT, Zellmer E, et al. Postgrafting immunosuppression with sirolimus and cyclosporine facilitates stable mixed hematopoietic chimerism in dogs given sublethal total body irradiation before marrow transplantation from DLA-identical littermates. Biol Blood Marrow Transplant. 2003;9(8):489-495. 87. Jagasia M, Perales MA, Schroeder MA, et al. Results from REACH1, a single-arm phase 2 study of ruxolitinib in combination with corticosteroids for the treatment of steroidrefractory acute graft-vs-host disease (abstract). Blood. 2018;132(Suppl 1):601. 88. Flowers MED, Apperley JF, van Besien K, et al. A multicenter prospective phase II randomized study of extracorporeal photopheresis for treatment of chronic graft-versus-host disease. Blood. 2008;112(7):26672674. 89. Sorror ML, Maris MB, Storb R, et al. Hematopoietic cell transplantation (HCT)specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. 90. Ferrara JLM, Chaudhry MS. GVHD: biology matters. Hematology Am Soc Hematol Educ Program. 2018;2018(1):221-227. 91. Storb R, Gyurkocza B, Storer BE, et al. Graftversus-host disease and graft-versus-tumor effects after allogeneic hematopoietic cell transplantation. J Clin Oncol. 2013;31(12): 1530-1538. 92. Granot N, Storer BE, Cooper JP, Flowers ME, Sandmaier BM, Storb R. Allogeneic hematopoietic cell transplantation in the outpatient setting. Biol Blood Marrow Transplant. 2019;25(11):2152-2159. 93. Kahl C, Storer BE, Sandmaier BM, et al. Relapse risk among patients with malignant diseases given allogeneic hematopoietic cell transplantation after nonmyeloablative conditioning. Blood. 2007;110(7):2744-2748. 94. Cooper JP, Storer BE, Granot N, et al. Allogeneic hematopoietic cell transplantation with non-myeloablative conditioning for patients with hematologic malignancies: improved outcomes over two decades. Haematologica. 2020 Jun 4. [Epub ahead of print]. 95. Ruutu T, Eriksson B, Remes K, et al. Ursodeoxycholic acid for the prevention of hepatic complications in allogeneic stem cell transplantation. Blood. 2002;100(6):19771983.

96. Hockenbery DM, Cruickshank S, Rodell TC, et al. A randomized, placebo-controlled trial of oral beclomethasone dipropionate as a prednisone-sparing therapy for gastrointestinal graft-versus-host disease. Blood. 2007;109(10):4557-4563. 97. Mielcarek M, Storer BE, Boeckh M, et al. Initial therapy of acute graft-versus-host disease with low-dose prednisone does not compromise patient outcomes. Blood. 2009;113(13):2888-2894. 98. Ullmann AJ, Lipton JH, Vesole DH, et al. Posaconazole or fluconazole for prophylaxis in severe graft-versus-host disease. N Eng J Med. 2007;356(4):335-347. 99. Storb R, Sandmaier BM. Nonmyeloablative allogeneic hematopoietic cell transplantation. Haematologica. 2016;101(5):521-530. 100. Luznik L, O'Donnell PV, Symons HJ, et al. HLA-haploidentical bone marrow transplantation for hematologic malignancies using nonmyeloablative conditioning and highdose, posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2008;14(6):641-650. 101. Gluckman E, Rocha V, Boyer-Chammard A, et al. Outcome of cord-blood transplantation from related and unrelated donors. N Eng J Med. 1997;337(6):373-381. 102. Brunstein CG, Gutman JA, Weisdorf DJ, et al. Allogeneic hematopoietic cell transplantation for hematological malignancy: relative risks and benefits of double umbilical cord blood. Blood. 2010;116(22):4693-4699. 103. Wagner JE Jr, Eapen M, Carter S, et al. Oneunit versus two-unit cord-blood transplantation for hematologic cancers. N Eng J Med. 2014;371(18):1685-1694. 104. Delaney C, Heimfeld S, Brashem-Stein C, Voorhies H, Manger RL, Bernstein ID. Notch-mediated expansion of human cord blood progenitor cells capable of rapid myeloid reconstitution. Nat Med. 2010;16 (2):232-237. 105. de Lima M, McNiece I, Robinson SN, et al. Cord-blood engraftment with ex vivo mesenchymal-cell coculture. N Eng J Med. 2012;367(24):2305-2315. 106. D'Souza A, Fretham C, Lee SJ, et al. Current use of and trends in hematopoietic cell transplantation in the United States. Biol Blood Marrow Transplant. 2020;26(8):e177-e182. 107. Petersdorf EW, Bengtsson M, De Santis D, et al. Role of HLA-DP expression in graft-versus-host disease after unrelated donor transplantation. J Clin Oncol. 2020;38(24):27122718. 108. Kanakry CG, O'Donnell PV, Furlong T, et al. Multi-institutional study of post-transplantation cyclophosphamide as single-agent graft-versus-host disease prophylaxis after allogeneic bone marrow transplantation using myeloablative busulfan and fludarabine conditioning. J Clin Oncol. 2014;32(31):3497-3505. 109. Wang Y, Liu DH, Liu KY, et al. Long-term follow-up of haploidentical hematopoietic stem cell transplantation without in vitro T cell depletion for the treatment of leukemia: nine years of experience at a single center. Cancer. 2013;119(5):978-985. 110. Patil S, Potter V, Mohty M. Review of conditioning regimens for haplo-identical donor transplants using post-transplant cyclophosphamide in recipients of G-CSF mobilised peripheral stem cell. Cancer Treat Rev. 2020;89:102071. 111. Sanz J, Montoro J, Solano C, et al. Prospective randomized study comparing myeloablative unrelated umbilical cord blood transplantation versus HLA-hap-

2727

N. Granot and R. Storb loidentical related stem cell transplantation for adults with hematologic malignancies. Biol Blood Marrow Transplant. 2020;26(2): 358-366. 112. Scott BL. Long-term follow up of BMT CTN 0901, a randomized phase III trial comparing myeloablative (MAC) to reduced intensity conditioning (RIC) prior to hematopoietic cell transplantation (HCT) for acute myeloid leukemia (AML) or myelodysplasia (MDS) (MAvRIC trial). Biol Blood Marrow Transplant. 2020;26(3):abstract S11. 113. Shadman M, Gauthier J, Hay KA, et al. Safety of allogeneic hematopoietic cell transplant in adults after CD19-targeted CAR Tcell therapy. Blood Adv. 2019;3(20):30623069. 114. Zhang X, Lu XA, Yang J, et al. Efficacy and safety of anti-CD19 CAR T-cell therapy in 110 patients with B-cell acute lymphoblastic leukemia with high-risk features. Blood Adv. 2020;4(10):2325-2338. 115. Park JH, Riviere I, Gonen M, et al. Long-term follow-up of CD19 CAR therapy in acute lymphoblastic leukemia. N Engl J Med. 2018;378(5):449-459. 116. Hamadini M, Litovich C, Khanal M, Kharfan-Dabaja MA, Ahn KW. Higher total body irradiation (TBI) dose-intensity in fludarabine (Flu)/TBI-based reduced-intensity conditioning (RIC) regimen is associated with inferior survival in non-Hodgkin lymphoma (NHL) patients undergoing allogeneic hematopoietic cell transplantation (alloHCT). Biol Blood Marrow Transplant. 2020;26(3 Suppl):S21-S22. 117. Greil C, Engelhardt M, Ihorst G, et al. Allogeneic transplantation of multiple myeloma patients may allow long-term survival in carefully selected patients with acceptable toxicity and preserved quality of life. Haematologica. 2019;104(2):370-379. 118. Gonsalves WI, Buadi FK, Ailawadhi S, et al. Utilization of hematopoietic stem cell transplantation for the treatment of multiple myeloma: a Mayo Stratification of Myeloma and Risk-Adapted Therapy (mSMART) consensus statement. Bone Marrow Transplant. 2019;54(3):353-367. 119. Gennery AR, Slatter MA, Grandin L, et al. Transplantation of hematopoietic stem cells and long-term survival for primary immunodeficiencies in Europe: entering a new century, do we do better? J Allergy Clin Immunol. 2010;126(3):602-610. 120. Pai SY, Logan BR, Griffith LM, et al. Transplantation outcomes for severe combined immunodeficiency, 2000-2009. N Engl J Med. 2014;371(5):434-446. 121. Sheth VS, Potter V, Gandhi SA, et al. Similar outcomes of alemtuzumab-based hematopoietic cell transplantation for SAA patients older or younger than 50 years. Blood Adv. 2019;3(20):3070-3079. 122. Rice C, Eikema DJ, Marsh JCW, et al. Allogeneic hematopoietic cell transplantation in patients aged 50 years or older with severe aplastic anemia. Biol Blood Marrow Transplant. 2019;25(3):488-495. 123. ElGohary G, El Fakih R, de Latour R, et al. Haploidentical hematopoietic stem cell transplantation in aplastic anemia: a systematic review and meta-analysis of clinical outcome on behalf of the severe aplastic anemia working party of the European Group for Blood and Marrow Transplantation (SAAWP of EBMT). Bone Marrow Transplant. 2020;55(10):1906-1917. 124. Xu LP, Wang SQ, Ma YR, et al. Who is the best haploidentical donor for acquired severe aplastic anemia? Experience from a

2728

multicenter study. J Hematol Oncol. 2019;12(1):87. 125. Kumar A, Reljic T, Hamadani M, Mohty M, Kharfan-Dabaja MA. Antithymocyte globulin for graft-versus-host disease prophylaxis: an updated systematic review and metaanalysis. Bone Marrow Transplant. 2019;54(7):1094-1106. 126. Marty FM, Ljungman P, Chemaly RF, et al. Letermovir prophylaxis for cytomegalovirus in hematopoietic-cell transplantation. N Engl J Med. 2017;377(25):2433-2444. 127. Papanicolaou GA, Silveira FP, Langston AA, et al. Maribavir for refractory or resistant cytomegalovirus infections in hematopoietic-cell or solid-organ transplant recipients: a randomized, dose-ranging, double-blind, phase 2 study. Clin Infect Dis. 2019;68(8):1255-1264. 128. Einsele H, Ljungman P, Boeckh M. How I treat CMV reactivation after allogeneic hematopoietic stem cell transplantation. Blood. 2020;135(19):1619-1629. 129. Einsele H, Kapp M, Grigoleit GU. CMV-specific T cell therapy. Blood Cells Mol Dis. 2008;40(1):71-75. 130. Barrett AJ, Bollard CM. The coming of age of adoptive T-cell therapy for viral infection after stem cell transplantation. Ann Transl Med. 2015;3(5):62. 131. Clift RA, Buckner CD, Appelbaum FR, et al. Allogeneic marrow transplantation in patients with acute myeloid leukemia in first remission: a randomized trial of two irradiation regimens. Blood. 1990;76(9): 1867-1871. 132. Scott BL, Pasquini MC, Logan BR, et al. Myeloablative versus reduced-intensity hematopoietic cell transplantation for acute myeloid leukemia and myelodysplastic syndromes. J Clin Oncol. 2017;35(11):11541161. 133. Morsink LM, Sandmaier BM, Othus M, et al. Conditioning intensity, pre-transplant flow cytometric measurable residual disease, and outcome in adults with acute myeloid leukemia undergoing allogeneic hematopoietic cell transplantation. Cancers (Basel). 2020;12(9):E2339. 134. Baker KS, Leisenring WM, Goodman PJ, et al. Total body irradiation dose and risk of subsequent neoplasms following allogeneic hematopoietic cell transplantation. Blood. 2019;133(26):2790-2799. 135. Biernacki MA, Bleakley M. Neoantigens in hematologic malignancies. Front Immunol. 2020;11:121. 136. Biernacki MA, Sheth VS, Bleakley M. T cell optimization for graft-versus-leukemia responses. JCI Insight. 2020;5(9):e134939. 137. Pagel JM, Gooley TA, Rajendran J, et al. Allogeneic hematopoietic cell transplantation after conditioning with 131I-anti-CD45 antibody plus fludarabine and low-dose total body irradiation for elderly patients with advanced acute myeloid leukemia or high-risk myelodysplastic syndrome. Blood. 2009;114(27):5444-5453. 138. Gopal AK, Guthrie KA, Rajendran J, et al. 90 Y-Ibritumomab tiuxetan, fludarabine, and TBI-based nonmyeloablative allogeneic transplantation conditioning for patients with persistent high-risk B-cell lymphoma. Blood. 2011;118(4):1132-1139. 139. Puronen CE, Cassaday RD, Stevenson PA, et al. Long-term follow up of 90Y-ibritumomab tiuxetan, fludarabine and total body irradiation-based non-myeloablative allogeneic transplant conditioning for persistent highrisk B-cell lymphoma. Biol Blood Marrow Transplant. 2018;24(11):2211-2215.

140. Mawad R, Gooley TA, Rajendran JG, et al. Radiolabeled anti-CD45 antibody with reduced-intensity conditioning and allogeneic transplantation for younger patients with advanced acute myeloid leukemia or myelodysplastic syndrome. Biol Blood Marrow Transplant. 2014;20(9):1363-1368. 141. Chen Y, Kornblit B, Hamlin DK, et al. Durable donor engraftment after radioimmunotherapy using alpha-emitter astatine211-labeled anti-CD45 antibody for conditioning in allogeneic hematopoietic cell transplantation. Blood. 2012;119(5):11301138. 142. O'Steen S, Comstock ML, Orozco JJ, et al. The alpha-emitter astatine-211 targeted to CD38 can eradicate multiple myeloma in a disseminated disease model. Blood. 2019;134(15):1247-1256. 143. Green DJ, Orgun NN, Jones JC, et al. A preclinical model of CD38-pretargeted radioimmunotherapy for plasma cell malignancies. Cancer Res. 2014;74(4):1179-1189. 144. Xuan L, Wang Y, Huang F, et al. Sorafenib maintenance in patients with FLT3-ITD acute myeloid leukaemia undergoing allogeneic haematopoietic stem-cell transplantation: an open-label, multicentre, randomised phase 3 trial. Lancet Oncol. 2020;21(9):12011212. 145. Maziarz RT, Fernandez H, Patnaik MM, et al. Radius: midostaurin (mido) plus standard of care (SOC) after allogeneic stem cell transplant (alloSCT) in patients (pts) with FLT3internal tandem duplication (ITD)–mutated acute myeloid leukemia (AML). Biol Blood Marrow Transplant. 2019;25(3 Suppl):S11S12. 146. Green DJ, Maloney DG, Storer BE, et al. Tandem autologous/allogeneic hematopoietic cell transplantation with bortezomib maintenance therapy for high-risk myeloma. Blood Adv. 2017;1(24):2247-2256. 147. Shadman M, Maloney DG, Storer B, et al. Rituximab-based allogeneic transplant for chronic lymphocytic leukemia with comparison to historical experience. Bone Marrow Transplant. 2020;55(1):172-181. 148. Granot N, Rezvani AR, Pender BS, et al. Impact of rituximab and host/donor Fc receptor polymorphisms after allogeneic hematopoietic cell transplantation for CD20+ B-cell malignancies. Biol Blood Marrow Transplant. 2020 Jul 18. [Epub ahead of print]. 149. Cutler CS, Koreth J, Ritz J. Mechanistic approaches for the prevention and treatment of chronic GVHD. Blood. 2017;129(1): 22-29. 150. MacDonald KP, Hill GR, Blazar BR. Chronic graft-versus-host disease: biological insights from preclinical and clinical studies. Blood. 2017;129(1):13-21. 151. Bleakley M, Heimfeld S, Jones LA, et al. Engineering human peripheral blood stem cell grafts that are depleted of naive T cells and retain functional pathogen-specific memory T cells. Biol Blood Marrow Transplant. 2014;20(5):705-716. 152. Locatelli F, Merli P, Pagliara D, et al. Outcome of children with acute leukemia given HLA-haploidentical HSCT after alphabeta T-cell and B-cell depletion. Blood. 2017;130(5):677-685. 153. Mathé G, Jammet H, Pendic B, et al. [Transfusions et greffes de moelle osseuse homologue chez des humains irradiés a haute dose accidentellement] Rev Fr Etud Clin Biol. 1959;4(3):226-238. 154. Snell GD. The Nobel Lectures in Immunology. Lecture for the Nobel Prize for

haematologica | 2020; 105(12)

History and future of hematopoietic cell transplantation Physiology or Medicine, 1980: studies in histocompatibility. Scand J Immunol. 1992;36(4):513-526. 155. Park I, Terasaki P. Origins of the first HLA specificities. Hum Immunol. 2000;61(3):185189. 156. Ferrebee JW, Lochte HL Jr, Jaretzki A III, Sahler OD, Thomas ED. Successful marrow homograft in the dog after radiation. Surgery. 1958;43(3):516-520. 157. Mannick JA, Lochte HL Jr, Ashley CA, Thomas ED, Ferrebee JW. Autografts of bone marrow in dogs after lethal total-body radiation. Blood. 1960;15:255-266. 158. Cavins JA, Scheer SC, Thomas ED, Ferrebee JW. The recovery of lethally irradiated dogs given infusions of autologous leukocytes preserved at -80 C. Blood. 1964;23:38-43. 159. Storb R, Epstein RB, Rudolph RH, Thomas ED. Allogeneic canine bone marrow transplantation following cyclophosphamide. Transplantation. 1969;7(3):378-386. 160. Santos GW, Owens AH Jr. Allogeneic marrow transplants in cyclophosphamide treated mice. Transplant Proc. 1969;1(1):44-46. 161. Uphoff DE. Alteration of homograft reaction by A-methopterin in lethally irradiated mice treated with homologous marrow. Proc Soc Exp Biol Med. 1958;99(3):651653. 162. Boak JL, Fox M, Wilson RE. Activity of lymphoid tissues from antilymphocyte-serumtreated mice. Lancet. 1967;1(493):750-752. 163. Brent L, Courtenay T, Gowland G. Immunological reactivity of lymphoid cells after treatment with anti-lymphocytic serum. Nature. 1967;215(109):1461-1464. 164. Thomas ED, Buckner CD, Rudolph RH, et al. Allogeneic marrow grafting for hematologic malignancy using HL-A matched donor-recipient sibling pairs. Blood. 1971;38(3):267-287. 165. Bortin MM, Rimm AA, Saltzstein EC, Rodey GE. Graft versus leukemia. III. Apparent independent antihost and antileukemic activity of transplanted immunocompetent cells. Transplantation. 1973;16(3):182-188. 166. Kapoor N, Kirkpatrick D, Oleske J, et al.

haematologica | 2020; 105(12)

Reconstitution of normal megakaryocytopoiesis and immunologic functions in Wiskott-Aldrich syndrome by marrow transplantation following myeloablation and immunosuppression with busulfan and cyclophosphamide. Blood. 1981;57(4):692696. 167. Sullivan KM, Shulman HM, Storb R, et al. Chronic graft-versus-host disease in 52 patients: adverse natural course and successful treatment with combination immunosuppression. Blood. 1981;57(2):267-276. 168. Sullivan KM, Witherspoon RP, Storb R, et al. Alternating-day cyclosporine and prednisone for treatment of high-risk chronic graft-versus-host disease. Blood. 1988;72(2):555-561. 169. Sullivan KM, Witherspoon RP, Storb R, et al. Prednisone and azathioprine compared with prednisone and placebo for treatment of chronic graft-versus-host disease: prognostic influence of prolonged thrombocytopenia after allogeneic marrow transplantation. Blood. 1988;72(2):546-554. 170. Kรถrbling M, Przepiorka D, Huh YO, et al. Allogeneic blood stem cell transplantation for refractory leukemia and lymphoma: potential advantage of blood over marrow allografts. Blood. 1995;85(6):1659-1665. 171. Petersdorf EW, Gooley TA, Anasetti C, et al. Optimizing outcome after unrelated marrow transplantation by comprehensive matching of HLA class I and II alleles in the donor and recipient. Blood. 1998;92(10):3515-3520. 172. Giralt S, Estey E, Albitar M, et al. Engraftment of allogeneic hematopoietic progenitor cells with purine analog-containing chemotherapy: harnessing graft-versusleukemia without myeloablative therapy. Blood. 1997;89(12):4531-4536. 173. Slavin S, Nagler A, Naparstek E, et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91(3):756-763. 174. Giralt S, Thall PF, Khouri I, et al. Melphalan

and purine analog-containing preparative regimens: reduced-intensity conditioning for patients with hematologic malignancies undergoing allogeneic progenitor cell transplantation. Blood. 2001;97(3):631-637. 175. McSweeney PA, Niederwieser D, Shizuru JA, et al. Hematopoietic cell transplantation in older patients with hematologic malignancies: replacing high-dose cytotoxic therapy with graft-versus-tumor effects. Blood. 2001;97(11):3390-3400. 176. Spitzer TR, McAfee S, Sackstein R, et al. Intentional induction of mixed chimerism and achievement of antitumor responses after nonmyeloablative conditioning therapy and HLA-matched donor bone marrow transplantation for refractory hematologic malignancies. Biol Blood Marrow Transplant. 2000;6(3A):309-320. 177. Boeckh M, Boivin G. Quantitation of cytomegalovirus: methodologic aspects and clinical applications. Clin Microbiol Rev. 1998;11(3):533-554. 178. Reisner Y, Kapoor N, Kirkpatrick D, et al. Transplantation for severe combined immunodeficiency with HLA-A,B,D,DR incompatible parental marrow cells fractionated by soybean agglutinin and sheep red blood cells. Blood. 1983;61(2):341-348. 179. Buckley RH, Schiff SE, Schiff RI, et al. Hematopoietic stem-cell transplantation for the treatment of severe combined immunodeficiency. N Engl J Med. 1999;340(7):508516. 180. Beatty PG, Clift RA, Mickelson EM, et al. Marrow transplantation from related donors other than HLA-identical siblings. N Eng J Med. 1985;313(13):765-771. 181. Aversa F, Terenzi A, Tabilio A, et al. Full haplotype-mismatched hematopoietic stem-cell transplantation: a phase II study in patients with acute leukemia at high risk of relapse. J Clin Oncol. 2005;23(15):3447-3454. 182. Anasetti C, Logan BR, Lee SJ, et al. Peripheral-blood stem cells versus bone marrow from unrelated donors. N Eng J Med. 2012;367(16):1487-1496.

2729

REVIEW ARTICLE Ferrata Storti Foundation

Why is it critical to achieve a deep molecular response in chronic myeloid leukemia? Susan Branford

Haematologica 2020 Volume 105(12):2730-2737

Department of Genetics and Molecular Pathology, Centre for Cancer Biology, SA Pathology; School of Pharmacy and Medical Science, Division of Health Sciences, University of South Australia; School of Medicine, Faculty of Health and Medical Sciences, University of Adelaide and School of Biological Sciences, University of Adelaide, Adelaide, Australia

ABSTRACT

Correspondence: SUSAN BRANFORD susan.branford@sa.gov.au Received: February 9, 2020. Accepted: April 29, 2020. Pre-published: September 17, 2020. doi:10.3324/haematol.2019.240739 ©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.

2730

he primary goal of tyrosine kinase inhibitor (TKI) therapy for patients with chronic myeloid leukemia is survival, which is achieved by the vast majority of patients. The initial response to therapy provides a sensitive measure of future clinical outcome. Measurement of BCR-ABL1 transcript levels using real-time quantitative polymerase chain reaction standardized to the international reporting scale is now the principal recommended monitoring strategy. The method is used to assess early milestone responses and provides a guide for therapeutic intervention. When patients successfully traverse the critical first 12 months of TKI therapy, most will head towards another milestone response, deep molecular response (DMR, BCR-ABL1 ≤0.01%). DMR is essential for patients aiming to achieve treatment-free remission and a prerequisite for a trial of TKI discontinuation. The success of discontinuation trials has led to new treatment strategies in order for more patients to reach this milestone response. DMR has been incorporated into endpoints of clinical trials and is considered by some expert groups as the optimal treatment response. But is DMR a stable response and does it provide the ultimate protection against TKI resistance and death? Do we need to increase the sensitivity of detection of BCR-ABL1 to better identify the patients who would likely remain in treatment-free remission after TKI discontinuation? Is it necessary to switch current TKI therapy to a more potent inhibitor if the goal is to achieve DMR? These are issues that I will explore in this review.

Introduction It has been 20 years since the first patients with newly diagnosed chronic myeloid leukemia (CML) were treated with imatinib in the International Randomized Study of Interferon and STI571 (IRIS) trial.1 The primary endpoint of that pivotal trial was the rate of progression, which included death, blast crisis or accelerated phase, loss of complete hematologic response, loss of major cytogenetic response or an increasing white cell count. Back then, in 2000, response was measured at the hematologic and cytogenetic levels. How things have changed for patients diagnosed with CML! Treatment response is now principally monitored at the molecular level by quantitative measurement of BCR-ABL1 transcripts. The prognostic value of molecular monitoring in patients treated with a tyrosine kinase inhibitor (TKI) was first demonstrated in the IRIS trial.2 The analysis was exploratory and in general performed only after the achievement of a complete cytogenetic response. The majority of imatinib-treated patients did indeed achieve a complete cytogenetic response and the term ‘major molecular response’ (MMR) was coined. This response was subsequently described as a ‘safe haven’ in which the risk of loss of response is low.3 The rate of MMR is now a primary or secondary outcome measure in CML clinical trials4,5 and is measured on an international reporting scale (IS) (≤0.1% BCR-ABL1 IS).6 A deep molecular response (DMR, BCR-ABL1 ≤0.01% IS) is also now an important milestone since it is a prerequisite for a trial of drug discontinuation with the aim of achieving treatment-free remishaematologica | 2020; 105(12)

The importance of a deep molecular response in CML

sion (TFR). The safety of TKI discontinuation for patients who achieve and maintain a DMR has been demonstrated in multiple clinical trials7-16 and international expert groups have incorporated TKI discontinuation into their recommendations and guidelines.17-20 The majority of CML patients will require lifelong TKI treatment and long-term molecular monitoring is recommended for all patients with CML.18-20 The European LeukemiaNet (ELN) mandates measurement of BCRABL1 transcripts according to the IS at least 3 monthly, even after MMR is confirmed because close monitoring is required to assess eligibility for TKI discontinuation.19 Most patients will eventually achieve a DMR.21,22 Whether or not DMR predicts survival has been debated and it has not been confirmed that it does.22-24 Achieving a sustained DMR is a prerequisite for TKI discontinuation, but is it also a biomarker for better clinical outcomes?23,25

Is deep molecular response the optimal molecular response for patients with chronic myeloid leukemia? In 2013, the ELN incorporated molecular monitoring using standardized real-time quantitative polymerase chain reaction (qRT-PCR) analysis into their recommendations for the management of CML.26 An optimal response was BCR-ABL1 ≤10%, ≤1% and ≤0.1% at 3, 6 and 12 months of TKI therapy. Treatment failure was defined as BCR-ABL1 >10% at 6 months and >1% at 12 months. These recommendations were based on strong evidence collected over many years27-35 and subsequent studies consolidated the recommendations.36-41 The importance of a one-log reduction of BCR-ABL1 by 3 months and two-log reduction by 6 months for progression-free survival was reported as early as 2003.28,29 The equivalent BCR-ABL1 transcript values on the IS are 10% and 1%, respectively. An update of molecular data generated in the IRIS trial was published in 2010, and used to examine the prognostic significance of early molecular response.30 Landmark analyses of BCR-ABL1 values at 6, 12 and 18 months of imatinib therapy established that event-free survival was inferior for patients with >10% at 6 months and >0.1% at 12 and 18 months. Progression to accelerated phase or blast crisis and overall survival were inferior for patients with BCR-ABL1 >10% at 6 months, and >1% at 12 and 18 months.30 In 2012, Hanfstein et al.31 and Marin et al.32 confirmed the strong association between BCR-ABL1 values at 3, 6 and 12 months and outcome. Marin et al. reported that the BCR-ABL1 value at 3 months was the only requirement for predicting outcome for patients treated with a TKI.32 Furthermore, a BCRABL1 value of ≤0.61% at 3 months was highly predictive of subsequent undetectable BCR-ABL1. The cumulative incidence of undetectable BCR-ABL1 at 8 years for patients with BCR-ABL1 ≤0.61% was 84.7% whereas it was 1.5% for those with >0.61% (P<0.001).32 This study highlighted the importance of rapid leukemic clearance for a subsequent DMR. The 2013 ELN recommendations for the management of CML26 were the first time that molecular response was incorporated into therapeutic decisions by an expert group, although quite wisely, caution was advised regarding the interpretation of the molecular values. An additional molecular test was recommended to confirm treatment failure. Numerous publications had confirmed the predictive value of molecular monitoring,27-35 but most of the studies had been performed in academic centers with haematologica | 2020; 105(12)

long-term experience in molecular monitoring. The ELN recognized that the standard of testing in these studies may not have represented the typical standard at that time. Widespread incorporation of molecular monitoring for clinical decisions was made possible by the introduction of the standardized IS for BCR-ABL1. This was coupled with harmonization of testing processes, standardization of the nomenclature for reporting molecular response and the development of reference material.42-50 The term complete molecular response was replaced by MR4 (BCRABL1 ≤0.01% IS) and MR4.5 (BCR-ABL1 ≤0.0032% IS).47 These terms apply to both detectable and undetectable BCR-ABL1 and incorporate the sensitivity achieved for individual samples. However, method standardization has been challenging and regular molecular monitoring on the IS is by no means available to all patients because economic circumstances may hinder its widespread use.51,52 Nevertheless, molecular monitoring is the principal recommended monitoring strategy.18-20 Furthermore, in countries with the most advanced standardized monitoring programs, multicenter, high-quality DMR assessment is achievable. This was demonstrated in a recent study conducted by the European Treatment and Outcome Study (EUTOS) group in which DMR was measured reliably by local laboratories in Europe, not just the key reference laboratories of individual countries.53 There are differences of opinion between experts regarding the early molecular response milestone values. Table 1 compares these values between the recent updated ELN recommendations19 and the National Comprehensive Cancer Network (NCCN) clinical practice guidelines.20 The NCCN guidelines have less stringent BCR-ABL1 cut-off values at 6 months (≤10%) and 12 months (≤1%) for TKI-sensitive disease (no change of therapy required). The ELN cut-off values for an optimal response are ≤1% at 6 months and ≤0.1% at 12 months (no change of therapy required). Furthermore, the ELN now considers a BCR-ABL1 value of ≤0.01% at any time as the optimal response for patients aiming for TFR. The ELN has a buffer response criterion of ‘warning’ between each milestone BCR-ABL1 cut-off value. A recommendation in cases of ‘warning’ is additional molecular monitoring if the kinetics of response is not clear. The trend of BCR-ABL1 decline over time can aid clinical decisions.54,55 The NCCN also suggests assessing the trend of decline for patients with BCR-ABL1 only slightly >10% at 3 months before making drastic decisions regarding the treatment strategy. The definition of TKI-resistant disease after 12 months of TKI therapy is less stringent in the NCCN guidelines than in the ELN recommendations: BCR-ABL1 cut-off >10% for the NCCN and >1% for the ELN (ELN ‘Failure’ category). The most recent NCCN guidelines differ from previous versions in which a BCR-ABL1 value of ≤0.1% at 12 months indicated TKI-sensitive disease.56 The value is now ≤1% at 12 months and TKI-resistant disease is defined as >1% at ≥15 months.20 However, the NCCN still recognizes the value of MMR at 12 months and a statement is included in the 2020 guidelines: “BCR-ABL1 0.1% at 12 months is associated with a very low probability of subsequent disease progression and a high likelihood of achieving a subsequent MR4.0, which may facilitate discontinuation of TKI therapy.”20 A recent analysis of the German CML-Study IV con2731

S. Branford

firmed the optimal response time to achieve 1% BCRABL1 at about 12 to 15 months for progression-free survival, with progression being development of accelerated phase, blast crisis or death.57 The study also investigated when it is necessary to regard lack of MMR as treatment failure, indicating that a switch of therapy is warranted. The landmark time point of 2.5 years to achieve MMR showed the largest difference between those with or without MMR with regard to progression-free survival.57 A specific time to achieve DMR for progression-free survival was not detected. The updated ELN recommendations now state a change of treatment may be considered if MMR is not reached by 36-48 months.19

How stable is deep molecular response? Multiple issues may contribute to loss of a DMR, including dose reduction, and cessation or non-adherence to therapy. A rise in BCR-ABL1 levels can be an exquisite indicator of non-adherence.58 A very rapid rise can indicate complete lack of kinase inhibition due to abrupt TKI cessation.58 Loss of a DMR is rarely associated with drug resistance. A recent single-center review of 450 patients demonstrated that sustained MR4 for at least 12 months represented a secure response threshold.59 This finding only applied to compliant patients with no history of previous TKI resistance who received standard-dose TKI. No such patient lost a MMR, whereas loss of MMR occurred in 25% of patients who had not achieved a MR4. Importantly, failure to sustain a MR4 was the only significant variable for loss of MMR in multivariate analysis.59 We also found that sustained undetectable BCR-ABL1 (MR4.5) was associated with sustained MMR.21 Conversely, MMR was lost in six of 22 (27%) patients with sustained detectable BCR-ABL1 and was associated with the acquisition of imatinib-resistant BCR-ABL1 kinase domain mutations in three of six patients. None of these three patients had achieved a DMR. Similarly, a recent study found BCR-ABL1 mutations in 26% of patients who lost a MMR, although it is not known whether any of these patients achieved a DMR.60 The molecular response levels use defined cut-off values and the inherent variability of the quantitative PCR assay means there will be fluctuations above or below the cut-off values.61,62 These fluctuations will be greater at low levels of BCR-ABL1. However, in some cases, fluctuations are an indication of subsequent relapse. A study of 208 patients treated with imatinib as their first-line therapy investigated the outcome of patients with fluctuating

BCR-ABL1 values according to the level of response achieved.63 A stable molecular response was defined as persistence of the same molecular response (MMR, MR4 or MR4.5) at three consecutive assessments. A fluctuation was the achievement of the molecular response and its subsequent loss. For patients who had not yet achieved DMR but had achieved MMR, 12.6% had at least one fluctuation. These patients had a significantly poorer failure-free survival compared to patients with no fluctuation: 82.4% versus 93.2%, respectively. Of patients with DMR and fluctuations, none acquired resistance or BCRABL1 kinase domain mutations. This study demonstrated that unstable MMR was associated with an increased risk of imatinib resistance, whereas fluctuations of deeper molecular responses did not influence outcome.63 Data suggest that when a DMR is achieved it is relatively stable and the risk of TKI resistance is low. However, vigilance and long-term molecular monitoring are recommended, even for patients with stable DMR. A rare case of late relapse associated with the acquisition of a BCR-ABL1 kinase domain mutation after long-term, stable, undetectable BCR-ABL1 (MR4.5) has been reported.64 A Y253H mutation was first detectable by Sanger sequencing more than 2 years after BCR-ABL1 transcripts became detectable, which was almost 9 years after commencing imatinib therapy.

Does real-time quantitative polymerase chain reaction analysis provide sufficient sensitivity? Multiple clinical trials have consistently confirmed that approximately half of the patients who stop TKI in a stable DMR have molecular recurrence.7-16 Despite many years having passed since results of the first discontinuation trials were reported, reliable prediction of molecular relapse has eluded researchers. The NCCN provides criteria for attempting TKI discontinuation, which include stable MR4 (BCR-ABL1 ≤0.01%) for at least 2 years.20 The French CML Study Group recommends MR4.5 (BCRABL1 ≤0.0032%) for at least 2 years.17 Although the difference in BCR-ABL1 levels seems minor, the slow kinetics of the BCR-ABL1 decline means that MR4.5 may not be reached until many months, or even years, after MR4.22,65 In the German CML-Study IV the estimated median time to reach MR4 was 3.1 years, and that to reach MR4.5 was 4.9 years.22 Horn et al. assessed the transcript dynamics of patients treated with 400 mg imatinib.65 They estimated that the median time to reach MR4 was 5.3-6.5 years while that to reach MR4.5 was 9.1-10.7 years. Similarly,

Table 1. Molecular response milestones according to international recommendations and guidelines.

Milestones

3 months 6 months 12 months ≥15 months Any time

ELN 202019 Optimal

NCCN 202020 TKI-sensitive disease

ELN 202019 Warning

ELN 202019 Failure

≤10%

NCCN 202020 Possible TKI resistance

NCCN 202020 TKI-resistant disease

>10%

≤1% ≤0.1%

≤10% ≤1% ≤1%

>1-10% >0.1-1%

NA >1-10% NA

>10% if confirmed within 1-3 months >10% >1%

≤0.1% or ≤0.01% with the aim to achieve TFR

>0.1-1%, loss of MMR indicates failure after TFR

>10% >10% >1%

>1%, resistance mutations and high-risk ACA

ELN: European LeukemiaNet; NCCN: National Comprehensive Cancer Network; TKI: tyrosine kinase inhibitor; NA: not applicable; TFR: treatment-free remission; MMR: major molecular response; ACA, additional chromosome abnormalities in Philadelphia-positive cells.

2732

haematologica | 2020; 105(12)

The importance of a deep molecular response in CML

for 528 patients treated with imatinib doses of 400, 600 or 800 mg in the first-line setting, we found an approximate 3-year difference between the times to reach MR4 and MR4.566 (Figure 1). These studies indicate that the criteria used for TKI discontinuation significantly influence the timing of discontinuation. A recent meta-analysis of 12 TKI discontinuation trials assessed the factors that influenced the rate of TFR.67 Trials with a molecular criterion of MR4.5 or better for selecting patients for TKI discontinuation documented higher rates of TFR at 24 months than those with a criterion of MR4.0: 57.2% versus 50.5%, respectively. This finding could be associated with the longer treatment duration needed to achieve MR4.5 or better. Longer treatment duration is associated with a higher probability of sustaining TFR.15 It could also indicate that lower

Figure 1. Time difference between achieving MR4 and MR4.5. We determined the cumulative incidence of achieving confirmed MR4 and MR4.5 in 528 patients consecutively treated in clinical trials of imatinib. There was an approximate 3-year difference in the median time to reach the deep molecular response levels in the total cohort. DMR: deep molecular response; CI: confidence interval.

BCR-ABL1 values before TKI cessation increase the chance of maintaining TFR. However, Ross et al. failed to predict TFR using a highly sensitive patient-specific DNA-based quantitative PCR technique (sensitivity106.2 68 ). None of the patients studied had detectable BCRABL1 mRNA transcripts, with a sensitivity of MR4.5, at TKI discontinuation. The DNA technique detected BCRABL1 in almost all patients, irrespective of whether TFR was maintained or not. Furthermore, DNA BCR-ABL1 continued to be detectable in remission. In a follow-up of nine patients in long-term TFR, residual DNA BCRABL1 was detected in most patients. In two of these nine patients, DNA BCR-ABL1 was persistently detectable at every measurement time-point over 6.5 and 10.5 years, and the level declined over time.69 The authors hypothesized that this could indicate gradual extinction of long-lived lineage-committed cells that lack self-renewal capacity or depletion of slowly proliferating leukemic precursor cells. In a recent analysis that included the same patients, the lineage of residual leukemic cells in patients who sustained TFR was investigated.70 Residual DNA BCR-ABL1 was detected predominantly in the lymphoid compartment and never in granulocytes. This was an important finding since the study demonstrated that the detection of residual BCRABL1 may not imply the persistence of multipotent leukemic cells. Additionally, lymphocytes were found to be part of the leukemic clone at the time of diagnosis of CML and BCR-ABL1 was expressed in both RNA and DNA. This may have implications for studies in which T cells are used as a source of non-leukemic cells to establish the somatic status of variants in next-generation sequencing studies.71 Digital PCR is being increasingly used to measure residual BCR-ABL1 DNA and RNA and may improve the precision and sensitivity of detection.13,72-82 Goh et al. improved the sensitivity of BCR-ABL1 transcript detection by pre-amplification prior to digital PCR,72 thereby introducing a semi-quantitative nested PCR approach.83 Importantly, BCR-ABL1 transcripts were detected by digital PCR in patients with stable undetectable BCR-ABL1 using the standard qRT-PCR method. This observation

Figure 2. Patients with MR4 at 3 years of imatinib treatment have a high probability of reaching MR4.5 with continued imatinib. Patients without a major molecular response or MR4 at 3 years of first-line imatinib therapy may benefit from a switch to a more potent tyrosine kinase inhibitor if the goal of therapy is to achieve MR4.5.66 MMR: major molecular response.

haematologica | 2020; 105(12)

2733

S. Branford

corroborates the studies reporting residual DNA BCRABL1 in similar patientsâ&#x20AC;&#x2122; samples.68,84 A critical factor for the reliability of data when using very sensitive PCR is controlling for contamination and false positive results. This is overcome by quantification of genomic DNA BCR-ABL1 breakpoints that are unique to each patient. However, RNA-based techniques are prone to contamination since common sequences are amplified for each patient. Negative controls for each step of the assay are essential to detect contamination and to establish the threshold for positivity. Optimization of a digital droplet PCR (ddPCR) method ensured a background false-positive rate of only 5% of samples and reliably detected BCR-ABL1 transcripts to MR5 (0.001%).78 Another group carefully evaluated the positivity threshold of a ddPCR method by testing 30 samples from non-CML patients.79 The method was used to measure residual disease in a substantial proportion of patients enrolled in the Stop Imatinib 2 (STIM2) study prior to TKI discontinuation. The patients all had undetectable BCR-ABL1 transcripts by qRT-PCR at a sensitivity level of MR4.5 for at least 2 years. Using the more sensitive ddPCR method, a cut-off value for detectable residual disease was established to predict TFR. A conversion factor to the IS was also established. Patients with residual BCR-ABL1 levels above the the threshold of 0.0023% IS had a higher probability of molecular recurrence at both 6 and 12 months after TKI discontinuation: 66% versus 44% and 68% versus 46%, respectively. However, purely on the basis of ddPCR positivity versus negativity, there was no predictive power for disease recurrence.79 Whether or not digital PCR can reliably identify the patients destined for TFR requires further analysis and at this stage the technique should not be used to select patients for a TFR attempt.79,85 Key factors for resolution are harmonization of methods and standardization to the IS. Importantly, laboratories must ensure that threshold levels above the background noise are carefully implemented. Furthermore, based on the ddPCR data from the STIM2 trial,79 Yan et al. estimated that if the ddPCR cut-off of <0.0023% IS had been added to the STIM2 entry criteria, the rate of TFR at 12 months would have been 54% instead of the reported 49% for the patients tested by ddPCR.85 Furthermore, some patients capable of maintaining TFR would be excluded. The authors of the STIM2 ddPCR study clearly stated that this degree of improvement was insufficient for implementation of ddPCR for selection of patients.79 However, digital PCR may become an important complement to qRT-PCR in decisions for attempting TFR, in particular, using a ddPCR method that reports BCR-ABL1 values on the IS.86 Factors other than the depth of BCR-ABL1 response at the time of stopping TKI may influence sustained TFR. The duration of DMR before stopping treatment and longer duration of therapy were factors in the EURO-SKI study, which is the largest discontinuation trial.15 The detection of BCR-ABL1 transcripts using sensitive PCR in patients with TFR demonstrates that elimination of the leukemic clone may not be necessary for sustained TFR.72 Immune surveillance may be an important factor.87,88 Sustained MR4.5 is a reasonable molecular response for a TFR attempt and methods should aim to reliably detect MR4.5, irrespective of whether the method used is qRTPCR or digital PCR. 2734

Strategies to improve the rates of MR4.5 The first-line use of the second-generation TKI nilotinib and dasatinib is associated with higher rates of MMR and MR4.5 than the rates following the use of imatinib and the responses are achieved earlier.89,90 However, deeper molecular responses did not translate into improved survival. Strategies have been explored to induce deeper molecular responses for imatinib-treated patients, including switching to a more potent inhibitor. The ENESTcmr clinical trial was a randomized study for imatinib-treated patients in complete cytogenetic response with detectable BCR-ABL1.91,92 Patients continued on imatinib or switched to nilotinib 400 mg twice daily. The primary endpoint of the study was undetectable BCR-ABL1 MR4.5 at 12 months. Higher rates of MMR and MR4.5 were achieved with nilotinib, although adverse events were more common. The cumulative incidence of MR4.5 following the switch to nilotinib was 32.7% at 12 months and 42.9% at 24 months.91 The cumulative incidence of MR4.5 among patients who continued imatinib therapy was 13.5% and 20.8% at 12 and 24 months, respectively. Consistent with other trials, cardiovascular events were more frequent among the nilotinib-treated patients.92 The potential for improved molecular responses with more potent TKI must be assessed in light of the potential for cardiovascular events.93 Furthermore, with the high cost of second-generation TKI in many countries, the incremental benefit of using these inhibitors to achieve DMR may not provide good value.94 The ELN had considerable discussion when revising the recommendations for managing CML in regards to the advisability of using a second-generation TKI in the first- or second-line setting to achieve DMR. However, there was no final consensus.19 We determined whether there was a level of BCR-ABL1 in imatinib-treated patients below which a switch to a more potent inhibitor may not be necessary in order to reach a timely MR4.5.66 Among 528 patients consecutively treated in clinical trials of imatinib, 147 had achieved a complete cytogenetic response, or its molecular equivalent of â&#x2030;¤1.0% BCR-ABL1,95 at 3 years of imatinib treatment. None of these patients had achieved MR4.5 at that time. Landmark analyses demonstrated that the patients without MMR at 3 years of imatinib therapy had a negligible probability of achieving MR4 or MR4.5 with up to 5 additional years of imatinib (Figure 2). These patients may benefit from a switch to a more potent inhibitor in order to achieve DMR. Similarly, patients with MMR but not MR4 at 3 years of imatinib therapy had a significantly lower cumulative incidence of MR4.5 compared to patients with MR4: MMR versus MR4, 61% versus 100% by 5 years after the landmark (P<0.0001). However, most patients with MR4 at 3 years of imatinib therapy did indeed achieve MR4.5 with 2 additional years of imatinib treatment. These findings may help clinical decisions when considering a switch of treatment to optimize TKI discontinuation options, while minimizing the additional risk of adverse events with more potent TKI.

Conclusion Reaching a DMR is now considered a goal of therapy by many clinicians. The importance of a DMR for patients aiming to achieve TFR is recognized by the inclusion of DMR as an endpoint measure in clinical trials. haematologica | 2020; 105(12)

The importance of a deep molecular response in CML

Additionally, TFR is increasingly acknowledged as an avenue to save money in over-stretched healthcare budgets.15,96 However, not all patients and clinicians consider TFR as the goal of therapy.97 Prolonged survival with minimal side-effects are equally important goals and the patientâ&#x20AC;&#x2122;s choice is central for treatment decisions. The take-home messages from this review are: (i) it can take many years to reach a DMR for some patients and earlier achievement is possible with a second-generation inhibitor; however, there is no consensus on the benefit of using a second-generation inhibitor to achieve a DMR and therapy choices must be made in the context of the

References 1. O'Brien SG, Guilhot F, Larson RA, et al. Imatinib compared with interferon and lowdose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med. 2003;348(11):994-1004. 2. Hughes TP, Kaeda J, Branford S, et al. Frequency of major molecular responses to imatinib or interferon alfa plus cytarabine in newly diagnosed chronic myeloid leukemia. N Engl J Med. 2003;349(15):1423-1432. 3. Hughes TP, Branford S. Monitoring disease response to tyrosine kinase inhibitor therapy in CML. Hematology Am Soc Hematol Educ Program. 2009;2009(1):477-487. 4. Lipton JH, Chuah C, Guerci-Bresler A, et al. Ponatinib versus imatinib for newly diagnosed chronic myeloid leukaemia: an international, randomised, open-label, phase 3 trial. Lancet Oncol. 2016;17(5):612-621. 5. Hughes TP, Mauro MJ, Cortes JE, et al. Asciminib in chronic myeloid leukemia after ABL kinase inhibitor failure. N Engl J Med. 2019;381(24):2315-2326. 6. Hughes T, Deininger M, Hochhaus A, et al. Monitoring CML patients responding to treatment with tyrosine kinase inhibitors: review and recommendations for harmonizing current methodology for detecting BCR-ABL transcripts and kinase domain mutations and for expressing results. Blood. 2006;108(1):28-37. 7. Mahon FX, Rea D, Guilhot J, et al. Discontinuation of imatinib in patients with chronic myeloid leukaemia who have maintained complete molecular remission for at least 2 years: the prospective, multicentre Stop Imatinib (STIM) trial. Lancet Oncol. 2010;11(11):1029-1035. 8. Ross DM, Branford S, Seymour JF, et al. Safety and efficacy of imatinib cessation for CML patients with stable undetectable minimal residual disease: results from the TWISTER study. Blood. 2013;122(4):515522. 9. Rea D, Nicolini FE, Tulliez M, et al. Discontinuation of dasatinib or nilotinib in chronic myeloid leukemia: interim analysis of the STOP 2G-TKI study. Blood. 2017;129(7):846-854. 10. Mahon F, Boquimpani C, Kim D, et al. Treatment-free remission after second-line nilotinib treatment in patients with chronic myeloid leukemia in chronic phase: results from a single-group, phase 2, open-label study. Ann Intern Med. 2018;168(7):461470. 11. Clark RE, Polydoros F, Apperley JF, et al. Deescalation of tyrosine kinase inhibitor therapy before complete treatment discontinua-

haematologica | 2020; 105(12)

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

individual patientâ&#x20AC;&#x2122;s comorbidities; (ii) a sustained MR4.5 prior to treatment discontinuation may be associated with higher rates of TFR than sustained MR4; and (iii) deeper molecular responses have so far not reliably predicted TFR. Regardless, sustained DMR is certainly a desirable outcome since it is associated with an extremely low risk of loss of response and TKI resistance, and is mandatory for any patient considering TFR. Acknowledgments The author is funded by the National Health and Medical Research Council of Australia, APP1117718.

tion in patients with chronic myeloid leukaemia (DESTINY): a non-randomised, phase 2 trial. Lancet Haematol. 2019;6 (7):e375-e383. Rousselot P, Charbonnier A, Cony-Makhoul P, et al. Loss of major molecular response as a trigger for restarting tyrosine kinase inhibitor therapy in patients with chronicphase chronic myelogenous leukemia who have stopped imatinib after durable undetectable disease. J Clin Oncol. 2014;32(5): 424-430. Lee SE, Choi SY, Song HY, et al. Imatinib withdrawal syndrome and longer duration of imatinib have a close association with a lower molecular relapse after treatment discontinuation: the KID study. Haematologica. 2016;101(6):717-723. Hochhaus A, Masszi T, Giles FJ, et al. Treatment-free remission following frontline nilotinib in patients with chronic myeloid leukemia in chronic phase: results from the ENESTfreedom study. Leukemia. 2017;31(7):1525-1531. Saussele S, Richter J, Guilhot J, et al. Discontinuation of tyrosine kinase inhibitor therapy in chronic myeloid leukaemia (EURO-SKI): a prespecified interim analysis of a prospective, multicentre, non-randomised, trial. Lancet Oncol. 2018;19(6): 747-757. Shah NP, Garcia-Gutierrez V, JimenezVelasco A, et al. Dasatinib discontinuation in patients with chronic-phase chronic myeloid leukemia and stable deep molecular response: the DASFREE study. Leuk Lymphoma. 2020;61(3):650-659. Rea D, Ame S, Berger M, et al. Discontinuation of tyrosine kinase inhibitors in chronic myeloid leukemia: recommendations for clinical practice from the French Chronic Myeloid Leukemia Study Group. Cancer. 2018;124(14):2956-2963. Hochhaus A, Saussele S, Rosti G, et al. Chronic myeloid leukaemia: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2017;28 (suppl_4):iv41-iv51. Hochhaus A, Baccarani M, Silver RT, et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia. 2020;34(4):966-984. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology: chronic myelogenous leukemia. Version 3, 2020. https://www.nccn.org/professionals/physician_gls/pdf/cml.pdf [Last accessed April 26, 2020]. Branford S, Seymour JF, Grigg A, et al. BCRABL messenger RNA levels continue to decline in patients with chronic phase chronic myeloid leukemia treated with ima-

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

tinib for more than 5 years and approximately half of all first-line treated patients have stable undetectable BCR-ABL using strict sensitivity criteria. Clin Cancer Res. 2007;13(23):7080-7085. Hehlmann R, Muller MC, Lauseker M, et al. Deep molecular response is reached by the majority of patients treated with imatinib, predicts survival, and is achieved more quickly by optimized high-dose imatinib: results from the randomized CML-study IV. J Clin Oncol. 2014;32(5):415-423. Falchi L, Kantarjian HM, Wang X, et al. Significance of deeper molecular responses in patients with chronic myeloid leukemia in early chronic phase treated with tyrosine kinase inhibitors. Am J Hematol. 2013;88(12):1024-1029. Kantarjian H, Cortes JE. Complete cytogenetic response, not deep molecular response, is associated with survival in chronic myeloid leukemia. J Clin Oncol. 2014;32(27):3077. Etienne G, Dulucq S, Nicolini F-E, et al. Achieving deeper molecular response is associated with a better clinical outcome in chronic myeloid leukemia patients on imatinib front-line therapy. Haematologica. 2014;99(3):458-464. Baccarani M, Deininger MW, Rosti G, et al. European LeukemiaNet recommendations for the management of chronic myeloid leukemia: 2013. Blood. 2013;122(6):872-884. Merx K, Muller MC, Kreil S, et al. Early reduction of BCR-ABL mRNA transcript levels predicts cytogenetic response in chronic phase CML patients treated with imatinib after failure of interferon alpha. Leukemia. 2002;16(9):1579-1583. Wang L, Pearson K, Ferguson JE, Clark RE. The early molecular response to imatinib predicts cytogenetic and clinical outcome in chronic myeloid leukaemia. Br J Haematol. 2003;120(6):990-999. Branford S, Rudzki Z, Harper A, et al. Imatinib produces significantly superior molecular responses compared to interferon alfa plus cytarabine in patients with newly diagnosed chronic myeloid leukemia in chronic phase. Leukemia. 2003;17(12):24012409. Hughes TP, Hochhaus A, Branford S, et al. Long-term prognostic significance of early molecular response to imatinib in newly diagnosed chronic myeloid leukemia: an analysis from the International Randomized Study of Interferon and STI571 (IRIS). Blood. 2010;116(19):3758-3765. Hanfstein B, Muller MC, Hehlmann R, et al. Early molecular and cytogenetic response is predictive for long-term progression-free and overall survival in chronic myeloid

2735

S. Branford

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

2736

leukemia (CML). Leukemia. 2012;26(9):2096-2102. Marin D, Ibrahim AR, Lucas C, et al. Assessment of BCR-ABL1 transcript levels at 3 months is the only requirement for predicting outcome for patients with chronic myeloid leukemia treated with tyrosine kinase inhibitors. J Clin Oncol. 2012;30 (3):232-238. Branford S, Kim D-W, Soverini S, et al. Initial molecular response at 3 months may predict both response and event-free survival at 24 months in imatinib-resistant or -intolerant patients with Philadelphia chromosomepositive chronic myeloid leukemia in chronic phase treated with nilotinib. J Clin Oncol. 2012;30(35):4323-4329. Neelakantan P, Gerrard G, Lucas C, et al. Combining BCR-ABL1 transcript levels at 3 and 6 months in chronic myeloid leukemia: implications for early intervention strategies. Blood. 2013;121(14):2739-2742. Marin D, Hedgley C, Clark RE, et al. Predictive value of early molecular response in patients with chronic myeloid leukemia treated with first-line dasatinib. Blood. 2012;120(2):291-294. Jain P, Kantarjian H, Nazha A, et al. Early responses predict better outcomes in patients with newly diagnosed chronic myeloid leukemia: results with four tyrosine kinase inhibitor modalities. Blood. 2013;121(24):4867-4874. Hughes TP, Saglio G, Kantarjian HM, et al. Early molecular response predicts outcomes in patients with chronic myeloid leukemia in chronic phase treated with frontline nilotinib or imatinib. Blood. 2014;123(9):1353-1360. Jabbour E, Kantarjian HM, Saglio G, et al. Early response with dasatinib or imatinib in chronic myeloid leukemia: 3-year follow-up from a randomized phase 3 trial (DASISION). Blood. 2014;123(4):494-500. Kim D, Hamad N, Lee HG, Kamel-Reid S, Lipton JH. BCR/ABL level at 6 months identifies good risk CML subgroup after failing early molecular response at 3 months following imatinib therapy for CML in chronic phase. Am J Hematol. 2014;89(6):626-632. Fava C, Rege-Cambrin G, Dogliotti I, et al. Early BCR-ABL1 reduction is predictive of better event-free survival in patients with newly diagnosed chronic myeloid leukemia treated with any tyrosine kinase inhibitor. Clin Lymphoma Myeloma Leuk. 2016;16 Suppl S96-S100. Zhang J, Wang Y, Wang J, et al. Early BCRABL1 decline in imatinib-treated patients with chronic myeloid leukemia: results from a multicenter study of the Chinese CML alliance. Blood Cancer J. 2018;8(7):61. Branford S, Fletcher L, Cross NCP, et al. Desirable performance characteristics for BCR-ABL measurement on an international reporting scale to allow consistent interpretation of individual patient response and comparison of response rates between clinical trials. Blood. 2008;112(8):3330-3338. Muller MC, Erben P, Saglio G, et al. Harmonization of BCR-ABL mRNA quantification using a uniform multifunctional control plasmid in 37 international laboratories. Leukemia. 2008;22(1):96-102. Muller MC, Cross NCP, Erben P, et al. Harmonization of molecular monitoring of CML therapy in Europe. Leukemia. 2009;23(11):1957-1963. White HE, Matejtschuk P, Rigsby P, et al. Establishment of the first World Health Organization International Genetic Reference Panel for quantitation of BCRABL mRNA. Blood. 2010;116(22):e111-117.

46. Cross NCP. Standardisation of molecular monitoring for chronic myeloid leukaemia. Best Pract Res Clin Haematol. 2009;22(3): 355-365. 47. Cross NCP, White HE, Muller MC, Saglio G, Hochhaus A. Standardized definitions of molecular response in chronic myeloid leukemia. Leukemia. 2012;26(10):21722175. 48. Cross NC, White HE, Colomer D, et al. Laboratory recommendations for scoring deep molecular responses following treatment for chronic myeloid leukemia. Leukemia. 2015;29(5):999-1003. 49. Cross NC, White HE, Ernst T, et al. Development and evaluation of a secondary reference panel for BCR-ABL1 quantification on the international scale. Leukemia. 2016;30(9):1844-1852. 50. Zhang J-W, Fu Y, Wu Q-S, et al. Standardization of BCR-ABL1 quantification on the international scale in China using locally developed secondary reference panels. Exp Hematol. 2020;81(e3):42-49. 51. Pagnano KBB. BCR-ABL1 level monitoring in chronic myeloid leukemia by real time polymerase chain reaction in Brazil - not so real. Rev Bras Hematol Hemoter. 2017;39(3):197-198. 52. Malhotra H, Radich J, Garcia-Gonzalez P. Meeting the needs of CML patients in resource-poor countries. Hematology. 2019;2019(1):433-442. 53. Mobius S, Schenk T, Himsel D, et al. Results of the European survey on the assessment of deep molecular response in chronic phase CML patients during tyrosine kinase inhibitor therapy (EUREKA registry). J Cancer Res Clin Oncol. 2019;145(6):16451650. 54. Hanfstein B, Shlyakhto V, Lauseker M, et al. Velocity of early BCR-ABL transcript elimination as an optimized predictor of outcome in chronic myeloid leukemia (CML) patients in chronic phase on treatment with imatinib. Leukemia. 2014;28(10):1988-1992. 55. Branford S, Yeung DT, Parker WT, et al. Prognosis for patients with CML and >10% BCR-ABL1 after 3 months of imatinib depends on the rate of BCR-ABL1 decline. Blood. 2014;124(4):511-518. 56. Radich JP, Deininger M, Abboud CN, et al. Chronic myeloid leukemia, version 1. 2019, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2018;16(9): 1108-1135. 57. Saussele S, Hehlmann R, Fabarius A, et al. Defining therapy goals for major molecular remission in chronic myeloid leukemia: results of the randomized CML Study IV. Leukemia. 2018;32(5):1222-1228. 58. Branford S, Yeung DT, Prime JA, et al. BCRABL1 doubling times more reliably assess the dynamics of CML relapse compared with the BCR-ABL1 fold rise: implications for monitoring and management. Blood. 2012;119(18):4264-4271. 59. Claudiani S, Gatenby A, Szydlo R, et al. MR4 sustained for 12 months is associated with stable deep molecular responses in chronic myeloid leukemia. Haematologica. 2019;104(11):2206-2214. 60. Etienne G, Dulucq S, Huguet F, et al. Incidence and outcome of BCR-ABL mutated chronic myeloid leukemia patients who failed to tyrosine kinase inhibitors. Cancer Med. 2019;8(11):5173-5182. 61. Branford S, Hughes TP. Practical considerations for monitoring patients with chronic myeloid leukemia. Semin Hematol. 2010;47 (4):327-334. 62. Branford S. Molecular monitoring in chronic

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

myeloid leukemia-how low can you go? ASH Education Program Book. 2016;2016 (1):156-163. Molica M, Breccia M, Colafigli G, et al. Long-term impact of molecular response fluctuations in chronic myeloid leukaemia patients treated with imatinib. Br J Haematol. 2018;181(2):275-278. Crampe M, Andrews C, Fortune A, Langabeer SE. Late emergence of an imatinib-resistant ABL1 kinase domain mutation in a patient with chronic myeloid leukemia. Case Rep Hematol. 2017;2017: 3548936. Horn M, Glauche I, MĂźller MC, et al. Model-based decision rules reduce the risk of molecular relapse after cessation of tyrosine kinase inhibitor therapy in chronic myeloid leukemia. Blood. 2013;121(2):378384. Branford S, Ross DM, Yeung DT, Braley J, Hughes TP. Achieving the deep molecular response levels required for an imatinib discontinuation trial is strongly associated with the BCR-ABL level at the first qualifying timepoint (Abstract). Blood. 2014;124(21): 4561. Kim J, Park J, Moon Y, et al. Effect of studylevel factors on treatment-free remission rate in patients with chronic myeloid leukemia: a systematic review and metaanalysis. Int J Hematol. 2019;110(6):683-689. Ross DM, Branford S, Seymour JF, et al. Patients with chronic myeloid leukemia who maintain a complete molecular response after stopping imatinib treatment have evidence of persistent leukemia by DNA PCR. Leukemia. 2010;24(10):17191724. Ross DM, Pagani IS, Shanmuganathan N, et al. Long-term treatment-free remission of chronic myeloid leukemia with falling levels of residual leukemic cells. Leukemia. 2018;32(12):2572-2579. Pagani IS, Dang P, Saunders VA, et al. Lineage of measurable residual disease in patients with chronic myeloid leukemia in treatment-free remission. Leukemia. 2020;34(4):1052-1061. Kim T, Tyndel MS, Kim HJ, et al. Spectrum of somatic mutation dynamics in chronic myeloid leukemia following tyrosine kinase inhibitor therapy. Blood. 2017;129(1):38-47. Goh HG, Lin M, Fukushima T, et al. Sensitive quantitation of minimal residual disease in chronic myeloid leukemia using nanofluidic digital polymerase chain reaction assay. Leuk Lymphoma. 2011;52(5): 896-904. Jennings LJ, George D, Czech J, Yu M, Joseph L. Detection and quantification of BCR-ABL1 fusion transcripts by droplet digital PCR. J Mol Diagn. 2014;16(2):174-179. Mori S, Vagge E, le Coutre P, et al. Age and dPCR can predict relapse in CML patients who discontinued imatinib: the ISAV study. Am J Hematol. 2015;90(10):910-914. Alikian M, Ellery P, Forbes M, et al. Nextgeneration sequencing-assisted DNA-based digital PCR for a personalized approach to the detection and quantification of residual disease in chronic myeloid leukemia patients. J Mol Diagn. 2016;18(2):176-189. Alikian M, Whale AS, Akiki S, et al. RTqPCR and RT-digital PCR: a comparison of different platforms for the evaluation of residual disease in chronic myeloid leukemia. Clin Chem. 2017;63(2):525-531. Wang W-J, Zheng C-F, Liu Z, et al. Droplet digital PCR for BCR/ABL(P210) detection of chronic myeloid leukemia: a high sensitive method of the minimal residual disease and

haematologica | 2020; 105(12)

The importance of a deep molecular response in CML

78.

79.

80.

81.

82.

83.

84.

disease progression. Eur J Haematol. 2018;101(3):291-296. Maier J, Lange T, Cross M, Wildenberger K, Niederwieser D, Franke GN. Optimized digital droplet PCR for BCR-ABL. J Mol Diagn. 2019;21(1):27-37. Nicolini FE, Dulucq S, Boureau L, et al. Evaluation of residual disease and TKI duration are critical predictive factors for molecular recurrence after stopping imatinib firstline in chronic phase CML patients. Clin Cancer Res. 2019;25(22):6606-6613. Bernardi S, Malagola M, Zanaglio C, et al. Digital PCR improves the quantitation of DMR and the selection of CML candidates to TKIs discontinuation. Cancer Med. 2019;8(5):2041-2055. Berdeja JG, Heinrich MC, Dakhil SR, et al. Rates of deep molecular response by digital and conventional PCR with frontline nilotinib in newly diagnosed chronic myeloid leukemia: a landmark analysis. Leuk Lymphoma. 2019;60(10):2384-2393. Franke GN, Maier J, Wildenberger K, et al. Comparison of real-time quantitative PCR and digital droplet PCR for BCR-ABL1 monitoring in patients with chronic myeloid leukemia. J Mol Diagn. 2020; 22(1):81-89. Ross DM, Branford S. Minimal residual disease: the advantages of digital over analog polymerase chain reaction. Leuk Lymphoma. 2011;52(7):1161-1163. Sobrinho-Simoes M, Wilczek V, Score J, Cross NCP, Apperley JF, Melo JV. In search of the original leukemic clone in chronic myeloid leukemia patients in complete molecular remission after stem cell trans-

haematologica | 2020; 105(12)

85.

86.

87.

88.

89.

90.

91.

plantation or imatinib. Blood. 2010;116(8): 1329-1335. Yan D, Pomicter AD, O'Hare T, Deininger MW. ddeeper than deep: can ddPCR predict successful imatinib cessation? Clin Cancer Res. 2019;25(22):6561-6563. Chung HJ, Hur M, Yoon S, et al. Performance evaluation of the QXDx BCRABL %IS droplet digital PCR assay. Ann Lab Med. 2020;40(1):72-75. Hughes A, Clarson J, White DL, Ross DM, Hughes TP, Yong AS. Enhanced natural killer and cytotoxic T lymphocyte responses, with decreased monocytic myeloid derived suppressor cells may promote treatment free remission in chronic myeloid leukaemia patients following tyrosine kinase inhibitor cessation. Blood. 2016;128(22):1122. Rea D, Henry G, Khaznadar Z, et al. Natural killer-cell counts are associated with molecular relapse-free survival after imatinib discontinuation in chronic myeloid leukemia: the IMMUNOSTIM study. Haematologica. 2017;102(8):1368-1377. Hochhaus A, Saglio G, Hughes TP, et al. Long-term benefits and risks of frontline nilotinib vs imatinib for chronic myeloid leukemia in chronic phase: 5-year update of the randomized ENESTnd trial. Leukemia. 2016;30(5):1044-1054. Kantarjian HM, Shah NP, Cortes JE, et al. Dasatinib or imatinib in newly diagnosed chronic-phase chronic myeloid leukemia: 2year follow-up from a randomized phase 3 trial (DASISION). Blood. 2012;119(5):11231129. Hughes TP, Lipton JH, Spector N, et al. Deep

92.

93.

94.

95.

96.

97.

molecular responses achieved in patients with CML-CP who are switched to nilotinib after long-term imatinib. Blood. 2014;124 (5):729-736. Hughes TP, Leber B, Cervantes F, et al. Sustained deep molecular responses in patients switched to nilotinib due to persistent BCR-ABL1 on imatinib: final ENESTcmr randomized trial results. Leukemia. 2017;31(11):2529-2531. Moslehi JJ, Deininger M. Tyrosine kinase inhibitor-associated cardiovascular toxicity in chronic myeloid leukemia. J Clin Oncol. 2015;33(35):4210-4218. Shih Y-CT, Cortes JE, Kantarjian HM. Treatment value of second-generation BCRABL1 tyrosine kinase inhibitors compared with imatinib to achieve treatment-free remission in patients with chronic myeloid leukaemia: a modelling study. Lancet Haematol. 2019;6(8):e398-e408. Ross DM, Branford S, Moore S, Hughes TP. Limited clinical value of regular bone marrow cytogenetic analysis in imatinib-treated chronic phase CML patients monitored by RQ-PCR for BCR-ABL. Leukemia. 2006;20 (4):664-670. Elias F, Gebran A, Said C, Beker RV, Ammar W. Budget impact of treatment-free remission in treating chronic-phase Philadelphiapositive chronic myeloid leukemia in lebanon. J Glob Oncol. 2019;5:1-7. Villemagne Sanchez LA, Oâ&#x20AC;&#x2122;Callaghan C, Gough K, et al. Patient perceptions of treatment-free remission in chronic myeloid leukemia. Leuk Lymphoma. 2018;59(2):406415.

2737

REVIEW ARTICLE Ferrata Storti Foundation

Discontinuation of tyrosine kinase inhibitors in chronic myeloid leukemia: when and for whom? Ehab Atallah1 and Charles A Schiffer2

1 Medical College of Wisconsin and 2Karmanos Cancer Institute, Wayne State University School of Medicine, Milwaukee, WI, USA

Haematologica 2020 Volume 105(12):2738-2745

ABSTRACT

reatment discontinuation is considered one of the main goals of therapy for patients with chronic myeloid leukemia. Several criteria are felt to be necessary to consider discontinuation, while others may predict a better chance of achieving treatment-free remission. Criteria for discontinuation include patients in chronic phase chronic myeloid leukemia, a minimum duration of tyrosine kinase inhibitor therapy of 3 years, sustained deep molecular response for at least 2 years and a molecular response of at least MR4. In addition, proper education of the patient on the need for more frequent monitoring, possible side effects related to stopping and having a reliable real-time quantitative polymerase chain reaction laboratory are paramount to the safety and success of treatment-free remission. Realistically though, a maximum of only 20-30% of newly diagnosed patients will be able to achieve a successful treatment-free remission. In this article we will review for whom and when a trial of discontinuation should be considered.

Introduction

Correspondence: EHAB ATALLAH eatallah@mcw.edu CHARLES A SCHIFFER schiffer@karmanos.org Received: July 10, 2020. Accepted: September 17, 2020. Pre-published: October 9, 2020. doi:10.3324/haematol.2019.242891 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.

2738

Although treatment with tyrosine kinase inhibitors (TKI) has led to remarkable improvement in the overall survival of patients with chronic myeloid leukemia (CML), until recently these drugs were not considered curative as patients needed to be treated indefinitely. Hematopoietic stem cell transplantation was considered to be the only curative therapy for patients with CML. Over the last decade, several studies have demonstrated that a proportion of patients may be able to stop TKI therapy without relapse, potentially rendering them cured of their disease.1-7 After discontinuing TKI therapy, the levels of BCR-ABL1 in the peripheral blood should be monitored closely by polymerase chain reaction (PCR), especially in the first year, and monitoring should continue indefinitely. This approach has been safe, with essentially all patients with a molecular relapse responding to re-initiation of TKI therapy and only a few case reports of disease progression to accelerated phase or blast crisis among thousands of patients discontinuing therapy. This rare phenomenon of sudden transformation to accelerated phase or blast crisis has also been reported in responding patients who remain on TKI.8 With discontinuation, approximately 50% of patients will stay in remission and will not need to restart therapy. This goal of being off drug therapy, also known as treatment-free remission (TFR), is emerging as the new aim of CML therapy. TFR has several benefits for patients, including improved quality of life, reduced cost to patients and society and decreased risk of long-term complications. Our aim in this article is to review the results of trials on stopping TKI in patients with CML and to discuss some clinical and biological questions raised by the observations.

Who is eligible to stop therapy? Current guidelines9,10 recommend considering stopping therapy for patients in chronic phase CML, who have been on TKI therapy for at least 3 years, and who have had a continued deep molecular response (DMR) lasting at least 2 years, with DMR defined as a BCR/ABL1 level of < 0.01% on the international scale (IS) (equivalent to a 4-log reduction in transcript level from baseline, MR4). Studies have only haematologica | 2020; 105(12)

TKI discontinuation in CML

included patients with chronic phase CML. The possibility of stopping therapy in patients with prior accelerated phase or blast phase is unknown at present. In general, we do not recommend stopping drug in such patients given the lack of data. The depth and duration of response that need to be achieved prior to discontinuation remain unclear. The required duration of DMR has varied from 1 to 3 years in different studies (Table 1). Most studies enrolled patients who were in a continued DMR for at least 2 years and the duration of DMR has been predictive of successful TFR (see later). Initially, studies enrolled only patients with undetectable BCR-ABL1 by PCR with an assay sensitivity of at least a 5 log reduction.11 This was followed by other trials such as the EURO-SKI6 and LAST studies12 in which patients with BCR/ABL1 <0.01% were enrolled. In the Euro-SKI trial, there was no difference in relapse rate between patients who had IS% <0.01% or 0.0032%.13 In contrast, the DESTINY trial conducted in the United Kingdom, enrolled patients with BCR-ABL1 <0.1% (defined as a major molecular response, MMR) and the rate of successful TFR only for patients in MMR was lower (33%). Based on that, a level of 0.01% or MR4 is considered an appropriate level to consider discontinuation outside of clinical trials. Of note, we and others have described that a significant number of patients who meet stopping criteria are actually not interested in stopping TKI therapy. These observations were made at a time when stopping TKI was still considered experimental and guidelines did not mention this option. Patients were hesitant to stop taking a drug mainly on the principle of “if it ain’t broke, don’t fix it”.14,15 In particular, patients who were not experiencing side effects from the TKI and who had good insurance coverage, so that out-of-pocket costs were minimal, had less obvious motivation to consider discontinuation. In contrast, more recently, many patients who have been told of their “excellent” responses have read about TFR on the Internet and are inquiring about stopping therapy, some-

times after relatively short exposure to treatment. Hence, the need for physicians to understand the implications of different levels of response on the IS as well as the requirement for many years of treatment for most patients. Although most trials only enrolled patients with no history of resistance to TKI, some patients who were resistant to frontline imatinib may stop TKI treatment successfully.7,16-18 The DASfree study enrolled patients on first- or second-line treatment with dasatinib.7 Of the 84 patients enrolled, 37 (44%) and 47 (56%) were receiving dasatinib as first- and second-line treatment, respectively. At 24 months, the rate of TFR was 51% and 42% in those treated with dasatinib first-line or as a subsequent line. The rate of successful TFR was 44% for patients who were either resistant or intolerant to first-line dasatinib. The ENEStop study enrolled patients intolerant of or resistant to imatinib. Of the 163 patients enrolled, 38 were resistant to imatinib. After 1 year of nilotinib consolidation therapy, 30 patients moved on to the TFR phase, of whom 16 (53%) remained off drug at 48 weeks16,19 (Table 2). A second attempt at TFR may be considered in some patients.20-22 The RE-STIM trial enrolled patients who had previously relapsed after TKI discontinuation and regained a DMR.21 Of the 70 patients, 35% remained off drug at 36 months. The median duration of TKI therapy after restarting was 5 months (range, 2-42 months). BCRABL1 kinetics at the time of first discontinuation was most predictive of success of a second TFR. The chances of a successful second TFR at 36 months were 46%, 31% and 0% for patients who had undetectable transcripts or >MR4.5 versus MR3-MR4.5 versus loss of MMR at 3 months after first discontinuation. In the Treatment-free Remission Accomplished by Dasatinib (TRAD) trial,20 patients who had confirmed loss of MR4 or loss of MMR after discontinuing imatinib were started on dasatinib 100 mg daily. Patients who achieved MR4 and sustained this level of response for at least 12 months had a second attempt at TFR. The TFR success rate at 6 months was 21.5%. Factors associated with a successful second TFR

Table 1. Key characteristics of selected studies of discontinuation of tyrosine kinase inhibitor therapy.

Eligibility criteria Reference

Median time in MR, months (range)

Median duration on TKI, months (range)

N. of patients

TKI(s)

Duration of TKI (y)

Depth of MR

Duration of MR (y)

ENESTfreedom4 STIM13

222 100

Nilotinib Imatinib#

2 3

MR4.5 MR5.0

1 2

NR 36.4 (24-107)

43.5 (32.9–88.7) 58.8 (35-112)

TWISTER5

Imatinib#

MR4.5

30*

70*

EURO-SKI6

868

MR4.0

56.4 (34.8-82.8)

90 (60-118.8)

DasFree7 LAST1

79 172

Imatinib or nilotinib or dasatinib Dasatinib Imatinib or dasatinib or nilotinib or bosutinib

2 3

MR4.5 MR4.0

1 2

28 (13–116) 56.6 (25.1 - 176.6)

9 (29–244) 82.7 (36.1 - 199)

Trigger to restart TKI

Loss of MMR At least two positive RT-PCR results showing an increase of one log or confirmed loss of MMR Loss of MMR or two consecutive positive samples at any value Loss of MMR

Loss of MMR Loss of MMR

MR: molecular response (with log reduction in BCR-ABL1 transcripts); TKI: tyrosine kinase inhibitor; y: years; NR: not reported; RT-PCR: reverse transcriptase polymerase chain reaction; MMR: major molecular response. *Patient on imatinib only, range not reported. #Prior interferon therapy allowed.

haematologica | 2020; 105(12)

2739

E. Atallah et al.

were: longer time to molecular relapse after imatinib discontinuation (>3 months vs. <3 months), deeper response to initial therapy (5.5 log reduction or deeper vs. others) and loss of MR4 (compared to loss of MMR).20 Both studies suggest that patients who have a rapid loss of MMR after a first discontinuation of a TKI have a negligible chance of achieving a seond TFR with TKI therapy alone (Table 3). It has been suggested that initial treatment with a second-generation TKI, rather than imatinib, might increase the fraction of patients eligible for consideration of stopping therapy because second-generation TKI do lead to a higher percentage of patients achieving MR4 or MR4.5. Whether this actually leads to a higher rate of successful TFR remains unknown as TFR outcomes are not available from the up-front randomized trials evaluating secondgeneration TKI. In the ENESTnd study of imatinib versus nilotinib, the rates of MR4 at 5 years were 42% and 66% for patients on imatinib and nilotinib, respectively, while the rates of MR4.5 at 5 years were 31% and 54%, respectively.23 Similarly, in the DASISION study of imatinib versus dasatinib, the rates of MR4.5 were 33% and 42%, respectively.24 However, only about 70% of patients remain on nilotinib or dasatinib in the long-term and the rates of significant side effects (predominantly cardiovascular and pleural effusions, respectively) are much higher with these drugs than with imatinib. The side effects associated with second-generation TKI may now be lower given better selection of patients, earlier identification of the side effects and the use of lower doses, as will be discussed later. Discontinuation trials in patients treated with nilotinib or dasatinb have shown the same ~50% rates of TFR as seen in the imatinib trials.7,25,26 Another strategy which has been evaluated is to switch imatinib-treated patients to a second-generation TKI to get patients to a deeper response and be eligible for discontinuation. In the ENESTcmr study, patients receiving imatinib who had not achieved MR4.5 were randomized

to either imatinib continuation or a switch to nilotinib.27 At 24 months, patients on imatinib who had detectable BCR-ABL1 were allowed to switch to nilotinib. At 24 months, the rate of MR4.5 was 54% for those who switched to nilotinib versus 32% for those who stayed on imatinib, reiterating prior observations that the depth of response often improves over time in responding patients receiving imatinib. The rate of successful TFR was not reported and a significant fraction of patients (18%) discontinued nilotinib because of adverse events, mostly cardiovascular in nature. Of the patients who switched to nilotinib after the 24 months, 37% achieved MR4.5. It should be noted however, that MR4 is considered sufficient for an attempt at TKI discontinuation, and it is not known whether there were important differences in the rates of sustained MR4 using the two strategies. In summary, the rates of achieving MR4.5 are approximately 10-20% higher with second-generation TKI than with imatinib whether the second-generation TKI is started as first-line therapy or switched to later. In addition, second-generation TKI lead to faster and deeper responses which could potentially lead to shorter time on treatment compared to the time on imatinib before a trial of discontinuation. However, this comes at an increased risk of clinical27,28 and financial toxicity.29 Hypothetically (because prospective data are not available), if 100 patients start imatinib, 40 will achieve MR4. Of those, approximately 20 will have a successful TFR. If a patient starts a secondgeneration TKI, approximately 60 will achieve MR4 and 30 will have a successful TFR.

When to restart? In the initial STIM111 and TWISTER5 discontinuation studies, criteria for restarting were loss of MR4.5; i.e., the patient restarted treatment as soon as BCR-ABL1 transcripts were detected. Subsequently, the A-STIM trial

Table 2. Treatment-free remission studies in patients resistant to first-line tyrosine kinase inhibitor therapy. 7

DASfree ENEStop16 DADI17 2G-TKI18

N. of patients

TFR (@months)

TKI

Indication to restart

25 30 13 9

40% (12) 26% (12) 8% (36) 33% (12)

Dasatinib Nilotinib Dasatinib Dasatinib or nilotinib

Loss of MMR Loss of MMR or confirmed loss of MR4 Stringent molecular relapse. Restarting at >0.0069% IS. Loss of MMR

TFR: treatment-free remission; TKI: tyrosine kinase inhibitor; MMR: major molecular response; MR4: molecular response with a 4-log reduction in BCR-ABL1 transcripts; IS: international scale. 2G-TKI: second-generation tyrosine kinase inhibitor.

Table 3. Completed and ongoing studies evaluating a second attempt to achieve treatment-free remission.

Number of patients

TFR (@months)

TRAD RE-STIM21

25 70

21% (6) 35% (36)

Matsuki22 Sweet Rousselot Olsson-Strรถmberg Spanish ELN

10 41 26 134 80 200

24% (24) NR NR NR NR NR

Notes Patients restarted dasatinib after MMR loss with imatinib TFR 72% vs. 36% for those who did vs. did not remain in DMR at 3 months with first attempt. 61 of 70 patients received the same TKI All patients on dasatinib. Ongoing study, adding ruxolitinib (NCT03610971) Ongoing study, adding pioglitazone (NCT02889003) Ongoing study, second attempt with dasatinib (NCT03573596) Ongoing study, second attempt with ponatinib (NCT04160546) Ongoing study, second attempt with nilotinib (NCT02917720)

TFR: treatment-free remission; MMR: major molecular response; DMR: deep molecular remission; TKI: tyrosine kinase inhibitor; NR: not reported; ELN: European LeukemiaNet.

2740

haematologica | 2020; 105(12)

TKI discontinuation in CML

demonstrated the safety of restarting treatment at loss of MMR30 and this is the recommended level for restarting in most current guidelines. A number of patients have fluctuating values ranging between 3-4.5 log reductions or undetectable transcripts followed by very low, but detectable values, who do not overtly relapse during long-term follow-up. It is intriguing that some patients can have detectable BCR-ABL1 transcripts by PCR analysis with no evidence of disease progression. In a study by Ross et al., seven of eight patients with undetectable BCR-ABL1 by real-time quantitative PCR had detectable BCR-ABL1 DNA and none of those relapsed.31 One explanation is that the BCRABL1-positivity arose from non-myeloid lineages such as memory lymphocytes. In a study by Pagani et al., blood from 20 patients in TFR was sorted into granulocytes, monocytes, B cells, T cells and natural killer (NK) cells. BCR-ABL1 was not detected in cells of myeloid lineage in any of the patients.32 Alternatively, transcripts may derive from a small, persistent population of immature but nonclonogenic myeloid precursors that slowly decrease with time.33 This could be a consequence of leukemia stem cell exhaustion as demonstrated by Caocci et al., who noted that patients with shorter telomeres in peripheral blood monocytes had a higher chance of successful TFR.34 Of note, in another study shorter telomeres at diagnosis were associated with molecular instability and worse overall outcome, possibly indicating that the role of telomeres varies with the stage of disease and might be a dynamic predictive marker.35

Success rate of treatment-free remission The success rate of TFR has been quite similar across studies from several countries despite the inclusion of patients receiving different TKI and some slight differ-

ences in enrollment criteria (Table 1). In addition, time to loss of MMR is very similar across studies with most patients relapsing in the first 6 months. The kinetics of molecular recurrence is similar across all TFR studies with the disease-free status from the STIM 1 study shown as an example in Figure 1.3 The molecular recurrence-free survival rate was 38% at 5 years. It is worth noting that patients who have had undetectable transcripts for many years can relapse so quickly once the suppressive pressure on the residual quiescent stem cells is released. The DASfree study enrolled patients receiving dasatinib as first- or second-line therapy. The TFR rate at 2 years was slightly higher for patients on first-line dasatinib compared to those given dasatinib as a second-line therapy (51% vs. 44%, respectively).7 The ENESTfreedom enrolled patients treated with nilotinib. Of the 190 patients who discontinued therapy, 44% remained in TFR at 192 weeks. The LAST1 and EURO-SKI studies6 enrolled patients on multiple TKI, with the majority of patients receiving imatinib. In the LAST study, with a median follow-up of 12 months, the molecular relapse-free survival rate was 66% at 24 months. In the EURO-SKI study with a median follow-up of 24 months, the molecular relapse-free survival was 50% at 24 months.

Monitoring As most molecular recurrences occur in the first 6 months after stopping TKI therapy, more intensive monitoring is recommended earlier after stopping the treatment. Currently, guidelines suggest monitoring with peripheral blood PCR analysis every 4 weeks for 6 to 12 months, then every 6 to 8 weeks for 12 to 18 months, then every 3 months thereafter. Whether this schedule of intensive upfront monitoring is really needed remains unclear and a recent mathematical modeling experiment suggests

Figure 1. Molecular recurrence-free survival after discontinuation of imatinib in patients (n=100) in the STIM1 study.3

haematologica | 2020; 105(12)

2741

E. Atallah et al.

that initial monitoring every 2 months followed by every 3 months, should transcripts remain undetectable after 6 months, is likely to be a safe approach.36 Monthly PCR evaluations increase patients’ anxiety and may present an economic hurdle for some patients or countries. The mechanisms by which the CML clone is either eliminated or perhaps more likely, suppressed are not fully understood. It has been suggested that small amounts of residual disease can be controlled by autologous immune suppression. A number of studies have evaluated T-cell and particularly NK-cell populations in patients in TFR, but none has been convincing and further studies are needed.37,38 It is conceivable that residual clonogenic cells remain in a relatively quiescent state which may take years to manifest clinically, perhaps because of an eventual failure of immune control. With few exceptions, we do not know the interval between the original acquisition of the BCR-ABL1 mutation and the development of clinically apparent CML. The median time to the development of “clinical” CML after the atomic bombing of Japan was approximately 10 years with cases continuing to be detected for years after that.39 Hence, the importance of long-term, continued PCR monitoring cannot be over emphasized as late relapses have been reported and remain a concern. Rousselot et al. identified nine patients relapsing more than 2 years (median 3.6 years) after stopping, with late relapses seeming to be more common in patients with fluctuating detectable transcript levels during earlier follow-up.40 Our policy is to continue to monitor patients about every 4-6 months after 3 years in TFR; it is likely that this approach could be modified with the availability of more long-term follow-up information. It should be noted that PCR standardization and monitoring are not uniformly available around the world and in the USA. It has been reported that approximately 50% of US physicians do not have adequate PCR monitoring and do not follow guidelines when it comes to TKI discontinuation, so further educational efforts are warranted.41

Are there factors that predict successful treatment-free remission? The duration of TKI treatment, Sokal risk score, duration of DMR, depth of response, the use of highly sensitive droplet digital PCR (ddPCR), number of NK cells, and type of BCR-ABL1 transcript have all been shown to predict the success rate of TFR to a varying degree in some, but not all studies. • The duration of TKI treatment has perhaps been the most reproducible predictor of successful therapy discontinuation. TKI therapy for more than 5 years has been associated in multiple studies with higher rates of TFR. For example in the STIM trial, the risk of molecular recurrence was 50% for those who were on imatinib >54 months versus 80% for those on imatinib <54 months.42 In the large EURO-SKI trial, the 6-month probability of maintaining a MMR was 63% for patients on imatinib treatment for >5.8 years versus 41% for those treated for <5.8 years. However, when the duration of TKI therapy and duration of DMR were entered into the same model, only one of them was predictive as both these factors were closely correlated.6 These observations emphasize 2742

the need for many years of treatment before most patients can be considered for discontinuation and should be part of the initial discussion with patients beginning treatment for chronic phase CML. It also indicates that caregivers must be skilled in managing the side effects of TKI therapy so as to maximize compliance. • The duration of DMR is less consistent in terms of predicting successful TFR. All studies of TKI discontinuation required a minimum of 2 years of documented DMR although the definition of DMR varied across studies (Table 1). The duration of deep remission was not predictive in STIM1, DASfree,7 or ENESTfreedom4 but was predictive in EURO-SKI.6 In the EURO-SKI trial, the 6-month probability of maintaining MMR was 61% for patients who maintained MR4 for >3.1 years versus 44% for those who maintained MR4 for <3.1 years. • Sokal risk score: a lower score was predictive of better outcomes in ENESTfreedom43 and EUROSKI;6 however, in everyday practice the Sokal risk score is currently calculated for very few patients, making this a less practical marker. Nonetheless, the rate of TFR is likely to be lower in patients with more “advanced” chronic phase CML. • Type of transcript: several studies have demonstrated that patients harboring the e14a2 transcript are more likely to achieve a DMR and have successful TFR compared to those with e13a2.44 In one study by D’Adda et al., the rates of a sustained DMR and maintaining TFR were 39.6% versus 19.4% and 61% versus 22% for patients with e14a2 versus e13a2 transcripts, respectively.45 • Depth of response is also emerging as an important factor in predicting successful TFR. In the DESTINY trial,2 in which patients with BCR-ABL1 <0.1% were enrolled, a dose de-escalation phase was included. After enrollment into the study, patients received 50% of the FDAapproved dose of their TKI for 1 year. Patients with at least one IS measurement of transcripts between 0.01% and 0.1% were considered in the MMR group, while patients who had all measurements <0.01% were considered in the MR4 group. At 3 years, the rate of successful TFR was 72% versus 33% in the MR4 and MMR groups, respectively.2 These results suggest that patients in continuous MMR may consider dose de-escalation or may even consider TKI discontinuation, understanding that the chance of successful TFR is approximately 30%. The ENESTfreedom study enrolled patients with chronic phase CML who had been on nilotinib for at least 2 years and achieved MR4.5. Patients who maintained MR4 during an additional year of follow-up discontinued the drug. Among the 190 patients who discontinued nilotinib, the rate of successful TFR was 50% for those who maintained MR4.5 versus 35% for those with at least one level >MR4.5.43 ddPCR is 1-2 logs more sensitive than real-time quantitative PCR46 and some studies have demonstrated that negative or undetectable transcripts by ddPCR are most predictive of successful TFR.47-49 In a study by Nicolini et al., of the 174 patients who discontinued therapy and for whom ddPCR results were available, 37 had BCR-ABL1 values ≥0.0023%. Loss of MMR occurred in 68% of patients who had BCR-ABL1 values ≥0.0023%, compared to 46% with levels <0.0023%. As for patients monitored by real-time quantitative PCR, patients with detectable BCR-ABL1 by ddPCR may still consider a TFR attempt and ddPCR BCR-ABL1 levels may further refine the selection of patients for an attempt at TKI discontinuation. haematologica | 2020; 105(12)

TKI discontinuation in CML

Currently ddPCR is not widely available and standardization across laboraties will be required prior to its use in routine clinical care. All these data suggest that the depth of response matters. The probability of successful TFR is highest for patients with negative ddPCR results, followed by those with MR4.5, then MR4 and finally, MR3 (Table 4). However, this distinction is not absolute, since relapse rates are still considerable at all these levels and patients and physicians may choose to consider a trial of TKI discontinuation at higher transcript levels with the understanding that the risks of molecular recurrence and the need to restart treatment are higher. • Other investigational studies have associated a higher percentage of innate T cells37 or NK cells38 with a better chance of TFR, suggesting that an immune surveillance mechanism is involved in maintaining molecular response.

Side effects of discontinuation As noted in the LAST study, which included a comprehensive evaluation of patient-reported outcomes after TKI discontinuation and at the time of restarting TKI, TKI discontinuation is associated with clinically significant improvements of fatigue, diarrhea, depression and sleep disturbances in many patients.50 Conversely, some patients who discontinue TKI and need to restart have significant anxiety51,52 with one patient describing it “as having leukemia all over again.” Musculoskeletal pain is a well-recognized side effect of TKI discontinuation.53,54 Approximately 30% of patients have an increase in musculoskeletal pain or periarticular stiffness, which is usually transient. These symptoms can be managed with non-steroidal anti-inflammatory drugs, and occasionally glucocorticosteroids and acetaminophen. Pain is more common in patients with a prior history of musculoskeletal pain and in those who have been on TKI therapy for a long period of time.53 A small minority of patients need to restart TKI for uncontrolled discomfort. Several mechanisms have been proposed to explain this phenomenon, including activation of other kinase-mediated pathways which had been inhibited by the TKI, such as release of c-kit inhibition,55 mast cell activation,56 or bone remodeling effects,57 but none has been proven yet.

Dose reduction For a large majority of patients who cannot achieve TFR, another approach to reduce side effects and costs and to maintain compliance would be to identify the lowest possible effective dose of TKI.58,59 Both dasatinib and nilotinib have demonstrated efficacy at doses lower than the FDA-approved dose. In a recently published phase II study from the MD Anderson Cancer Center,60 dasatinib was given at a dose of 50 mg daily, instead of the FDAapproved dose of 100 mg, to newly diagnosed patients. Of the 81 patients enrolled, 81% had achieved a MMR at 12 months according to the follow-up data at the time of the publication. This is higher than in the DASISION study, in which the MMR rate was 46% at 1 year with dasatinib.24 The 50 mg dose was better tolerated with a low cumulative incidence of pleural effusions of 6%. haematologica | 2020; 105(12)

Table 4. Studies reporting the rate of successful treatment-free remission based on depth of response.

Study Bernardi49

Nicolini48 ENEST freedom43 DESTINY2

Level

TFR %

ddPCR <0.468 ddPCR >0.468 MR4 MR 4.5 - 5 ddPCR <0.0023% ddPCR >0.0023% MR4 MR4.5 <MR4 MR3

83 52 74 80 68 46 35 51 72 33

2 years 2 years 2 years 2 years 1 year 1 year 2 years 2 years 3 years 3 years

TFR: treatment-free remission; MR: molecular response (with log reduction in BCRABL1 transcripts); ddPCR: droplet digital polymerase chain reaction analysis.

The NiloRED study was an observational study evaluating dose reductions for patients on nilotinib61 Of the 67 evaluable patients, 68.6% were receiving nilotinib as firstline therapy and 31.4% as second-line treatment. The median duration of MMR was 25 months at the time of dose reduction. The nilotinib dose was reduced to 450 mg daily, 400 mg daily and 300 mg daily in 87%, 10% and 3% patients, respectively. Only two patients lost the MMR, although they regained MR4 later on without a dose increase. The DESTINY trial, detailed earlier, provides a perfect example illustrating the safety of dose reduction.2 Of the 174 patients enrolled, 148, 16 and 10 were receiving imatinib, nilotinib and dasatinib, respectively. At study entry, the TKI dose was reduced to half the standard dose for 12 months: imatinib 200 mg daily, dasatinib 50 mg daily, or nilotinib 200 mg twice daily. After 1 year, 19% of patients in the group with MMR had lost their MMR compared to 2% of those starting with deeper responses. After 1 year of half-dose therapy, patients who had not “relapsed” discontinued therapy as discussed earlier. It is unclear what would have happened if those patients had stayed on a lower dose and had not attempted treatment discontinuation. The INTERIM trial was another study that demonstrated the safety of imatinib dose reduction for patients in a complete cytogenetic remission (CCyR).62 In that study, imatinib was reduced to 1 month on/1 month off in 67 patients who were >65 years of age and had attained CCyR. Of those patients, 17% and 18% lost their CCyR and MMR, respectively. All patients who restarted higher dose therapy attained CCyR and MMR again. In summary, these data demonstrate that, particularly for patients experiencing chronic symptoms, dose reduction can be an option for those on a stable dose of TKI with adequate response, including those who are not eligible for or do not want an attempt at treatment discontinuation, and those in whom previous TKI discontinuation failed. More frequent monitoring is advisable for the first 6 months after dose reduction to detect the unusual individual who rapidly loses response. In this regard, Fassoni and colleagues, using data from large clinical trials, modeled the kinetics of transcript levels after a 50% reduction in dose in patients in stable MMR.58 Importantly, their model predicts a transient increase in BCR-ABL1 transcripts shortly after dose reduction which decreases to baseline or lower, without a return to the original dose. Thus, the model suggests that clinicians should 2743

E. Atallah et al.

not reflexively increase the dose should the transcript level increase, but rather that they should continue to observe the patient closely.59 Indeed, this phenomenon was noted in some patients in the NiloRed trial mentioned above.54 Additional prospective, systematic studies of dose reduction would be welcomed in the future.

Future directions An attempt at treatment discontinuation should be considered the standard of care for patients with CML who have a sustained DMR. Realistically, only 40-50% of newly diagnosed patients with CML will achieve this threshold, and only half of those (~20% of all patients) will achieve a successful TFR. These estimates are extrapolated from the results of clinical trials and may overstate the actual number of patients treated in general practice who may achieve the required depth and duration of response and be eligible for a trial of TKI discontinuation.

References 1. Atallah E, Schiffer CA, Radich JP, et al. Results from the U.S. Life after Stopping TKIs (LAST) study. Blood. 2017;130(Suppl 1):2903. 2. Clark RE, Polydoros F, Apperley JF, et al. Deescalation of tyrosine kinase inhibitor therapy before complete treatment discontinuation in patients with chronic myeloid leukaemia (DESTINY): a non-randomised, phase 2 trial. Lancet Haematol. 2019;6(7): e375-e83. 3. Etienne G, Guilhot J, Rea D, et al. Long-term follow-up of the French Stop Imatinib (STIM1) study in patients with chronic myeloid leukemia. J Clin Oncol. 2017;35(3): 298-305. 4. Giles FJ, Masszi T, Casares MTG, et al. Treatment-free remission (TFR) following frontline (1L) nilotinib (NIL) in patients (pts) with chronic myeloid leukemia in chronic phase (CML-CP): 192-week data from the ENESTfreedom study. J Clin Oncol. 2019;37 (15_suppl):7013. 5. Ross DM, Branford S, Seymour JF, et al. Safety and efficacy of imatinib cessation for CML patients with stable undetectable minimal residual disease: results from the TWISTER study. Blood. 2013, 2013;122(4): 515-522. 6. Saussele S, Richter J, Guilhot J, et al. Discontinuation of tyrosine kinase inhibitor therapy in chronic myeloid leukaemia (EURO-SKI): a prespecified interim analysis of a prospective, multicentre, non-randomised, trial. Lancet Oncol. 2018;19(6): 747-757. 7. Shah NP, García-Gutiérrez V, JiménezVelasco A, et al. Dasatinib discontinuation in patients with chronic-phase chronic myeloid leukemia and stable deep molecular response: the DASFREE study. Leuk Lymphoma. 2020;61(3):650-659. 8. Tantiworawit A, Power MM, Barnett MJ, et al. Long-term follow-up of patients with chronic myeloid leukemia in chronic phase developing sudden blast phase on imatinib therapy. Leuk Lymphoma. 2012;53(7):13211326. 9. NCCN guidelines Clinical Practice Guidelines in Oncology. Version 3.2020January 30, 2020. https://www.nccn.org/

2744

Many unanswered questions remain. How can we better identify patients who will achieve DMR and TFR? Should these patients be considered “cured” of their disease and what are the mechanisms by which the disease is controlled/eliminated? A better understanding of these mechanisms might suggest new approaches to prevent relapse. How can more patients achieve a DMR in order to consider a trial to achieve TFR? The limitations of the currently available TKI have been defined but a new inhibitor, ABL001, which inhibits signaling by blocking the myristoyl pocket of the enzyme,63,64 has recently been developed and combinations with traditional TKI will be conducted in the hopes of achieving deeper responses as the pathway to higher rates of TFR. Nonetheless, when the extraordinary results using imatinib first became available, there was little expectation that the responses would be so durable and that it might eventually be possible to avoid lifetime treatment. CML investigators should be proud of these remarkable accomplishments.

professionals/physician_gls [Accessed April 10, 2020]. 10. Hochhaus A, Baccarani M, Silver RT, et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia. 2020;34(4):966-984. 11. Mahon FX, Rea D, Guilhot J, et al. Discontinuation of imatinib in patients with chronic myeloid leukaemia who have maintained complete molecular remission for at least 2 years: the prospective, multicentre Stop Imatinib (STIM) trial. Lancet Oncol. 2010;11(11):1029-1035. 12. Atallah E, Schiffer CA, Weinfurt KP, et al. Design and rationale for the Life After Stopping Tyrosine kinase inhibitors (LAST) study, a prospective, single-group longitudinal study in patients with chronic myeloid leukemia. BMC Cancer. 2018;18 (1):359. 13. Mahon F-X, Richter J, Guilhot J, et al. Interim analysis of a pan European stop tyrosine kinase inhibitor trial in chronic myeloid leukemia: the EURO-SKI study. Blood. 2014;124(21):151. 14. Flynn KE, Myers JM, D'Souza A, Schiffer CA, Thompson JE, Atallah E. Exploring patient decision making regarding discontinuation of tyrosine kinase inhibitors for chronic myeloid leukemia. Oncologist. 2019;24(9):1253-1258. 15. Goldberg S, Hamarman S. Patients with chronic myelogenous leukemia may not want to discontinue tyrosine kinase inhibitor therapy. Blood. 2015;126(23): 1584. 16. Hughes TP, Boquimpani CM, Takahashi N, et al. Treatment-free remission in patients with chronic myeloid leukemia in chronic phase according to reasons for switching from imatinib to nilotinib: subgroup analysis from ENESTop. Blood. 2016;128(22): 792. 17. Okada M, Imagawa J, Tanaka H, et al. Final 3-year results of the dasatinib discontinuation trial in patients with chronic myeloid leukemia who received dasatinib as a second-line treatment. Clin Lymphoma Myeloma Leuk. 2018;18(5):353-360.e1. 18. Rea D, Rousselot P, Guilhot F, et al. Discontinuation of second generation (2G) tyrosine kinase inhibitors (TKI) in chronic

phase (CP)-chronic myeloid leukemia (CML) patients with stable undetectable BCR-ABL transcripts. Blood. 2012;120(21): 916. 19. Hughes TP, Boquimpani C, Takahashi N, et al. ENESTop 192-week results: treatmentfree remission (TFR) in patients (pts) with chronic myeloid leukemia in chronic phase (CML-CP) after stopping second-line (2L) nilotinib (NIL). J Clin Oncol. 2019;37 (15_suppl):7005. 20. Kim DDH, Busque L, Forrest DL, et al. Second attempt of TKI discontinuation with dasatinib for treatment-free remission after failing first attempt with imatinib: Treatment-free Remission Accomplished by Dasatinib (TRAD) trial. Blood. 2018;132 (Supplement_1):787. 21. Legros L, Nicolini FE, Etienne G, et al. Second tyrosine kinase inhibitor discontinuation attempt in patients with chronic myeloid leukemia. Cancer. 2017;123(22): 4403-4410. 22. Matsuki E, Sakurai M, Karigane D, et al. Second attempt to discontinue TKI in CML patients who have sustained CMR for over 2 years is rarely successful even with the use of second generation TKIs. Blood. 2016;128(22):1887. 23. Hochhaus A, Saglio G, Hughes TP, et al. Long-term benefits and risks of frontline nilotinib vs imatinib for chronic myeloid leukemia in chronic phase: 5-year update of the randomized ENESTnd trial. Leukemia. 2016;30(5):1044-1054. 24. Cortes JE, Saglio G, Kantarjian HM, et al. Final 5-year study results of DASISION: the dasatinib versus imatinib study in treatment-naïve chronic myeloid leukemia patients trial. J Clin Oncol. 2016;34(20): 2333-2340. 25. Hochhaus A, Masszi T, Giles FJ, et al. Treatment-free remission following frontline nilotinib in patients with chronic myeloid leukemia in chronic phase: results from the ENESTfreedom study. Leukemia. 2017;31(7):1525-1531. 26. Mahon FX, Nicolini FE, Noël M-P, et al. Preliminary report of the STIM2 study: a multicenter stop imatinib trial for chronic phase chronic myeloid leukemia de novo patients on imatinib. Blood. 2013;122

haematologica | 2020; 105(12)

TKI discontinuation in CML

(21):654. 27. Hughes TP, Leber B, Cervantes F, et al. Sustained deep molecular responses in patients switched to nilotinib due to persistent BCR-ABL1 on imatinib: final ENESTcmr randomized trial results. Leukemia. 2017;31(11):2529-2531. 28. Cole AL, Wood WA Jr, Muluneh B, et al. Comparative safety and health care expenditures among patients with chronic myeloid leukemia initiating first-line imatinib, dasatinib, or nilotinib. JCO Oncol Pract. 2020;16(5):e443-e455. 29. Shih YT, Cortes JE, Kantarjian HM. Treatment value of second-generation BCRABL1 tyrosine kinase inhibitors compared with imatinib to achieve treatment-free remission in patients with chronic myeloid leukaemia: a modelling study. Lancet Haematol. 2019;6(8):e398-e408. 30. Rousselot P, Charbonnier A, Cony-Makhoul P, et al. Loss of major molecular response as a trigger for restarting tyrosine kinase inhibitor therapy in patients with chronicphase chronic myelogenous leukemia who have stopped imatinib after durable undetectable disease. J Clin Oncol. 2014;32(5): 424-430. 31. Ross DM, Branford S, Seymour JF, et al. Patients with chronic myeloid leukemia who maintain a complete molecular response after stopping imatinib treatment have evidence of persistent leukemia by DNA PCR. Leukemia. 2010;24(10):17191724. 32. Pagani IS, Dang P, Saunders VA, et al. Lineage of measurable residual disease in patients with chronic myeloid leukemia in treatment-free remission. Leukemia. 2020; 34(4):1052-1061. 33. Ross DM, Pagani IS, Shanmuganathan N, et al. Long-term treatment-free remission of chronic myeloid leukemia with falling levels of residual leukemic cells. Leukemia. 2018;32(12):2572-2579. 34. Caocci G, Greco M, Delogu G, et al. Telomere length shortening is associated with treatment-free remission in chronic myeloid leukemia patients. J Hematol Oncol. 2016;9(1):63. 35. Wenn K, Tomala L, Wilop S, et al. Telomere length at diagnosis of chronic phase chronic myeloid leukemia (CML-CP) identifies a subgroup with favourable prognostic parameters and molecular response according to the ELN criteria after 12 months of treatment with nilotinib. Leukemia. 2015;29(12):2402-2404. 36. Shanmuganathan N, Braley JA, Yong ASM, et al. Modeling the safe minimum frequency of molecular monitoring for CML patients attempting treatment-free remission. Blood. 2019;134(1):85-89. 37. Cayssials E, Jacomet F, Piccirilli N, et al. Sustained treatment-free remission in chronic myeloid leukaemia is associated with an increased frequency of innate CD8(+) Tcells. Br J Haematol. 2019;186(1):54-59. 38. Ilander M, Olsson-Stromberg U, Schlums H, et al. Increased proportion of mature NK cells is associated with successful imatinib

haematologica | 2020; 105(12)

discontinuation in chronic myeloid leukemia. Leukemia. 2017;31(5):1108-1116. 39. Hsu W-L, Preston DL, Soda M, et al. The Incidence of leukemia, lymphoma and multiple myeloma among atomic bomb survivors: 1950â&#x20AC;&#x201C;2001. Radiat Res. 2013;179(3): 361-382. 40. Rousselot P, Loiseau C, Delord M, Cayuela JM, Spentchian M. Late molecular recurrences in patients with chronic myeloid leukemia experiencing treatment-free remission. Blood Adv. 2020;4(13):3034-3040. 41. Atallah EL, Sadek I, Cao X, et al. Tyrosine kinase inhibitor therapy discontinuation in clinical practice in chronic myeloid leukemia - a US physician survey conducted after guideline updates. Blood. 2019;134 (Suppl_1):2208. 42. Etienne G, Nicolini NE, Dulucq S, et al. Achieving a complete molecular remission under imatinib therapy is associated with a better outcome in chronic phase chronic myeloid leukaemia patients on imatinib frontline therapy. Blood. 2012;120(21):3754. 43. Ross DM, Masszi T, GĂłmez Casares MT, et al. Durable treatment-free remission in patients with chronic myeloid leukemia in chronic phase following frontline nilotinib: 96-week update of the ENESTfreedom study. J Cancer Res Clin Oncol. 2018;144(5): 945-954. 44. Baccarani M, Rosti G, Soverini S. Chronic myeloid leukemia: the concepts of resistance and persistence and the relationship with the BCR-ABL1 transcript type. Leukemia. 2019;33(10):2358-2364. 45. D'Adda M, Farina M, Schieppati F, et al. The e13a2 BCR-ABL transcript negatively affects sustained deep molecular response and the achievement of treatment-free remission in patients with chronic myeloid leukemia who receive tyrosine kinase inhibitors. Cancer. 2019;125(10):1674-1682. 46. Jennings LJ, George D, Czech J, Yu M, Joseph L. Detection and quantification of BCR-ABL1 fusion transcripts by droplet digital PCR. J Mol Diagn. 2014;16(2):174-179. 47. Mori S, Vagge E, le Coutre P, et al. Age and dPCR can predict relapse in CML patients who discontinued imatinib: the ISAV study. Am J Hematol. 2015;90(10):910-914. 48. Nicolini FE, Dulucq S, Boureau L, et al. Evaluation of residual disease and TKI duration are critical predictive factors for molecular recurrence after stopping imatinib firstline in chronic phase CML patients. Clin Cancer Res. 2019;25(22):6606-6613. 49. Bernardi S, Malagola M, Zanaglio C, et al. Digital PCR improves the quantitation of DMR and the selection of CML candidates to TKIs discontinuation. Cancer Med. 2019;8(5):2041-2055. 50. Flynn KE, Weinfurt KP, Lin L, et al. Patientreported outcome results from the U.S. Life after Stopping TKIs (LAST) study in patients with chronic myeloid leukemia. Blood. 2019;134(Suppl_1):705. 51. Sogawa R, Kimura S, Yakabe R, et al. Anxiety and depression associated with tyrosine kinase inhibitor discontinuation in patients with chronic myeloid leukemia. Int

J Clin Oncol. 2018;23(5):974-979. 52. Sharf G, Marin C, Bradley JA, et al. Treatment-free remission in chronic myeloid leukemia: the patient perspective and areas of unmet needs. Leukemia. 2020;34(8):2102-2112. 53. Berger MG, Pereira B, Rousselot P, et al. Longer treatment duration and history of osteoarticular symptoms predispose to tyrosine kinase inhibitor withdrawal syndrome. Br J Haematol. 2019;187(3):337-346. 54. Diab M, Schiffer CA. The spectrum of musculoskeletal symptoms in patients with chronic myeloid leukemia after stopping tyrosine kinase inhibitors. Leuk Res. 2019;79:1-2. 55. Ceko M, Milenkovic N, le Coutre P, Westermann J, Lewin GR. Inhibition of c-Kit signaling is associated with reduced heat and cold pain sensitivity in humans. Pain. 2014;155(7):1222-1228. 56. Ramanujam D, McNicholl F, Furby D, et al. Dramatic resolution of respiratory symptoms with imatinib mesylate in patients with chronic myeloid leukemia presenting with lower airway symptoms resembling asthma. Leuk Lymphoma. 2009;50(10): 1721-1722. 57. Breccia M, Alimena G. The metabolic consequences of imatinib mesylate: changes on glucose, lypidic and bone metabolism. Leuk Res. 2009;33(7):871-875. 58. Fassoni AC, Baldow C, Roeder I, Glauche I. Reduced tyrosine kinase inhibitor dose is predicted to be as effective as standard dose in chronic myeloid leukemia: a simulation study based on phase III trial data. Haematologica. 2018;103(11):1825-1834. 59. Schiffer JT, Schiffer CA. To what extent can mathematical modeling inform the design of clinical trials? The example of safe dose reduction of tyrosine kinase inhibitors in responding patients with chronic myeloid leukemia. Haematologica. 2018;103(11): 1756-1757. 60. Naqvi K, Jabbour E, Skinner J, et al. Longterm follow-up of lower dose dasatinib (50 mg daily) as frontline therapy in newly diagnosed chronic-phase chronic myeloid leukemia. Cancer. 2020;126(1):67-75. 61. Rea D, Cayuela J-M, Dulucq S, Etienne G. Molecular responses after switching from a standard-dose twice-daily nilotinib regimen to a reduced-dose once-daily schedule in patients with chronic myeloid leukemia: a real life observational study (NILO-RED). Blood. 2017;130(Suppl_1):318. 62. Russo D, Martinelli G, Malagola M, et al. Effects and outcome of a policy of intermittent imatinib treatment in elderly patients with chronic myeloid leukemia. Blood. 2013;121(26):5138-5144. 63. Hughes TP, Mauro MJ, Cortes JE, et al. Asciminib in chronic myeloid leukemia after ABL kinase inhibitor failure. N Engl J Med. 2019;381(24):2315-2326. 64. Schoepfer J, Jahnke W, Berellini G, et al. Discovery of asciminib (ABL001), an allosteric inhibitor of the tyrosine kinase activity of BCR-ABL1. J Med Chem. 2018;61(18):8120-8135.

2745

ARTICLE Ferrata Storti Foundation

Hematopoiesis

Hematopoietic stem and progenitor cells use podosomes to transcellularly cross the bone marrow endothelium

Timo Rademakers,1* Marieke Goedhart,2* Mark Hoogenboezem,1 Alexander GarcĂa Ponce,3 Jos van Rijssel,1 Maryna Samus,4 Michael Schnoor,3 Stefan Butz,4 Stephan Huveneers,5 Dietmar Vestweber,4 Martijn A. Nolte,2 Carlijn Voermans2 and Jaap D. van Buul1

Haematologica 2020 Volume 105(12):2746-2756

*TR and MG contributed equally as co-first authors

Department of Plasma Proteins, Sanquin Research and Landsteiner Laboratory, Academic Medical Center, Amsterdam, the Netherlands; 2Department of Hematopoiesis, Sanquin Research and Landsteiner Laboratory, Academic Medical Center, Amsterdam, the Netherlands; 3Department of Molecular Biomedicine, Center of Research and Advanced Studies (CINVESTAV-IPN), Mexico-City, Mexico; 4Max Planck Institute for Molecular Biomedicine, MĂźnster, Germany and 5Department of Medical Biochemistry, Academic Medical Center, Amsterdam, the Netherlands 1

ABSTRACT

Correspondence: JAAP D. VAN BUUL j.vanbuul@sanquin.nl Received: May 4, 2018. Accepted: January 20, 2020. Pre-published: January 23, 2020. doi:10.3324/haematol.2018.196329 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.

2746

one marrow endothelium plays an important role in the homing of hematopoietic stem and progenitor cells (HSPC) upon transplantation, but surprisingly little is known on how the bone marrow (BM) endothelial cells regulate local permeability and hematopoietic stem and progenitor cells transmigration. We show that temporal loss of vascular endothelial-cadherin function promotes vascular permeability in BM, even upon low-dose irradiation. Loss of vascular endothelial-cadherin function also enhances homing of transplanted HSPC to the BM of irradiated mice although engraftment is not increased. Intriguingly, stabilizing junctional vascular endothelial-cadherin in vivo reduced BM permeability, but did not prevent HSPC cells migration into the BM, suggesting that HSPC use the transcellular migration route to enter BM. Indeed, using an in vitro migration assay, we show that human HSPC cells predominantly cross BM endothelium in a transcellular manner in homeostasis by inducing podosome-like structures. Taken together, vascular endothelial-cadherin is crucial for BM vascular homeostasis but dispensable for the homing of HSPC. These findings are important in the development of potential therapeutic targets to improve HSPC homing strategies.

Introduction Hematopoietic stem cell transplantation is used to restore hematopoiesis in patients with (hematological) malignancies and disorders after chemotherapy and/or irradiation. The first step to therapeutic success of hematopoietic stem cell transplantation critically depends on the homing of sufficient numbers of hematopoietic stem- and progenitor cells (HSPC) to the bone marrow (BM).1 An important step in homing is the actual transmigration of re-infused HPSC across the BM endothelium to the underlying parenchyma. This extravasation event requires a set of specific molecular interactions that mediate the firm adhesion of the HSPC to the BM endothelial cells, and subsequent transmigration across the endothelial lining.2,3 For the diapedesis step, HSPC have to cross the endothelial barrier. It has been described that immune cells can cross the endothelial barrier on two different routes: the paracellular route, i.e., through the cell-cell junctions, or the transcellular pathway, i.e., through the cell body.4 It appears that neutrophils and monocytes prefer the paracellular route. However, T lymphocytes may have a preference for the transcellular route.5 Not only the immune cell type seems to matter here, but also the state of the endothelium: the brain endothelium allows more cells to traffic in a transcellular fashion, whereas in the lungs, all immune cells seem to prefer the junctions, i.e., the paracellular route.6,7 In addition to that, haematologica | 2020; 105(12)

HSPC migrate transcellularly

the amount of ICAM-1, one of the crucial adhesion molecules expressed by the inflamed endothelium can steer the route preference: overexpression of ICAM-1 drives lymphocytes towards the transcellular route.8-10 Up to now, it is unclear why one immune cell prefers the paracellular route whereas the other crosses in a transcellular way. Upon HPSC transplantation, it is desired to have as many HSPC to home to the bone marrow as possible. Understanding how these cells would cross the endothelium may open new opportunities to promote HPSC homing. It has been postulated that the route of the least resistance may play a role here.6 One important mediator of this hypothesis would be the barrier function of the endothelium. VE-cadherin is an important mediator of the endothelial barrier.11 It specifically controls the leakage in lung and skin tissues.7 In the BM, sinusoid lining appears to be continuous.12-14 This was demonstrated by scanning electron microscopy. The number of fenestrations found without the presence of a migrating immune cell was very limited. The role for VE-cadherin in the regulation of BM vascular integrity is not clear. As VE-cadherin is an important regulator for the barrier function as well as for transmigration, we investigated to what extend VE-cadherin regulates the integrity of the BM vasculature and the homing of HSPC in that respect. Our data show that VE-cadherin regulates the vascular integrity in the BM upon low-dose irradiation conditions. Blocking of VE-cadherin with interfering antibodies increases permeability and promotes the homing of HPSC to the BM. Using the VE-cadherin-α-catenin chimera knock-in animals, we could show that HPSC cross the BM endothelium in a transcellular manner, although basal and irradiation-induced permeability was reduced. Additional experiments implicate podosome structures on the HSPC to be involved in the transmigration event. Interestingly, blocking VE-cadherin did promote the number of HPSC that home to the BM. However, we did not find a beneficial effect for the long-term engraftment of these cells in the BM. Together, we conclude that HSPC cross the BM endothelium in a transcellular fashion using podosomes. Temporal targeting of the BM-endothelial VE-cadherin will most likely result in a disturbance of the BM homeostasis and will not result in a faster recovery of the BM population upon irradiation. These findings are important in the development for potential therapeutic target to improve HSPC homing strategies.

Methods Mice The following strains were used: C57BL/6, C57BL/6-Ly5.1, VEcadherin/α-catenin7 and VE-cadherin-GFP.15 The VE-cadherin/αcatenin mice were characterized previously and showed no clear differences in hematopoiesis.16 Mice were maintained on a C57BL/6 background in the animal facilities of the Netherlands Cancer Institute (Amsterdam, the Netherlands) and the Max Planck Institute for Molecular Biomedicine (Münster, Germany) in specific pathogen-free conditions. All animal experiments were approved by the local Animal Ethical Committee in accordance with national regulations.

Multi-photon imaging of vascular permeability Mice were injected with PBS containing GS-I17,18 or VE-cadherin haematologica | 2020; 105(12)

blocking or IgG1 isotype control antibody.19 See the Online Supplementary Material and Methods for further details.

Murine HSPC homing assay See the Online Supplementary Material and Methods for details.

Flow cytometry See the Online Supplementary Material and Methods for details.

Confocal microscopy of murine BM sections See the Online Supplementary Material and Methods for details.

Human HSPC migration assays Cord blood (CB) was collected according to the guidelines of Eurocord Nederland and CD34+ cells were isolated as previously described.20 Generation of HBMEC cell lines was previously described.21 See the Online Supplementary Material and Methods for details.

Physiological flow assays Physiological flow experiments were performed as previously described.22 See the Online Supplementary Material and Methods for details.

Podosome formation assays Human dendritic cells (DC) were generated as described previously.23 See the Online Supplementary Material and Methods for details.

VE-cadherin internalization assays See the Online Supplementary Material and Methods for details.

Immunoprecipitation and Western blot analysis Immunoprecipitation and Western blotting were performed as previously described.24 See the Online Supplementary Material and Methods for further details.

Statistics See the Online Supplementary Material and Methods for details.

Results BM vasculature is highly permeable for small molecules To test the vascular barrier in the BM during homeostasis, fluorescently-labeled 10 kDa dextrans were intravenously administered to mice together with the vascular marker GS-I and allowed to circulate for 5 minutes, after which the mice were sacrificed (Figure 1A). To define vascular permeability, we employed whole mount multiphoton imaging of several organs and measured the fluorescent intensity of the dextrans in the vascular tissue microenvironment within a perimeter of one cell layer (8 mm) around individual blood vessels (Figure 1B). To correct for potential loss of intensity at greater tissue depth, fluorescence intensity of the dextrans was normalized to intensity values of the GS-I vascular staining (Figure 1C). We determined the vascular permeability of BM, liver, lung and heart as a ratio to the vascular permeability of the brain, where vascular permeability is exceptionally low (Figure 1D-E). Vascular permeability in the BM was comparable to that of the liver, and approximately 2-3 times higher than that of the lung and heart, respectively (Figure 1D, 1E). Thus, BM vessels are permeable for small dex2747

T. Rademakers et al. A

Figure 1. Bone marrow vasculature is highly permeable for small molecules. (A) Schematic summary of experimental set up. (B) Bone marrow (BM) imaging after injection of fluorescent dyes with in green GS-I for vessel labeling and in red 10 kDa dextran (panel 1). Panel 2-4 show identification of a vessel (solid line) with red fluorescence detection outside of the vessel (dotted line), as a measurement of vascular permeability. V indicates the vessel lumen. (C) Mathematical equation for vascular permeability: Fluorescence intensity of dextran (solid line in 1B) is divided by the fluorescence intensity of the vessel stain in the selected area (dotted line in 1B). (D) Detection of vascular permeability in several organs as indicated. (E) Quantification of the vascular permeability per organ, calculated as described in (C) (n=2). (F) Detailed analysis of vascular permeability in specific BM regions: Diaphysis and metaphysis and discrimination between arterioles and sinusoids, based on intensity of GS-I stainings (n=5). Scale bars: 25 mm.

2748

haematologica | 2020; 105(12)

HSPC migrate transcellularly

trans, similar to fenestrated sinusoids of the liver. Within the BM, we discriminated between sinusoids and arterioles in the metaphysis and diaphysis area, based on CD31 intensity staining, according to Bixel and coworkers.4 We found that basal permeability in sinusoids was significantly higher than in arterioles (Figure 1F). These data indicate that baseline permeability differs per type of vessel in the BM.

VE-cadherin regulates BM vascular permeability in homeostatic conditions and after irradiation As VE-cadherin is recognized as the major regulator of vascular permeability,25 we investigated whether vascular permeability in the BM is also regulated by VE-cadherin. To block homotypic binding of VE-cadherin at the endothelial junctions, we injected mice intravenously with a blocking antibody against VE-cadherin (clone 7519) 4 hours prior to the injection of fluorescently labeled 10 kD dextran (Figure 2A). It has previously been shown that VE-cadherin-blocking antibodies increased vascular permeability of the heart, lung and lymph nodes in vivo,19,26 but less is known about their effect on the integrity of the BM vasculature. We found that loss of VE-cadherin function resulted in increased vascular permeability in both sinusoids and arterioles in the BM (Figure 2B-C) compared to non-treated conditions. Blocking VE-cadherin resulted in increased permeability in the heart as well (Online Supplementary Figure S1A). Interestingly, the effects of VEcadherin blockage in the BM on the permeability of 500 kDa dextran did not result in a significant increase (Figure 2B-C and Online Supplementary Figure S1A), indicating that the blockage does not massively disrupt endothelial junctions. When locking VE-cadherin-based junctions, using the VE-cadherin-alpha-catenin chimera knock-in animals, we found that under homeostatic conditions, the BM vascular permeability, measured with 10kDa dextrans, was significantly reduced compared to WT controls (Figure 2D-E). These data indicate that VE-cadherin regulates vascular permeability during homeostasis in a paracellular fashion. From a clinical perspective, total body irradiation is commonly applied before hematopoietic stem cell transplantation (HSCT) to enable homing of sufficient numbers of HSPC.27 We measured vascular permeability of BM sinusoids of low dose irradiated mice with 10 kDa dextrans (Figure 3A) and found that vascular permeability in the BM is significantly increased after irradiation (Figure 3B-C), in line with previous studies.28-30 To substantiate the role of VE-cadherin in regulating the integrity of the BM vasculature, we examined permeability in VE-cadherin-Îącatenin chimera mice upon low dose irradiation. The effect of stabilizing endothelial junctions on BM vascular permeability was still apparent after low dose irradiation of VE-Îą-catenin fusion mice, as BM vascular permeability was significantly lower compared to irradiated controls, in sinusoids as well as in arterioles (Figure 3D-E). Taken together, these data show that VE-cadherin is important in regulating vascular permeability in the BM. Using the VEcadherin-GFP transgenic knock-in mouse model, we did not observe clear changes in VE-cadherin distribution in the BM after low-dose irradiation, suggesting that the physical structure of the BM vasculature is not affected by low dose irradiation (Figure 3F). Also VE-cadherin expression was unaltered, as determined by Western blotting (data not shown). VE-cadherin function is known to be haematologica | 2020; 105(12)

regulated by phosphorylation.31,32 Consequently, VE-cadherin is internalized. To study in more detail if irradiation alters VE-cadherin internalization or phosphorylation levels that undermine the endothelial cell-cell junction integrity, we used in vitro human umbilical vein endothelial cells (HUVEC) and found that low-dose irradiation promoted internalization of VE-cadherin (Online Supplementary Figure S1B-D). However, no clear increase in tyrosine phosphorylation of VE-cadherin or a loss of the interaction of p120-catenin with VE-cadherin was observed, although both events are known to be involved in VE-cadherin internalization31,33 (Online Supplementary Figure S1E-F). Although serine phosphorylation may also be involved in VE-cadherin internalization,32 we did not study this. From these data, we concluded that low-dose irradiation does not affect the overall physical structure of endothelial junctions in the BM vasculature, but rather induces a change in the regulation of vascular permeability, possibly by increasing the internalization of VE-cadherin.

Loss of VE-cadherin function increases HSPC homing to the BM We next examined whether VE-cadherin also regulates homing of HSPC to the BM. C-kit+ HSPC were adoptively transferred into low dose-irradiated mice in the presence or absence of a blocking VE-cadherin antibody (Figure 4A). After 16 hours, mice were sacrificed and the presence of donor HSPC in the BM, lung, spleen, and liver was determined (Figure 4B). We found that in the mice in which VE-cadherin function was blocked, homing of lineage-Sca-1+c-kit+ (LSK) cells to the BM was significantly increased 2-fold, compared to controls (Figure 4C). There was also a tendency towards increased homing of donor HSPC to spleen in anti-VE-cadherin-injected mice albeit not significant (Figure 4C). Homing of HSPC to the lung and liver was not affected by anti-VE-cadherin antibodies. The reason for this may be that percentages of donor HSPC in other organs than the BM were >40 fold lower than in the BM (data not shown), implying that the presence of adoptively transferred HSC in these organs was not due to directed migration. This is not surprising considering that CXCL12, the most important chemokine for HSPC, is expressed in the BM at much higher levels than in any other organ.34 To study if increased homing also resulted in increased engraftment, we measured the blood content after 1, 2 and 3 months of blocking of VE-cadherin function, as described above. The results showed a significant increase in myeloid cell numbers in the blood after 3 months (Online Supplementary Figure S2A). These data were supported by the BM cell phenotyping after 3 months (Online Supplementary Figure S2B). Also, here we found the myeloid cell fraction to be significantly increased, whereas the other leukocyte types, including the LSK cells were not (Online Supplementary Figure S2B). Shimoto and colleagues recently reported that under nonirradiation conditions, large numbers of HSPC need to be transplanted to reach normal homing of stem cells to the BM.35 Additionally, homing is improved without total body irradiation preconditioning.36 To test if opening the endothelial junctions by blocking VE-cadherin may promote this, we performed this experiment. However, the results showed that also under non-irradiated conditions, no increase in engraftment upon treatment with the VEcadherin antibody was observed (Online Supplementary 2749

T. Rademakers et al.

Figure S2C). Together, these data show that blocking VEcadherin promotes initial homing of HSPC but does not have a long-term effect on engraftment. As we observed a clear correlation between increased BM vascular permeability and increased homing of HSPC to the BM upon loss of VE-cadherin function, we expected

decreased homing of HSPC to the BM of VE-cadherin-αcatenin fusion mice that have decreased BM vascular permeability. Surprisingly, we did not observe decreased homing of HSPC to the BM of VE-cadherin-α-catenin fusion mice compared to WT controls (Figure 4D), indicating that permeability and transmigration in the BM are

Figure 2 Anti-VE-cadherin antibodies increase bone marrow vascular permeability in homeostatic conditions and after irradiation. (A) Schematic summary of experimental set up. (B) Bone marrow (BM) imaging after injection of fluorescent dyes with in green GS-I for vessel labeling and in red 10 and 500 kDa dextran as indicated. M: megakaryocytes. (C) Quantification of vascular permeability in arterioles and sinusoids in metaphysis and diaphysis after blocking VE-cadherin (10 kD: n=7-8 per group, 500 kD: n=4 per group). (D) BM imaging of wild-type and VE-cadherin-α-catenin chimera mice after injection of fluorescent dyes with in green GSI for vessel labeling and in red 10 kDa dextran (n=3-4 per group). (E) Quantification of vascular permeability in arterioles and sinusoids in wild-type and VE-cadherinα-catenin chimera mice as indicated. Scale bars: 25 mm, ctrl: control.

2750

haematologica | 2020; 105(12)

HSPC migrate transcellularly

regulated separately, in line with our previous finding where we showed that the endothelium actively limits vascular permeability during neutrophil extravasation.37 There was also no significant difference in HSPC homing to the spleen and liver in VE-α-catenin fusion mice, although a trend for more homing to the lungs was detected (Figure 4D). To verify that the donor HSPC which we identified in the BM of VE-α-catenin fusion mice by fluorescence activated cell sorting (FACS) analysis were not trapped in the vascular bed of the BM, cryosections of the BM of transplanted WT and VE-cadherin-α-catenin fusion mice were analyzed (Online Supplementary Figure S2D). Here, we found that the majority of adoptively transferred

HSPC were located inside the BM parenchyma (Online Supplementary Figure S2E). Moreover, the antibody effect was gone after 16 hours of treatment, based on an increase in BM vascular permeability (Online Supplementary Figure S2D,F). These data showed that stabilizing endothelial junctions does not impair the homing of HSPC to the BM.

HSPC predominantly use the transcellular route of migration over BM endothelium and form podosome-like structures Our finding that homing of HPSC in VE-cadherin-αcatenin fusion mice is unaltered suggested that these cells crossed the BM endothelium in a transcellular manner,

D E

Figure 3. Reduced bone marrow vascular permeability in VE-cadherinα-catenin fusion mice. (A) Schematic summary of experimental set up. (B) Bone marrow (BM) imaging of wild-type, VE-cadherin-α-catenin chimera mice and supplemented with anti-VE-cadherin antibody after injection of fluorescent dyes with in green GS-I for vessel labeling and in red 10 kDa dextran (n=3-4 per group). (C) Quantification of vascular permeability in sinusoids in wild-type, VE-cadherin-α-catenin chimera mice and supplemented with anti-VE-cad antibody as indicated. (D) BM imaging of wild-type and VE-cadherin-α-catenin chimera mice after 5 Gy irradiation after injection of fluorescent dyes with in green GS-I for vessel labeling and in red 10 kDa dextran (n=3-4 per group). (E) Quantification of vascular permeability in arterioles and sinusoids in wild-type and VE-cadherin-α-catenin chimera mice upon irradiation as indicated. (F) BM vasculature using a VE-cadherin-GFP knock-in mice showed proper VE-cadherin lining in control and upon irradiation conditions. Images on right are zoom from dotted box. Scale bars: 25 mm; ctrl: control.

haematologica | 2020; 105(12)

2751

T. Rademakers et al.

rather than through the cell-cell junctions. As it is technically challenging to discriminate between both routes using in vivo imaging of the BM vasculature, we employed an in vitro approach, where we used human CD34+ CB HSPC and immortalized human BM endothelial cells (HBMEC).21 To properly discriminate between the para- or transcellular route, HBMEC were transfected with LifeactGFP to visualize F-actin, activated for 4 hours with IL-1b21 and 30 minutes prior to adding the CD34+ cells, HBMEC were incubated with fluorescently-labeled, non-blocking VE-cadherin antibodies to visualize endothelial junctions (Figure 5A).22 Interestingly, we found that the vast majority (~75%) of HSPC crossed the HBMEC transcellularly (Figure 5B). Pretreatment of the endothelial cells with a

blocking VE-cadherin antibody promoted paracellular migration, indicating that these cells can cross through the junctions (Figure 5B). Although it is not fully understood how the transcellular migration route is regulated, it is believed that membrane fusion events are required to form a so-called transcellular pore.38 Therefore, we blocked membrane fusion activity in endothelial cells using the SNARE-complex inhibitor N-Ethylmaleimide (NEM).38 HSPC migration across a confluent HBMEC layer that was treated with NEM showed a significant inhibition of the number of HSPC transmigrating towards CXCL12 (Figure 5C). These data indicated that the formation of transcellular migration pores is the major route for HSPC to cross HBMEC. It should be

Figure 4. Anti-VE-cadherin antibodies increase hematopoietic stem and progenitor cells homing to the bone marrow. (A) Schematic summary of experimental set up for two groups: anti-VE-cadherin antibody administration and VE-cadherin/alpha-catenin chimera (VE-Îą-cat). Experiment also includes control group. (B) Gating strategy to select for LSK cells (murine hematopoietic stem and progenitor cells [HSPC]) using flow cytometry. (C) Normalized homing efficiency of HSPC to different organs after 16 hours (hrs) of transplantation in control and anti-VE-cadherin antibody group (n=4-5 per group, representative of two independent experiments). (D) Normalized homing efficiency of HSPC to different organs after 16 hrs of transplantation in wild-type and VE-Îą-cat groups (n=4-5 per group).

2752

haematologica | 2020; 105(12)

HSPC migrate transcellularly

haematologica | 2020; 105(12)

Figure 5. Hematopoietic stem and progenitor cells predominantly use the transcellular route of migration across bone marrow endothelium by forming podosome-like structures. (A) IL1b-pretreated immortalized human BM endothelial cells (HBMEC) were cultured and transfected with GFP-LifeAct (green) to monitor F-actin dynamics during hematopoietic stem and progenitor cells (HSPC) interactions. Both migration routes were detected for HSPC migration across HBMEC: paracellular with junctions marked by VE-cadherin (magenta), and transcellular, nuclei stained in blue and DIC shows overview. White dotted lines indicate localization of human HSPCs with #1 representing paracellular migration and #2 representing transcellular migration. Lower panel shows XZ projection with position HSPC 1 for paracellular migration (illustrated by the two magenta stains at the side) and position HSPC 2 for transcellular migration. Scale bar: 20 mm. (B) Quantification of transmigration route of HSPC. Open bar represents percentage of cells that cross transcellularly, and closed bars represent percentage that crosses paracellularly (mean ± standard error of the mean [SD], n=3 per group). Anti-VE-cadherin indicates that VE-cadherin function is blocked and promotes paracellular pathway. (C) Blocking of membrane fusion events in HBMEC by N-Ethylmaleimide (NEM) impairs HSPC transendothelial migration towards CXCL12 in an in vitro Transwell system (Mean ± SD, n=3 per group). (D) Migration tracks of cells under flow conditions show highly motile neutrophils and passive HSPC, quantified by distance traveled on the apical surface of HBMEC in µm (E) or average velocity (µm/s) (F). These experiments were independently carried out at least three times. (G) HSPC were treated for 30 minutes with PMA showing the induction of podosomes, based on typical podosome markers: vinculin (green) in the circle and HS-1 (magenta) and F-actin (red) in the core. Scale bar: 3 mm. (H) Orthogonal projection shows the distribution of the podosomes at the basolateral surface (open arrowheads) with F-actin in red, vinculin in green and the nucleus in blue.

2753

T. Rademakers et al.

noted that NEM treatment did also reduce the endothelial cell monolayer barrier. The fact that HPSC prefer the transcellular route is a surprising finding, considering that most immune cells use the paracellular route for transmigration.39 Although the exact mechanism of diapedesis is still under debate, transand paracellular migration are most likely regulated by different molecular mechanisms. On the endothelial side, VE-cadherin plays a key role in paracellular migration but not in transcellular migration.40 From the leukocyte, Mac1/CD11b is thought to be essential for crawling towards the nearest endothelial junction, thus facilitating paracellular migration.41 Interestingly, only a small subset of HSPC expresses Mac-1/CD11b,42 suggesting that the majority of HSPC is incapable to efficiently migrate to the nearest endothelial junctions. To test this hypothesis, we used in vitro physiological flow conditions and live-cell real-time imaging22 to quantify HSPC crawling over HBMEC monolayers. HSPC did not crawl on HBMEC monolayers, whereas neutrophils did (Figure 5C). Quantification of the migration tracks showed that both the distance and velocity of HSPC was drastically reduced compared to neutrophils (Figure 5D-E). Off note, neutrophils crawl equally well on HUVEC as on HBMEC monolayers (Online Supplementary Figure S3A-C). Thus, our data suggest that HPSC are not equipped to crawl on the endothelium to locate the nearest endothelial junction, but rather start their transmigration event immediately upon adhesion, potentially in a transcellular manner. As of yet, nothing is known on how HSPC may initiate transcellular migration.43 For transcellular migration of T-lymphocytes, it has been suggested that these cells make invasive protrusions, or so-called ‘podosomes’ to palpate the surface of the endothelium and initiate transcellular pore formation.38 Other cellular projections such as uropod and filopodia have been observed in HPSC,44 but the formation of podosomes by HSPC has not yet been documented. Podosomes consist of a protrusive actin core and an adhesive ring of integrins and actinadaptor proteins, such as vinculin and cortactin, or its hematopoietic homologue, hematopoietic lineage cellspecific protein 1 (HS-1).45,46 To determine whether HSPC can form podosomes, we allowed HSPC to adhere to fibronectin-coated coverslips and exposed them to phorbol 12-myristate 13-acetate (PMA) for 30 minutes. As a positive control we induced podosome formation in cultured dendritic cells, where we indeed observed clusters of podosomes with a ring of vinculin and a core containing actin and HS-1, as previously described46 (Online Supplementary Figure S3D-E). Remarkably, we could identify similar, podosome-like structures in human HPSC (Figure 5G-H), indicating that HSPC are equipped with the machinery to induce podosomes that may allow them to migrate through endothelial cells in a transcellular manner. Additional experiments were added to show that murine LSK cells can also induce podosomes as well (Online Supplementary Figure S3F).

Discussion We show that VE-cadherin regulates the vascular integrity of the BM vasculature and inhibition of VE-cadherin function can promotes the homing of HPSC, although this has no consequence for the engraftment 2754

potential. Moreover, we show that human as well as murine HSPC cross the BM endothelium in a transcellular manner under homeostatic conditions, and can initiate podosome-like structures that may assist these cells on their way through. Our in vivo data suggest that murine HSPC cross the vasculature in a transcellular manner. Previous observations using electron microscopy already indicated that leukocytes crossed the vasculature of the BM in a transcellular manner.12,13 The fact that HSPC do not crawl on endothelial monolayers, presumably because they predominantly express LFA-1, known to regulate firm adhesion, and not so much Mac-1, required for directional migration,47 also points towards the effective use of the transcellular pathway. It remains unclear why HSPC prefer the trans- over the paracellular route. Interestingly, our data furthermore show that the route to cross the endothelium is not exclusively regulated. When opening cell-cell junctions, by blocking VE-cadherin interactions, and thereby increasing vascular permeability, more HSPC cross the endothelium. These data show that junctional resistance also plays a role in the molecular decision for HSPC transmigration. Putatively, transcellular migration of HSPC continues upon administration of a VE-cadherin antibody, and the observed increase in HSPC homing to the BM can be explained by an increase of paracellular migration of HSPC. The fact that HSPC predominantly use the transcellular route to cross the endothelium may also explain that HSPC homing to the BM is not impaired in VE-cadherin-α-catenin fusion mice. It was previously shown that, in contrast to our data on HSPC homing, neutrophil migration to inflamed cremaster, lung, and skin is significantly impaired in VE-cadherin-α-catenin fusion mice.7 However, this is not surprising considering that neutrophils almost exclusively use the paracellular route of transmigration.37,39,41,48 Interestingly, also in these studies, it was shown that lymphocyte homing was not affected in the VE-cadherin-α-catenin fusion mice.7 Although we have shown that anti-VE-cadherin antibodies increase the homing of HSPC to the BM, we did not measure an increase in the engraftment. Together with the notion that HSPC cross the endothelium in a transcellular fashion, our data indicate that VE-cadherin is not the preferred target to increase HSPC transplantations after irradiation. Rather, a target for transcellular migration would be more beneficial. Unfortunately, until today, it is not understood how leukocytes migrate through the endothelial cell body. One potential mechanism may be the induction of podosomes. Podosomes have been described for DC and are believed to degrade the extracellular matrix.49 For T lymphocytes, it is recognized that podosomes may help these cells to cross in a transcellular manner.5,50 The fact that T cells and HSPC show the same transmigration kinetics, i.e., they firmly adhere to the initial point of attachment and stabilize, indicate that they may use the same molecular mechanisms to cross the endothelium. Indeed, blocking endothelial vesicle transport using NEM in both cases reduces the number of cells that transmigrates. Thus, we put forward that HSPC, like T lymphocytes, use podosomes to cross the vascular endothelium in a transcellular manner. Taken together, we report that although HSPC predominantly use the transcellular route for transendothelial migration in homeostasis, blocking VE-cadherin homotypic interactions favors paracellular migration of HSPC and increases homing of HSPC into the BM. This does not haematologica | 2020; 105(12)

HSPC migrate transcellularly

result in increased long-term engraftment. Our work offers valuable insight into the mechanisms of HSPC migration across the BM endothelium. Acknowledgments The authors thank Simon Tol, Erik Mul, and Aafke de Ligt for technical assistance and the staff of the animal facilities of the NKI and the Max Planck Institute for Molecular Biomedicine for excellent animal care. The authors thank Alessandra Cambi and Koen van den Dries for sharing their expertise on podosomes. AGP received a pre-doctoral scholarship from the Mexican

References 1. Dercksen MW, Rodenhuis S, Dirkson MK, et al. Subsets of CD34+ cells and rapid hematopoietic recovery after peripheralblood stem-cell transplantation. J Clin Oncol. 1995;13(8):1922-1932. 2. Voermans C, van Hennik PB, van der Schoot CE. Homing of human hematopoietic stem and progenitor cells: new insights, new challenges? J Hematother Stem Cell Res. 2001;10(6):725-738. 3. Lapidot T, Dar A, Kollet O. How do stem cells find their way home? Blood. 2005;106(6):1901-1910. 4. Carman CV. Mechanisms for transcellular diapedesis: probing and pathfinding by `invadosome-like protrusions'. J Cell Sci. 2009;122(17):3025-3035. 5. Carman CV, Sage PT, Sciuto TE, de la Fuente MA, et al. Transcellular diapedesis is initiated by invasive podosomes. Immunity. 2007; 26(6):784-797. 6. Martinelli R, Zeiger AS, Whitfield M, et al. Probing the biomechanical contribution of the endothelium to lymphocyte migration: diapedesis by the path of least resistance. J Cell Sci. 2014;127(Pt 17):3720-3734. 7. Schulte D, Kuppers V, Dartsch N, et al. Stabilizing the VE-cadherin-catenin complex blocks leukocyte extravasation and vascular permeability. EMBO J. 2011;30(20):41574170. 8. Yang L, Froio RM, Sciuto TE, Dvorak AM, Alon R, Luscinskas FW. ICAM-1 regulates neutrophil adhesion and transcellular migration of TNF-alpha-activated vascular endothelium under flow. Blood. 2005; 106(2):584-592. 9. Abadier M, Haghayegh JN, Cardoso AL, et al. Cell surface levels of endothelial ICAM1 influence the transcellular or paracellular T-cell diapedesis across the blood-brain barrier. Eur J Immunol. 2015;45(4):10431058. 10. Millan J, Hewlett L, Glyn M, Toomre D, Clark P, Ridley AJ. Lymphocyte transcellular migration occurs through recruitment of endothelial ICAM-1 to caveola- and F-actinrich domains. Nat Cell Biol. 2006;8(2):113123. 11. Vestweber D, Broermann A, Schulte D. Control of endothelial barrier function by regulating vascular endothelial-cadherin. Curr Opin Hematol. 2010;17(3):230-236. 12. De Bruyn PP, Michelson S, Thomas TB. The migration of blood cells of the bone marrow through the sinusoidal wall. J Morphol. 1971;133(4):417-437. 13. Muto M. A scanning and transmission electron microscopic study on rat bone marrow sinuses and transmural migration of blood cells. Arch Histol Jpn. 1976;39(1):51-66. 14. Sarin H. Physiologic upper limits of pore

haematologica | 2020; 105(12)

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

National Council for Science and Technology (Conacyt, 369767); and a travel fellowship grant from the Journal of Cell Science, The Company of Biologists Limited (www. biologists.com) to finance a research stay in the laboratory of JDvB. Work in the laboratory of MS is funded by Conacyt (grant 233395). Funding This work was supported by Sanquin Research (PPOC13030P grant). The authors declare no competing financial interests.

size of different blood capillary types and another perspective on the dual pore theory of microvascular permeability. J Angiogenes Res. 2010;2:14. Winderlich M, Keller L, Cagna G, et al. VEPTP controls blood vessel development by balancing Tie-2 activity. J Cell Biol. 2009;185(4):657-671. Dartsch N, Schulte D, Hagerling R, Kiefer F, Vestweber D. Fusing VE-cadherin to alphacatenin impairs fetal liver hematopoiesis and lymph but not blood vessel formation. Mol Cell Biol. 2014;34(9):1634-1648. Winderlich M, Keller L, Cagna G, et al. VEPTP controls blood vessel development by balancing Tie-2 activity. J Cell Biol. 2009; 185(4):657-671. Halai K, Whiteford J, Ma B, Nourshargh S, Woodfin A. ICAM-2 facilitates luminal interactions between neutrophils and endothelial cells in vivo. J Cell Sci. 2014;127(Pt 3):620-629. Gotsch U, Borges E, Bosse R, et al. VE-cadherin antibody accelerates neutrophil recruitment in vivo. J Cell Sci. 1997;110( Pt 5):583-588. Klamer SE, Kuijk CG, Hordijk PL, et al. BIGH3 modulates adhesion and migration of hematopoietic stem and progenitor cells. Cell Adh Migr. 2013;7(5):434-449. Rood PM, Calafat J, von dem Borne AE, Gerritsen WR, van der Schoot CE. Immortalisation of human bone marrow endothelial cells: characterisation of new cell lines. Eur J Clin Invest. 2000;30(7):618629. Kroon J, Daniel AE, Hoogenboezem M, van Buul JD. Real-time imaging of endothelial cell-cell junctions during neutrophil transmigration under physiological flow. J Vis Exp. 2014;(90):e51766. de Vries IJ, Eggert AA, Scharenborg NM, et al. Phenotypical and functional characterization of clinical grade dendritic cells. J Immunother. 2002;25(5):429-438. Timmerman I, Heemskerk N, Kroon J, et al. A local VE-cadherin and Trio-based signaling complex stabilizes endothelial junctions through Rac1. J Cell Sci. 2015;128(18):3514. Giannotta M, Trani M, Dejana E. VE-cadherin and endothelial adherens junctions: active guardians of vascular integrity. Dev Cell. 2013;26(5):441-454. Corada M, Mariotti M, Thurston G, et al. Vascular endothelial-cadherin is an important determinant of microvascular integrity in vivo. Proc Natl Acad Sci U S A. 1999; 96(17):9815-9820. Storb R, Sandmaier BM. Nonmyeloablative allogeneic hematopoietic cell transplantation. Haematologica. 2016;101(5):521-530. Daldrup-Link HE, Link TM, Rummeny EJ, et al. Assessing permeability alterations of the blood-bone marrow barrier due to total

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40. 41.

body irradiation: in vivo quantification with contrast enhanced magnetic resonance imaging. Bone Marrow Transplant. 2000; 25(1):71-78. Shirota T, Tavassoli M. Alterations of bone marrow sinus endothelium induced by ionizing irradiation: implications in the homing of intravenously transplanted marrow cells. Blood Cells. 1992;18(2):197-214. Zhou BO, Ding L, Morrison SJ. Hematopoietic stem and progenitor cells regulate the regeneration of their niche by secreting Angiopoietin-1. Elife. 2015; 4:e05521. Wessel F, Winderlich M, Holm M, et al. Leukocyte extravasation and vascular permeability are each controlled in vivo by different tyrosine residues of VE-cadherin. Nat Immunol. 2014;15(3):223-230. Gavard J, Gutkind JS. VEGF controls endothelial-cell permeability by promoting the beta-arrestin-dependent endocytosis of VE-cadherin. Nat Cell Biol. 2006;8(11):12231234. Xiao K, Allison DF, Buckley KM, et al. Cellular levels of p120 catenin function as a set point for cadherin expression levels in microvascular endothelial cells. J Cell Biol. 2003;163(3):535-545. Uhlen M, Fagerberg L, Hallstrom BM, et al. Proteomics. Tissue-based map of the human proteome. Science. 2015; 347(6220): 1260419. Shimoto M, Sugiyama T, Nagasawa T. Numerous niches for hematopoietic stem cells remain empty during homeostasis. Blood. 2017;129(15):2124-2131. Quesenberry PJ, Colvin G, Abedi M. Perspective: fundamental and clinical concepts on stem cell homing and engraftment: a journey to niches and beyond. Exp Hematol. 2005;33(1):9-19. Heemskerk N, Schimmel L, Oort C, et al. Factin-rich contractile endothelial pores prevent vascular leakage during leukocyte diapedesis through local RhoA signalling. Nat Commun. 2016;7:10493. Carman CV, Sage PT, Sciuto TE, et al. Transcellular diapedesis is initiated by invasive podosomes. Immunity. 2007;26(6):784797. Woodfin A, Voisin MB, Beyrau M, et al. The junctional adhesion molecule JAM-C regulates polarized transendothelial migration of neutrophils in vivo. Nat Immunol. 2011;12(8):761-769. Vestweber D. How leukocytes cross the vascular endothelium. Nat Rev Immunol. 2015;15(11):692-704. Phillipson M, Heit B, Colarusso P, Liu L, Ballantyne CM, Kubes P. Intraluminal crawling of neutrophils to emigration sites: a molecularly distinct process from adhesion in the recruitment cascade. J Exp Med.

2755

T. Rademakers et al. 2006;203(12):2569-2575. 42. Dercksen MW, Gerritsen WR, Rodenhuis S, et al. Expression of adhesion molecules on CD34+ cells: CD34+ L-selectin+ cells predict a rapid platelet recovery after peripheral blood stem cell transplantation. Blood. 1995;85(11):3313-3319. 43. Magnon C, Frenette PS. Hematopoietic stem cell trafficking. StemBook [Internet]. Cambridge (MA): Harvard Stem Cell Institute; 2008. 44. Dykstra B, Ramunas J, Kent D, et al. Highresolution video monitoring of hematopoietic stem cells cultured in single-cell arrays identifies new features of self-renewal. Proc

2756

Natl Acad Sci U S A. 2006;103(21):81858190. 45. Dehring DA, Clarke F, Ricart BG, et al. Hematopoietic lineage cell-specific protein 1 functions in concert with the WiskottAldrich syndrome protein to promote podosome array organization and chemotaxis in dendritic cells. J Immunol. 2011;186(8):4805-4818. 46. van den Dries K, Schwartz SL, Byars J, et al. Dual-color superresolution microscopy reveals nanoscale organization of mechanosensory podosomes. Mol Biol Cell. 2013;24(13):2112-2123. 47. Heit B, Colarusso P, Kubes P. Fundamentally

different roles for LFA-1, Mac-1 and alpha4integrin in neutrophil chemotaxis. J Cell Sci. 2005;118(Pt 22):5205-5220. 48. Feng D, Nagy JA, Pyne K, Dvorak HF, Dvorak AM. Neutrophils emigrate from venules by a transendothelial cell pathway in response to FMLP. J Exp Med. 1998; 187(6):903-915. 49. van den Dries K, Bolomini-Vittori M, Cambi A. Spatiotemporal organization and mechanosensory function of podosomes. Cell Adh Migr. 2014;8(3):268-272. 50. Carman CV, Springer TA. Transcellular migration: cell-cell contacts get intimate. Curr Opin Cell Biol. 2008;20(5):533-540.

haematologica | 2020; 105(12)

ARTICLE

Immunodeficiency

Somatic mutations and T-cell clonality in patients with immunodeficiency

Ferrata Storti Foundation

Paula Savola,1,2 Timi Martelius,3 Matti Kankainen,1,2,4,5 Jani Huuhtanen,1,2 Sofie Lundgren,1,2 Yrjö Koski,1,2 Samuli Eldfors,4 Tiina Kelkka,1,2 Mikko A.I. Keränen,1,2 Pekka Ellonen,4 Panu E. Kovanen,6 Soili Kytölä,7 Janna Saarela,4 Harri Lähdesmäki,8 Mikko R.J. Seppänen2,3,9 and Satu Mustjoki1,2,10

Hematology Research Unit Helsinki, University of Helsinki and Department of Hematology, HUS Helsinki University Hospital Comprehensive Cancer Center, Helsinki; 2 Translational Immunology Research Program, University of Helsinki, Helsinki; 3 Adult Immunodeficiency Unit, Infectious Diseases, Inflammation Center, University of Helsinki, HUS Helsinki University Hospital, Helsinki; 4Institute for Molecular Medicine Finland (FIMM), HILIFE, University of Helsinki, Helsinki; 5Medical and Clinical Genetics, University of Helsinki and HUS Helsinki University Hospital, Helsinki; 6Department of Pathology, University of Helsinki and HUSLAB, HUS Helsinki University Hospital, Helsinki; 7 Laboratory of Genetics, HUSLAB, HUS Helsinki University Hospital, Helsinki; 8Department of Computer Science, Aalto University School of Science, Espoo; 9Rare Diseases Center and Pediatric Research Center, Children and Adolescents, University of Helsinki and HUS Helsinki University Hospital, Helsinki and 10Department of Clinical Chemistry and Hematology, University of Helsinki, Helsinki, Finland 1

Haematologica 2020 Volume 105(12):2757-2768

ABSTRACT

ommon variable immunodeficiency (CVID) and other late-onset immunodeficiencies often co-manifest with autoimmunity and lymphoproliferation. The pathogenesis of most cases is elusive, as only a minor subset harbors known monogenic germline causes. The involvement of both B and T cells is, however, implicated. To study whether somatic mutations in CD4+ and CD8+ T cells associate with immunodeficiency, we recruited 17 patients and 21 healthy controls. Eight patients had late-onset CVID and nine patients other immunodeficiency and/or severe autoimmunity. In total, autoimmunity occurred in 94% and lymphoproliferation in 65%. We performed deep sequencing of 2,533 immune-associated genes from CD4+ and CD8+ cells. Deep T-cell receptor b-sequencing was used to characterize CD4+ and CD8+ T-cell receptor repertoires. The prevalence of somatic mutations was 65% in all immunodeficiency patients, 75% in CVID, and 48% in controls. Clonal hematopoiesis-associated variants in both CD4+and CD8+ cells occurred in 24% of immunodeficiency patients. Results demonstrated mutations in known tumor suppressors, oncogenes, and genes that are critical for immune- and proliferative functions, such as STAT5B (2 patients), C5AR1 (2 patients), KRAS (one patient), and NOD2 (one patient). Additionally, as a marker of T-cell receptor repertoire perturbation, CVID patients harbored increased frequencies of clones with identical complementarity determining region 3 sequences despite unique nucleotide sequences when compared to controls. In conclusion, somatic mutations in genes implicated for autoimmunity and lymphoproliferation are common in CD4+ and CD8+ cells of patients with immunodeficiency. They may contribute to immune dysregulation in a subset of immunodeficiency patients.

Introduction Immunodeficiencies often manifest with autoimmune disease, as shown in the most common primary immunodeficiency in the adult population: common variable immunodeficiency (CVID).1 Although the main manifestations of CVID are recurrent infections and low levels of plasma immunoglobulins, patients often present with immune-mediated blood-cell cytopenias, enteropathy, arthritis, lymphoproliferation, and/or granulomatous disease.1 Currently known monogenic causes account for 2-10% of CVID cases.2 In the absence of disease-associated haematologica | 2020; 105(12)

Correspondence: SATU MUSTJOKI satu.mustjoki@helsinki.fi Received: March 14, 2019. Accepted: December 18, 2019. Pre-published: December 19, 2019. doi:10.3324/haematol.2019.220889 ©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.

2757

P. Savola et al.

germline variants, the etiology of CVID and other delayed-onset immunodeficiency remains largely elusive. Although a B-cell defect apparently causes hypogammaglobulinemia, T-cell abnormalities also occur in CVID patients. Decreased levels of regulatory T cells and naïve CD4+ T cells, an increase in CD8+ expansion, disturbed cytokine secretion, and large granular lymphocyte expansions have all been described.1,3,4 Altered immune homeostasis may promote both immunodeficiency and infections. In CVID, the prevalence of autoimmune disease is 30%,1,3,5 lymphoproliferation 50%,3 and patients also have an increased risk for malignant diseases. The standardized incidence ratio for lymphomas and stomach cancer is 10-12.1,6 Somatic mutations play a key role in malignant transformation, while recent discoveries have highlighted associations between somatic mutations and non-malignant disease.7,8 Somatic STAT3 mutations occur in diseases with autoimmune manifestations, such as large granular lymphocyte (LGL) leukemia and aplastic anemia.9-12 In LGL leukemia, patients present with immune-mediated cytopenias and rheumatoid arthritis, and LGL leukemia patients with multiple STAT3 mutations more often have RA (46%) than patients without mutations (6%).13 Somatic, activating STAT3 mutations have also been discovered in intraepithelial lymphocytes of refractory celiac disease patients14 and in patients with Felty syndrome (long-lasting RA and neutropenia).15 Recent reports by us and others have shown that somatic mutations in genes other than STAT3 occur in mature T cells.16,17 For example, patients with untreated RA have somatic mutations in mature CD8+ T cells in 20% of cases.16 In addition to mutations in mature T cells, hematopoietic progenitor cells may also harbor mutations: somatic mutations originating from hematopoietic progenitor cells occur in 10% of elderly individuals, and these mutations confer a risk for blood cancer and cardiovascular disease.1820 These stem-cell-derived mutations have effects beyond malignant transformation. For example, somatic loss-offunction Tet2 mutations in myeloid cells lead to proinflammatory cytokine production and increased atherosclerosis in mice.21 As another example, somatic mutations in the FAS gene in hematopoietic precursors can lead to autoimmune lymphoproliferative syndrome (ALPS).22 Thus, somatic events in blood cells have the potential to modulate immune-mediated diseases. To study whether patients with late-onset CVID and other immunodeficiency and/or severe autoimmunity have somatic mutations in CD4+ and CD8+ cells, and whether they associate with disease phenotype, we used a customized deep sequencing panel covering 2,533 genes. In addition, to characterize the CD4+and CD8+ T-cell repertoire in more detail, we investigated the T-cell clone size, clonality, and characteristics by deep T-cell receptor b-chain (TCRB) sequencing.

Methods Patients We recruited eight patients with late-onset CVID and nine patients with other types of immunodeficiency or severe autoimmunity from the Helsinki University Hospital infectious disease clinic (Table 1). CVID patients were included in a previously described CVID cohort.3 Five of the patient cases with other types of immune deficiency or severe autoimmunity have 2758

been previously described: patient 17 has a germline mutation in ADA2 (p.Arg169Gln/p.Arg169Gln) and somatic STAT3 D661V and N647I mutations in T cells;23 patients 9-11 have germline gain-of-function mutations in STAT3;24 and patient 12 has a compound immunodeficiency and arr11q24.2q25 (126,074,297134,927,114) x 1 (Jacobsen syndrome).25 Blood donor buffy coat samples from the Red Cross Blood Service (Online Supplementary Table S1) served as healthy controls. This study was approved by the local ethics committee and was conducted according to the principles of the Declaration of Helsinki. All patients signed informed consent.

Sample preparation CD4+ and CD8+ cells were purified via positive selection with magnetic beads (Miltenyi Biotech) from peripheral blood mononuclear cells. Sample material for patient 9 was bone marrow rather than peripheral blood.

Identification of somatic mutations To discover somatic variants, we designed a custom sequencing panel that comprised the coding areas of 2,533 genes. The gene selection was based on the InnateDB database (http://www.innatedb.ca/) and on other genes important for hematopoietic cells, adaptive immune responses, and autoimmunity (Online Supplementary Table S2). All samples (CD4+ and CD8+ cells of 17 immunodeficiency patients and 21 healthy controls) were sequenced with HiSeq2500 (Illumina). Somatic variants were identified with the GATK toolkit and MuTect2 using a panel of 21 healthy controls (Online Supplementary Methods). To discover variants occurring only in CD4+ or CD8+ cells, variant calling was performed as a pairedsample analysis, using CD4+ cells as a germline control and CD8+ as a 'tumor' sample, and vice versa. In addition, we performed variant calling without a germline control sample from the same patient to discover variants that occur in both CD4+ and CD8+ cells. Variants identified by single-sample calling method were required to occur in both CD4+ and CD8+ cells in a known hematopoietic tumor suppressor or oncogene.18-20 All variants were required to pass multiple quality filters (Online Supplementary Methods). Only mutations that altered the aminoacid coding sequence were reported, but non-coding mutations were included in mutation signature analyses. Signaling pathways were annotated using the Reactome public database and gene with Gene Ontology (GO) Consortium biological process terms (Online Supplementary Table S3).

T-cell receptor b-chain sequencing

T-cell receptor b-chain (TCRB) sequencing was performed from CD8+ and CD4+ cells’ genomic DNA with the human TCRB immunoSEQ assay (Adaptive Biotechnologies) according to the manufacturer’s instructions. The TCRB repertoires of seven CVID patients and 27 healthy controls were compared. Only productive T-cell receptor (TCR) sequences were analyzed. To search for public, previously reported TCR sequences associated with pathogens, autoimmunity, or cancer, we queried TCR against the manually curated McPAS database. “Convergent TCR” were defined as clones that share an amino-acid CDR3 sequence although they have distinct nucleotide CDR3 sequences. A T-cell clone was defined by a unique nucleotide sequence (Online Supplementary Figure S1).

Data sharing Original sequencing data are available from the corresponding author upon reasonable request due to ethical permit constraints. haematologica | 2020; 105(12)

Somatic mutations in immunodeficiency Table 1. Patient phenotypes. ID Germline Disease variant?

Age at sampling

Age at dg

Sex

CVID

TACI S68fsX11

CVID

4 5

No exome seq No

CVID CVID

70 70

59 68

M F

6 7

No exome seq No

CVID CVID

67 35

64 32

F F

CVID

STAT3 K658N

STAT3 GOF

STAT3 K392R

STAT3 GOF

STAT3 M394T

STAT3 GOF

Jacobsen sdr

13 14

No exome seq, arr11q24.2q25 (126,074,297134,927,114) x 1 No No

Other Other

73 70

67 NA

F F

Other

No exome seq

Good sdr

ADA2 def

17 ADA2 R169Q/R169Q

Phenotype

Details

Infections, autoimmunity, Atrophic gastritis, colitis, GLILD, bronchiectasis, enteropathy, lymphoproliferation liver cirrhosis, splenomegaly, lymphadenopathy Infections, autoimmunity, ITP, GLILD, splenomegaly, granulomatous disease, lymphadenopathy lymphoroliferation Infections, autoimmunity, GLILD, bronchiectasis, granulomatous disease, cholestasis due to hepatic lymphoproliferation LGL infiltration, chronic gastritis, sicca, psoriasis, portal hypertension, LGL lymphoproliferation, thrombocytopenia, splenomegaly, lymphadenopathy Infections, autoimmunity Hyperthyroidism, asthma, chronic gastritis Infections, autoimmunity, Atrophic gastritis, psoriasis, lymphoproliferation asthma, splenomegaly infections, autoimmunity Hypothyroidism, vitamin D malabsorption Infections, autoimmunity, Neutropenia, anterior uveitis, granulomatous disease, sicca, GLILD, lymphoproliferation lymphadenopathy Infections, autoimmunity, ITP, AIHA, hyperthyroidism, granulomatous disease, sialadenitis, GLILD, splenomegaly, lymphoproliferation lymphadenopathy Autoimmunity, enteropathy, Autoimmune enteropathy, lymphoproliferation, infections bronchiectasis, asthma, cryptogenic organizing pneumonia, dermatitis, lymphadenopathy, splenomegaly, AIHA, sicca, uveitis, sterile pleuritis, short stature, bone-marrow eosinophilia, hypogammaglobulinemia Autoimmunity, infections, Neonatal diabetes, celiac disease, lymphoproliferation rudimentary pancreas, T-cell LGL leukemia, desquamative interstitial pneumonitis, AIHA, bone-marrow eosinophilia, severe allergy, short stature, hypogammaglobulinemia Autoimmunity, ITP, lymphocytic colitis, lymphoproliferation, splenomegaly, lymphadenopathy, infections hypogammaglobulinemia Infections, ITP, neutropenia, hypothyroidism, autoimmunity, condylomas, angioedema, asthma, lymphoproliferation splenomegaly, lymphadenopathy, hypogammaglobulinemia Hypogammaglobulinemia, recurrent stomatitis Infections, autoimmunity Recurrent EBV meningoencephalitis, palindromic seronegative arthritis Autoimmnunity Periorbital adul-onset xantogranulomatosis, psoriatic arthritis, ulcerative colitis, asthma, alopecia, stroke Infections, autoimmunity LGL lymphoproliferation, bronchiectasis, sicca, seronegative rheumatoid arthritis/tenosynovitis, hypogammaglobulinemia Infections, autoimmunity, ITP, AIHA, neutropenia, lymphoproliferation LGL lymphoproliferation, hypogammaglobulinemia, pulmonary hypertension, splenomegaly

Phenotypes of the patients included in the study. Eight were late-onset common variable immunodeficiency (CVID) patients, other patients had another type of immunodeficiency or severe autoimmunity. The column â&#x20AC;&#x153;Germline variantâ&#x20AC;? shows the results of clinical exome sequencing: if a pathogenic variant was identified, it is shown. ID: identifier; dg: diagnosis; M: male; F: female; ITP: immune thrombocytopenic purpura; GLILD: granulomatous-lymphocytic interstitial lung disease; LGL: large granular lymphocyte; AIHA: autoimmune hemolytic anemia; EBV: Epstein-Barr virus; seq: sequencing; GOF: gain-of function; sdr: syndrome; def: deficiency; COSMIC: Catalogue of Somatic Mutations in Cancer.

haematologica | 2020; 105(12)

2759

P. Savola et al.

Results

Somatic mutations in CD4+ and CD8+ cells

Patients' characteristics All patients but one included in the study had autoimmune disease manifestations, and 11 of 17 (65%) had lymphoproliferation (Table 1). All CVID patients (patients 1-8) suffered from recurrent infections. Lymphoproliferation, as defined by lymphadenopathy and/or splenomegaly, occurred in 75% of CVID patients, granulomatous disease in 12.5%, and enteropathy in 12.5%. Because we included only late-onset CVID in the study, the median age at diagnosis was 59 years for CVID patients. According to the EUROClass classification system, 25% (2 of 8) of CVID patients did not have B cells (B-) (Online Supplementary Table S4). In the remaining B+ patients, switched memory-B cells (CD19+CD27+IgDnegIgMneg) were low in 50% (3 of 6) of cases. None had high frequencies of transitional B cells (CD19+CD38++IgM++), and 67% (4 of 6) had increased CD21low cells (CD19+CD21lowCD38low) according to the EUROClass system. Other patients were similarly classified, although they did not have CVID (Online Supplementary Table S4).

The custom gene panel covered the coding areas of 2,533 genes, spanning over approximately 5.2 million base-pairs. Immunodeficiency patients harbored 45 and healthy controls 28 somatic mutations that existed in either CD4+ or CD8+ cells, identified by paired-sample variant calling (Table 2, Figure 1 and Online Supplementary Table S5). In two cases, we had access to non-hematopoietic-tissue-derived DNA; the mutations did not exist in these samples (Online Supplementary Table S6). The prevalence of somatic mutations was 65% (11 of 17) in immunodeficiency patients, 75% (6 of 8) in CVID, and 48% (10 of 21) in healthy controls. The median ages of patients and controls were similar (Figure 2A and Online Supplementary Table S1). There was no statistically significant difference in the prevalence of mutations between patients and controls (Figure 2B) or in the number of identified mutations (Online Supplementary Figure S2). The median number of mapped bases was higher in immunodeficiency samples, but the number of mutations did not correlate with mapped bases in CD8+ cells, in which most mutations were discovered (Online Supplementary

Table 2. Selected somatic mutations identified in CD4+ and CD8+ cells.

Pt. ID Disease 1 2

CVID CVID

CVID

4 12

CVID Jacobsen sdr

13 15 16

Other Other Good sdr

17 HC1 HC3

ADA2 def Healthy Healthy

HC11

Healthy

HC16

Healthy

HC17

Healthy

HC20 HC21

Healthy Healthy

Cells +

CD8 CD4+ CD4+ CD4+ CD4+ CD4+ CD8+ CD8+ CD4+ CD4+ CD4+ CD8+ CD4+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD8+ CD4+

HGVS

AA change

COSMIC identifier

1:g.1804503C>G NM_001282539:exon6:c.G346C:p.G116R 4:g.105235713delC NM_001127208:exon3:c.1771delC:p.Q591fs 4:g.105243618G>T NM_001127208:exon6:c.G3643T:p.E1215X COSM3719016 16:g.50699619G>T NM_022162:exon2c.G205T:p.E69X 4:g.105269703C>T NM_001127208:exon9:c.C4138T:p.H1380Y COSM87161 17:g.42217380A>T NM_012448:exon10:c.T1254A:p.N418K 7:g.100677602C>G NM_005273:exon6:c.C372G:p.Y124X 11:g.75290010T>G NM_004041:exon2:c.A50C:p.K17T 9:g.5073770G>T NM_001322195:exon13:c.G1849T:p.V617F COSM12600 22:g.36929726C>T NM_000395:exon6:c.C637T:p.R213W 19:g.47319962C>T NM_001736:exon2:c.C185T:p.T62M COSM400357 X:g.21432736G>A NM_014927:exon3:c.G353A:p.R118Q 17:g.42210194G>C NM_012448:exon15:c.C1883G:p.T628S COSM6022929 22:g.25667739C>G NM_005160:exon6:c.C442G:p.P148A 1:g.56692436C>T NM_006252:exon4:c.C409T:p.H137Y 12:g.25227351G>A NM_004985:exon3:c.C173T:p.T58I COSM87288, COSM5490513 19:g.47320366C>T NM_001736:exon2:c.C589T:p.R197W COSM4667798 5:g.138466921G>A NM_001964:exon2:c.G472A:p.V158I 9:g.35707732T>C NM_006289:exon35:c.A4631G:p.K1544R 15:g.41056702G>A NM_017553:exon17:c.C1990T:p.R664C COSM5383245, COSM5383244 2:g.25240306A>G NM_175629:exon19:c.T2318C:p.L773P COSM1583115 9:g.130880097T>C NM_007313:exon9:c.T1510C:p.S504P 3:g.184365148T>G NM_001278698:exon3:c.T173G:p.L58W 4:g.38797118T>C NM_003263:exon4:c.A1714G:p.M572V 5:g.148827319T>G NM_000024:exon1:c.T488G:p.L163R 15:g.74036026C>G NM_033247:exon5:c.C1274G:p.A425G 16:g.31266064A>T NM_001145808:exon5:c.A344T:p.E115V 21:g.31251838C>G NM_003253:exon6:c.G1315C:p.A439P 21:g.31266560T>A NM_003253:exon5:c.A413T:p.D138V 11:g.129870009T>C NM_006165:exon22:c.A3091G:p.I1031V 1:g.114716123C>T NM_002524:exon2:c.G38A:p.G13D COSM573

Gene

VAF

SIFT Polyphen2 ExAC All

GNB1 TET2 TET2 NOD2 TET2 STAT5B GNB2 ARRB1 JAK2 CSF2RB C5AR1 CNKSR2 STAT5B GRK3 PRKAA2 KRAS

0.07 0.056 0.053 0.06 0.027 0.036 0.051 0.025 0.017 0.026 0.02 0.03 0.032 0.034 0.032 0.079

D NA NA NA D D NA D D D T T D P D D

D NA NA NA D D NA D D D D D D P D D

NA NA NA NA 0.000092 NA NA NA 0.0007 NA 0.000049 NA 0.000015 NA NA NA

C5AR1 EGR1 TLN1 INO80

0.13 0.023 0.021 0.068

T D T D

P P P D

0.000033 NA NA NA

DNMT3A ABL1 POLR2H TLR1 ADRB2 PML ITGAM TIAM1 TIAM1 NFRKB NRAS

0.043 0.059 0.032 0.021 0.057 0.025 0.05 0.029 0.047 0.054 0.018

D D D T D T D T D T D

D P D B D B P D P P B

NA NA NA NA NA NA NA NA NA NA 0.0000082

Selected mutations identified in healthy controls and immunodeficiency patients in CD4+ and CD8+ cells. The mutations were identified in paired-sample analyses in either CD4+ or CD8+ cells. Patient (Pt.) 17 had also somatic STAT3 mutations in T cells (D661V and N647I), but they did not pass all MuTect2 filters due to clustered events and are thus not listed in the table. CVID: common variable immunodeficiency; ID: identifier; AA: amino acid; VAF: variant allele frequency; SIFT: SIFT prediction; PolyPhen2: Polyphen2 HDIV prediction; T: tolerated; D: deleterious (SIFT)/probably damaging (PolyPhen2); B: benign; P: possibly damaging; ExAC all: the variant frequency in the general Exome Aggregation Consortium; sdr: syndrome; def: deficiency; COSMIC: Catalogue of Somatic Mutations in Cancer.

2760

haematologica | 2020; 105(12)

Somatic mutations in immunodeficiency

Table S7 and Online Supplementary Figure S3). The vertical and horizontal coverage of the assay is shown in Online Supplementary Figure S4. In immunodeficiency patients, the identified mutations comprised of 38 missense, 5 nonsense, and 2 frameshift mutations (Figure 2C). Immunodeficiency patients har-

bored a larger proportion of damaging mutations than healthy controls (56% vs. 21%, Fisherâ&#x20AC;&#x2122;s exact test P=0.0069) (Figure 2D). For CVID patients, the proportion of damaging mutations was even higher (70%; P=0.0007 when compared with healthy controls). Damaging mutations were defined as frameshift, nonsense and as muta-

Figure 1. Discovered somatic mutations in CD4+ and CD8+ cells in patients and controls. Selected somatic mutations in patients and healthy controls identified in paired-sample variant calling analyses are shown, with different colors representing different mutation types (missense, nonsense, frameshift, splicing; see color code). Mutations were grouped according to Reactome pathways, but one gene was grouped only to one pathway for visual clarity. For immunodeficiency patients, the patient phenotypes are also shown at the top of the figure. Patients 1-8 had common variable immunodeficiency (CVID). Oncogenes and tumor suppressors relevant for hematopoietic tissue were derived from previous publications.18-20 The mean RNA expression of the mutated genes in healthy individuals' CD4+ (n=3) and CD8+ (n=5) is shown as log2-transformed counts per million (cpm) values, and error bars represent standard deviation. B2M and ACTB are shown as highly-expressed housekeeping gene references for expression levels. HC: healthy control; ACTB: actin beta; B2M: beta-2-microglobulin; cpm: counts per million.

haematologica | 2020; 105(12)

2761

P. Savola et al. A

E D

tions with damaging predictions by both Polyphen-2 and SIFT. Mutations in known tumor suppressors or oncogenes18–20 occurred in six patients (35%) and two controls (9.5%) (Figure 1). Gene Ontology (GO) term annotations revealed that patients harbored variants in genes that associate with lymphocyte functions (5 patients, 29%), inflammation (7 patients, 41%), and proliferation (8 patients, 47%). Most genes that harbored mutations were expressed in healthy CD4+ and CD8+ cells (Figure 1), suggesting functional relevance. Mutational signature analysis revealed that the most dominant signature in immunodeficiency patients’ CD4+ and CD8+ cells was age-related signature 1. Mismatch-repair-associated signature 15 was more dominant in healthy controls’ CD4+ cells (Figure 2E). 2762

Figure 2. The prevalence and consequences of somatic mutations in immunodeficiency patients and healthy controls. (A) There was no statistically significant difference (Mann-Whitney test) in median age (years) between immunodeficiency patients and healthy controls. (B) The prevalence of somatic mutations in CD8+ cells, CD4+ cells, and in either CD4+ or CD8+ (overall) in patients and controls. (C) The percentages of different mutation types in immunodeficiency patients and healthy controls (see color code). (D) Immunodeficiency patients harbored a larger proportion of damaging mutations than healthy controls (Fisher exact test, P=0.0069). Damaging mutations were defined by being either a nonsense, frameshift, or missense mutations with damaging predictions by both PolyPhen2 and SIFT. (E) Mutational signatures in healthy controls’ and immunodeficiency patients’ CD4+ and CD8+ cells. Both non-coding and protein-altering mutations were included in the analyses. Signatures with significant weights (>0.06) are shown; others are classified as “Other”. (F) The number of mutations in paired analyses in CD4+ and CD8+ cells were compared between immunodeficiency patients who harbored clonal hematopoiesis variants in both CD4+ and CD8+ cells. The clonal hematopoiesis variants themselves were not included in the analyses, but their existence was used as a grouping factor. Patients with clonal hematopoiesis variants in both CD4+ and CD8+ cells (clonal hematopoiesis) harbored more somatic mutations in CD4+ cells than patients without clonal hematopoiesis variants (Mann-Whitney test, P=0.03439), but there was no statistically significant difference in CD8+ cells. CVID: common variable immunodeficiency.

Clonal hematopoiesis variants in CD4+ and CD8+ cells To complement the findings in paired-sample analyses by discovering variants that occur in both CD4+ and CD8+ cells, we performed single-sample variant calling analysis in genes associated with clonal hematopoiesis.1820 In the immunodeficiency cohort, 24% of patients (4 of 17) had in total seven clonal hematopoiesis-associated variants while none was observed in healthy controls (Table 3 and Figure 1). Patients who had clonal hematopoiesis variants harbored more mutations in CD4+ cells than patients without clonal hematopoiesis (Figure 2F). Deep sequencing from flow cytometry-sorted cells assessed the cellular lineage of these variants. Clonal hematopoiesis-associated variants were found in sorted, pure CD3+ T cells, and, in some cases, in CD19+ haematologica | 2020; 105(12)

Somatic mutations in immunodeficiency Table 3. Clonal hematopoiesis variants occurring in both CD4+ and CD8+ cells.

Pt. ID

AA change

Consequence

COSMIC identifier

Cell type +

DNMT3A:NM_022552:exon20:c.2388dupT:p.N797_L798delinsX

Nonsense

TET2:NM_001127208:exon3:c.1771delC:p.Q591fs

Frameshift

TET2:NM_001127208:exon6:c.G3643T:p.E1215X

Nonsense

COSM3719016

TET2:NM_001127208:exon11:c.C4889G:p.S1630X

Nonsense

COSM5945066

DNMT3A: NM_022552:exon14:c.G1591A:p.D531N

Missense

COSM1583077

DNMT3A: NM_022552:exon14:c.G1591A:p.D531N

Missense

COSM1583077

PRPF8:NM_006445:exon39:c.A6272G:p.Y2091C

Missense

CD4 CD8+ CD4+ CD8+ CD4+ CD8+ CD4+ CD8+ CD4+ CD8+ CD4+ CD8+ CD4+ CD8+

VAF 0.043 0.038 0.056 0.0065 0.053 0.027 0.041 0.032 0.013 0.024 0.022 0.027 0.019 0.023

Variants that occurred in clonal hematopoiesis-associated genes and in both CD4+ and CD8+ cells in single-sample variant calling are shown. Pt. ID: patient identifier;VAF: variant allele frequency; AA: amino acid; COSMIC: Catalogue of Somatic Mutations in Cancer.

B cells. There were no mutations in CD14+ monocytes (Online Supplementary Table S8).

Somatic mutations in key immunologic pathways Overall, mutations occurred in genes that are linked with cell proliferation and hematologic malignancies. Patient 2 harbored the highest number of mutations: seven in CD4+ cells and two in CD8+ cells. This patient had a STAT5B (N418K) mutation in CD4+ cells with 3.3% variant allele frequency (VAF), but also a NOD2 and four different TET2 mutations in CD4+ cells (Tables 2 and 3, and Figure 1). The patient had late-onset CVID with multiple co-morbidities: infections, idiopathic thrombocytopenic purpura (ITP), granulomatous-lymphocytic interstitial lung disease (GLILD), splenomegaly, lymphadenopathy, and granulomas (Table 1). Patient 13 also harbored a STAT5B T628S mutation with 3.2% variant allele frequency (VAF) in CD4+ cells (Table 2 and Figure 1). In addition, she had a truncating clonal hematopoiesis-associated DNMT3A variant in both CD4+ and CD8+ cells (Table 3). Her disease phenotype comprised of hypogammaglobulinemia and recurrent, non-infectious stomatitis. Patient 16 harbored seven mutations in CD8+ cells and one in CD4+ cells, occurring in genes involved in RAF/MAPK and GPCR signaling, TP53 regulation and inflammasome activity (KRAS T58I, C5AR1, PRKAA2 and MEFV) (Figure 1). In addition, this patient had two clonal hematopoiesis-associated variants in both CD4+ and CD8+ cells (Table 3). This patient's immunodeficiency presented as Good syndrome with LGL lymphoproliferation, bronchiectasis, sicca, seronegative RA/tenosynovitis, hypogammaglobulinemia, and infections (Table 1). Patient 12 harbored five mutations in CD4+ cells and three in CD8+ cells. Mutations included a C5AR1 mutation, and two mutations in RAF/MAPK signaling cascade genes (CNKSR2 and CSF2RB) (Table 2 and Figure 1). The patient had Jacobsen syndrome with infections, hypogammaglobulinemia, lymphoproliferation, and multiple autoimmune manifestations (Table 1). To validate a subset of mutation findings, we perhaematologica | 2020; 105(12)

formed Amplicon re-sequencing of interesting mutations (23 different positions in patients) and negative control positions (Online Supplementary Methods and Online Supplementary Table S9) with a multiplexed customized amplicon panel. All mutation positions covered by sufficient coverage (20 of 23, 87%) were successfully validated (20 of 20, 100%). As expected, mutations were only detected in the original patient and not in other individuals. All control positions yielded negative results. Mutations presented with similar VAF in re-sequencing as in the original sequencing assay (Online Supplementary Table S10).

CD4+ clonality correlates with CD8+ clonality and B-cell phenotypes There was no difference in clonality indices of either CD4+ or CD8+ cells between patients and controls (Figure 3A). Overall, CD8+ cells showed higher clonality than CD4+ cells (P<0.0001 in healthy controls; P=0.0135 in CVID) (Figure 3A). The size of the largest T-cell clone (as frequency of all productive rearrangements) and the proportion of productive sequences of all TCRB rearrangements were similar in patients and controls (Online Supplementary Figure S5). CD4+ and CD8+ clonality showed a positive non-linear correlation in both CVID and healthy controls (Figure 3B; Spearman correlation P=0.0019 and P=0.0234, respectively). An increased proportion of CD8+ cells was associated with higher clonality in CD4+ and CD8+ cells (Figure 3C; Spearman correlation P=0.0238 CD4+ clonality and 0.0341 CD8+ clonality). The largest T-cell clone size was greater in individuals with mutations than in individuals without mutations (Online Supplementary Figure S6). Due to the importance of helper T cells in B-cell maturation, we correlated CD4+ clonality indices with the frequencies of different B-cell subsets in CVID patients. Higher CD4+ clonality correlated with lower frequencies of memory and switched memory B cells (P=0.008 and P=0.0238) (Figure 3D and E). CD4+ clonality did not associate with the frequency of total B cells, marginal-zone, activated, or transitional B cells (Online Supplementary Figure S7). 2763

P. Savola et al. A P<0.0001

P=0.0135

CD45+CD19+CD27+ B-memory cells (% of B cells)

P=0.0008

Spearman P=0.0019 Spearman P=0.0234

CD4+

P=0.01268

CD45+CD19+CD27+IgDâ&#x20AC;&#x201C;IgMâ&#x20AC;&#x201C; Switched B-memory cells (% of B cells) P=0.0238

CD8+

P=0.02147

T-cell receptor b-chain gene segment usage and junctional diversity

T-cell receptor b-chain V- and J-gene usage often serves as a marker for structural similarity of TCR. In total productive CD4+ and CD8+ TCR repertoires, there was no difference in TCRB V-gene family and J-gene usage between CVID patients and controls (Online Supplementary Figure S8). When comparing junctional diversity, no differences were found between patients and controls in the sums of base deletions and non-templated base insertions in the V-D and D-J junctions (Online Supplementary Figure S9). All non-templated base edition comparisons between CVID patients and controls are shown in Online Supplementary Figure S10.

Common variable immunodeficiency patients harbor higher frequencies of convergent T-cell receptor clones than healthy controls Distinct T-cell clones have unique nucleotide-level CDR3 sequences, but different nucleotide sequences can translate into the same functional CDR3 amino-acid sequences (TCR convergence). CVID patients showed 2764

Spearman P=0.0238 Spearman P=0.0341

Figure 3. T-cell repertoire characteristics in common variable immunodeficiency (CVID) patients. (A) Clonality indices of CD4- and CD8-cell productive T-cell receptor (TCR) rearrangements. No statistically significant differences were seen between patients and healthy controls, but CD8+ cells were more clonal than CD4+ cells. Other: immunodeficiency patients other than CVID. Statistical testing comprised a Kruskall-Wallis test as an omnibus test and Dunn multiple comparison tests as post-hoc tests. Medians and interquartile ranges are shown. (B) CD4+ and CD8+ clonality indices correlate with each other, but not in a linear fashion. (C) Low CD4/CD8 ratios are associated with higher CD8 clonality. (D) Increased CD4+ clonality is associated with decreased frequency of memory B cells in CVID patients (Spearman correlation, P=0.0008). (E) Increased CD4+ clonality is associated with decreased frequency of switched memory B cells in CVID patients (P=0.0238). (F) CVID patients show a higher frequency of convergent TCR of all TCR amino-acid rearrangements (Mann-Whitney test, P=0.013 for CD4+ and P=0.021for CD8+ ). Vertical lines: median; box hinges: interquartile ranges; whiskers: reasonable extremes of the data; ns: not significant; HC: healthy control; CVID: common variable immunodeficiency; Other: immunodeficiency other than CVID. ****P<0.0001; *P<0.05.

more convergence in CD4+ and CD8+ cells than healthy controls (CD4+ P=0.0123 and CD8+ P=0.0215) (Figure 3F). This was not due to higher frequencies of previously reported pathogen-specific TCR, such as Epstein-Barr virus (EBV) or cytomegalovirus (CMV)-specific TCR, in the CVID convergent TCR sequences (Online Supplementary Figure S11). However, convergent clones did not show higher abundance in CVID (abundance being defined as the sum of all convergent clone rearrangement frequencies) (Online Supplementary Figure S12). We also investigated the physico-chemical properties of highly convergent and highly expanded TCR sequences with a generic string kernel algorithm (GSKernel) and unsupervised clustering analyses: CD8+ cells had 275 expanded and 3,738 highly convergent clones, and CD4+ cells had 342 expanded clones and 848 highly convergent clones. In CD8+ cells, these selected TCR formed 24 clusters based on the TCR physico-chemical similarity (Figure 4A). Six clusters were enriched with CVID TCR (Figure 4B). A major determinant driving the clustering was J-gene usage, but not V-gene usage (Online Supplementary Figure S13). In CD4+ cells, selected TCR formed 20 clusters with haematologica | 2020; 105(12)

Somatic mutations in immunodeficiency

Figure 4. Convergent T-cell receptors (TCR) form structurally similar clusters, and some clusters enrich with common variable immunodeficiency (CVID) patients. Convergent and highly expanded TCR were analyzed for physico-chemical structure similarity and visualized with t-distributed stochastic neighbor embedding (t-SNE). Only highly convergent TCR and large clones (defined in “Methods”) were included in the analyses. Each dot represents a single TCR, and each TCR can be mapped to one patient. (A) Analyzed TCR formed 24 clusters in CD8+ cells, of which six were enriched with CVID TCR, as analyzed with Fisher's test (#). (B) The same clustering analysis, with CD8+ CVID or healthy control (HC) TCR with colors. (C) Analyzed TCR formed 20 clusters in CD4+ cells, of which two were enriched with CVID TCR (#). (D) The same clustering analysis, with CD4+ CVID or HC TCR with colors.

similar results (Figure 4C and Online Supplementary Figure S14). Interestingly, convergent CD8+ TCR in CVID had slightly different properties when compared to control TCR: CVID TCR showed a difference in polarity and hydrophobicity when compared to controls (Online Supplementary Table S11), whereas no differences were seen in CD4+ TCR.

CD8+ cells show higher frequencies of antigen-specific T cells than CD4+ cells Lastly, we investigated whether CVID patients had haematologica | 2020; 105(12)

higher frequencies or abundances of previously reported TCR, with a focus on pathogen-specific TCR (see “Methods” section). There was no difference in frequencies of antigen-specific TCR between CVID patients and controls (Online Supplementary Figures S15 and S16; healthy CD4+ vs. CVID CD4+, and healthy CD8+ vs. CVID CD8+). However, in healthy controls, the frequencies of pathogen-specific, cancer-specific, and autoimmune-associated TCR were higher in CD8+ cells than CD4+ cells, but these differences were not statistically significant in CVID (Online Supplementary Figures S15 and S16). The abun2765

P. Savola et al.

dance of pathogen-specific, but not cancer or autoimmune TCR, was also increased in CD8+ when compared to CD4+ cells in controls.

T-cell clone size and somatic mutation variant allele frequency Individuals with mutations had larger CD8+ T-cell clones, but similar association was not observed among CD4+ T cells (Online Supplementary Figure S6). We plotted the frequency of the largest T-cell clone and mutation VAF (Online Supplementary Figure S17) to investigate the relationship between the clone size and VAF. If a somatic mutation occurs only in one individual T-cell clone, the mutation VAF should be half the clone size or less (assuming that most somatic mutations are heterozygous). In CD8+ T cells this was true, and mutation VAF correlated well with the clone size (Online Supplementary Figure S17). However, in some cases, the mutation VAF in CD4+ cells were higher than expected (Online Supplementary Table S12 and Online Supplementary Figure S17), suggesting that mutations occur also in other cells rather than just in one CD4+ T-cell clone. Using flow cytometry-mediated sorting and sequencing, we confirmed that the mutations were still confined to the CD4+ T-cell fraction as they did not occur in other cell types, such as CD14+ cells or CD3+CD8+ cells (patients with STAT5B mutations) (Online Supplementary Table S13). However, clonal hematopoiesisassociated variants occurred in both CD4+ and CD8+ cells (Table 3 and Online Supplementary Table S8).

Discussion In our small cohort (n=17) of late-onset CVID and other immunodeficiency patients, we show that somatic mutations occur in CD4+ and CD8+ cells in 75% of the CVID patients and in 65% of all patients. Lymphoproliferation (65%) and autoimmune disease manifestations were highly prevalent (94%) in our study cohort. Healthy controls harbored mutations in 45% of cases. Considering also the data on somatic mutations occurring in either CD4+ or CD8+ cells, we identified clonal hematopoiesis variants in both CD4+ and CD8+ cells in 24% of immunodeficiency patients. CVID patients also harbored higher frequencies of convergent TCR than healthy controls, and clustering analyses on selected TCR identified structurally similar TCR clusters that were enriched with CVID TCR. Somatic mutations occur throughout life at an estimated rate of 3.5*10-9â&#x20AC;&#x201C;1.6*10-7 per base-pair per mitosis in many somatic tissues.26 Since somatic mutations accumulate with age, the discovery of somatic mutations in CD8+ and CD4+ cells is not unexpected.27-29 The majority of the identified mutations occurred in CD8+ cells, which could be due to the higher clonality of CD8+ T cells. In agreement with this, our results showed that patients with mutations had larger CD8+ T-cell clones. However, it should be noted that although the overall CD4+ clonality was low, many immunodeficiency patients harbored mutations also in CD4+ cells. Interestingly, the most dominant mutational signatures in T cells were signatures 1 and 15. Signature 1 correlates with age and derives from spontaneous deamination of 5-methylcytosine;30 this signature was especially dominant in CD8+ cells and in immunodeficiency patients. Signature 15, which associates with defective mismatch repair,31 was more prominent in healthy control CD4+ cells. 2766

Many of the genes with somatic mutations in immunodeficiency patients link to malignant disorders, autoimmunity, or lymphoproliferative disorders. Somatic mutations in known hematologic tumor suppressors and/or oncogenes18-20 occurred in 35% of immunodeficiency patients. CVID patients harbored these mutations in 38% (3 of 8) of cases and healthy controls in 9.5% (2 of 21). Two patients harbored somatic STAT5B mutations (patient 2, N418K; patient 13, T628S) in CD4+ cells. The T628S mutation in patient 13 results in increased STAT5 phosphorylation, transcriptional activity, and cell proliferation.32 Activating STAT5B mutations in lymphocytes occur especially in CD4+ T-cell large-granular lymphocyte leukemia and T-prolymphocytic leukemia.32,33 Patient 16 harbored a KRAS T58I mutation in CD8+ cells with VAF 7.9%. KRAS is mutated in nearly 30% of cancers,34 and somatic activating KRAS mutations in both lymphoid and myeloid cells are observed in Ras-associated autoimmune leukoproliferative disorder.35 Of note, the patient had a major T-cell clone in CD8+ cells comprising 51% of all CD8+ cells. Two patients had C5AR1 (complement C5a receptor 1) mutations. Recent data indicate that C5AR1 activation promotes the proinflammatory Th1 responses in CD4+ cells via reactive oxygen species and inflammasome activation.36 Due to the lack of statistical power and differences in sequencing coverage, we cannot draw conclusions about the differences in mutation prevalence and mutation counts between immunodeficiency patients and controls. Although immunodeficiency patients had potentially pathogenic variants, also healthy controls harbored mutations in tumor suppressors and oncogenes, such as in NRAS (G13D) and DNMT3A (L773P), which is not unexpected due to the high prevalence of clonal hematopoiesis in 60-69 year olds.37-39 Nevertheless, immunodeficiency patients had mutations that do not generally occur in clonal hematopoiesis in the general population, such as mutations in STAT5B. These mutations, or a subset of them, could contribute to lymphoproliferation in these patients, despite showing low VAF. Although the function of the discovered mutations was not analyzed in our study, previous studies by us and others suggest that somatic mutations may alter the phenotype and function of mutated T cells.16,40 Interestingly, Fraietta et al. has shown that a hypomorphic TET2 mutation can lead to the expansion and persistence of chimeric antigen receptor (CAR)-T cells.40 Tet2 inactivation in mice results in altered T- and B-cell differentiation, and TET2 mutations occur in T-cell lymphomas.41 DNMT3A mutations may also facilitate lymphoproliferation, as shown by Dnmt3a deficiency leading to both myeloid and lymphoid malignancies in mice.42 DNMT3A is also commonly mutated in human acute T-cell lymphoblastic leukemias.42 The clinical relevance of DNMT3A and TET2 mutations and the possible survival advantage they may confer in immunodeficiency requires more studies, as the prevalence of these mutations is relatively high also in healthy T cells. The majority of the detected mutations occurred exclusively in CD4+ or CD8+ cells. The origin of these mutations is elusive. Do these somatic mutations occur in one T-cell clone or in multiple clones? Do they occur in every cell of the same clone or only in a subset? Any conclusive answer could only be obtained through in-depth singlecell sequencing studies. In our previous reports, we haematologica | 2020; 105(12)

Somatic mutations in immunodeficiency

showed that mutations are confined to specific T-cell populations in CD8+ cells.16 Similarly, in immunodeficiency patient mutation, VAF and clone sizes matched well in CD8+ T cells, suggesting that mutations occur late during T-cell maturation. However, the analysis of CD4+ T cells suggested that mutations may also occur in precursor cells, as mutation VAF were higher than individual CD4+ T-cell clone sizes. In addition to mutations found exclusively in one cell type, we also detected mutations which were found in both CD4+ and CD8+ compartments, suggesting that they could have originated from a hematopoietic stem cell, or from multipotent or lineage-specific progenitors. Our results are in line with previous publications showing that clonal hematopoiesis-associated mutations (such as DNMT3A and TET2) can occur in T cells.37,43,44 These mutations likely originate from progenitor cells, because in our cases they were found both in CD4+ and CD8+ T cells, and in CD19+ B cells, but not in myeloid CD14+ cells. However, myeloid cells can harbor the same mutations as B cells (as in monoclonal B-cell lymphocytosis and indolent chronic lymphocytic leukemia) or T cells (in individuals without hematologic disease).37,44-47 In the present study, we detected somatic mutations with relatively low VAF by using high sequencing coverage in the assays. To avoid exhaustion of the unique DNA molecules, a sufficient amount of DNA (100-500 ng) was used based on the calculations that one nanogram of DNA comprises approximately 300 copies of unique DNA molecules. Assay sensitivity is also determined by other factors, such as sequencer performance and bioinformatics approaches. In general, Illumina sequencers show a low error rate,48 and we applied standard variant detection practices with multiple quality filters (such as VAF threshold and the requirement for at least seven supporting reads for the variant), similar to reports by others using capture sequencing technologies.7,18,19,49,50 Although a previous study had suggested increased T-cell clonality in CVID patients,51 we did not observe any difference in clonality between CVID and controls. The heterogenous phenotypes of CVID patients, ages of controls, and sample material may explain the difference in results. Interestingly, however, we noted increased TCR

References 1. Bonilla FA, Barlan I, Chapel H, et al. International consensus document (ICON): common variable immunodeficiency disorders. J Allergy Clin Immunol Pract. 2016; 4(1):38-59. 2. Bogaert DJA, Dullaers M, Lambrecht BN, Vermaelen KY, De Baere E, Haerynck F. Genes associated with common variable immunodeficiency: one diagnosis to rule them all? J Med Genet. 2016;53(9):575-590. 3. Selenius JS, Martelius T, Pikkarainen S, et al. Unexpectedly high prevalence of common variable immunodeficiency in Finland. Front Immunol. 2017;81190. 4. Wong GK, Huissoon AP. T-cell abnormalities in common variable immunodeficiency: the hidden defect. J Clin Pathol. 2016;69(8):672-676. 5. Gathmann B, Mahlaoui N, Gerard L, et al. Clinical picture and treatment of 2212 patients with common variable immunode-

haematologica | 2020; 105(12)

10.

convergence in CVID. Previous studies have shown that TCR convergence associates with clone size52 and public TCR in mice.53 We hypothesize that convergence is due to enrichment mediated by antigen stimulation. The latter is further supported by the finding that CVID convergent TCR showed some structural similarity (Figure 4), which could be related to common antigen stimulation. In conclusion, we discovered somatic mutations and increased TCR convergence in mature T cells in patients with immunodeficiency. TCR convergence may be due to a chronic antigen response, and somatic variants may contribute to immune dysregulation. Somatic mutations may originate either in mature T cells or hematopoietic stem cells. Although healthy controls also harbor some variants with potentially pathogenic consequences, our results demonstrate that mutations in genes that have been associated with autoimmunity and lymphoproliferation (such as STAT5B) occur in immunodeficiency patients. Further studies are needed to discover the full spectrum of somatic mutations and their role as possible regulators of immune responses both in normal and diseased T cells. Acknowledgments We would like to acknowledge the personnel at Hematology Research Unit Helsinki for their assistance, and the FIMM Technology Centre sequencing unit for their assistance on sequencing and data analysis. The HUSLAB flow cytometry unit is acknowledged for their expertise on the B-cell phenotyping assays. The Biomedicum Flow cytometry core unit (HiLIFE) is acknowledged for their assistance on cell sorting. Funding This work was supported by the European Research Council (M-IMM project), the ERAPerMed consortium ‘JAKSTAT-TARGET, Academy of Finland, Finnish special governmental subsidy for health sciences, research and training, the Sigrid Juselius Foundation, the Instrumentarium Science Foundation, the Finnish Cancer Societies, the Finnish Cancer Institute, the Biomedicum Helsinki Foundation, the Finnish Medical Foundation, the Orion Research Foundation, the Juhani Aho Foundation, the K. Albin Johansson Foundation, the Paulo Foundation, and the Foundation for Pediatric Research.

ficienc. J Allergy Clin Immunol. 2014; 134(1):116-126. Mellemkjær L, Hammarström L, Andersen V, Yuen J, Heilmann C, Barington T. Cancer risk among patients with IgA deficiency or common variable immunodeficiency and their relatives : a combined Danish and Swedish study. Clin Exp Immunol. 2002; 130(3):495-500. Nikolaev SI, Vetiska S, Bonilla X, et al. Somatic activating KRAS mutations in arteriovenous malformations of the brain. N Engl J Med. 2018;378(3):250-261. Anglesio MS, Papadopoulos N, Ayhan A, et al. Cancer-associated mutations in endometriosis without cancer. N Engl J Med. 2017;376(19):1835-1848. Koskela HLM, Eldfors S, Ellonen P, et al. Somatic STAT3 mutations in large granular lymphocytic leukemia. N Engl J Med. 2012; 366(20):1905-1913. Jerez A, Clemente MJ, Makishima H, et al. STAT3 mutations unify the pathogenesis of chronic lymphoproliferative disorders of NK

11.

12.

13.

14.

cells and T cell large granular lymphocyte leukemia. Blood. 2012;120(15):3048-3058. Andersson E, Kuusanmäki H, Bortoluzzi S, et al. Activating somatic mutations outside the SH2-domain of STAT3 in LGL leukemia. Leukemia. 2016;30(5):1204-1208. Jerez A, Clemente MJ, Makishima H, et al. STAT3-mutations indicate the presence of subclinical T cell clones in a subset of aplastic anemia and myelodysplastic syndrome patients. Blood. 2013;122(14):24532459. Rajala HLM, Olson T, Clemente MJ, et al. The analysis of clonal diversity and therapy responses using STAT3 mutations as a molecular marker in large granular lymphocytic leukemia. Haematologica. 2015; 100(1):91-99. Ettersperger J, Montcuquet N, Malamut G, et al. Interleukin-15-dependent T-cell-like innate intraepithelial lymphocytes develop in the intestine and transform into lymphomas in celiac disease. Immunity. 2016; 45(3):610-625.

2767

P. Savola et al. 15. Savola P, Brück O, Olson T, et al. Somatic STAT3 mutations in Felty syndrome: an implication for a common pathogenesis with large granular lymphocyte leukemia. Haematologica. 2018;103(2):304-312. 16. Savola P, Kelkka T, Rajala HLM, et al. Somatic mutations in clonally expanded cytotoxic T lymphocytes in patients with newly diagnosed rheumatoid arthritis. Nat Commun. 2018;8:15869. 17. Valori M, Jansson L, Kiviharju A, et al. A novel class of somatic mutations in blood detected preferentially in CD8+ cells. Clin Immunol. 2017;175:75-81. 18. Jaiswal S, Fontanillas P, Flannick J, et al. Agerelated clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014; 371(26):2488-2498. 19. Jaiswal S, Natarajan P, Silver AJ, et al. Clonal hematopoiesis and risk of atherosclerotic cardiovascular disease. N Engl J Med. 2017; 377(2):111-121. 20. Genovese G, Kähler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-2487. 21. Fuster JJ, MacLauchlan S, Zuriaga MA, et al. Clonal hematopoiesis associated with Tet2 deficiency accelerates atherosclerosis development in mice. Science. 2017; 355(6327):842-847. 22. Holzelova E, Vonarbourg C, Stolzenberg MC, et al. Autoimmune lymphoproliferative syndrome with somatic Fas mutations. N Engl J Med. 2004;351(14):1409-1418. 23. Trotta L, Martelius T, Siitonen T, et al. ADA2 deficiency: clonal lymphoproliferation in a subset of patients. J Allergy Clin Immunol. 2018;141(4):1534-1537. 24. Haapaniemi EM, Kaustio M, Rajala HLM, et al. Autoimmunity, hypogammaglobulinemia, lymphoproliferation, and mycobacterial disease in patients with activating mutations in STAT3. Blood. 2015;125(4):639-649. 25. Seppänen M, Koillinen H, Mustjoki S, Tomi M, Sullivan KE. Terminal deletion of 11q with significant late-onset combined immune deficiency. J Clin Immunol. 2014; 34(1):114-118. 26. Werner B, Sottoriva A. Variation of mutational burden in healthy human tissues suggests non-random strand segregation and allows measuring somatic mutation rates. PLoS Comput Biol. 2018;14(6): e1006233. 27. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264-278. 28. Blokzijl F, De Ligt J, Jager M, et al. Tissue-

2768

29.

30.

31.

32.

33.

34. 35.

36.

37.

38.

39.

40.

41.

specific mutation accumulation in human adult stem cells during life. Nature. 2016; 538(7624):260-264. Franco I, Johansson A, Olsson K, et al. Somatic mutagenesis in satellite cells associates with human skeletal muscle aging. Nat Commun. 2018;9(1):800. Alexandrov LB, Jones PH, Wedge DC, et al. Clock-like mutational processes in human somatic cells. Nat Genet. 2015;47(12):14021407. Petljak M, Alexandrov LB. Understanding mutagenesis through delineation of mutational signatures in human cancer. Carcinogenesis. 2016;37(6):531-540. Kiel MJ, Velusamy T, Rolland D, et al. Integrated genomic sequencing reveals mutational landscape of T-cell prolymphocytic leukemia. Blood. 2014;124(9):14601473. Andersson EI, Tanahashi T, Sekiguchi N, et al. High incidence of activating STAT5B mutations in CD4-positive T-cell large granular lymphocyte leukemia. Blood. 2016; 128(20):2465-2468. Hobbs GA, Der CJ, Rossman KL. RAS isoforms and mutations in cancer at a glance. J Cell Sci. 2016;129(7):1287-1292. Niemela JE, Lu L, Fleisher TA, et al. Somatic KRAS mutations associated with a human nonmalignant syndrome of autoimmunity and abnormal leukocyte homeostasis. Blood. 2012;117(10):2883-2886. Arbore G, West EE, Spolski R, et al. T helper 1 immunity requires complement-driven NLRP3 inflammasome activity in CD4+ T cells. Science. 2016;352(6292):aad1210. Arends CM, Galan-Sousa J, Hoyer K, et al. Hematopoietic lineage distribution and evolutionary dynamics of clonal hematopoiesis. Leukemia. 2018;32(9):1908-1919. Acuna-Hidalgo R, Sengul H, Steehouwer M, et al. Ultra-sensitive sequencing identifies high prevalence of clonal hematopoiesisassociated mutations throughout adult life. Am J Hum Genet. 2017;101(1):50-64. Buscarlet M, Provost S, Zada YF, et al. DNMT3A and TET2 dominate clonal hematopoiesis and demonstrate benign phenotypes and different genetic predispositions. Blood. 2017;130(6):753-762. Fraietta JA, Nobles CL, Sammons MA, et al. Disruption of TET2 promotes the therapeutic efficacy of CD19-targeted T cells. Nature. 2018;558(7709):307-312. Quivoron C, Couronné L, Della Valle V, et al. TET2 inactivation results in pleiotropic hematopoietic abnormalities in mouse and is a recurrent event during human lym-

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

phomagenesis. Cancer Cell. 2011;20(1):2538. Bowman RL, Busque L, Levine RL. Clonal hematopoiesis and evolution to hematopoietic malignancies. Cell Stem Cell. 2018; 22(2):157-170. Buscarlet M, Provost S, Feroz Zada Y, et al. Lineage restriction analyses in CHIP indicate myeloid bias for TET2 and multipotent stem cell origin for DNMT3A. Blood. 2018; 132(3):277-280. Young AL, Challen GA, Birmann BM, Druley TE. Clonal haematopoiesis harbouring AML-associated mutations is ubiquitous in healthy adults. Nat Commun. 2016; 7:12484. Damm F, Mylonas E, Cosson A, et al. Acquired initiating mutations in early hematopoietic cells of CLL patients. Cancer Discov. 2014;4(9):1088-1101. Condoluci A, Rossi D. Age-related clonal hematopoiesis and monoclonal B-cell lymphocytosis / chronic lymphocytic leukemia: a new association? Haematologica. 2018; 103(5):751-752. 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. Caspar SM, Dubacher N, Kopps AM, Meienberg J, Henggeler C, Matyas G. Clinical sequencing: from raw data to diagnosis with lifetime value. Clin Genet. 2018; 93(3):508-519. Desai P, Mencia-Trinchant N, Savenkov O, et al. Somatic mutations precede acute myeloid leukemia years before diagnosis. Nat Med. 2018;24(7):1015-1023. Abelson S, Collord G, Stanley W, et al. Prediction of acute myeloid leukaemia risk in healthy individuals. Nature. 2018; 559(7714):400-404. Ramesh M, Hamm D, Simchoni N, Cunningham-Rundles C. Clonal and constricted T cell repertoire in common variable immune deficiency. Clin Immunol. 2017; 178:1-9. Venturi V, Quigley MF, Greenaway HY, et al. A mechanism for TCR sharing between T cell subsets and individuals revealed by pyrosequencing. J Immunol. 2011; 186(7):4285-4294. Madi A, Shifrut E, Reich-Zeliger S, et al. Tcell receptor repertoires share a restricted set of public and abundant CDR3 sequences that are associated with self-related immunity. Genome Res. 2014;24(10):1603-1612.

haematologica | 2020; 105(12)

ARTICLE

Red Cell Biology & its Disorders

No evidence of hemoglobin damage by SARS-CoV-2 infection

Ferrata Storti Foundation

Anthony W. DeMartino,1* Jason J. Rose,1,2,3* Matthew B. Amdahl,1 Matthew R. Dent,1 Faraaz A. Shah,2,4.5 William Bain,2,4.5 Bryan J. McVerry,2,5 Georgios D. Kitsios,2,5 Jesús Tejero1,2,3,6 and Mark T. Gladwin1,2,3

Heart, Lung, Blood, and Vascular Medicine Institute, University of Pittsburgh; 2Division of Pulmonary, Allergy and Critical Care Medicine, University of Pittsburgh; 3Department of Bioengineering, University of Pittsburgh; 4VA Pittsburgh Healthcare System; 5The Acute Lung Injury Center of Excellence, University of Pittsburgh and 6Department of Pharmacology and Chemical Biology, University of Pittsburgh, Pittsburgh, PA, USA 1

*AWD and JJR contributed equally as co-first authors

Haematologica 2020 Volume 105(12):2769-2773

ABSTRACT

he severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) disease (COVID-19) has affected over 22 million patients worldwide as of August 2020. As the medical community seeks better understanding of the underlying pathophysiology of COVID-19, several theories have been proposed. One widely shared theory suggests that SARS-CoV-2 proteins directly interact with human hemoglobin (Hb) and facilitate removal of iron from the heme prosthetic group, leading to the loss of functional hemoglobin and accumulation of iron. Herein, we refute this theory. We compared clinical data from 21 critically ill COVID-19 patients to 21 non-COVID-19 acute respiratory distress syndrome (ARDS) patient controls, generating hemoglobin-oxygen dissociation curves from venous blood gases. This curve generated from the COVID-19 cohort matched the idealized oxygen-hemoglobin dissociation curve well (Pearson correlation R2=0.97, P<0.0001; a coefficient of variation of the root-mean-square deviation [CV(RMSD)] =7.3%). We further analyzed hemoglobin, total bilirubin, lactate dehydrogenase, iron, ferritin, and haptoglobin levels. For all analyzed parameters, patients with COVID-19 had similar levels compared to patients with ARDS without COVID-19. These results indicate that patients with COVID-19 do not exhibit any hemolytic anemia or a shift in the normal hemoglobin-oxygen dissociation curve. We therefore conclude that COVID-19 does not impact oxygen delivery through a mechanism involving red cell hemolysis and subsequent removal of iron from the heme prosthetic group in hemoglobin.

Introduction The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pandemic has progressed around the globe with over 22 million cases at the time of this writing. In the early stages of the SARS-CoV-2 disease (COVID-19) pandemic, there were reports of unique clinical phenotypes in affected patients. One study of the initial experiences of critically ill patients in Italy described an atypical form of viral pneumonia-induced acute respiratory distress syndrome (ARDS) with normal lung compliance and low ventilation-to-perfusion ratio.1 Initial reports had suggested very high mortality rate (>80%)2 for patients with respiratory failure from SARSCoV-2 disease (COVID-19), compared to pre-COVID-19 ARDS mortality rates in the range of 30-40%.3 A clinical syndrome of “silent” or “happy” hypoxemia has been widely observed, with patients exhibiting minimal dyspnea or signs of neurocognitive dysfunction despite severe hypoxemia detected via pulse oximetry.4 A number of hospitalized COVID-19 patients appear to have significant hemostatic activation, with 25-31% prevalence of venous thromboembolism observed in some cohorts.5,6 More recent clinical studies have shown that the mechanical ventilation requirements in COVID-19 patients are similar to populations of patients enrolled in ARDS clinical trials without COVID-19.7,8 Further observational studies haematologica | 2020; 105(12)

Correspondence: JASON J. ROSE rosejj@upmc.edu Received: June 24, 2020. Accepted: August 27, 2020. Pre-published: September 10, 2020. doi:10.3324/haematol.2020.264267 ©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.

2769

A.W. DeMartino et al.

Figure 1. Description of pathophysiology of SARS-CoV-2 infection. Infection via the lungs (open arrows) resulting in hypoxemia (left) vs. direct ‘attack’ by COVID-19 viral proteins on red blood cell (RBC) hemoglobin (Hb) as proposed by Liu and Li13 (right, hashed arrows/gray boxes). Hypothesized pathways crossed out by red X’s and highlighted in gray boxes are not supported by available clinical evidence, in vitro/in vivo studies, our results, nor mechanisms of viral interaction with RBC. See text for details.

have reported mortality rates for COVID-19 in the range of 25-32%, similar to the mortality rates for non-COVID19 ARDS.9,10 While the exact pathophysiology of COVID19 remains a topic of active investigation, growing evidence suggests that respiratory failure in COVID-19 patients behaves similarly to respiratory failure in patients with severe viral pneumonia that triggers ARDS.11,12 The uncertainty at the outset of this pandemic as well as some persistently unique features of the disease (e.g., increased thromboembolic risks), have led to a large number of proposed hypotheses regarding the pathophysiological mechanisms of SARS-CoV-2. One widely proposed theory holds that viral proteins directly interact with human hemoglobin (Hb) and facilitate removal of iron from the protein’s heme prosthetic group, resulting in the loss of functional hemoglobin and the toxic accumulation of iron. This theory originated from a manuscript by Liu and Li13 posted in the preprint server ChemRxiv (over 1.89 million manuscript views as of August 2020). This work by Liu and Li,13 which has not been peer-reviewed and continues to be cited,14,15 uses in silico approaches to model interactions between several viral proteins and hemoglobin. In brief, the manuscript suggests that viral proteins ORF1ab, ORF10, and ORF3a are derived from infected plasma cells and work in concert to remove heme from the b chain of hemoglobin, strip iron from heme, and sequester the resulting iron-free protoporphyrin IX (PPIX). The authors speculate that this coordinated "attack" of hemoglobin occurs in the plasma following immune hemolysis, resulting in the release of toxic amounts of iron, diminished functional hemoglobin levels, and disrupted heme metabolism. Finally, Liu and 2770

Li13 further purport that the consequences of such hemoglobin degradation account for some of the irregular clinical characteristics reported early in the pandemic (Figure 1). While this work has received significant attention from scientists, physicians, the press, and the general public, the study advances a theory of viral interaction with hemoglobin that is inconsistent with well-characterized mechanisms of physiological heme degradation,16 and, critically, advances a pathophysiological mechanism inconsistent with the clinical presentation of COVID-19 patients. Herein, we compare clinical laboratory data from confirmed COVID-19 patients admitted to the intensive care unit (ICU) at University of Pittsburgh Medical Center (UPMC) to patients with ARDS but without COVID-19. These empirical data challenge the theory that SARS CoV2 proteins directly remove iron from human hemoglobin as a pathophysiological mechanism of COVID-19.

Methods We reviewed laboratory data from 21 patients with known COVID-19 (PCR positive for SARS-CoV-2) admitted to the ICU at UPMC and compared with 21 patients with non-COVID-19 ARDS from different etiologies, who had been enrolled in the Acute Lung Injury Registry (ALIR) and Biospecimen Repository at UPMC. Patient data were de-identified. This study was approved by the University of Pittsburgh Institutional Reviewer Board. We recorded clinically available venous blood gas values of partial pressure of carbon dioxide (PvCO2), partial pressure of oxygen (PvO2), pH, and venous oxygen saturation of hemoglobin (SvO2). We recorded initial hemoglobin and total bilirubin levels. Values haematologica | 2020; 105(12)

No hemoglobin damage by SARS-CoV-2

COVID-19

ARDS

Standard curve

PvO2 (mmHg) Figure 2. Summary of clinical data of COVID-19 and non-COVID-19 acute respiratory distress syndrome patient cohorts. Oxygen-hemoglobin dissociation curve (left): standard curve as described by JW Severinghaus17 (black squares, gray line), values of oxygenation saturation plotted versus PvO2 for patients admitted to intensive care unit (ICU) with COVID-19 (blue diamonds) and ICU patients with acute respiratory distress syndrome (ARDS) but without COVID-19 (red circles). Individual replicates (i.e., same PvO2) are displayed. Corresponding laboratory values of hemoglobin, total bilirubin, ferritin, iron, and lactate dehydrogenase (LDH) were similar between ICU patients with COVID-19 and those with ARDS without COVID-19 (right, no significant differences: P=0.76, P=0.28, P=0.82, P=0.52, P=0.94, respectively). Few patients had haptoglobin data, but available COVID-19 values were similar to ARDS (mean 202 vs. 283, respectively). Error bars represent standard error of mean.

for iron, ferritin, haptoglobin, and lactate dehydrogenase (LDH) levels (whenever available during the inpatient admission) were collected as well. Hemoglobin and total bilirubin levels between COVID-19 and non-COVID-19 patients were compared using a Student’s t-test. For iron, ferritin and LDH, Mann-Whitney testing was used. Values for SvO2 versus PvO2 were plotted to determine a line of best fit. The deviation of the PvO2 by COVID-19 and ARDS cohorts from the theoretical-standard-predicted-PvO2 (described by JW Severinghaus17) was evaluated using a Pearson correlation and a coefficient of variation of the root-mean-square deviation or CV(RMSD). All data variance is described as standard error of mean, except for age, which is described as standard deviation.

Results Oxygen-hemoglobin dissociation curves were generated using COVID-19 patient data and non-COVID-19 ARDS patient data and compared against the theoretical standard curve generated by the Severinghaus model.17 The average age of the COVID-19 cohort was 62 ± 9 years versus 47 ± 17 years for the ARDS cohort. The distribution of sex was 12 males and 9 females in the COVID-19 cohort and 10 males and 11 females in the non-COVID-19 ARDS cohort. The oxygen-hemoglobin dissociation curve generated from patients with COVID-19 was similar to the curve generated from patients with ARDS without COVID-19. The fitting generated from COVID-19 patient data matched the ideal oxygen-hemoglobin dissociation curve well (Pearson correlation R2=0.97, P<0.0001; CV(RMSD)=7.3%). The fitting generated from nonCOVID-19 ARDS patients also matches the ideal oxygenhemoglobin dissociation curve reasonably well (Pearson correlation R2=0.92, P<0.0001; CV(RMSD)=9.4%, Figure 2). This comparison of oxygen-hemoglobin dissociation curves suggests that hemoglobin oxygen affinity is not altered in patients with COVID-19 admitted to the ICU. Patients with COVID-19 had similar total hemoglobin, haematologica | 2020; 105(12)

total bilirubin, ferritin, iron and LDH levels compared to patients with ARDS without COVID-19 (none were significantly different, Figure 2). Few patients had haptoglobin levels available. In those COVID-19 patients that did, the average haptoglobin level was similar to that of ARDS patients (mean 205 mg/dL, n=2, vs. 283±19, n=6). There was no laboratory evidence of on-going red blood cell hemolysis or degradation of hemoglobin.

Discussion Liu and Li13 generated computational results obtained by sequence analysis, molecular modeling, and docking approaches to propose a novel model of viral degradation of hemoglobin-derived heme. However, the authors employed methodologies and docking simulations that have been heavily criticized, as thoroughly discussed in a recent report by Read.18 It is important to note that Liu and Li13 do not present any experimental support for their theories, and even though they have revised the initial manuscript and vastly changed their calculated parameters (as of version 9), their new calculations are speculative at best, as Read has recently addressed in an addendum to his manuscript (version 2).18 Moreover, the hypotheses originally put forth by Liu and Li13 remain unchanged. Our work thus focuses on the unique mechanism of SARSCoV-2 proposed by Liu and Li:13 direct virus triggered hemoglobin degradation. We highlight that this hypothesis is not supported by existing evidence, known pathologies of coronaviruses, Liu and Li own docking calculations,18 or the clinical data presented herein. Most nonstructural viral proteins of coronaviruses are not found in large amounts in plasma but rather localize in infected cells where they play important roles in RNA replication.19 Thus, these proteins are unlikely to access appreciable amounts of hemoglobin. There is no evidence suggesting that the virus enters erythrocytes, where heme concentrations are 15-20 mM, and these highly specia2771

A.W. DeMartino et al.

lized cells lack the requisite cellular machinery needed to produce viral proteins.20 As a result, intraerythrocytic hemoglobin is likely protected from exposure to viral proteins. Liu and Li13 posit that the interaction between viral proteins and hemoglobin may occur in the plasma after immune hemolysis, but significant hemolysis has not been documented in COVID-19 patients,8,21,22 nor did we observe hemolysis in our patients. Any elevation in LDH levels seen in COVID-19 patients is likely derived from hepatocellular injury and not intravascular hemolysis. Liu and Li13 suggest that the virus could infect plasma cells via the ACE2 receptor and induce secretion of viral proteins from infected plasma cells. However, there is no evidence of SARS-CoV-2 infecting plasma cells.23,24 Further, secretion of viral proteins from any infected cells is extremely rare and does not occur in any viruses related to SARSCoV-2.25 The manuscript’s flawed18 and experimentally unverified docking models lead the authors to suggest that a non-structural viral protein, ORF10, binds to heme and releases heme-derived iron. In this putative mechanism, ORF1ab and ORF3a bind the hemoglobin protein and cause conformational changes that expose the heme to ORF10, which subsequently breaks down the cofactor into iron and PPIX; the latter is then theoretically captured by ORF1ab. This model would represent an entirely novel mechanism of hemoglobin degradation, as hemoglobin has not been documented to undergo large conformational shifts as a result of protein-protein interactions.26 Further, the removal of iron from PPIX by ORF10, a protein of only 38 amino acids, is unlikely considering that heme degradation is catalyzed by significantly larger and more complex, well-characterized heme-oxygenase proteins.16 As the authors’ hypothesis represents a completely novel and unexpected model of heme degradation, careful in vitro and in vivo studies would be required to confirm such a mechanism. It is worth noting that we do not observe a statistical difference in iron, bilirubin, or ferritin levels between COVID-19 patients and ARDS controlpatients (Figure 2), indicating that heme breakdown is not occurring above typical catabolic levels in the COVID-19 cohort. We further assert that the clinical syndrome observed in COVID-19 patients is not consistent with Liu and Li’s13 model of heme degradation. The clinical evidence presented here does neither suggest hemolysis, hemoglobin degradation, removal of iron from the heme molecule, nor altered oxygen affinity of hemoglobin (Figure 2). The oxygen dissociation curve calculated from real-life patient data fits quite well with the idealized standard curve in both the COVID and ARDS cohorts (R2=0.97, R2=0.92,

References 1. Gattinoni L, Chiumello D, Caironi P, et al. COVID-19 pneumonia: different respiratory treatments for different phenotypes? Intensive Care Med. 2020;46(6):1099-1102. 2. Yang X, Yu Y, Xu J, et al. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study. Lancet Respir Med. 2020; 8(5):475-481. 3. Bellani G, Laffey JG, Pham T, et al.

2772

respectively). There is minimal deviation from the idealized curve CV(RMSD) of 7.3 and 9.4%, respectively). This finding is in agreement with another recent study that uses hemoglobin isolated from the red blood cells of COVID-19 patients demonstrating normal hemoglobinoxygen dissociation properties ex vivo.27 The clinical data presented here suggests no significant hemolysis or abnormal hemoglobin-oxygen dissociation characteristics. As aforementioned, newer clinical reports of critically ill COVID-19 patients suggest both mortality rates and mechanical ventilation requirements similar to other forms of ARDS.2,7–10 There has been no evidence to suggest a unique hemoglobin-specific mechanism such as large-scale hemoglobin degradation. In addition to the work here, there has been no other evidence of significant anemia or iron overload.8,21,22 While the finding of elevated transaminase levels has been widely described, this condition exists without hyperbilirubinemia, which would be a signal of excessive hemolysis and release of hemoglobin into the plasma.26,28 Increased risk of thromboembolism is observed in COVID-19 patients;5,6 however, acute infections are known to increase the risk of thromboembolism29 and thromboembolism is not generally driven by a hemoglobin-based toxicity. Neither “silent hypoxia”,4 nor patients who maintain a normal work of breathing before rapid onset of ARDS1 suggest an undetected hemoglobin-based toxicity. Any removal of iron and/or heme from hemoglobin would not have an effect on pulse oximetry measurements, as the resulting heme-free hemoglobin does not absorb light in the wavelength range used by these detectors and thus would not interfere with such measurements.30,31 Hemoglobin desaturation correlates well with decreases in partial pressure of oxygen in patients with COVID-19 respiratory failure: in our cohort, we did not observe gross abnormalities in the partial pressure of oxygen versus hemoglobin oxygen saturation. The world community is rapidly working to understand the pathophysiology of COVID-19 in an effort to better prevent the spread of the disease, manage patients, and ultimately develop definitive therapies. There is no suggestion that patients with COVID-19 exhibit a hemolytic anemia or a shift in the normal hemoglobin-oxygen dissociation curve. Thus, COVID-19 does not impact oxygen delivery through a mechanism involving red cell hemolysis and removal of iron from the heme prosthetic group in hemoglobin. Funding This work is funded by 2 P01 HL114453-06 to BJM and 1 K08 HL136857 to JJR.

Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries. JAMA. 2016;315(8):788-800. 4. Xie J, Tong Z, Guan X, Du B, Qiu H, Slutsky AS. Critical care crisis and some recommendations during the COVID-19 epidemic in China. Intensive Care Med. 2020;46(5):837-840. 5. Cui S, Chen S, Li X, Liu S, Wang F. Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia. J Thromb Haemost. 2020; 18(6):1421-1424.

6. Klok FA, Kruip MJHA, van der Meer NJM, et al. Confirmation of the high cumulative incidence of thrombotic complications in critically ill ICU patients with COVID-19: an updated analysis. Thromb Res. 2020;191:148-150. 7. Bhatraju PK, Ghassemieh BJ, Nichols M, et al. Covid-19 in critically ill patients in the Seattle region - case series. N Engl J Med. 2020;382(21):2012-2022. 8. Guan W, Ni Z, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in China. N Engl J Med. 2020;382(18):17081720.

haematologica | 2020; 105(12)

No hemoglobin damage by SARS-CoV-2

9. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA. 2020;323(20):2052-2059. 10. Public Health Scotland Communications, editor. Scottish Intensive Care Society Audit Group report on COVID-19. 13 May 2020. 11. Ziehr DR, Alladina J, Petri CR, et al. Respiratory pathophysiology of mechanically ventilated patients with COVID-19: a cohort study. Am J Respir Crit Care Med. 2020;201(12):1560-1564. 12. Bos LD, Paulus F, Vlaar APJ, Beenen LFM, Schultz MJ. Subphenotyping ARDS in COVID-19 patients: consequences for ventilator management. Ann Am Thorac Soc. 2020 Sep;17(9):1161-1163. 13. Liu W, Li H. COVID-19: Attacks the 1-Beta Chain of Hemoglobin and Captures the Porphyrin to Inhibit Human Heme Metabolism. ChemRxiv. 2020 July 7, v9. [Epub ahead of print]. 14. Soy M, Keser G, Atagündüz P, Tabak F, Atagündüz I, Kayhan S. Cytokine storm in COVID-19: pathogenesis and overview of anti-inflammatory agents used in treatment. Clin Rheumatol. 2020;39(7):20852094. 15. Baidya A, Singh SK, Bajaj S, et al. Diabetes and COVID-19: A Review. J ASEAN Fed Endocr Soc. 2020;35(1):40-48. 16. Gozzelino R, Jeney V, Soares MP.

haematologica | 2020; 105(12)

17.

18.

19.

20.

21.

22.

23.

Mechanisms of cell protection by heme oxygenase-1. Annu Rev Pharmacol Toxicol. 2010;50:323-354. Severinghaus JW. Simple, accurate equations for human blood O2 dissociation computations. J Appl Physiol. 1979; 46(3):599-602. Read R. Flawed methods in “COVID-19: attacks the 1-beta chain of hemoglobin and captures the porphyrin to inhibit human heme metabolism.” ChemRxiv. 2020 May 5, v2. [Epub ahead of print]. Snijder EJ, Decroly E, Ziebuhr J. The nonstructural proteins directing Coronavirus RNA synthesis and processing. Adv Virus Res. 2016;96:59-126. Asher DR, Cerny AM, Finberg RW. The erythrocyte viral trap: transgenic expression of viral receptor on erythrocytes attenuates coxsackievirus B infection. Proc Natl Acad Sci U S A. 2005;102(36):1289712902. Fan BE, Chong VCL, Chan SSW, et al. Hematologic parameters in patients with COVID-19 infection. Am J Hematol. 2020; 95(6):e131-134. Mitra A, Dwyre DM, Schivo M, et al. Leukoerythroblastic reaction in a patient with COVID-19 infection. Am J Hematol. 2020;95(8):999-1000. Thevarajan I, Nguyen THO, Koutsakos M, et al. Breadth of concomitant immune responses prior to patient recovery: a case

24.

25. 26.

27.

28.

29.

30.

31.

report of non-severe COVID-19. Nat Med. 2020;26(4):453-455. Zhu Y, Jiang M, Gao L, Huang X. Single cell analysis of ACE2 expression reveals the potential targets for 2019-nCoV. Preprints. 2020 Feb 16. [Epub ahead of print]. Schatz M, Tong PBV, Beaumelle B. Unconventional secretion of viral proteins. Semin Cell Dev Biol. 2018;83:8-11. Stites WE. Protein−protein interactions: interface structure, binding thermodynamics, and mutational analysis. Chem Rev. 1997;97(5):1233-1250. Daniel Y, Hunt BJ, Retter A, et al. Haemoglobin oxygen affinity in patients with severe COVID-19 infection. Br J Haematol. 2020;190(3):e126-127. Fan Z, Chen L, Li J, et al. Clinical features of COVID-19-related liver functional abnormality. Clin Gastroenterol Hepatol. 2020; 18(7):1561-1566. Smeeth L, Cook C, Thomas S, Hall AJ, Hubbard R, Vallance P. Risk of deep vein thrombosis and pulmonary embolism after acute infection in a community setting. Lancet. 2006;367(9516):1075-1079. Pires IS, Belcher DA, Palmer AF. Quantification of active apohemoglobin heme-binding sites via dicyanohemin incorporation. Biochemistry. 2017; 56(40):5245-5259. Jubran A. Pulse oximetry. Crit Care Lond Engl 2015;19(1):272.

2773

ARTICLE Ferrata Storti Foundation

Red Cell Biology & its Disorders

Distal and proximal hypoxia response elements co-operate to regulate organ-specific erythropoietin gene expression Ilaria M. C. Orlando,1,2 Véronique N. Lafleur,3 Federica Storti,1,2 Patrick Spielmann,1,2 Lisa Crowther,1,2 Sara Santambrogio,1,2 Johannes Schödel,4 David Hoogewijs,2,5 David R. Mole3 and Roland H. Wenger1,2

Haematologica 2020 Volume 105(12):2774-2784

1 Institute of Physiology, University of Zürich, Zürich, Switzerland; 2National Center of Competence in Research “Kidney.CH”, Zürich, Switzerland; 3NDM Research Building, University of Oxford, Oxford, UK; 4Department of Nephrology and Hypertension, Universitätsklinikum Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany and 5Department of Medicine/Physiology, University of Fribourg, Fribourg, Switzerland

ABSTRACT

Correspondence:

hile it has been well-established that distal hypoxia response elements (HRE) regulate hypoxia-inducible factor (HIF) target genes such as erythropoietin (Epo), an interplay between multiple distal and proximal (promoter) HRE has not yet been described. Hepatic Epo expression is regulated by an HRE located downstream of the EPO gene, but this 3' HRE is not essential for renal EPO gene expression. We previously identified a 5' HRE and could show that both HRE direct exogenous reporter gene expression. Here, we show that, whereas in hepatic cells the 3' but not the 5' HRE is required, in neuronal cells both the 5' and 3' HRE contribute to endogenous Epo induction. Moreover, two novel putative HRE were identified in the EPO promoter. In hepatoma cells, HIF interacted mainly with the distal 3' HRE, but in neuronal cells, HIF most strongly bound the promoter, bound to a lesser extent the 3' HRE, and did not bind the 5' HRE. Interestingly, mutation of either of the two distal HRE abrogated HIF binding to the 3' and promoter HRE. These results suggest that a canonical functional HRE can recruit multiple transcription factors (not necessarily HIF) to mediate HIF binding to different distant HRE in an organ-specific manner.

ROLAND H. WENGER roland.wenger@access.uzh.ch

Introduction

Received: August 27, 2019. Accepted: December 18, 2019. Pre-published: December 19, 2019. doi:10.3324/haematol.2019.236406 ©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.

2774

Erythropoietin (Epo) is essential for the maintenance of a normal blood oxygen concentration and Epo synthesis is induced under anemic and hypoxic conditions.1,2 In the adult, Epo production by the kidney and liver accounts for approximately 90% and 10%, respectively, of total circulating Epo, but ectopic Epo expression was also found in the brain, uterus and testis, among others, where it is thought to have pleiotropic, organ-restricted and tissue protective functions.3-6 Renal Epo is synthesized by peritubular pericytes with fibroblastic and neuronal features, located in the juxtamedullary cortex.7-9 These cells respond to a decrease in tissue oxygen partial pressure by hypoxia-inducible factor (HIF)-2α stabilization and HIF-2dependent transcriptional induction of EPO gene expression.10 HIF bind to a hypoxia response element (HRE) containing the essential core consensus sequence 5'-RCGTG-3' which, however, is not sufficient to confer hypoxia-inducible gene expression.11 Distinct regulatory DNA elements enhance EPO promoter activity in liver and kidney. Transgenic mouse models showed that in the liver, 0.7 kb of the immediate 3'-flanking region is required, whereas in the kidney, the essential regulatory element resides between -14 and -6 kb in the distal 5'-region.12-15 The 3' HRE is well established and has been shown to be necessary and sufficient for liver-specific EPO gene expression after embryonic day 14.5.16,17 However, the DNA element responsible for kidney-specific EPO gene expression has been far less well characterized and nothing is known about the endogenous HRE in Epo-producing cells of other tissues. We recently discovered a strongly conserved distal 5' HRE and suggested that it haematologica | 2020; 105(12)

Multiple HRE co-operate in EPO gene expression

might contribute to oxygen-regulated EPO expression.18 This 5' HRE resides within a DNaseI hypersensitive site 9.2 kb upstream of the EPO transcriptional start site, contains both the 5'-ACGTG-3' core HIF DNA binding site as well as the ancillary 5'-CACA-3' element,11 and confers hypoxia-inducible exogenous reporter gene expression in Epo expressing and non-expressing cell lines.18 However, the organ-specific relevance of the endogenous 5' and 3' HRE, as well as their functional interaction with the EPO promoter, remained unknown. Considering the neuronal features that we and others reported for the renal Epo-producing cells,8,9 we therefore analyzed the relative contribution of the 5' and 3' HRE in Kelly cells, a human neuroblastoma cell line that has previously been shown to regulate the EPO gene in an oxygen and HIF-dependent manner.19 Endogenous HRE function was investigated by gene editing and HIF-DNA interaction studies. Intriguingly, novel promoter HRE were identified and profound differences in the functional co-operation between the 5', 3' and promoter HRE in Kelly cells were found when compared with the well-established human Hep3B hepatoma cell model.20

Methods

Luciferase reporter gene assays Reporter gene constructs have been described previously.18 Canonical 5'-RCGTG-3' HRE were replaced by 5'-RAAAG-3' using site-directed mutagenesis. Following transfection with lipofectamine, cells were incubated under normoxic or hypoxic (0.2% O2) conditions for 24 hours (h). Reporter gene assays were performed as described before21 using a luciferase assay kit (Promega, Madison, WI, USA). Luciferase activity was normalized to the protein content as determined by the Bradford assay.24

Hypoxia-inducible factor chromatin immunoprecipitation Chromatin immunoprecipitation (ChIP) experiments were performed as described previously,21,25,26 using the following rabbit polyclonal antibodies: anti-HIF-1α (PM14), anti-HIF-2α (PM9), anti-HIF-b (NB100-110; Novus Biologicals, Littleton, CO, USA), normal rabbit serum (X0902; Dako, Glostrup, Denmark). Co-precipitated DNA was quantified by real-time qPCR using the primers listed in Online Supplementary Table S1.

Statistical analysis All data are shown as mean+standard error of mean. Unpaired two-tailed Student t-tests were applied. P<0.05 was considered statistically significant.

Cell culture Human Hep3B and HepG2 hepatocellular carcinoma (American Type Culture Collection, LGC Standards, Wesel, Germany) and Kelly neuroblastoma (kindly provided by J. Fandrey, Essen, Germany) cell lines were cultured in high glucose Dulbecco's modified Eagle's medium (DMEM) and RPMI-1640 medium, respectively (Sigma Aldrich, Saint Louis, MO, USA), supplemented with 10% heat-inactivated fetal bovine serum (Thermo Fisher Scientific, Waltham, MA, USA), 50 IU/mL penicillin and 50 mg/mL streptomycin (Sigma Aldrich). Cells were exposed to hypoxic conditions using a workstation as previously described.21

Gene editing Clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated 9 (Cas9) gene editing was performed as reported22 and is described in detail in the Online Supplementary Methods.

RNA and protein analyses RNA was extracted and quantified by reverse-transcription (RT) real-time quantitative (q) polymerase chain reaction (PCR) as previously described.21,23 In brief, RT was performed with 2 mg total RNA and AffinityScript reverse transcriptase (Agilent), and the cDNA quantified using SYBR Green qPCR reagent kit (Kapa Biosystems, London, UK) in a MX3000P light cycler (Agilent). Transcripts levels were calculated through comparison with calibrated standard curves and normalized to human ribosomal protein L28 mRNA. Primers used for RT-qPCR are listed in Online Supplementary Table S1. Epo protein was detected by ELISA according to the manufacturer's protocol (R&D Systems, Minneapolis, MN, USA). Immunoblotting was performed as previously described21 using the following primary antibodies: mouse monoclonal anti-HIF-1α (#610959; BD Transduction Laboratories, San Jose, CA, USA), rabbit monoclonal anti-HIF2α (#PAB12124; Abnova, Taipei, Taiwan), mouse monoclonal anti-HIF-b (D28F3; Cell Signaling Technology, Danvers, MA, USA), mouse monoclonal anti-b-actin (A5441; Sigma Aldrich). Secondary antibodies were HRP-conjugated goat polyclonal anti-rabbit or anti-mouse IgG (#31460 and #31430, respectively; Thermo Fisher Scientific). haematologica | 2020; 105(12)

Results Mutation of the EPO 5' and 3' hypoxia response elements by gene editing CRISPR-Cas9 technology was used for the specific destruction of the endogenous EPO HRE to dissect the relative contribution of the -9.2 kb 5' HRE and the +3.0 kb 3' HRE to hypoxia-inducible EPO gene expression in neuronal and hepatic cell lines (Figure 1A). sgRNAs were designed to target the HIF-binding core sequence 5'ACGTG-3' (Figure 1B).11 The restriction enzyme TaiI was used to assess the presence of the 5'-ACGT-3' sequence.27 Successful HRE destruction confers TaiI resistance to the PCR products as exemplified for Kelly cells in Figure 1C and D. Monoclonal cell lines were obtained by limiting dilution cloning of initial partially gene edited polyclonal cell pools. The HRE regions were again amplified by PCR, re-tested for complete TaiI resistance, and cloned into plasmid vectors. Multiple independent plasmids were sequenced to ensure biallelic HRE inactivation. Alternatively, amplicons were directly deep sequenced. Clone verification and sequence information is provided in Online Supplementary Figure S1. Note that, in some cases (Kelly-3'B2, Hep3B-5'H11), more than two mutant alleles were detected, which may be due to either polyclonal cell lines, polyploidy, gene amplification or genetic drift. Other clones maintained wild-type HRE sequences (Kelly5'C4, Kelly-3'C4, Hep3B-5'A5, HepG2-5'H8, HepG23'D6; confirmed by sequencing of PCR products) and were included as additional controls in subsequent experiments.

The 5' hypoxia response element contributes to hypoxic EPO induction in neuronal but not hepatic cells The neuroblastoma cell line Kelly has been reported to induce endogenous Epo mRNA levels, at 24 h, by more than 71-fold under hypoxic conditions (3% O2) and more than 238-fold under anoxic conditions.19 Kelly cells therefore recapitulate the well-known hypoxia-inducible EPO 2775

I.M.C. Orlando et al. A

Figure 1. EPO 5' and 3' hypoxia response elements (HRE) gene editing. (A) Scheme depicting the human EPO locus. HSS: hypersensitivity site; TSS: transcriptional start site; KIE: kidney inducible element; NRE: negative regulatory element; LIE: liver inducible element; NRLE: negative regulatory liver element. (B) Single-guided (sg) RNA sequences designed to target the EPO 5' and 3' HRE by CRISPR-Cas9. In bold: consensus HIF binding site; arrows: polymerase chain reaction (PCR) primers used to amplify the HRE regions; fwd: forward; rev: reverse; wt: wild-type; mt: mutant. Mutation of the consensus HRE confers resistance to TaiI restriction digestion as shown by agarose gel analysis of digested or undigested PCR products using as template genomic DNA isolated from Kelly neuroblastoma cells following 5' (C) or 3' (D) HRE gene editing. M: marker; PL: polyclonal pool of cells; B4 and C3: 5' HRE mutant clones; T1 and F3: 3' HRE mutant clones; ctr: non-template control.

gene expression in the brain.4 Epo mRNA and secreted protein were induced by 1,603-fold and 22-fold, respectively, following exposure of wild-type Kelly cells to hypoxia (0.2% O2) for 24 h (Figure 2). However, due to the very low basal Epo transcript levels, the fold induction factor is subject to variation and must be interpreted with caution. Indeed, the wild-type HRE subclones Kelly-5'C4 and Kelly-3'C4 showed reduced Epo mRNA induction, even though the hypoxic Epo mRNA levels were not significantly different from the maternal Kelly cells (Figure 2A). These results were confirmed on the protein level where the normoxic levels also varied greatly but the hypoxic levels were indistinguishable from the maternal cell line (Figure 2B). Therefore, only hypoxic Epo levels were considered for subsequent analyses. Interestingly, all mutant 5' and 3' HRE Kelly clones showed significantly reduced hypoxic Epo expression, on the mRNA (Figure 2A) as well as on the protein (Figure 2B) level. This is in striking contrast to mutant Hep3B (Figure 2C and D) and HepG2 (Figure 2E) clones where only 3' but not 5' HRE mutations strongly reduced hypoxic Epo mRNA and protein levels. Note that while Kelly and Hep3B cells were exposed to hypoxia for 24 h, HepG2 cells showed maximal Epo mRNA induction already after 8 h, with approximately 10-fold lower mRNA levels, which was not sufficient for detectable Epo protein accu2776

mulation in the supernatant. To ensure that no general offtarget effects caused these results, the mRNA levels of the HIF-1 and HIF-2 target genes CAIX and LOXL2 (Kelly) or PAI1 (Hep3B and HepG2), respectively,28 were measured in the same samples. EPO HRE mutations did not significantly reduce the expression of these genes under hypoxic conditions (Online Supplementary Figure S2). Moreover, also hypoxic HIF-1Îą and HIF-2Îą protein stabilization was not altered by the EPO HRE mutations (Online Supplementary Figure S3). These results suggest that both the 5' and 3' HRE contribute to hypoxia-inducible EPO gene expression in neuronal cells whereas in hepatic cells only the 3' HRE is required.

5' and 3' EPO hypoxia response elements co-operatively enhance hypoxic reporter gene expression in neuronal cells We had previously shown that the EPO 5' HRE confers hypoxia-inducible expression to a heterologous SV40 promoter-driven reporter gene in both Hep3B and Kelly cells.18 These results stand in apparent contrast to the 5' HRE mutation data presented above and imply differences between exogenous bacterial reporter gene plasmids and endogenous chromatin regulation. In fact, similar reporter gene results have been obtained even in non-Epo-expressing HeLa and HK-2 cells18 as well as in Hek293 and CHO haematologica | 2020; 105(12)

Multiple HRE co-operate in EPO gene expression

E Figure 2. Hypoxia-inducible EPO gene expression in 5' or 3' hypoxia response elements (HRE) mutant neuroblastoma and hepatoma cells. EPO expression under normoxic or hypoxic conditions (0.2% O2; 24 hours [h] for Kelly and Hep3B, 8 h for HepG2) was measured in Kelly (A and B), Hep3B (C and D) and HepG2 (E) cells by reverse-transcription (RT) real-time quantitative (q) polymerase chain reaction (PCR) on the mRNA level (A, C and E) and by supernatant ELISA on the protein level (B and D). Epo mRNA levels were normalized to ribosomal protein L28 mRNA levels. Numbers above the columns indicate hypoxic induction factors. Data are shown as mean+standard error of mean of three independent experiments. Student t-tests were used to statistically evaluate the difference between hypoxic gene edited cells and hypoxic wild-type (WT) cells. *P<0.05, **P<0.01, ***P<0.001. n.d.: not detectable; Pool: polyclonal pool of cells; +/+: subclones containing two wild-type alleles; -/-: subclones with biallelic HRE mutation.

cells (data not shown). We therefore expanded these experiments by transiently transfecting Kelly cells with minimal (138 bp) EPO promoter-driven reporter genes, enhanced by various DNA fragments containing the 5' and/or 3' HRE. In contrast to our previous results obtained with Hep3B cells,18 EPO promoter-driven luciferase activity in Kelly cells was significantly further increased when the 100 bp fragment containing the 5' HRE was combined with a 126 bp fragment containing the 3' HRE (Figure 3A). While a longer 3 kb DNA fragment containing the 5' HRE reduced both hypoxic and normoxic reporter gene expression driven by the EPO promoter/3' HRE, it actually led to even higher hypoxic induction factors. These data imply co-operation between the EPO 5' and 3' HRE specifically in neuronal cells, and suggest that additional distal and proximal 5' flanking elements contribute to tissue-specific and conditional EPO regulation. haematologica | 2020; 105(12)

A novel EPO promoter hypoxia response element contributes to hypoxic reporter gene expression in neuronal cells Interestingly, a slight but significant hypoxic induction of the minimal EPO promoter could be seen in Kelly cells (Figure 3A) which we previously had not observed in Hep3B cells.18 Inspection of the 147 bp fragment containing the 138 bp EPO promoter revealed a tandem dimeric repeat with two previously unreported putative HRE (Figure 3B). These putative HRE are highly conserved and locate close to WT1 and GATA binding sites (Online Supplementary Figure S4). In combination with the 5' and 3' HRE enhancers, luciferase expression driven by this promoter fragment was induced 152-fold by hypoxia in this experimental series. Mutation of either promoter HRE (pHRE) 1 or pHRE2 significantly reduced hypoxic luciferase expression levels, whereas the effect on the 2777

I.M.C. Orlando et al.

Figure 3. EPO promoter-driven and hypoxia response elements (HRE)-enhanced luciferase reporter gene expression. Kelly (A and C) and Hep3B (D) cells were transiently transfected with the indicated reporter gene constructs and exposed for 24 hours to normoxic or hypoxic (0.2% O2) conditions. Luciferase activities were normalized to the protein content and shown as mean+standard error of mean of six (A and C) or three (D) independent experiments. Numbers on the right of the bars indicate hypoxic induction factors. Student t-tests were used to statistically assess hypoxic EPO promoter activities (EpoProm, wild-type; mtpHRE, mutant) and the effects of combining the 5' with the 3' HRE. *P<0.05, **P<0.01, ***P<0.001. (B) Sequence of the 147 bp human EPO promoter fragment (starting 138 bp upstream of the transcription start site; TSS), indicating a conserved tandem dimeric repeat (underlined), each containing a putative promoter HRE (pHRE1 and pHRE2).

2778

haematologica | 2020; 105(12)

Multiple HRE co-operate in EPO gene expression

Figure 4. Hypoxia-inducible factor (HIF) binding to the EPO locus in 5' or 3' hypoxia response elements (HRE) mutant neuroblastoma and hepatoma cells. (A) Scheme depicting the regions of the EPO locus analyzed for HIF interaction by chromatin immunoprecipitation (ChIP), including the 5' and 3' HRE, the minimal promoter and a negative (-Ctrl) control region devoid of any HIF binding. Kelly (B) and Hep3B (C) cells were exposed for 24 hours to normoxic or hypoxic (0.2% O2) conditions, followed by ChIP using antibodies derived against HIF-1α, HIF-2α or HIF-b, or a negative control serum. The promoter HRE of the established HIF target genes NDRG1 and PAI1 served as control for ChIP efficiency. The amount of co-precipitated DNA was determined by real-time quantitative (q) polymerase chain reaction (PCR) and displayed relative to input. (D) These experiments were repeated with hypoxic wild-type (WT), 5' or 3' HRE mutant Hep3B cells. Mean+standard error of mean of three (Kelly) or four (Hep3B) independent experiments are shown. Student t-tests were performed to statistically evaluate the difference between the hypoxic ChIP samples and hypoxic serum controls. *P<0.05, **P<0.01, ***P<0.001. (E) Immunoblot confirmation of similar HIF-b protein levels in the hypoxic Hep3B subclones used above. b-Actin was used as loading and blotting control.

hypoxic induction factors was less conclusive (Figure 3C). Double pHRE1 and pHRE2 mutation did not further decrease reporter gene expression, suggesting that this tandem dimeric sequence in the EPO promoter acts as a single HRE. Similar results were obtained with Hep3B cells (Figure 3D), though absolute as well as hypoxicallyinduced reporter gene levels were much lower than in Kelly cells.

Hypoxia-inducible factor interacts with the EPO promoter and the 3', but not the 5', hypoxia response element in neuronal cells To directly analyze the interaction between HIF and the haematologica | 2020; 105(12)

various HRE of the endogenous EPO locus, ChIP followed by qPCR experiments were performed. Because of the contribution of the promoter to the hypoxic activation of reporter gene expression observed above, we also included the proximal 5' region in these ChIP-qPCR experiments (Figure 4A). Unexpectedly, Kelly cells did not show any significant binding of the HIF subunits HIF-1α, HIF-2α or HIF-b to the 5' HRE, whereas a significant hypoxic increase in HIF-2α/HIF-b binding to the 3' HRE could be detected (Figure 4B). Surprisingly, however, the strongest HIF interaction in hypoxic Kelly cells was observed with the EPO promoter region, again with a preference for HIF2α/HIF-b. In contrast, in hypoxic Hep3B cells the strongest 2779

I.M.C. Orlando et al.

HIF-2α/HIF-b interaction was found at the 3' HRE and only a weak interaction with the promoter region could be observed (Figure 4C). Similar findings had previously been reported for hypoxic HepG2 cells using ChIPsequencing,29 whereas non-Epo-producing cell lines did

not show any HIF binding to the EPO locus (Online Supplementary Figure S5). To control for ChIP efficiency, the presence of DNA fragments containing the HRE of the HIF-1/2 target genes NDRG1 and PAI121,25 was quantified by qPCR in the same samples. HIF binding to the estab-

Figure 5. Hypoxia-inducible factor (HIF) binding to the EPO locus and Epo regulation in 3'/5' hypoxia response elements (HRE) double-mutant neuroblastoma cells. (A and B) of hypoxic wild-type (WT) or mutant (as indicated) Kelly cells using antibodies derived against HIF-b or negative control serum. The amount of co-precipitated DNA was determined by real-time quantitative polymerase chain reaction ((RT-qPCR) covering the indicated regions of the EPO locus (A) or the HRE of the NDRG1 control locus (B) and displayed relative to input. Data are shown as mean+standard error of mean of four independent experiments. Student t-tests were performed to statistically evaluate the difference to hypoxic WT (*P<0.05) or 3' B2 (#P<0.05) or 5' C3 cells (§P<0.05, §§P<0.01). (C) Immunoblot confirmation of similar HIF-α protein levels in the hypoxic Kelly subclones used above. b-Actin was used as loading and blotting control. (D and E) Epo production in WT, single 3' B2 or double 3' B2 5' HRE mutant Kelly cells exposed for 24 hours to normoxia or hypoxia (0.2% O2) was measured by RT-qPCR on the mRNA level (D) or by ELISA on the protein level (E). Data are shown as mean+standard error of mean of 3-6 (D) or 5-10 (E) independent experiments. Student t-tests were used to statistically evaluate the difference to hypoxic WT (**P<0.01, ***P<0.001) or 3' B2 cells (##P<0.01; ###P<0.001; ns: not significant; n.d.: not detectable).

2780

haematologica | 2020; 105(12)

Multiple HRE co-operate in EPO gene expression

Figure 6. EPO promoter/enhancer-driven luciferase reporter gene expression following single and combined hypoxia response elements (HRE) mutations. Kelly (A and B) and Hep3B (C and D) cells were transiently transfected with the indicated reporter gene constructs and exposed for 24 hours to normoxic or hypoxic (0.2% O2) conditions. Luciferase activities were normalized to the protein content and shown as mean+standard error of mean of four independent experiments. Numbers on the right of the bars indicate hypoxic induction factors. Student t-tests were used to statistically assess hypoxic EPO promoter activities. *P<0.05, **P<0.01, ***P<0.001.

haematologica | 2020; 105(12)

2781

I.M.C. Orlando et al.

lished NDRG1 and PAI1 promoter HRE was comparable to the Epo HRE, and anti-HIF-b antibodies precipitated roughly additive amounts of chromatin seen with the HIF1α and HIF-2α antibodies (Figures 4B and C).

Functional distal 3' and 5' hypoxia response elements are required for remote control of hypoxia-inducible factor binding to the EPO promoter Because ChIP resolution is insufficient to detect the precise protein-DNA interaction site, we next wanted to confirm that HIF actually bound the core sequence of the investigated HRE. Therefore, Hep3B 3' and 5' mutant HRE subclones were exposed to hypoxia and used for ChIP with anti-HIF-b antibodies to precipitate both HIF-1 and HIF-2. As expected, HIF bound the wild-type 3' HRE but did not interact with the mutant 3' HRE (Figure 4D). Of note, the 3' HRE mutation apparently also decreased the weak HIF interaction with the EPO promoter region. Consistent with its non-functionality in Hep3B cells, mutation of the 5' HRE did not influence HIF binding to any of the tested HRE (Figure 4D). Neither HIF-b binding to the PAI1 control HRE (Figure 4D), nor HIF-b protein levels (Figure 4E), were altered in these Hep3B subclones, confirming that HIF-b expression is not generally affected by the 3' and 5' HRE mutations, and that HIF-b ChIP reliably detects HIF-DNA interactions. To further investigate the unexpected inhibitory effect of the distal 3' HRE mutation on the HIF interaction with the proximal EPO promoter region in Hep3B cells, these experiments were repeated with Kelly 3' and 5' mutant HRE subclones. Intriguingly, the 3' HRE mutation (clone B2) not only completely abrogated the interaction between HIF and the 3' HRE, but also strongly decreased the interaction with the promoter region (Figure 5A). Moreover, the 5' HRE mutation (clone C3) similarly impaired HIF interaction with the 3' HRE and the promoter region (Figure 5A), while comparable HIF binding to the NDRG1 control HRE (Figure 5B) and HIF-b protein levels (Figure 5C) could be detected in these mutant Kelly subclones.

reversed the inhibitory effect of the 3' HRE mutation on HIF promoter interaction. HIF-b binding to the NDRG1 control HRE (Figure 5B), as well as HIF-b protein levels (Figure 5C), remained at similar levels, confirming that HIF-b expression is not generally affected in these Kelly subclones, and that HIF-b ChIP reliably detects HIF-DNA interactions. Consistent with the ChIP data, the additional larger or smaller 5' HRE deletions further decreased or partly reversed, respectively, Epo mRNA (Figure 5D) and protein (Figure 5E) levels in hypoxic 3' HRE mutant Kelly cells.

The core hypoxia-inducible factor-binding sites of all EPO hypoxia response elements are required for hypoxic reporter gene expression To further dissect the relevance of the various EPO HRE, the conserved core HIF-binding sites were mutated (5'CGTG-3' to 5'-AAAG-3') in EPO promoter/enhancer-driven reporter gene constructs. Mutation of either the 5' or the 3' HRE significantly reduced hypoxic reporter gene expression in Kelly cells, with the 5' HRE showing a stronger effect than the 3' HRE. Simultaneous mutation of both sites completely abrogated hypoxic reporter gene expression (Figure 6A). Despite the very weak activity of the minimal EPO promoter (Figure 3A), even the double mutation of the two EPO promoter HRE significantly reduced hypoxic reporter gene expression in these constructs (Figure 6B). Additional 5' or 3' HRE single or double mutations further reduced or abrogated, respectively, hypoxic reporter gene expression (Figure 6B). In contrast to the endogenous EPO 5' and/or 3' HRE deletions by gene editing (see above), both HRE contribute to hypoxic EPO promoter induction in exogenous reporter genes transfected into Hep3B cells (Figure 6C), as we have previously reported for these cells as well as for non-Epoproducing cells.18 Both absolute and hypoxically-induced reporter gene levels were much lower than in Kelly cells. Again, double mutation of the EPO promoter HRE significantly reduced hypoxic reporter gene expression, which could only be completely abolished when the 5' and 3' HRE were additionally mutated (Figure 6D).

Large but not small deletions of the 5' hypoxia response element abrogate Epo induction in 3'/5' hypoxia response element double-mutant neuronal cells

Discussion

Because there was residual HIF binding to the EPO promoter in the single 3' and 5' HRE mutant Kelly subclones, we wondered whether 3'/5' HRE double mutations would lead to a further decrease in the HIF interaction with the EPO promoter. The introduction of a secondary mutation was rather inefficient and worked only in the B2 3' HRE mutant Kelly subclone. Deep sequencing of the PCR products derived from the mutant 5' HRE revealed 3 subclones with large (31 bp and 18 bp; varying ratios) and 3 subclones with small (1, 2 or 3 bp; biallelic) deletions (Online Supplementary Figure S6A). Despite some reduction in CAIX mRNA and HIF-1α protein in the subclones containing the larger deletions, the double-mutant subclones did not show any major general differences in mRNA levels of the HIF target genes CAIX and LOXL2, or in hypoxic HIF-1α and HIF-2α protein levels (Online Supplementary Figure S6B-E). In hypoxic Kelly cells, only the larger but not the smaller 5' HRE deletions further decreased HIF binding to the EPO promoter in the 3'/5' HRE double-mutant subclones (Figure 5A). The small core 5' HRE mutations partly

The data reported here reveal a complex and organ-specific interplay between various HRE at the EPO locus. The previously described mechanisms of EPO regulation in hepatic cells were mostly confirmed by our experiments, including a strong functional dependence on the endogenous 3' HRE with a robust HIF interaction, and a complete lack of requirement for the endogenous 5' HRE despite a strong hypoxic enhancer function in exogenous reporter gene experiments. In neuronal cells, the endogenous 5' HRE seems to be as important as the 3' HRE for hypoxic induction of EPO gene expression, and the 5' and 3' HRE co-operate in hypoxic enhancement of exogenous reporter gene expression, but the 5' HRE does not directly interact with HIF. In contrast to hepatic cells, where the EPO promoter is only slightly bound by HIF and does not confer hypoxic reporter gene induction,18 in neuronal cells it is the EPO promoter that most strongly interacts with HIF, consistent with a weak hypoxic induction of reporter genes. EPO minimal promoter fragments spanning at least -91 bp upstream of the transcriptional start site have

2782

haematologica | 2020; 105(12)

Multiple HRE co-operate in EPO gene expression

previously been shown to be hypoxia responsive in reporter gene experiments also in Hep3B cells.30,31 It is still not clear why we repeatedly found significant hypoxic promoter induction only in Kelly but not in Hep3B cells. Inspection of the minimal EPO promoter region revealed two potential HRE characterized by the presence of evolutionary conserved 5'-GCGTG-3' HIF binding motifs. Mutation of either of these two HRE in Kelly cells reduced both basal and hypoxia-inducible promoter activity enhanced by the 5' and 3' HRE. Tissue-specific transcriptional co-activators and/or co-repressors may be involved in the difference in hypoxic promoter activity and HIF binding between hepatic and neuronal cells. Indeed, a co-operative interaction with additional transcription factors binding in close proximity to HIF is typical for HRE and has also been found in other HIF target genes.21 While the ancillary 5'-CACA-3' element11 is missing, its supportive function may be replaced by the neighboring DNA binding sites for GATA factors and the Wilms tumor gene (WT1) product, which are well-known to regulate the EPO promoter.32-35 It has recently been shown that WT1 is itself hypoxia-inducible and restricted to neuroblastoma and kidney-derived cell lines,9,36 which might contribute to both hypoxia-inducible as well as tissue-specific EPO promoter activity. Alternatively, it is possible that binding of HIF to the promoter and/or transcriptional activation may be blocked by binding of an as yet unknown tissue-specific factor in Hep3B cells. Of note, we have previously shown that ATF/CREB family members are able to directly interact with the HRE core motif,37 and it may be that such an interaction also occurs at the promoter HRE. How can it be explained that a conserved HRE consensus core sequence, including the ancillary element, is functionally relevant for hypoxic induction of gene expression but not directly bound by HIF, as appears to be the case for the EPO 5' HRE? We have previously shown for PAG1, another HIF-2 target gene, that a single distal 82 kb 5' HRE resides in an isolated DNA region, bound by many additional transcription factors, and forms multiple chromatin loops both locally and over a long distance with the promoter region.21 While in this case HIF-2Îą did interact with the HRE, neither hypoxia nor the presence of HIF was needed for the long-range chromatin interaction with the promoter region. Only the presence of the core 5'-ACGTG-3' HIF binding DNA sequence was required to maintain this interaction, suggesting that preformed chromatin loops enable oxygen-regulated conditional gene regulation. This model has subsequently been confirmed for many other HIF target genes by genomewide approaches.38,39 Therefore, the EPO 5' HRE might well be functionally required for hypoxia-inducible gene expression by maintaining a constitutive chromatin architecture that supports promoter activity in a cell type-specific manner. The finding that only additional large but not small 5' HRE deletions fully abrogated endogenous Epo induction and HIF:promoter interaction in 3' HRE mutant Kelly cells suggests that, like in the case of the PAG1 gene, additional transcription factors bind close to the consensus HRE sequence and are involved in chromatin looping and trans-activation of the EPO promoter. We still have no explanation why a small 5' HRE deletion alone inhibited HIF-promoter interaction, but in combination with a 3' HRE mutation partially rescued the inhibitory effect of the 3' HRE mutation. Nonetheless, it haematologica | 2020; 105(12)

is without precedent that the extended 5' HRE strongly cis-enhanced HIF binding to the promoter and 3' HRE. The differences between Epo-producing hepatic and neuronal cells raise the question of how EPO is regulated in renal Epo-producing (REP) cells, the main source of circulating Epo. The overlap of neuronal markers with genetically tagged mouse REP cells8,9 suggests that Kelly cells may represent a better model than hepatoma cells to recapitulate human oxygen-regulated EPO gene expression in the kidney, at least concerning the enigmatic 5' regulatory regions. Indeed, deletion of the endogenous 3' HRE in the mouse suppressed only hepatic but not renal Epo expression, suggesting the presence of one or more additonal HRE in REP cells.17 Consistently, deletion of a 17.4 to -3.6 kb Epo 5' region in transgenic mouse models abrogated Epo gene expression specifically in REP cells whereas it was dispensable for Epo expression in the brain.31 Mutation of the mouse -8.3 kb 5' HRE (corresponds to the human -9.2 kb 5' HRE) within this Epo 5' region did not abrogate transgenic Epo gene expression in the kidney (brain was not analyzed) and a minimal 0.3 kb fragment containing this 5' HRE was not sufficient to drive transgenic GFP expression in the mouse kidney.31 Our results on the co-operation between 5', 3' and promoter HRE in neuronal and hepatic cell lines may help to explain these findings in transgenic mice: while the deletion of the 3' HRE in the liver is sufficient to abrogate hypoxic Epo expression, only the combined deletion of the extended 5' and the minimal 3' (and maybe the promoter) HRE may affect Epo expression in the kidney. Altogether, these results demonstrate that several HRE are involved in oxygen-regulated EPO gene expression. While in the liver the 3' HRE is both necessary and sufficient, in the kidney the 3' HRE is not essential and a 5' HRE acts in concert with additional long-range enhancer elements, and probably local chromatin structure. Although transgenic mouse models are still lacking, our results suggest that in the brain, both of these distal HRE functionally co-operate with HIF binding to promoter HRE. Our results further illustrate that not all HRE act in a canonical way by directly binding HIF, and that HREpromoter co-operations in oxygen-regulated gene expression need to be analyzed in a cell type-specific context and cannot be generalized based on the results obtained with a single cell line. Similar results have recently been reported for the MALAT1 locus, expressing a lncRNA which is strongly hypoxia-inducible,40 where HIF-dependent and independent long-range interactions contribute to hypoxia-inducibility in a cell-type specific manner.41 Finally, our findings are also relevant for disease-associated polymorphisms that either create or delete potential HRE26,42,43 and may therefore interfere with long-range chromatin interactions. Acknowledgments The authors wish to thank J. Fandrey for providing cells, KA Nolan for critical comments and R. Hunkeler for technical assistance. Funding This work was supported by the National Centre of Competence in Research "Kidney.CH" and the Swiss National Science Foundation (310030_184813 to RHW), Cancer Research UK (A416016 to DRM) and the National Institute for Health Research (NIHR-RP-2016-06-004 to DRM). 2783

I.M.C. Orlando et al.

References 16. 1. Wenger RH, Hoogewijs D. Regulated oxygen sensing by protein hydroxylation in renal erythropoietin-producing cells. Am J Physiol Renal Physiol. 2010;298(6):F12871296. 2. Wenger RH, Kurtz A. Erythropoietin. Compr Physiol. 2011;1:1759-1794. 3. Tan CC, Eckardt KU, Ratcliffe PJ. Organ distribution of erythropoietin messenger RNA in normal and uremic rats. Kidney Int. 1991; 40(1):69-76. 4. Marti HH, Wenger RH, Rivas LA, et al. Erythropoietin gene expression in human, monkey and murine brain. Eur J Neurosci. 1996;8(4):666-676. 5. Yasuda Y, Masuda S, Chikuma M, Inoue K, Nagao M, Sasaki R. Estrogen-dependent production of erythropoietin in uterus and its implication in uterine angiogenesis. J Biol Chem. 1998;273(39):25381-25387. 6. Magnanti M, Gandini O, Giuliani L, et al. Erythropoietin expression in primary rat Sertoli and peritubular myoid cells. Blood. 2001;98(9):2872-2874. 7. Maxwell PH, Osmond MK, Pugh CW, et al. Identification of the renal erythropoietinproducing cells using transgenic mice. Kidney Int. 1993;44(5):1149-1162. 8. Obara N, Suzuki N, Kim K, Nagasawa T, Imagawa S, Yamamoto M. Repression via the GATA box is essential for tissue-specific erythropoietin gene expression. Blood. 2008;111(10):5223-5232. 9. Imeri F, Nolan KA, Bapst AM, et al. Generation of renal Epo-producing cell lines by conditional gene tagging reveals rapid HIF-2 driven Epo kinetics, cell autonomous feedback regulation, and a telocyte phenotype. Kidney Int. 2019;95(2):375-387. 10. Kapitsinou PP, Liu Q, Unger TL, et al. Hepatic HIF-2 regulates erythropoietic responses to hypoxia in renal anemia. Blood. 2010;116(16):3039-3048. 11. Wenger RH, Stiehl DP, Camenisch G. Integration of oxygen signaling at the consensus HRE. Sci STKE. 2005;2005(306):re12. 12. Semenza GL, Dureza RC, Traystman MD, Gearhart JD, Antonarakis SE. Human erythropoietin gene expression in transgenic mice: multiple transcription initiation sites and cis-acting regulatory elements. Mol Cell Biol. 1990;10(3):930-938. 13. Semenza GL, Nejfelt MK, Chi SM, Antonarakis SE. Hypoxia-inducible nuclear factors bind to an enhancer element located 3' to the human erythropoietin gene. Proc Natl Acad Sci U S A. 1991;88(13): 5680-5684. 14. Madan A, Lin C, Hatch SLI, Curtin PT. Regulated basal, inducible, and tissue-specific human erythropoietin gene expression in transgenic mice requires multiple cis DNA sequences. Blood. 1995;85(10): 2735-2741. 15. Köchling J, Curtin PT, Madan A. Regulation of human erythropoietin gene induction by

2784

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

upstream flanking sequences in transgenic mice. Br J Haematol. 1998;103(4):960-968. Semenza GL, Wang GL. A nuclear factor induced by hypoxia via de novo protein synthesis binds to the human erythropoietin gene enhancer at a site required for transcriptional activation. Mol Cell Biol. 1992;12(12):5447-5454. Suzuki N, Obara N, Pan X, et al. Specific contribution of the erythropoietin gene 3' enhancer to hepatic erythropoiesis after late embryonic stages. Mol Cell Biol. 2011;31 (18):3896-3905. Storti F, Santambrogio S, Crowther L, et al. A novel distal upstream hypoxia response element regulating oxygen-dependent erythropoietin gene expression. Haematologica. 2014;99(4):e45-e48. Stolze I, Berchner-Pfannschmidt U, Freitag P, et al. Hypoxia-inducible erythropoietin gene expression in human neuroblastoma cells. Blood. 2002;100(7):2623-2628. Goldberg MA, Glass GA, Cunningham JM, Bunn HF. The regulated expression of erythropoietin by two human hepatoma cell lines. Proc Natl Acad Sci U S A. 1987;84(22):7972-7976. Schörg A, Santambrogio S, Platt JL, et al. Destruction of a distal hypoxia response element abolishes trans-activation of the PAG1 gene mediated by HIF-independent chromatin looping. Nucl Acids Res. 2015;43(12):5810-5823. Harms DW, Quadros RM, Seruggia D, et al. Mouse genome editing using the CRISPR/Cas system. Curr Protoc Hum Genet. 2014;83:15. Wollenick K, Hu J, Kristiansen G, et al. Synthetic transactivation screening reveals ETV4 as broad coactivator of hypoxiainducible factor signaling. Nucl Acids Res. 2012;40(5):1928-1943. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding. Anal Biochem. 1976;72:248254. Schödel J, Oikonomopoulos S, Ragoussis J, Pugh CW, Ratcliffe PJ, Mole DR. High-resolution genome-wide mapping of HIF-binding sites by ChIP-seq. Blood. 2011; 117(23):e207-217. Schödel J, Bardella C, Sciesielski LK, et al. Common genetic variants at the 11q13.3 renal cancer susceptibility locus influence binding of HIF to an enhancer of cyclin D1 expression. Nat Genet. 2012;44(4):420-425. Wenger RH, Kvietikova I, Rolfs A, Camenisch G, Gassmann M. Oxygen-regulated erythropoietin gene expression is dependent on a CpG methylation-free hypoxia-inducible factor-1 DNA-binding site. Eur J Biochem. 1998;253(3):771-777. Stiehl DP, Bordoli MR, Abreu-Rodríguez I, et al. Non-canonical HIF-2α function drives autonomous breast cancer cell growth via an AREG-EGFR/ErbB4 autocrine loop. Oncogene. 2012;31(18):2283-2297. Smythies JA, Sun M, Masson N, et al.

30.

31.

32.

33.

34. 35.

36.

37.

38.

39.

40.

41.

42.

43.

Inherent DNA-binding specificities of the HIF-1α and HIF-2α transcription factors in chromatin. EMBO Rep. 2019;20(1). Blanchard KL, Acquaviva AM, Galson DL, Bunn HF. Hypoxic induction of the human erythropoietin gene: cooperation between the promoter and enhancer, each of which contains steroid receptor response elements. Mol Cell Biol. 1992;12(12):5373-5385. Hirano I, Suzuki N, Yamazaki S, et al. Renal anemia model mouse established by transgenic rescue with an erythropoietin gene lacking kidney-specific regulatory elements. Mol Cell Biol. 2017;37(4):e00451-00416. Imagawa S, Yamamoto M, Miura Y. Negative regulation of the erythropoietin gene expression by the GATA transcription factors. Blood. 1997;89(4):1430-1439. Imagawa S, Suzuki N, Ohmine K, et al. GATA suppresses erythropoietin gene expression through GATA site in mouse erythropoietin gene promoter. Int J Hematol. 2002;75(4):376-381. Dame C, Sola MC, Lim KC, et al. Hepatic erythropoietin gene regulation by GATA-4. J Biol Chem. 2004;279(4):2955-2961. Dame C, Kirschner KM, Bartz KV, Wallach T, Hussels CS, Scholz H. Wilms tumor suppressor, Wt1, is a transcriptional activator of the erythropoietin gene. Blood. 2006; 107(11):4282-4290. Krueger K, Catanese L, Sciesielski LK, Kirschner KM, Scholz H. Deletion of an intronic HIF-2α binding site suppresses hypoxia-induced WT1 expression. Biochim Biophys Acta Gene Regul Mech. 2019; 1862(1):71-83. Kvietikova I, Wenger RH, Marti HH, Gassmann M. The transcription factors ATF-1 and CREB-1 bind constitutively to the hypoxia-inducible factor-1 (HIF-1) DNA recognition site. Nucl Acids Res. 1995;23(22): 4542-4550. Platt JL, Salama R, Smythies J, et al. CaptureC reveals preformed chromatin interactions between HIF-binding sites and distant promoters. EMBO Rep. 2016;17(10):1410-1421. Niskanen H, Tuszynska I, Zaborowski R, et al. Endothelial cell differentiation is encompassed by changes in long range interactions between inactive chromatin regions. Nucl Acids Res. 2018;46(4):1724-1740. Lelli A, Nolan KA, Santambrogio S, et al. Induction of long noncoding RNA MALAT1 in hypoxic mice. Hypoxia (Auckl). 2015;3:45-52. Stone JK, Kim JH, Vukadin L, et al. Hypoxia induces cancer cell-specific chromatin interactions and increases MALAT1 expression in breast cancer cells. J Biol Chem. 2019; 294(29):11213-11224. Grampp S, Platt JL, Lauer V, et al. Genetic variation at the 8q24.21 renal cancer susceptibility locus affects HIF binding to a MYC enhancer. Nat Commun. 2016;7:13183. Grampp S, Schmid V, Salama R, et al. Multiple renal cancer susceptibility polymorphisms modulate the HIF pathway. PLoS Genet. 2017;13(7):e1006872.

haematologica | 2020; 105(12)

ARTICLE

Myelodysplastic Syndromes

Eltrombopag monotherapy can improve hematopoiesis in patients with low to intermediate risk-1 myelodysplastic syndrome Alana Vicente,1* Bhavisha A. Patel,1* Fernanda Gutierrez-Rodrigues,1 Emma M. Groarke,1 Valentina Giudice,1 Jennifer Lotter,1 Xingmin Feng,1 Sachiko Kajigaya,1 Barbara Weinstein,1 Evette Barranta,1 Matthew J. Olnes,1 Ankur R. Parikh,1 Maher Albitar,2 Colin O. Wu,3 Ruba Shalhoub,3 Katherine R. Calvo,4 Danielle M. Townsley,1 Phillip Scheinberg,5 Cynthia E. Dunbar,1 Neal S. Young1 and Thomas Winkler1

Hematology Branch, National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health (NIH), Bethesda, MD, USA; 2Genomics Testing Cooperative Laboratories, Irvine, CA, USA; 3Office of Biostatistics Research, National Institutes of Health (NIH), Bethesda, MD, USA; 4Hematology Section, Department of Laboratory Medicine, Clinical Center, National Institutes of Health (NIH), Bethesda, MD, USA and 5 Hematology Department, Hospital A Beneficiencia Portuguesa, Sao Paulo, Brazil 1

Ferrata Storti Foundation

Haematologica 2020 Volume 105(12):2785-2794

*AV and BAP contributed equally as co-first authors.

ABSTRACT

yelodysplastic syndromes (MDS) are a group of clonal myeloid disorders characterized by low blood counts and a propensity to develop acute myeloid leukemia. The management of lowerrisk (LR) MDS with persistent cytopenias remains suboptimal. Eltrombopag, a thrombopoietin-receptor agonist, can improve platelet counts in LR-MDS and trilineage hematopoiesis in aplastic anemia. We conducted a phase II dose modification study to investigate the safety and efficacy of eltrombopag in LR-MDS. The eltrombopag dose was escalated from 50 mg/day to a maximum of 150 mg/day over a period of 16 weeks. The primary efficacy endpoint was hematologic response at 16-20 weeks. Eleven of 25 (44%) patients responded; five and six patients had uni- or bi-lineage hematologic responses, respectively. The predictors of response were presence of a paroxysmal nocturnal hemoglobinuria clone, marrow hypocellularity, thrombocytopenia, and elevated plasma thrombopoietin levels at study entry. The safety profile was consistent with that found in previous eltrombopag studies in aplastic anemia; no patients discontinued the drug due to adverse events. Three patients developed reversible grade 3 liver toxicity and one patient had increased reticulin fibrosis. Ten patients discontinued eltrombopag after achieving a robust response (median time 16 months); four of them reinitiated eltrombopag because of declining blood counts, and all attained a second robust response. Six patients had disease progression not associated with expansion of mutated clones and no patient progressed to develop acute myeloid leukemia on study. In conclusion, eltrombopag was well-tolerated and effective in restoring hematopoiesis in some patients with low or intermediate-1 risk MDS. This study was registered at clinicaltrials.gov as #NCT00932156.

Correspondence: NEAL S. YOUNG youngns@nhlbi.nih.gov Received: March 24, 2020. Accepted: May 21, 2020. Pre-published: May 21, 2020. doi:10.3324/haematol.2020.249995 ÂŠ2020 NIH (National Institutes of Health)

Introduction Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal myeloid disorders characterized by ineffective hematopoiesis and cytopenias, with variable risks of progression to acute myelogenous leukemia (AML).1,2 The natural history of the disease is divergent between lower-risk (LR) and higher-risk MDS patients, evidenced by differences in clinical course, treatment efficacy, and overall survival. Higher-risk MDS appears close in pathophysiology to AML3 whereas LRhaematologica | 2020; 105(12)

2785

A. Vicente et al.

MDS is a more diverse group containing not only welldefined World Health Organization (WHO) classified categories but also subtypes that overlap with bone marrow failure syndromes, such as hypoplastic MDS (hypoMDS), MDS and paroxysmal nocturnal hemoglobinuria (PNH), and MDS evolved from aplastic anemia (AA). In these subtypes, T-cell-mediated suppression of hematopoiesis similar to that occurring in AA has been described.4-6 The prognosis of patients with MDS is determined using the International Prognostic Scoring System (IPSS) and the revised IPSS (IPSS-R), based on the degree of cytopenias, bone marrow blast percentage, and presence of specific cytogenetic abnormalities.7,8 Targeted nextgeneration sequencing has identified somatic variants in candidate genes associated with myeloid malignancies in more than 80% of MDS patients.9,10 Although the implications of these somatic variants in MDS have been extensively studied in the past years, most are not yet included in scoring systems. MDS therapy is guided by IPSS risk stratification, with goals of treatment and tolerance of drug toxicity differing for higher risk-MDS and LR-MDS. In contrast to higher risk-MDS, supportive measures such as red blood cell transfusions, growth factors (erythropoiesis-stimulating agents and granulocyte colony-stimulating factor), and lenalidomide for patients with del(5q) are common first options for LR-MDS.11-13 In addition, immunosuppressive treatments have demonstrated efficacy in LR-MDS, most notably in patients who are younger, HLA-DR15-positive, and have a more limited transfusion history.14,15 The treatment options for cytopenias in non-responders, especially for thrombocytopenia, are very limited, and such patients are often managed with long-term transfusion support. They remain at high risk of bleeding, developing infections, and having an overall poor quality of life. Eltrombopag, a thrombopoietin-receptor agonist, was first used to treat thrombocytopenia in patients with idiopathic thrombocytopenic purpura,16 but has also been shown to improve hematologic response in patients with refractory severe AA and to increase overall and complete responses when combined with standard immunosuppression in treatment-naïve severe AA.17-20 In MDS, monotherapy with thrombopoietin agonists has only been tested in two studies, in which increased platelet counts were seen in nearly 50% of the patients.21,22 A randomized, double-blind study with romiplostim versus placebo for LR-MDS was stopped early due to an apparent increased risk of AML progression, which was not confirmed with long-term follow up.23,24 When eltrombopag was added to azacitidine to improve treatmentrelated thrombocytopenia in intermediate/high-risk MDS, it resulted in worse platelet recovery and increased progression to AML.25 In this study, we investigated the safety and efficacy of eltrombopag monotherapy in LR-MDS and any cytopenia in a non-randomized phase II, investigator-initiated clinical trial.

Methods Patients and eligibility Subjects 18 years or older with LR-MDS were enrolled into this phase II, dose modification study of oral eltrombopag 2786

between March, 2011 and July, 2017. The protocol was approved by the Institutional Review Board of the National Heart, Lung, and Blood Institute, and monitored by an independent Data Safety and Monitoring Board. The initial version of the protocol only included patients with platelet counts ≤30x109/L or platelet-transfusion dependence. After accrual of the first five patients, the inclusion criteria were broadened to enroll patients with any cytopenia. The revised inclusion criteria were: hemoglobin ≤9.0 g/dL or red blood cell transfusion-dependence (at least 4 units of red blood cells at 8 weeks prior to enrollment); platelet counts ≤30x109/L or platelet transfusion-dependence; or absolute neutrophil count (ANC) ≤0.5x109/L. Patients with refractory anemia with excess blasts, AML, treatment-related MDS, or chronic myelomonocytic leukemia were excluded.

Treatment plan and study endpoints Patients received eltrombopag for 16-20 weeks. Eltrombopag was initiated at a dose of 50 mg daily and the dose was increased to a maximum of 150 mg, unless toxicity-related stopping rules were met, dose reduction laboratory values occurred (Online Supplementary Table S1A, B), or hematologic response was achieved (Figure 1A). The primary safety endpoint was assessed using the National Cancer Institute’s Common Terminology Criteria for Adverse Events version 4.0 (CTCAE v4.0). The primary efficacy endpoint was hematologic response at 16 weeks, defined as either: (i) a platelet count increase of ≥20x109/L above the baseline or stable platelet counts with transfusion-independence for ≥8 weeks; (ii) a hemoglobin increase of ≥1.5 g/dL or a reduction in red blood cell transfusion requirements by at least 50% over the preceding 8 weeks; (iii) ≥100% increase in ANC for those with a pretreatment ANC of <0.5x109/L or an absolute increase >0.5x109/L. If patients had a clinical response in any lineage at 16 weeks but did not yet meet full primary endpoint criteria, eltrombopag was continued for another 4 weeks and response was assessed at 20 weeks. Responding patients could receive eltrombopag on the extension arm until they met the criteria for a robust response (platelet count >50x109/L, hemoglobin >10 g/dL, and ANC >1.0x109/L), at which time eltrombopag was discontinued. Eltrombopag was restarted in patients with blood counts falling below platelets <30x109/L, hemoglobin <9 g/dL, or ANC <0.5x109/L. Secondary endpoints were progression to higher-risk MDS, changes in serum thrombopoietin levels measured at the primary endpoint by magnetic multiplex assays (Luminex),26 eltrombopag discontinuation due to the achievement of a robust response, or grade 2 or higher bleeding events. International Working Group (IWG) criteria were used to determine the cytogenetic response and progression of disease.27 We screened all patients at baseline, at the primary endpoint, and at the time of disease progression for somatic variants in 63 candidate genes associated with myeloid malignancies using a targeted next-generation sequencing panel (Online Supplementary Table S2).28

Statistics In this intention-to-treat study, summary statistics were used for patients’ demographics and laboratory measurements. Covariate effects on the response rates and the distributions of survival time were evaluated using univariable logistic regression and Cox proportional hazard models, respectively. Further details on methods can be found in the Online Supplementary Methods. haematologica | 2020; 105(12)

Eltrompobag in myelodysplastic syndrome

Figure 1. Study design and clinical outcomes of 30 patients with myelodysplastic syndrome. (A) Study flowchart. All patients enrolled on study received eltrombopag (EPAG) at a starting oral dose of 50 mg/day, increased up to a maximum dose of 150 mg/day for 16 weeks. Primary efficacy endpoint was assessed as hematologic improvement at 16-20 weeks. Non-responders were taken off study while responders continued EPAG in the extension phase until they achieved a protocol-defined robust response. If patients achieved a robust response, EPAG was discontinued and their blood counts and bone marrow values were monitored for 2 years. (B) Venn diagrams showing the number of patients with single lineage and multilineage responses to EPAG at the primary endpoints, and best responses in the extension phase. Laboratory parameters are also represented under the individual Venn diagrams. (C) Hematologic improvement of all responders, including the ten robust responders. The median neutrophil counts, hemoglobin concentration, and platelet levels are shown in the figure at the indicated time-points. (D) Thrombopoietin (TPO) levels of responders and non-responder at baseline and at the primary endpoint of the study. (E) TPO levels measured in patients with hypoplastic myelodysplastic syndrome (h-MDS), with myelodysplastic syndrome evolved from aplastic anemia (evolved from AA), and with normo- and hypercellular myelodysplastic syndrome (MDS) at study entry. NS: not statistically significant.

haematologica | 2020; 105(12)

2787

A. Vicente et al.

Results Patientsâ&#x20AC;&#x2122; characteristics and disposition A total of 30 patients were enrolled in the study and received eltrombopag. The first five subjects enrolled (UPN-1 to UPN-5) were entered when eligibility criteria included only thrombocytopenia. They were not included in the efficacy analysis set, as requested by the Institutional Review Board, but were included for secondary endpoint and sensitivity analyses. In our cohort, 90% of patients were classified as IPSS intermediate-1 risk and as IPSS-R very low to intermediate risk (Table 1). Twenty-two patients (73%) had either refractory cytopenia with multilineage dysplasia or refractory cytopenia with unilineage dysplasia. At enrollment, 11 patients had bicytopenia, ten had anemia (hemoglobin <9.0 g/dL) or were red blood cell transfusion-dependent, and nine had thrombocytopenia (platelets <30x109/L) or were platelet-transfusion dependent (Table 2). Median blood counts for patients with anemia were hemoglobin 8.2 g/dL (range, 7.1-11); with thrombocytopenia, platelets 11x109/L (range, 4-28), and neutropenia, 0.38x109/L. Twelve patients (40%) had received at least one prior treatment other than supportive care and were considered to have relapsed/refractory disease. Prior therapies included lenalidomide, azacitidine, erythropoiesis-stimulating agents, and immunosuppressive treatments (Table 1). Four patients discontinued the study before the primary endpoint evaluation: UPN13 opted for supportive care, UPN-23 died from acute respiratory distress syndrome and mycobacterial infection, and UPN-19 and UPN-20 had worsening cytopenias with disease progression (described in more detail below and in Table 3).

Safety In 25 of 30 (83%) patients eltrombopag was escalated to the maximum dose (150 mg in all patients except the 3 patients of East or South-Asian origin). Of the five remaining patients, two had thrombocytosis at 75 mg/day requiring dose reduction (to 37.5 mg/day), one patient had grade 3 elevated liver enzymes (alanine transaminase and aspartate transaminase >5 times the reference value) at a dose of 75 mg/day which improved at a lower dose of 50 mg/day, and two achieved platelet responses at lower doses (75 mg/day and 125 mg/day) so that dose escalation was halted per protocol. At the maximum dose of 150 mg, two patients experienced grade 3 reversible increases in liver transaminases, requiring dose interruption. After normalization of transaminases, eltrombopag was restarted at the lower dose level (125 mg/day) in both patients (UPN-4, UPN-18). The most frequent treatment-related adverse events were nausea and vomiting (20%), skin lesions (20%), headaches (17%), and discoloration of the sclerae (17%) (Online Supplementary Table S3). There were no serious adverse events attributed to eltrombopag at the time of the data cut (Online Supplementary Table S4). One patient (UPN-24) with no response to treatment at the primary endpoint had increased reticulin fibrosis (from 1+ to 3+). Five patients (17%) had grade 2 or higher bleeding adverse events at a median of 1 month (range, 0.34-4.5 months), which were not deemed to be related to eltrombopag, but to diseaseassociated thrombocytopenia. There were no eltrom2788

bopag-related deaths, thrombotic events, or progression to AML on study. One patient died due to acute respiratory distress syndrome unrelated to eltrombopag.

Hematologic response Eleven of 25 patients (44%) achieved a hematologic response at the primary endpoint; ten had been classified

Table 1. Baseline characteristics of the patients.

Baseline characteristics

Cohort (n = 30)

Age, years Median (range) 65 (35-85) Sex, n (%) Male 21 (70) Female 9 (30) Ethnicity, n (%) Asian 3 (10) Black or African American 5 (16.7) White 21 (70) Other 1 (3.3) WHO classification, n (%) RCUD 11 (36.7) RCMD 11 (36.7) MDS-U 6 (20) RARS 2 (6.7) IPSS risk, n (%) Low 2 (6.7) Intermediate-1 27 (90) Intermediate-2 1 (3.3) IPSS-R risk, n (%) Very low 1 (3.3) Low 8 (26.7) Intermediate 18 (60) High 3 (10) IPSS cytogenetic risk classification*, n (%) Good 16 (53.3) Intermediate 13 (43.3) Poor 1 (3.3) Types of previous systemic therapy for MDS, n (%) Lenalidomide 5 (17) Azacitidine 6 (20) Erythropoietin-stimulating agents 12 (40) Immunosuppressive therapy 1 (3) Laboratory parameters Neutrophil count, x109/L Median (range) 0.995 (0.26-3.77) Platelet count, x109/L Median (range) 23 (4-256) Hemoglobin, g/dL Median (range) 8.95 (6.2-12) PNH clones, n (%) â&#x2030;Ľ 1.0% 11 (36.7) < 1.0% 19 (63.3) Thrombopoietin, pg/mL Median (range) 2119 (71-4817) Transfusion dependency, n(%) Platelets 16 (53.3) Red cells 22 (73.3)

Cohort (n = 25) 63 (35-85) 17 (68) 8 (32) 3 (12) 4 (16) 17 (68) 1 (4) 10 (40) 9 (36) 4 (16) 2 (8) 2 (8) 23 (92) 0 (0) 0 (0) 8 (32) 15 (60) 2 (8) 14 (56) 11 (44) 0 (0) 5 (20) 3 (12) 9 (36) 0 (0)

1.06 (0.26-3.77) 26 (4-256) 8.80 (6.2-12) 10 (40) 15 (60) 2080 (71-4817) 11 (44) 18 (76)

All patients included in the study are shown in the left column (n=30), whereas the right column (n=25) shows the patients included in the primary endpoint analysis. WHO: World Health Organization; RCUD: refractory cytopenia with unilineage dysplasia; RCMD: refractory cytopenia with multilineage dysplasia; MDS-U: dysplasia; myelodysplastic syndrome-unclassifiable; RARS:, refractory anemia with ringed sideroblasts; IPSS: International Prognostic Scoring System; IPSSR: Revised International Prognostic Scoring System; MDS: myelodysplastic syndrome: PNH: paroxysmal nocturnal hemoglobinuria.

haematologica | 2020; 105(12)

Eltrompobag in myelodysplastic syndrome

at baseline as IPSS intermediate-1 risk and one as low risk. The median time to response was 16 weeks (range, 16-20 weeks). Both unilineage (5/11, 46%) and bilineage (6/11, 55%) responses were seen at 16 weeks (Figure 1B and Table 2). Eight of 11 responders were transfusion dependent for platelets and/or red blood cells before eltrombopag treatment and six of them became transfusion independent at 16 weeks. Three of 11 responders (27%) showed normalization of a previously abnormal karyotype (trisomy 6, trisomy 15, and deletion 13q) at a median time of 20 months (range, 9-21 months) (Online Supplementary Table S5). Additionally, three of the first five patients enrolled (UPN-1, UPN-4, UPN-5) excluded from the efficacy analysis set achieved a platelet response and continued eltrombopag on the extension arm. A total of 14 patients, including three of the first five

enrolled patients excluded from the efficacy analysis set, continued to receive eltrombopag in the extension phase of the study at a median dose of 150 mg/day (range, 37.5– 150 mg). All 14 patients experienced further hematologic improvement (robust response or single lineage response) with longer treatment. At the primary endpoint, the median increase in hemoglobin levels was 1.4 g/dL (range, 1.2–3.3), platelet numbers 14 x109/L (range, -12 – 67) and neutrophil counts 0.71x109/L (range, -0.2 – 2.52) (Figure 1C). At best response, the median increase from baseline for hemoglobin was 4.45 g/dL for platelets 53.5x109/L and the median increase in neutrophils was 1.14x109/L. A robust response was achieved by ten of 14 patients with median drug administration of 16 months (range, 9-42 months) (Figure 1C), and eltrombopag was discontinued per protocol (Figure 2). Of these, four sus-

Table 2. Patients’ characteristics and hematologic response to eltrombopag.

Age

Responders 1 65 4 79 5 46 6 53 7 73 11 35 14 85 16 62 17 54 18 59 25 72 26 36 27 47 30 47 Non-responders 2 76 3 74 8 47 9 67 10 76 12 76 13 54 15 64 19 55 20 63 21 68 22 72 23 76 24 63 28 79 29 76

Sex

WHO subtype

PNH clone

IPSS

Bone marrow cellularity

Cytopenias at study entry ANC Hb Plt x

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

RCMD RCUD MDS-U RCMD RCUD RCMD RCUD RCMD RCUD RCUD RCMD MDS-U MDS-U MDS-U

<1% <1% 4.60% 8.60% <1% 3.30% <1% 3% <1% 38.20% <1% 5% 5.80% 4.80%

Int-1 Int-2 Int-1 Int-1 Int-1 Int-1 Low Int-1 Int-1 Int-1 Int-1 Int-1 Int-1 Int-1

35% 5% <10% 45% 15-20% 30% 25% 50% 40% 5% 80% 20% 30% 5%

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

MDS-U RCMD MDS-U RCUD RCMD RARS RCUD RCUD RCUD RCUD RCUD RARS RCMD RCMD RCMD RCMD

<1% <1% <1% <1% <1% <1% 1.50% 1.70% <1% <1% 47.10% <1% <1% <1% <1% <1%

Int-1 Int-1 Int-1 Int-1 Int-1 Low Int-1 Int-1 Int-1 Int-1 Int-1 Int-1 Int-1 Int-1 Int-1 Int-1

90% 50% 5% 5% 50% 90% 40% 5% 50% 60% 40% 70% 30% 40% 70% 40%

x x x x x x x

x x x x x x x x x x

x x

Lineage responses at primary endpoint ANC Hb Plt

x x x

x x

x x x x x

x x x x x x x x x

Robust cell counts Months ANC

Plt

20 16 12 12

x x x x

19 9 42

x x x

22 14

x x

x x x x x x x x x x x x x x x x

x x

Cytopenia at baseline, response at primary endpoint and time-point of a robust response are shown in this table for all 30 patients. UPN: unique patient number; WHO: World Health Organization; PNH: paroxysmal nocturnal hemoglobinuria; IPSS: International Prognostic Scoring System; ANC: absolute neutrophil count; Hb: hemoglobin; Plt: platelets; M: male; F: female; NR: non-response; RCMD: refractory cytopenia with multilineage dysplasia; RCUD: refractory cytopenia with unilineage dysplasia; MDS-U: myelodysplastic syndrome-unclassifiable; RARS: refractory anemia with ringed sideroblasts; int-1: intermediate 1; int-2: intermediate 2.

haematologica | 2020; 105(12)

2789

A. Vicente et al.

Figure 2. Clinical outcomes of the 30 patients with myelodysplastic syndrome enrolled in the study. Swimmer plot with the clinical outcomes of responders (n=14) and non-responders (n=16). Bars represent the follow-up time for each patient. On the timeline, black bars represent the start of eltrombopag (EPAG) treatment until the primary outcome; blue bars represent the time for which patients continued EPAG on the extension arm; green bars represent the time that patients relapsed and restarted EPAG; pink triangles indicate the time of disease progression; dashed blue bars indicate when patients went off drug due to robust response. An asterisk indicates that the patient withdrew from the study, while a cross indicates that the patient died. MDS: myelodysplastic syndrome.

tained a hematologic response with a median follow up of 15 months off drug (range, 12-21 months). Declining counts were noted in four patients and eltrombopag was restarted at the last effective dose; all patients achieved a second robust response after a median of 12 months (range, 9-14 months) of additional eltrombopag treatment. At the time of data cut, eltrombopag was being tapered in all of these four patients. Of the remaining two robust responders, one developed a PNH clone and intravascular hemolysis, and another patient had disease progression. Four of 14 patients who had achieved single lineage response at the primary endpoint sustained their response on the extension arm but discontinued treatment: one was lost to follow-up, one remained refractory to reinitiation of eltrombopag which was originally discontinued due to thrombocytosis, and two had progressive disease according to IWG criteria.

Predictors of response On univariate analysis, the presence of more than 1% glycosylphosphatidylinositol-deficient neutrophils (P=0.036), thrombopoietin levels â&#x2030;Ľ2219 pg/mL (P=0.008), thrombocytopenia with or without other cytopenia (P=0.015), and hypocellular marrow (P=0.036) at baseline correlated with response to eltrombopag (Online 2790

Supplementary Table S6A, B). Other baseline features such as age, absolute reticulocyte count, and ANC were not predictive. At baseline, median thrombopoietin plasma levels were significantly higher in patients who achieved response compared with levels in non-responders (median 2766 pg/mL vs. 562 pg/mL, P=0.018) (Figure 1D). Among the responders, the two subjects with low thrombopoietin levels failed to achieve a robust response. At the primary endpoint, thrombopoietin levels remained elevated in responders compared to the levels in non-responders (median 2565 pg/mL vs. 1840 pg/mL). High thrombopoietin levels were also associated with better survival according to Cox regression analysis (hazard ratio <1; P=0.024) (Online Supplementary Table S6C). We also compared thrombopoietin levels among MDS patients whose disease evolved from AA, who had hypo-MDS at diagnosis, or who had hyper/normocellular MDS. Hypo-MDS was defined as bone marrow cellularity <30% in patients younger than 70 years or <20% in those older than 70 years. Thrombopoietin levels in patients whose MDS evolved from AA were significantly higher than those in patients with de novo MDS at baseline and at the primary endpoint (P=0.0067) (Figure 1E). The difference in thrombopoietin levels between hypo-MDS compared to hyper/normocellular MDS was not statistically significant (P=0.12) (Figure 1E). Response rates in patients who had haematologica | 2020; 105(12)

Eltrompobag in myelodysplastic syndrome Table 3. Clinical characteristics of patients who progressed on study.

Age, years Sex IPSS Baseline Cytogenetics Bone marrow blasts (%) Baseline or best response ANC (x109/L) Hemoglobin (g/dL) Platelets (x109/L) Disease progression* Cytogenetic

Bone marrow blasts (%) ANC (x109/mL) Hemoglobin (g/dL) Platelets (x109/mL) Time on eltrombopag (months) Time to progression (months) Present status

UPN-4

Responders UPN-6

UPN-14

UPN-19

Non-responders UPN-20**

UPN-24

79 M Int-2

53 M Int-1

85 M Low

55 M Int-1

63 M Int-1

63 F Int-1

45,XY,-7[20]

46,XY[20]

<2%

<5%

46,XY,t(1;9) (p34;q22)[20] <5%

46,XX[20]

<5%

46,XY,del (5) (q13q33)[20] <5%

0.71 8.6 70

3.6 12.9 64

4.9 12.9 93

1.62 8.9 202

1.64 9.1 93

1.1 11 20

45,XY,-7[20]

47,XY,+21[11]/46,XY[9]

8% 0.4 9.9 22 9 9 Deceased

6% 1.92 13 46 28 35 Alive

NA 3.9 13.1 10 9 9 Deceased

46,XY[3]/46,XY,del(5) 46,XY,t(1;9) (q13q33)[9]/47,idem, (p34;q22)[20] +21[5],46,idem, I(21)(q10)[5] <5% <5% 1.13 2.1 8.9 9.1 66 43 3 3 3 3 Alive Alive

<5%

46,XX[20]

<5% 0.82 10 7 4 4 Alive

*According to the modified 2006 International Working Group criteria; **UPN-20 was noted to have peripheral blasts at the time of progression; UPN: unique patient number; M: male; F: female; IPSS: International Prognostic Scoring System; Int-1: intermediate 1; Int-2: intermediate; NA: not available; BM: bone marrow; ANC: absolute neutrophil count; Hb: hemoglobin.

been previously treated were 20% after lenalidomide, 33% after hypomethylating agents and 50% after erythropoiesis-stimulating agents (Online Supplementary Table S7).

Disease progression Of all 30 patients enrolled, six (20%) had disease progression with a median time to progression of 6.5 months (range, 3-35 months). Three responding patients progressed during the extension phase of the study with a median time to progression of 9 months (range, 9-35 months) (Table 3). UPN-4, who presented with IPSS intermediate-2 and deletion 7q at baseline, was deemed a responder at the primary endpoint but platelet counts later declined and myeloblasts increased from <5% to 8% after 9 months of eltrombopag treatment. The patient died from infectious complications after discontinuation of eltrombopag while receiving supportive care. Platelets and ANC declined in another responding patient 7 months after discontinuation of eltrombopag because of the patientâ&#x20AC;&#x2122;s robust response; evaluation of the bone marrow revealed an increase in blasts and acquisition of trisomy 21. This patient underwent successful allogeneic stem cell transplant. In UPN-14, platelet counts fell more than 50% at 9 months on eltrombopag, and the patient died from bleeding 1 month after stopping the drug; we were unable to evaluate his bone marrow at the time of disease progression. Among the non-responders, three patients had disease progression at the time of the primary endpoint evaluation based on a decline in platelet counts by more than 50% when compared to the laboratory values at study entry (Table 3). None of these patients had increased blast percentage. In addition, UPN-19 had acquired a complex haematologica | 2020; 105(12)

karyotype at the primary endpoint assessment. UPN-19 and UPN-20 underwent allogeneic stem cell transplantation and are alive. UPN-24 remained dependent on platelet transfusions at the 6-month follow up after discontinuing eltrombopag. Furthermore, two non-responding patients with an abnormal baseline karyotype developed additional chromosome abnormalities (monosomy 7 in UPN-2 and a complex karyotype in UPN-3) at 16 weeks but did not meet IWG criteria for disease progression. UPN-2 died from AML 5 years after acquiring monosomy 7 and UPN3 died of bleeding complications 9 months after going off study.

Somatic variants in myeloid candidate genes At baseline, 23 of 29 patients (52%) were identified with somatic variants in genes recurrently mutated in myeloid malignancies. The most commonly mutated genes were related to epigenetic regulators and splicing factors, such as ASXL1 (21%), TET2 (17%), and SF3B1 (14%) (Figure 3A). At the primary endpoint, variants were found in six responders and seven non-responders (13 of 24 patients; 54%) (Figure 3A). Novel variants were identified in two responders (UPN-14 and UPN-4) and in three non-responders (UPN-2, UPN-9, and UPN-12). Moreover, somatic variants in DNMT3A, BCOR, SETBP1, and ASXL1 at baseline were no longer detected at the primary endpoint in three non-responders (UPN-24, UPN-8, and UPN-2) (Figure 3A). We investigated whether eltrombopag promoted the expansion of clones identified at baseline. We found no difference in the allele frequencies of variants detected 2791

A. Vicente et al.

before and after eltrombopag, regardless of the patientsâ&#x20AC;&#x2122; response (P=0.85) (Figure 3B, C). No particular gene was associated with either expansion or reduction in the size of clones. Among four patients who progressed on study, according to IWG criteria, and had samples available for longitudinal analysis, only two acquired novel clones at progression (UPN-6 and UPN-20) (Online Supplementary Figure S2). At progression, a novel ASXL1 clone was found in UPN-20. UPN-6 progressed with trisomy 21 7 months after eltrombopag had been halted because of robust response, with concomitant expansion of the

ASXL1 clone (variant allele frequency of 24%-39%) and acquisition of a RUNX1 variant (variant allele frequency of 54%). Other clones identified at baseline remained stable or were no longer detected after progression (Figure 3A and Online Supplementary Figure S2).

Discussion Our prospective, phase II study shows the efficacy of eltrombopag in inducing multilineage hematologic

Figure 3. Clonal dynamics of variants identified in myeloid genes during the study. We screened 29 patients at baseline and 24 patients with samples available at the primary endpoint for somatic variants in genes recurrently mutated in myeloid neoplasia by targeted next-generation sequencing (Online Supplementary Table S2). (A) Summary of somatic variants identified in enrolled patients at baseline and at the primary endpoint. For each patient, white and gray columns represent the baseline and primary endpoints (16 or 20 weeks) on eltrombopag (EPAG), respectively. For each time-point, the trilineage responses, normal or abnormal cytogenetics, the frequency of somatic variants, and disease progression status are represented by different colors specified under the panel. Patients were grouped into responders and non-responders after EPAG treatment and then sorted by variant numbers (highest to lowest). The variant names, the number of variants, and gene functions are also indicated in the panel for each patient. Of note, other than a single nonsense variant, all the ASXL1 mutations were frameshift. Also, both patients with mutated SF3B1 had the K700E variant. (B) Variant allele frequencies of the somatic variants identified at baseline (pre-EPAG) and at the primary endpoint (post-EPAG) in the entire cohort. (C) Variant allele frequencies of somatic clones identified pre- and post-EPAG according to the patientsâ&#x20AC;&#x2122; response.

2792

haematologica | 2020; 105(12)

Eltrompobag in myelodysplastic syndrome

responses in about half of LR-MDS patients. Moreover, peripheral blood cell counts continued to improve with longer treatment duration and were sustained in some patients after discontinuation of eltrombopag. Our results confirm and extend observations of previous studies with thrombopoietin agonists, eltrombopag and romiplostim, which demonstrated platelet responses and reduction of thrombocytopenia-related adverse events in patients with LR-MDS and low platelet counts.21,22 The quality of the hematologic response, with one-third of patients achieving a robust resposne, is encouraging, particularly considering that 40% (12/30) of patients had failed more than two lines of prior therapies. Furthermore, counts remained stable even after eltrombopag was discontinued and all patients who restarted eltrombopag achieved a second response. Remarkably, 20% (3/14) of the responding patients in our cohort achieved a major cytogenetic response according to IWG 2006 criteria. Although this response was noted in small clones with abnormal karyotypes (Online Supplementary Table S5), these findings may indicate that eltrombopag preferentially stimulates normal hematopoietic stem and progenitor cells. The toxicity profile in this study is comparable to that in previous studies in bone marrow failure,17-20 with only a few instances of temporary dose interruptions because of transient elevations of liver transaminases. Increased reticulin fibrosis (grade 1 to grade 3) was noted in one patient with disease progression, which could not be clearly attributed to either the study drug eltrombopag or underlying disease. Baseline characteristics of a PNH clone, elevated thrombopoietin levels, thrombocytopenia with or without another cytopenia, and low marrow cellularity correlated with response to eltrombopag are novel findings in our study. Patients with a previous history of AA or hypoMDS at diagnosis may benefit from eltrombopag treatment more than do patients with more typical hyper/normocellular MDS (Online Supplementary Table S6B). The efficacy of immunosuppressive treatments and eltrombopag in AA and a group of LR-MDS patients suggests the existence of similar pathological mechanisms in these syndromes.14,29 Eltrombopag has been reported to modulate T regulatory cells, restore Fc-γ receptor balance in phagocytes, and to mobilize intracellular iron,30-32 but the exact mechanism of any interaction between eltrombopag and the immune system needs further investigation. Despite the benefit of eltrombopag in improving cytopenias in patients with LR-MDS, one major concern regarding the use of thrombopoietin mimetics in myeloid malignancies is the expansion and stimulation of malignant clones. We found no correlation between patients’ somatic gene mutation profile and hematologic response or progression of disease in our study. Our cohort included a large number of patients with hypo-MDS at diagnosis and whose MDS evolved from AA, some with the other features of immune-mediated marrow failure (PNH clone, elevated thrombopoietin levels, marrow hypocellularity), but overall the somatic mutation profile was representative of MDS. Frequently mutated genes were ASXL1 (21%), TET2 (17%) and SF3B1 (14%), a different profile from that typical of AA (BCOR, BCORL1, PIGA, and DNMT3A).33 No patient progressed to AML on study. One patient

haematologica | 2020; 105(12)

developed AML after having been off study for 5 years, consistent with the natural history of MDS, and this event was most likely not due to the earlier brief course of eltrombopag. Similar to our results, eltrombopag monotherapy was also not associated by others with an increased progression to AML in LR-MDS patients,21 being reported only in high-risk patient populations (those with refractory anemia with excess blasts-1 and -2 with romiplostim treatment), with a higher dose of eltrombopag (300 mg/day), and combination therapy with azacytidine.25,34 Six patients progressed on study (20%), comprising three responders and three non-responders. The rate of progression observed in our study is similar to that in a previous eltrombopag trial (12%).21 There was a difference in the timing and the type of progression between responders and non-responders; in non-responders the only criterion for progression before or at the time of the primary efficacy assessment was a decline in platelets, whereas responders had both cytopenia and an increased percentage of blasts during the extension phase. While eltrombopag at 150 mg/day did not appear to result in progression to AML in LR-MDS patients, caution is indicated in the treatment of individual patients and further clinical studies are warranted. Our conclusions do not apply to eltrombopag in patients with high-risk IPSS scores or high-risk cytogenetics irrespective of IPSS. Until further data are available, close monitoring of peripheral blood counts, and frequent bone marrow and cytogenetic evaluations should be performed while patients are on eltrombopag. The appearance of transient cytogenetically abnormal clones was observed in patients during the extension arm, a phenomenon that has also been reported in treatmentnaïve AA patients after immunosuppressive treatment alone and with eltrombopag monotherapy for refractory AA.35 In MDS, some transient clones seem to be associated with better outcomes and may reflect momentary episodes of genetic instability, not of long-term clinical significance.36 In addition, no clonal expansion was noted after treatment with eltrombopag in either responders or non-responders in our trial. In conclusion, our results indicate that eltrombopag as monotherapy is well tolerated and can be effective treatment for patients with low to intermediate-1 risk MDS. Our study not only confirmed the previously reported platelet response27 but showed robust and durable trilineage responses. Hypocellular marrow, elevated thrombopoietin, and a PNH clone predicted response to eltrombopag treatment. The main limitations of the study are the small sample size and the unique patients’ characteristics resulting from the referral pattern of our institution. Further larger, prospective and controlled studies are warranted to better define the role of eltrombopag in the treatment of LR-MDS. Acknowledgments The authors gratefully acknowledge NIH physicians, nurses and other patient care providers involved. We thank co-investigators Dr. Elaine Sloand, Dr. Zhijie Wu and Dr. Carrie Diamond for their participation and contributions; and the Flow Cytometry Core at NHLBI. This research was funded by the Intramural Research Program of the NHLBI. Eltrombopag was provided by GlaxoSmithKline (Collegeville, PA, USA) and subsequently by Novartis (East Hanover, NJ, USA).

2793

A. Vicente et al.

References 1. Goasguen JE, Bennett JM. Classification and morphologic features of the myelodysplastic syndromes. Semin Oncol. 1992;19(1):4-13. 2. Ganser A, Hoelzer D. Clinical course of myelodysplastic syndromes. Hematol Oncol Clin North Am. 1992;6(3):607-618. 3. Sekeres MA, Cutler C. How we treat higher-risk myelodysplastic syndromes. Blood. 2014;123(6):829-836. 4. Huang TC, Ko BS, Tang JL, et al. Comparison of hypoplastic myelodysplastic syndrome (MDS) with normo-/hypercellular MDS by International Prognostic Scoring System, cytogenetic and genetic studies. Leukemia. 2008;22(3):544-550. 5. Young NS. Aplastic anemia. N Engl J Med. 2018;379(17):1643-1656. 6. Kasahara S, Hara T, Itoh H, et al. Hypoplastic myelodysplastic syndromes can be distinguished from acquired aplastic anaemia by bone marrow stem cell expression of the tumour necrosis factor receptor. Br J Haematol. 2002;118(1):181-188. 7. Greenberg P, Cox C, LeBeau MM, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood. 1997;89(6):2079-2088. 8. Greenberg PL, Tuechler H, Schanz J, et al. Revised international prognostic scoring system for myelodysplastic syndromes. Blood. 2012;120(12):2454-2465. 9. Bejar R. Clinical and genetic predictors of prognosis in myelodysplastic syndromes. Haematologica. 2014;99(6):956-964. 10. Nazha A, Seastone D, Radivoyevitch T, et al. Genomic patterns associated with hypoplastic compared to hyperplastic myelodysplastic syndromes. Haematologica. 2015;100(11):e434-437. 11. List A, Dewald G, Bennett J, et al. Lenalidomide in the myelodysplastic syndrome with chromosome 5q deletion. N Engl J Med. 2006;355(14):1456-1465. 12. Almeida A, Fenaux P, List AF, Raza A, Platzbecker U, Santini V. Recent advances in the treatment of lower-risk non-del(5q) myelodysplastic syndromes (MDS). Leuk Res. 2017;52:50-57. 13. Fenaux P, Adès L. How we treat lower-risk myelodysplastic syndromes. Blood. 2013; 121(21):4280-4286. 14. Stahl M, DeVeaux M, de Witte T, et al. The use of immunosuppressive therapy in MDS: clinical outcomes and their predictors in a

2794

large international patient cohort. Blood Adv. 2018;2(14):1765-1772. 15. Molldrem JJ, Caples M, Mavroudis D, Plante M, Young NS, Barrett AJ. Antithymocyte globulin for patients with myelodysplastic syndrome. Br J Haematol. 1997;99(3):699705. 16. Bussel JB, Cheng G, Saleh MN, et al. Eltrombopag for the treatment of chronic idiopathic thrombocytopenic purpura. N Engl J Med. 2007;357(22):2237-2247. 17. Olnes MJ, Scheinberg P, Calvo KR, et al. Eltrombopag and improved hematopoiesis in refractory aplastic anemia. N Engl J Med. 2012;367(1):11-19. 18. Desmond R, Townsley DM, Dumitriu B, et al. Eltrombopag restores trilineage hematopoiesis in refractory severe aplastic anemia that can be sustained on discontinuation of drug. Blood. 2014;123(12):1818-1825. 19. Townsley DM, Scheinberg P, Winkler T, et al. Eltrombopag added to standard immunosuppression for aplastic anemia. N Engl J Med. 2017;376(16):1540-1550. 20. Winkler T, Fan X, Cooper J, et al. Treatment optimization and genomic outcomes in refractory severe aplastic anemia treated with eltrombopag. Blood. 2019;133(24): 2575-2585. 21. Oliva EN, Alati C, Santini V, et al. Eltrombopag versus placebo for low-risk myelodysplastic syndromes with thrombocytopenia (EQoL-MDS): phase 1 results of a single-blind, randomised, controlled, phase 2 superiority trial. Lancet Haematol. 2017;4(3):e127-e136. 22. Kantarjian H, Fenaux P, Sekeres MA, et al. Safety and efficacy of romiplostim in patients with lower-risk myelodysplastic syndrome and thrombocytopenia. J Clin Oncol. 2010;28(3):437-444. 23. Giagounidis A, Mufti GJ, Fenaux P, et al. Results of a randomized, double-blind study of romiplostim versus placebo in patients with low/intermediate-1-risk myelodysplastic syndrome and thrombocytopenia. Cancer. 2014;120(12):1838-1846. 24. Kantarjian HM, Fenaux P, Sekeres MA, et al. Long-term follow-up for up to 5 years on the risk of leukaemic progression in thrombocytopenic patients with lower-risk myelodysplastic syndromes treated with romiplostim or placebo in a randomised double-blind trial. Lancet Haematol. 2018;5(3):e117-e126. 25. Dickinson M, Cherif H, Fenaux P, et al. Azacitidine with or without eltrombopag for first-line treatment of intermediate- or

high-risk MDS with thrombocytopenia. Blood. 2018;132(25):2629-2638. 26. Feng X, Scheinberg P, Wu CO, et al. Cytokine signature profiles in acquired aplastic anemia and myelodysplastic syndromes. Haematologica. 2011;96(4):602-606. 27. Cheson BD, Greenberg PL, Bennett JM, et al. Clinical application and proposal for modification of the International Working Group (IWG) response criteria in myelodysplasia. Blood. 2006;108(2):419-425. 28. Albitar A, Townsley D, Ma W, et al. Prevalence of somatic mutations in patients with aplastic anemia using peripheral blood cfDNA as compared with BM. Leukemia. 2018;32(1):227-229. 29. Saunthararajah Y, Nakamura R, Wesley R, Wang QJ, Barrett AJ. A simple method to predict response to immunosuppressive therapy in patients with myelodysplastic syndrome. Blood. 2003;102(8):3025-3027. 30. Liu XG, Liu S, Feng Q, et al. Thrombopoietin receptor agonists shift the balance of Fcγ receptors toward inhibitory receptor IIb on monocytes in ITP. Blood. 2016;128(6):852-861. 31. Schifferli A, Nimmerjahn F, Kühne T. Immunomodulation in primary immune thrombocytopenia: a possible role of the Fc fragment of romiplostim? Front Immunol. 2019;10:1196. 32. Roth M, Will B, Simkin G, et al. Eltrombopag inhibits the proliferation of leukemia cells via reduction of intracellular iron and induction of differentiation. Blood. 2012;120(2):386-394. 33. Yoshizato T, Dumitriu B, Hosokawa K, et al. Somatic mutations and clonal hematopoiesis in aplastic anemia. N Engl J Med. 2015;373 (1):35-47. 34. Kantarjian HM, Giles FJ, Greenberg PL, et al. Phase 2 study of romiplostim in patients with low- or intermediate-risk myelodysplastic syndrome receiving azacitidine therapy. Blood. 2010;116(17):3163-3170. 35. Teramura M, Kimura A, Iwase S, et al. Treatment of severe aplastic anemia with antithymocyte globulin and cyclosporin A with or without G-CSF in adults: a multicenter randomized study in Japan. Blood. 2007;110(6):1756-1761. 36. Schanz J, Cevik N, Fonatsch C, et al. Detailed analysis of clonal evolution and cytogenetic evolution patterns in patients with myelodysplastic syndromes (MDS) and related myeloid disorders. Blood Cancer J. 2018;8(3):28.

haematologica | 2020; 105(12)

ARTICLE

Acute Myeloid Leukemia

Dasatinib response in acute myeloid leukemia is correlated with FLT3/ITD, PTPN11 mutations and a unique gene expression signature

Ferrata Storti Foundation

Sigal Tavor,1 Tali Shalit,2 Noa Chapal Ilani,3 Yoni Moskovitz,3 Nir Livnat,3 Yoram Groner,4 Haim Barr,2 Mark D. Minden,5,6,7,8 Alexander Plotnikov,2 Michael W. Deininger,9,10 Nathali Kaushansky3,# and Liran I. Shlush3,5,11,#

Hemato-Oncology Department, Assuta Medical Center, Tel Aviv, Israel; 2G-INCPM, Weizmann Institute of Science, Rehovot, Israel; 3Department of Immunology, Weizmann Institute of Science, Rehovot, Israel; 4Department of Molecular Genetics, Weizmann Institute of Science, Rehovot, Israel; 5Princess Margaret Cancer Centre, University Health Network Toronto, Ontario, Canada; 6Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada; 7Department of Medicine, University of Toronto, Toronto, Ontario, Canada; 8Division of Medical Oncology and Hematology, University Health Network, Toronto, Ontario, Canada; 9Division of Hematology and Hematologic Malignancies, University of Utah, Salt Lake City, UT, USA; 10Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA and 11Division of Hematology, Rambam Healthcare Campus, Haifa, Israel 1

Haematologica 2020 Volume 105(12):2795-2804

NK and LIS contributed equally as co-senior authors.

ABSTRACT

ovel targeted therapies improve the survival of specific subgroups (defined by genetic variants) of patients with acute myeloid leukemia (AML), validating the paradigm of molecularly targeted therapy. However, identifying correlations between AML molecular attributes and effective therapies is challenging. Recent advances in highthroughput, in vitro drug sensitivity screening applied to primary AML blasts were used to uncover such correlations; however, these methods cannot predict the response of leukemic stem cells. Our study aimed to predict in vitro response to targeted therapies, based on molecular markers, with subsequent validation in leukemic stem cells. We performed ex vivo screening of sensitivity to 46 drugs on 29 primary AML samples at diagnosis or relapse. Using unsupervised hierarchical clustering analysis we identified a group with sensitivity to several tyrosine kinase inhibitors, including the multi-tyrosine kinase inhibitor, dasatinib, and searched for correlations between the response to dasatinib, exome sequencing and gene expression in our dataset and in the Beat AML dataset. Unsupervised hierarchical clustering analysis of gene expression resulted in clustering of dasatinib responders and non-responders. In vitro response to dasatinib could be predicted based on gene expression (area under the curve=0.78). Furthermore, mutations in FLT3/ITD and PTPN11 were enriched in the dasatinib-sensitive samples as opposed to mutations in TP53 which were enriched in resistant samples. Based on these results, we selected FLT3/ITD AML samples and injected them into NSG-SGM3 mice. Our results demonstrate that in a subgroup of FLT3/ITD AML (4 out of 9) dasatinib significantly inhibited leukemic stem cell engraftment. In summary we show that dasatinib has an anti-leukemic effect both on bulk blasts and, more importantly, on leukemic stem cells from a subset of AML patients that can be identified based on mutational and expression profiles. Our data provide a rational basis for clinical trials of dasatinib in a molecularly selected subset of AML patients. haematologica | 2020; 105(12)

Correspondence: LIRAN I. SHLUSH liran.shlush@weizmann.ac.il Received: October 16, 2019 Accepted: April 30, 2020. Pre-published: May 21, 2020. doi:10.3324/haematol.2019.240705 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.

2795

S. Tavor et al.

Introduction AML is an aggressive myeloid neoplasm with complex and heterogeneous genetics that influence prognosis and treatment response. Furthermore, AML is a multi-stage disease with preleukemic, leukemic and late stages.1 Identification of AML-inducing mutations is required for accurate diagnosis and to tailor therapy according to the genetic profile of individual patients. Recent advances in AML targeted therapy based on driver mutations have improved overall survival.2,3 Increased understanding of AML molecular pathophysiology provided new opportunities to target specific mutants, such as FLT3/ITD (FMSlike tyrosine kinase-3, internal tandem duplications) and IDH1/2.4-6 Moreover, the BCL2 inhibitor, venetoclax, has increased activity in patients with mutated NPM1c and RAD21,7 and in combination with hypomethylating agents targets leukemia stem cells (LSC).8 While these novel therapies can achieve prolonged remissions, most patients eventually relapse. Our recent studies suggest that the origins of AML relapse are heterogeneous, but relapsing clones invariably exhibit stem cell properties.9 Relapse usually originates from leukemic subclones that were present before therapy, and were either selected due to resistance mechanisms9,10 or gained competitiveness through further evolution, possibly due to exposure to chemotherapy.11 Targeted therapies can fail to prevent relapse by addressing one leukemic cell clone, while allowing other clones to expand and cause relapse.12 In order to prevent relapse, it is crucial to eradicate LSC in the early stages of treatment (induction/consolidation) before they expand and acquire additional mechanisms of resistance. It has been suggested that combination therapies that simultaneously target multiple leukemic subclones may overcome AML heterogeneity. However, combination therapy approaches are limited by tolerability, particularly in older individuals, and it is important to maximize efficacy and minimize toxicity by selecting the right drug for the right set of patients. Given the multitude of therapeutic options, it is essential to design clinical trials based on molecular markers that select for those patients likely to benefit. Our latest studies demonstrated that, in certain AML patients in whom relapse originated from primitive cells,9 the cells from the time of diagnosis capable of leukemic engraftment in immunodeficient mice were identical to those that caused relapse. In these patients the bulk of cells were responsive to chemotherapy, but rare LSC that expanded in xenografts could drive relapse. Accordingly, drugs that inhibit the engrafting clones should be useful in preventing relapse, but screening a large number of drugs in a large number of samples in xenografts is not feasible. An alternative approach is to screen for drugs with known mechanistic effects against large and heterogeneous AML cohorts in vitro, which allows for high throughput screening of many drugs to establish correlations between response and molecular attributes of AML. Disadvantages include the lack of a microenvironment and immune system and the short duration of exposure to the drug, which reflects the inability of AML cells to grow in vitro. Short-term in vitro drug sensitivity assays have become feasible in recent years by mirroring normal hematopoietic stem and progenitor cell culture conditions. So far in vitro drug testing of primary AML samples have focused on predicting clinical outcome,13,14 and, in some cases, guiding individ2796

ualized therapy.15 Such personalized therapy is particularly important for patients with relapsed/refractory disease.16 In vitro drug studies identified biomarkers of response to specific drugs17 and clusters of AML samples with similar response profiles.18 In the current study, we studied in vitro drug sensitivity of primary AML samples from patients who achieved complete remission and compared our results to those of the Beat AML study. Our global aim was to identify therapies and biomarkers that could predict which drugs might be added to induction therapy to prevent relapse in specific subtypes of AML and then validate the results in vivo. We discovered that samples with in vitro sensitivity to dasatinib (a multikinase inhibitor) had a specific in vitro drug sensitivity pattern and gene expression signature, and were enriched for FLT3/ITD and PTPN11 mutations. Xenograft studies confirmed that dasatinib targets LSC in vivo.

Methods Primary acute myeloid leukemia cells Frozen mononuclear cells from bone marrow or peripheral blood of AML patients at diagnosis and relapse, when available, were used. Inclusion criteria for sample selection were achievement of complete remission. Patients with a low peripheral blood blast count and acute promeylocytic leukemia were excluded (Online Supplementary Table S1). Ethics approval: UHN 01-0573.

Drug library We selected Food and Drug Administration-approved cancer drugs with demonstrated clinical and/or preclinical efficacy in leukemia and in other cancers. The 46 drugs and their mechanisms of action are detailed in Online Supplementary Table S2.

Drug screening technology Cells were thawed, washed, counted and suspended in RPMI supplemented with 10% fetal calf serum and human cytokines: stem cell factor (100 ng/mL), interleukin 3 (10 ng/mL), interleukin 6 (20 ng/mL) and thrombopoietin (10 ng/mL). Sensitivity profiling was determined by dose-response over several log ranges (Online Supplementary Figure S1). Assay-ready plates in 384-well format were arrayed in 12 concentrations for each compound (2 nM -50 mM), using a Labcyte Echo 555 acoustic dispenser, and frozen until use. Patientsâ&#x20AC;&#x2122; samples were dispensed at approximately 1,000 cells per well and cultured for 2 days. Viable cells were quantified using Cell Titer Glo reagent (Promega Madison, WI, USA). Data were normalized using Genedata Screener software: dimethylsulfoxide-treated cells corresponded to 100% viability and samples without cells corresponded to 0% viability. Normalized data were loaded into CDD Vault software to calculate the half maximal inhibitory concentration (IC50), and minimum and maximum responses for each drug and patient. We compared the in vitro response to dasatinib in our study with the results of the Beat AML drug sensitivity and resistance testing (DSRT). In the Beat AML study in vitro cultures were without cytokines.17

RNA and exome sequencing The workflow for RNA sequencing was performed as previously described9 and the methods for RNA sequencing, exome sequencing, and gene expression and mutation analyses are detailed in the Online Supplementary Methods. For pathway analyses we used enrichr.19,20 haematologica | 2020; 105(12)

Response to dasatinib in AML

Prediction of response to dasatinib Dasatinib responders were defined in our cohort and in the Beat AML cohort as having an IC50 <0.01, and dasatinib non-responders in both cohorts as having an IC50 >0.8 (Online Supplementary Table S3). To identify predictors of response to dasatinib we initially selected genes that were differentially expressed (DE) between dasatinib responders and non-responders. Furthermore, other gene sets were selected based on pathway analysis of the DE genes and genes known to be targets of dasatinib as reported by the KinMap (Online Supplementary Table S4, genes with a Kd(nM) <7 [37 genes]). To refine feature selection, we tested which dasatinib target genes were DE between dasatinib responders and nonresponders from the Beat AML study (Online Supplementary Figure S3, Online Supplementary Table S4). Various machine-learning models were tested, all with a cross-validation of 10-fold, on the Beat AML data only. The best prediction model was the k-nearest neighbor algorithm with cosine similarity. This model was validated on the Israel National Center for Personalized Medicine (INCPM) gene expression data (Matlab statistical tool box).

In vivo dasatinib treatment SGM3 mice were given an intrafemoral injection of 1x106 CD3depleted AML human cells. From day 35 after the injection the mice were treated with 50 mg/kg dasatinib orally for 21 days and sacrificed on day 56. Fluorescence-activated cell sorting analysis was performed. Bone marrow cell suspensions were analyzed for expression of human CD45 (BV521, clone HL30), CD34 (APCcy7, clone 581), CD38 (PE cy7, clone 303516), CD15 (BV421, clone W6D3) (products from Biolegend), CD33 (APC), CD3 (FITC) and CD19 (PE) markers (products from Becton Dickinson). Human engraftment was defined as >0.1% of cells positive for human CD45 at the time of sacrifice. The percent engraftment of human cells was compared between drug- versus control-treated mice using the Wilcoxon test. The Institutional Animal Care and Use Committee of the Weizmann Institute approved the experiments, which were performed in accordance with its relevant guidelines and regulations (11790319-2).

Statistical analysis Heatmap and clustering of drug sensitivity analysis was performed with R package ComplexHeatmap. Differences between the defined groups were validated by statistical analysis. Enrichment analyses were performed using the Fisher exact test. Analysis of the DE genes was performed using the DESeq2 package (1.10.1)21 with the betaPrior, coks cutoff and independent filtering parameters set to “False”. Raw P values were adjusted for multiple testing using the procedure of Benjamini and Hochberg. DE genes were determined by an adjusted P<0.05, absolute fold changes >2 and maximum raw counts >30 (Figure 2, Online Supplementary Tables S5 and S6). Overlap between DE genes in the Beat AML and INCPM was tested by the hypergeometric distribution using the following R function sum (dhyper[t:b, a, n - a, b]). List a contains the DE genes in the Beat AML dataset, and b the DE genes in INCPM. The total number of expressed genes in both datasets was 20,055. The intersection between a and b is t (Figure 3A).

Results Classification of acute myeloid leukemia samples based on drug response, and mutation profile The overall goal of the current study was to identify drugs that can be safely given at diagnosis and target LSC. To achieve this goal, we first tried to identify drugs that haematologica | 2020; 105(12)

were more effective at relapse than at diagnosis. Such drugs would presumably target LSC, which are enriched at relapse. Comparing the DSRT between paired AML diagnostic and relapse samples we found that for the majority of tested compounds the IC50 was higher for the relapse samples (Figure 1A), although the difference was statistically insignificant in most cases. Clinically relevant drugs that are being used to treat patients at relapse (etoposide,22 midostaurin23 and fludarabine24) were significantly less effective at relapse (Figure 1A). Next, we took an unbiased approach to identify subgroups of AML sensitive to specific drugs in order to identify patients likely to benefit from specific targeted therapies. We screened 29 primary AML samples by DSRT using a test set of 46 drugs. Five compounds (BEZ235, lenalidomide, visomodegib, EPZ-5676 and AG-120) were excluded from the analysis because of the low variability in the response among the different samples (almost all samples were resistant to these drugs). Unsupervised hierarchical clustering analysis of IC50 values demonstrated two distinct drug response patterns (Figure 1B). One group displayed high sensitivity to TKI (dasatinib, MEK inhibitors, PI3K inhibitor [BKM120], pim inhibitor, ruxolitinib, mTORC1/2 inhibitor [INK128], quizartinib, midostaurin), BET inhibitors (OTX015, JQ1) and several other drugs. This group was termed sensitive. The other group was resistant to most compounds. The “resistant” group displayed heterogeneous sensitivity to BCL2 inhibitors (ABT737 and venetoclax), purine analogs (clofarabine, fludarabine,) and the hypomethylating agent decitabine. Interestingly, samples sensitive in vitro to BCL2 inhibitors (ABT-737 and venetoclax) did not cluster in a specific group, consistent with clinical observations (Figure 1B).8,25 To define the molecular pathways underlying the drug sensitivity of the clustering groups, we analyzed the somatic mutations (exome sequencing) and gene expression of the AML samples. We focused our genetic analysis on genes that are recurrently mutated in AML.26 Mutations in FLT3/ITD were observed in seven of 17 patients in the “sensitive” group and none of 12 in the “resistant” group (P=0.023). No other significant enrichment was identified in the current dataset.

Gene expression and mutation profile of dasatinib-sensitive acute myeloid leukemia cells Analysis of the in vitro DSRT assay identified a subgroup of AML samples sensitive to a broad range of TKI (Figure 1B). Among the TKI, the sensitivity to dasatinib was different between the two groups. In particular, 52% (9 of 17) of the AML samples in the group of “responders” displayed sensitivity to dasatinib (defined as a 10-fold difference in IC50 between sensitive and non-sensitive samples) whereas in the group of “non-responders” none of the samples showed in vitro sensitivity to dasatinib (P=0.003) (Figure 1B). Based on these results we focused our analysis on dasatinib, which has been safely used for many years as monotherapy or in combination with chemotherapy for Philadelphia-positive leukemia. Next, we tried to define those molecular features that could predict which AML samples would be sensitive to dasatinib. As our cohort was small, we compared our results to those of the Beat AML study.17 We classified dasatinib responders (n=61) and non-responders (n=134) based on IC50 values (Online Supplementary Table S3). Other parameters, including age, gender, leukocyte number, time of sample [diag2797

S. Tavor et al. A

2798

Figure 1. In vitro drug resistance. (A) In vitro drug resistance to most drugs is acquired after relapse. An average drug response of seven couples of primary acute myeloid leukemia (AML) samples, at diagnosis (DX) and relapse (REL) is shown, comparing the half maximal inhibitory concentration (IC50) of 41 drugs. Each dot represents the response to a specific drug, calculated by the median IC50 ratio of diagnosis vs. relapse. (B) Drug sensitivity and resistance testing (DSRT) of primary AML cells. Hierarchical clustering using Pearson dissimilarity and complete linkage was performed. Data are the log2 IC50 +0.001, standardized for each compound by reducing the mean. DSRT for 41 drugs of clinical and preclinical use in AML shows two clustered groups of patients. The age and gender of the patients and the origin of the patientsâ&#x20AC;&#x2122; diagnostic or relapse sample, the name of the drugs, the class of the drug, karyotype and mutation status for FLT3/ITD and NPM1 are shown.

haematologica | 2020; 105(12)

Response to dasatinib in AML

Figure 2. Transcriptome profiling of dasatinib responder samples. Gene expression analysis of whole transcriptome mRNA sequencing comparing dasatinib sensitive to non-sensitive samples of acute myeloid leukemia (AML). (A) Differentially expressed genes between "responders", and "non-responders". (B) In order to extend the tested samples we also applied the same analysis to AML patientsâ&#x20AC;&#x2122; samples from the Beat AML dataset.

Figure 3. Analysis of differentially expressed genes between dasatinib responders and non-responders from the INCPM and Beat AML datasets. (A) Intersection of upregulated (upper panel) and downregulated (lower panel) differentially expressed genes from the INCPM and Beat AML cohorts. (B) Pathway analysis of intersecting upregulated genes in both cohorts identified significant enrichment of genes that are co-expressed with several dasatinib targets. Dasatinib targets (CSF1R, SRC, BLK) are shown in red and the asterisk marks significant enrichment (false discovery rate<0.05).

haematologica | 2020; 105(12)

2799

S. Tavor et al.

nosis/relapse], type of AML [de novo/transformed], karyotype and mutation profile were also tested for correlation with response to dasatinib. Differential gene expression analysis of responders and non-responders to dasatinib in our cohort identified 297 genes significantly overexpressed and 404 genes downregulated in the sensitive group (false discovery rate <0.05) (Online Supplementary Table S5). In the Beat AML cohort, we identified 300 significantly upregulated and 720 downregulated genes in the dasatinib-sensitive group as compared to the resistant group (Online Supplementary Table S6). Unsupervised hierarchical clustering of all expressed genes enriched dasatinib responders in specific clusters both in our cohort (Figure 2A) and in the Beat AML cohort (Figure 2B). Intersecting the upregulated and downregulated DE genes from the Beat AML and the INCPM cohorts resulted in a significant number of intersecting genes

2800

(P<10-23 and P<10-21, respectively) (Figure 3A). Pathway analysis of intersecting upregulated genes in both cohorts identified significant enrichment of genes that are coexpressed with dasatinib targets (CSF1R, SRC) (Figure 3B, Online Supplementary Table S10). On the other hand, the same pathway analysis of intersecting downregulated genes did not enrich for dasatinib targets. As we identified upregulation of genes that are co-expressed with dasatinib targets we next performed a focused analysis on the expression of actual dasatinib targets. We identified all known dasatinib targets and performed unsupervised hierarchical clustering on their gene expression in both our cohort and the Beat AML cohort. In both cohorts a cluster enriched with dasatinib responders was identified. In the Beat AML cohort 13 of 21 samples (59%) were responders in this cluster as opposed to 61 out of 195 (31.3%) responders in the entire Beat AML cohort (P=0.016) (Online

Figure 4. Prediction model to identify dasatinib responders based on gene expression. (A, B) The k-nearest neighbor (KNN) algorithm with cosine similarity was used to predict the response to dasatinib. The greatest accuracy was achieved with 10fold cross-validation applied to the differentially downregulated genes from the Beat AML dataset. Accordingly, a sensitivity of 0.85 and a specificity of 0.7 were achieved. (C) Validation of the KNN cosine prediction model on INCPM data.

haematologica | 2020; 105(12)

Response to dasatinib in AML

Supplementary Figure S2). In both cohorts the dasatinib target cluster included three SRC family tyrosine kinases FGR, HCK and LYN as well as PTK6, CSK, GAK and EPHB2. Several of these genes were significantly overexpressed in dasatinib responders in the Beat AML cohort (Online Supplementary Figure S3). Next, we used the data from the gene expression analysis to establish a prediction model for identifying dasatinib responders in vitro based on gene expression. For all experiments we used a cross-validation approach, training the model only on the Beat AML data and validating results on the INCPM data. We tested the following genes sets from the Beat AML data for prediction: all genes, all DE genes, all upregulated genes, all downregulated genes, all dasatinib targets and DE dasatinib targets. The best prediction model, with a sensitivity of 0.85 and specificity of 0.7 (area under the curve=0.78) (Figure 4A,B), was achieved when using the Beat AML downregulated genes. When testing the same model on the INCPM data the positive predictive value was 100% and the false discovery rate was 60% (Figure 4C). These data provide evidence that gene expression can predict in vitro response to dasatinib, albeit with limited accuracy. As gene expression did not provide an ideal prediction for dasatinib response, we tested whether any other parameters from the Beat AML could predict response to dasatinib. We analyzed molecular and clinical data for enrichment in dasatinib responders. We found that samples carrying mutations in FLT3/ITD (P=0.011) (Online Supplementary Table S7), IDH2 and PTPN11 (P<0.05) were enriched in dasatinib responders in the Beat AML cohort. The IC50 for dasatinib in carriers of FLT3/ITD and PTPN11 was significantly lower than the IC50 of non-carriers (P<0.0001 and P=0.02, respectively) (Figure 5A, B and Online Supplementary Table S8), whereas only samples with mutations in TP53 had significantly higher IC50 compared to wild-type samples (P=0.0003) (Figure 5C and Online Supplementary Table S9). Adding the mutation status to the gene expression prediction model did not improve accuracy. Collectively, these data suggest that molecular features (mutations and gene expression profile) identify a group of AML cases responsive to dasatinib in vitro. Specifically, the fact the FLT3/ITD mutant cases are enriched in dasatinib responders in both the Beat AML and INCPM cohorts prompted us to test the in vivo efficacy of dasatinib in this specific group of patients. As our ultimate goal was to test whether in vitro response to dasatinib would predict in vivo AML response, we chose to focus our in vivo dasatinib response studies on samples known to carry FLT3/ITD. The focus on FLT3/ITD also stemmed from the fact that FLT3/ITD AML samples are known to be good AML engrafters.27

nine samples, engraftment was significantly reduced in the dasatinib-treated group. This result suggests that dasatinib has anti-LSC activity in a subset of FLT3/ITD positive AML cases.

Dasatinib reduces leukemia stem cells in mice engrafted with FLT3/ITD-mutant acute myeloid leukemia We injected NSG mice (n=5-10/sample) with cells from nine patients with FLT3/ITD AML. On day 35 the animals were randomized to dasatinib or a carrier control. Following 3 weeks of treatment we measured human engraftment in the treated and control groups (Figure 6). Fluorescence-activated cell sorting of bone marrow demonstrated leukemic engraftment (CD33+ >90% of human CD45+ cells) in mice from both groups. For four of haematologica | 2020; 105(12)

Figure 5. Correlation between FLT3/ITD, PTPN11 and TP53 mutations in acute myeloid leukemia samples and response to dasatinib in vitro in the Beat AML study. (A-C) Median dasatinib half maximal inhibitory concentration (IC50) values in cases with mutated FLT3/ITD (A), PTPN11 (B) or TP53 (C) and wild-type (WT) samples. Differences between groups were validated by the Wilcoxon rank-sum test.

2801

S. Tavor et al.

Discussion We provide evidence that a subset of FLT3/ITD+ AML cases and AML subtypes with a unique gene expression pattern are sensitive to dasatinib in vitro and that this sensitivity translates into sensitivity in vivo, at least in FLT3/ITD+ cases. We initially discovered sensitivity to dasatinib using DSRT and validated the DSRT results

based on the Beat AML study. DSRT can only measure toxicity to bulk blast cells and cannot detect effects on differentiation and LSC. Accordingly, we tested the response of FLT3/ITD+ primary AML samples to dasatinib in a xenograft model and confirmed response in vivo. FLT3/ITD is one of the most common mutations in AML, and is associated with high risks of relapse and mor-

Figure 6. Dasatinib delays development of human FLT3/ITD-mutated acute myeloid leukemia in transplanted mice. SGM3 mice were treated with dasatinib for 3 weeks after transplantation of human cells from nine patients with FLT3/ITD acute myeloid leukemia (AML). The percentage of human CD45+ (hCD45) cells in engrafted murine bone marrow with/without dasatinib treatment is shown after staining for CD45, CD34, CD33, CD38, CD15, CD3 and CD19 to determine myeloid lineage cells and analyzed by fluorescence-activated cell sorting immunostaining. Wilcoxon rank sum test, *P<0.05; **P<0.005; ***P<0.0005. The mutation status of the AML samples was: 140005: FLT3/ITD+, FLT3/TKD-, NPM1+; 40094: NPM1+, FLT3/ITD+; 150809: NPM1-, FLT3/ITD+; FLT3/ITD+, 150279: 40034: NPM1-, FLT3/ITD+, FLT3/TKD-, NPM1+; 130695: + + NPM1 , FLT3/ITD , FLT3/TKD ; 160436: NPM1+, FLT3/ITD+; 160406: FLT3/ITD+, NPM1+; 130607: FLT3/ITD+, NPM1+.

2802

haematologica | 2020; 105(12)

Response to dasatinib in AML

tality.28-31 Recently the introduction of FLT3 TKI has improved overall survival;32,33 however, not all FLT3/ITD+ patients respond to FLT3 inhibitors, whereas some FLT3/ITD- patients do.34 We show that FLT3/ITD AML cells are sensitive to dasatinib both in vitro and in xenograft models. FLT3/ITD results in constitutive activation of FLT3 and activates several signal transduction pathways such as RAS-MAPK, PI3K, JAK-STAT and SRC. This broad signal transduction activation might explain the heterogeneity observed in the response to FLT3 inhibitors. The current standard of care for patients with FLT3/ITD is midostaurin combined with chemotherapy. However, in the RATIFY study only 46.4% of patients achieving complete remission after midostaurin remained disease-free after 4 years with an ongoing risk of relapse, especially during the first year after treatment.23 The high failure rate even following combination therapy indicates that additional therapy is needed upfront. Previous studies by Weisberg et al. demonstrated synergistic effects for the combination of midostaurin and dasatinib in FLT3/ITDmutated AML cell lines.35 Our in vivo data demonstrate that the anti-leukemic activity of dasatinib extends to FLT3/ITD LSC, at least for a subset of patients, and suggest that addition of dasatinib to the current standard of care may be beneficial for FLT3/ITD AML patients. Our results also show that a unique gene expression signature correlates with response to dasatinib. Consistent with mechanistic predictions dasatinib responders overexpressed multiple dasatinib targets. SRC family kinase (SFK) genes, such as LYN, HCK36 and FGR (which are all dasatinib targets) are overexpressed in dasatinib responders in the Beat AML dataset (Figure 3B, Online Supplementary Figure S3), and were shown to be overexpressed in AML-LSC and to contribute to the survival and proliferation of these cells.37 Exposure to dasatinib treatment inhibited SFK phosphorylation in primitive and committed AML progenitors. In this study the combination of dasatinib and chemotherapy enhanced LSC targeting by p53 signaling.37 Furthermore recent evidence suggests that FLT3 activates SFK and these kinases in turn regulate the activity of the RAS/ERK pathways.38 Altogether, our data suggest that dasatinib may be active not only in FLT3/ITD AML but also in other subtypes of AML sharing a similar dasatinib responsive gene expression signature. Similarly, we showed that dasatinib may be effective in treating PTPN11-mutated AML, but likely not all PTPN11-mutated AML will respond to dasatinib

References 1. Tuval A, Shlush LI. Evolutionary trajectory of leukemic clones and its clinical implications. Haematologica. 2019;104(5):872-880. 2. McMahon CM, Perl AE. Gilteritinib for the treatment of relapsed and/or refractory FLT3mutated acute myeloid leukemia. Expert Rev Clin Pharmacol. 2019;12(9):841-849. 3. Perl AE. The role of targeted therapy in the management of patients with AML. Blood Adv. 2017;1(24):2281-2294. 4. Stone RM, Larson RA, Dohner H. Midostaurin in FLT3-mutated acute myeloid leukemia. N Engl J Med. 2017;377(19):1903. 5. DiNardo CD. Ivosidenib in IDH1-mutated acute myeloid leukemia. N Engl J Med. 2018;379(12):1186.

haematologica | 2020; 105(12)

and a larger cohort is needed to identify additional biomarkers. Mutations in PTPN11 are also found in 35% of patients with juvenile myelomonocytic leukemia and acute lymphocytic leukemia, and in 4-7% of AML patients. PTPN11 encodes the Shp2 cytoplasmatic protein tyrosine phosphatase, and SHP2 mutations cause increased phosphatase activity that contributes to leukemogenesis by upregulating the RAS, JAK-STAT, and PI3K pathways, leading to aberrant proliferation of myeloid progenitors.39,40 A recent study demonstrated that TNK2 activates PTPN11; furthermore, mutant myeloid leukemic cells carrying PTPN11-activating mutations were sensitive to TNK inhibition by dasatinib in vitro.41 Future studies are needed to better define the subset of PTPN11-mutated AML likely to respond to dasatinib and to test the activity of dasatinib activity in vivo. Our results suggest that the addition of dasatinib to the current standard of care for FLT3/ITD AML (an FLT3 inhibitor in combination with cytarabine/anthracycline) may benefit a subset of FLT3/ITD+ patients and should be tested in a clinical trial. Similarly, patients ineligible for intensive induction therapy may benefit from combining dasatinib, FLT3 TKI, hypomethylating agents and venetoclax. Clearly the potential toxicity of such combinations needs to be taken into account. Another possible clinical trial would be to investigate the addition of dasatinib to standard of care in patients with PTPN11-mutated AML. A unique gene expression signature, as detected by RNAsequencing, may predict response to dasatinib, suggesting that selection of patients based on RNA-sequencing may be included in future clinical trials. Given dasatinibâ&#x20AC;&#x2122;s activity against LSC, there is hope that addition of dasatinib during initial therapy will reduce the risk of subsequent relapse. Acknowledgments This research was supported by a grant from the Nancy and Stephen Grand Israel National Center for Personalized Medicine to HB and by the EU horizon 2020 grant project MAMLE ID: 714731, LLS grant ID: RTF6005-19, and IMOS-712843 awarded to LIS. LIS is an incumbent of the Ruth and Louis Leland career development chair. NK is an incumbent of the Applebaum Foundation Research Fellow Chair. This research was also supported by the Sagol Institute for Longevity Research, the Barry and Eleanore Reznik Family Cancer Research Fund, Steven B. Rubenstein Research Fund for Leukemia and Other Blood Disorders, the Rising Tide Foundation and the Applebaum Foundation.

6. Stein EM, DiNardo CD, Pollyea DA, et al. Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood. 2017;130(6):722-731. 7. Bisaillon R, Moison C, Thiollier C, et al. Genetic characterization of ABT-199 sensitivity in human AML. Leukemia. 2020;34(1): 63-74. 8. Pollyea DA, Stevens BM, Jones CL, et al. Venetoclax with azacitidine disrupts energy metabolism and targets leukemia stem cells in patients with acute myeloid leukemia. Nat Med. 2018;24(12):1859-1866. 9. Shlush LI, Mitchell A, Heisler L, et al. Tracing the origins of relapse in acute myeloid leukaemia to stem cells. Nature. 2017;547(7661):104-108. 10. Shlush LI, Chapal-Ilani N, Adar R, et al. Cell lineage analysis of acute leukemia relapse

uncovers the role of replication-rate heterogeneity and microsatellite instability. Blood. 2012;120(3):603-612. 11. Ding L, Ley TJ, Larson DE, et al. Clonal evolution in relapsed acute myeloid leukaemia revealed by whole-genome sequencing. Nature. 2012;481(7382):506-510. 12. Estey E, Levine RL, Lowenberg B. Current challenges in clinical development of "targeted therapies": the case of acute myeloid leukemia. Blood. 2015;125(16):2461-2466. 13. Eriksson A, Osterroos A, Hassan S, et al. Drug screen in patient cells suggests quinacrine to be repositioned for treatment of acute myeloid leukemia. Blood Cancer J. 2015;5:e307. 14. Snijder B, Vladimer GI, Krall N, et al. Imagebased ex-vivo drug screening for patients with aggressive haematological malignan-

2803

S. Tavor et al. cies: interim results from a single-arm, openlabel, pilot study. Lancet Haematol. 2017;4(12):e595-e606. 15. Pemovska T, Kontro M, Yadav B, et al. Individualized systems medicine strategy to tailor treatments for patients with chemorefractory acute myeloid leukemia. Cancer Discov. 2013;3(12):1416-1429. 16. Bose P, Vachhani P, Cortes JE. Treatment of Relapsed/Refractory Acute Myeloid Leukemia. Curr Treat Options Oncol. 2017;18(3):17. 17. Tyner JW, Tognon CE, Bottomly D, et al. Functional genomic landscape of acute myeloid leukaemia. Nature. 2018;562(7728): 526-531. 18. Gerstung M, Papaemmanuil E, Martincorena I, et al. Precision oncology for acute myeloid leukemia using a knowledge bank approach. Nat Genet. 2017;49(3):332-340. 19. Chen EY, Tan CM, Kou Y, et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics. 2013;14:128. 20. Kuleshov MV, Jones MR, Rouillard AD, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 2016;44(W1): W90-97. 21. Anders S, McCarthy DJ, Chen Y, et al. Count-based differential expression analysis of RNA sequencing data using R and Bioconductor. Nat Protoc. 2013;8(9):17651786. 22. Megias-Vericat JE, Martinez-Cuadron D, Sanz MA, et al. Salvage regimens using conventional chemotherapy agents for relapsed/refractory adult AML patients: a systematic literature review. Ann Hematol. 2018;97(7):1115-1153. 23. Stone RM, Mandrekar SJ, Sanford BL, et al. Midostaurin plus chemotherapy for acute myeloid leukemia with a FLT3 mutation. N Engl J Med. 2017;377(5):454-464. 24. Fiegl M, Unterhalt M, Kern W, et al. Chemomodulation of sequential high-dose cytarabine by fludarabine in relapsed or refractory acute myeloid leukemia: a ran-

2804

domized trial of the AMLCG. Leukemia. 2014;28(5):1001-1007. 25. DiNardo CD, Pratz K, Pullarkat V, et al. Venetoclax combined with decitabine or azacitidine in treatment-naive, elderly patients with acute myeloid leukemia. Blood. 2019;133(1):7-17. 26. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209-2221. 27. Sanchez PV, Perry RL, Sarry JE, et al. A robust xenotransplantation model for acute myeloid leukemia. Leukemia. 2009;23(11): 2109-2117. 28. Kottaridis PD, Gale RE, Frew ME, et al. The presence of a FLT3 internal tandem duplication in patients with acute myeloid leukemia (AML) adds important prognostic information to cytogenetic risk group and response to the first cycle of chemotherapy: analysis of 854 patients from the United Kingdom Medical Research Council AML 10 and 12 trials. Blood. 2001;98(6):17521759. 29. Yanada M, Matsuo K, Suzuki T, et al. Prognostic significance of FLT3 internal tandem duplication and tyrosine kinase domain mutations for acute myeloid leukemia: a meta-analysis. Leukemia. 2005;19(8):13451349. 30. Thiede C, Steudel C, Mohr B, et al. Analysis of FLT3-activating mutations in 979 patients with acute myelogenous leukemia: association with FAB subtypes and identification of subgroups with poor prognosis. Blood. 2002;99(12):4326-4335. 31. Kayser S, Schlenk RF, Londono MC, et al. Insertion of FLT3 internal tandem duplication in the tyrosine kinase domain-1 is associated with resistance to chemotherapy and inferior outcome. Blood. 2009;114(12):23862392. 32. Sasaki K, Kantarjian HM, Kadia T, et al. Sorafenib plus intensive chemotherapy improves survival in patients with newly diagnosed, FLT3-internal tandem duplication mutation-positive acute myeloid

leukemia. Cancer. 2019;125(21):3755-3766. 33. Uy GL, Mandrekar SJ, Laumann K, et al. A phase 2 study incorporating sorafenib into the chemotherapy for older adults with FLT3-mutated acute myeloid leukemia: CALGB 11001. Blood Adv. 2017;1(5):331340. 34. Rollig C, Serve H, Huttmann A, et al. Addition of sorafenib versus placebo to standard therapy in patients aged 60 years or younger with newly diagnosed acute myeloid leukaemia (SORAML): a multicentre, phase 2, randomised controlled trial. Lancet Oncol. 2015;16(16):1691-1699. 35. Weisberg E, Liu Q, Nelson E, et al. Using combination therapy to override stromalmediated chemoresistance in mutant FLT3positive AML: synergism between FLT3 inhibitors, dasatinib/multi-targeted inhibitors and JAK inhibitors. Leukemia. 2012;26(10):2233-2244. 36. Saito Y, Kitamura H, Hijikata A, et al. Identification of therapeutic targets for quiescent, chemotherapy-resistant human leukemia stem cells. Sci Transl Med. 2010;2(17):17ra9. 37. Dos Santos C, McDonald T, Ho YW, et al. The Src and c-Kit kinase inhibitor dasatinib enhances p53-mediated targeting of human acute myeloid leukemia stem cells by chemotherapeutic agents. Blood. 2013;122 (11):1900-1913. 38. Kazi JU, Ronnstrand L. The role of SRC family kinases in FLT3 signaling. Int J Biochem Cell Biol. 2019;107:32-37. 39. Li L, Modi H, McDonald T, et al. A critical role for SHP2 in STAT5 activation and growth factor-mediated proliferation, survival, and differentiation of human CD34+ cells. Blood. 2011;118(6):1504-1515. 40. Schubbert S, Lieuw K, Rowe SL, et al. Functional analysis of leukemia-associated PTPN11 mutations in primary hematopoietic cells. Blood. 2005;106(1):311-317. 41. Jenkins C, Luty SB, Maxson JE, et al. Synthetic lethality of TNK2 inhibition in PTPN11-mutant leukemia. Sci Signal. 2018;11(539):eaao5617.

haematologica | 2020; 105(12)

ARTICLE

Non-Hodgkin Lymphoma

Treatment of patients with MYC rearrangement positive large B-cell lymphoma with R-CHOP plus lenalidomide: results of a multicenter phase II HOVON trial

Martine E.D. Chamuleau,1 Coreline N. Burggraaff,1 Marcel Nijland,2 Katerina Bakunina,3 Rogier Mous,4 Pieternella J. Lugtenburg,5 Daan Dierickx,6 Gustaaf W. van Imhoff,2 Joost S.P. Vermaat,7 Eric A.F.Marijt,7 Otto Visser,8 Caroline Mandigers,9 Yavuz M. Bilgin,10 Aart Beeker,11 Mark F. Durian,12 Bas P. van Rees,13 Lara H. Bohmer,14 Lidwine W. Tick,15 Rinske S. Boersma,16 Tjeerd J.F. Snijders,17 Harry C. Schouten,18 Harry R. Koene,19 Eva de Jongh,20 Nathalie Hijmering,21 Arjan Diepstra,22 Anke van de Berg,22 Anne I.J. Arens,23 Julia Huijbregts,24 Otto Hoekstra,25 Josee M. Zijlstra,1 Daphne de Jong21 and Marie José Kersten26

Department of Hematology, Amsterdam UMC, VU University Amsterdam, Cancer Center Amsterdam, Amsterdam, the Netherlands; 2Department of Hematology, UMC Groningen, University of Groningen, Groningen, the Netherlands; 3Department of Hematology, HOVON Data Centre, Erasmus MC Cancer Institute, Rotterdam, the Netherlands; 4Department of Hematology, UMC Utrecht Cancer Centre, University Medical Centre Utrecht, Utrecht, the Netherlands; 5Department of Hematology Erasmus MC Cancer Institute, Rotterdam, the Netherlands; 6Department of Hematology, University Hospitals Leuven, Leuven, Belgium; 7 Department of Hematology, Leiden University Medical Centre, Leiden, the Netherlands; 8 Department of Hematology, Oncology Centre Isala, Zwolle, the Netherlands; 9Department of Hematology, Canisius-Wilhelmina Hospital, Nijmegen, the Netherlands; 10Department of Internal Medicine, Admiraal de Ruijter Hospital Goes, the Netherlands; 11Department of Internal Medicine, Spaarne Gasthuis, Hoofddorp, the Netherlands; 12Department of Internal Medicine, Tweesteden Hospital, Tilburg, the Netherlands; 13Department of Internal Medicine, Tjongerschans Hospital, Heerenveen, the Netherlands; 14Department of Internal Medicine, Haga Hospital, Den Haag, the Netherlands; 15Department of Internal Medicine, Máxima Medisch Centrum, Veldhoven, the Netherlands; 16Department of Internal Medicine, Amphia Hospital, Breda, the Netherlands; 17Department of Hematology, Medisch Spectrum Twente, Enschede, the Netherlands; 18Department of Hematology, Maastricht UMC, Maastricht, the Netherlands; 19Department of Internal Medicine, St Antonius Hospital, Nieuwegein, the Netherlands; 20Department of Internal Medicine, Albert Schweitzer Hospital, Dordrecht, the Netherlands; 21Department of Pathology, Amsterdam UMC, location VU University Amsterdam, Amsterdam, the Netherlands; 22 Department of Pathology and Medical Biology, UMC Groningen, University of Groningen, Groningen, the Netherlands; 23Department of Radiology and Nuclear Medicine, Radboud University Medical Centre, Nijmegen, the Netherlands; 24Department of Radiology and Nuclear Medicine, Gelre Hospital, Apeldoorn, the Netherlands; 25Department of Radiology and Nuclear Medicine, Amsterdam UMC, VU University Amsterdam, Amsterdam, the Netherlands and 26Department of Hematology, Amsterdam UMC, University of Amsterdam, Cancer Center Amsterdam, Amsterdam, the Netherlands

Ferrata Storti Foundation

Haematologica 2020 Volume 105(12):2805-2812

ABSTRACT

atients with MYC-rearrangement positive large B-cell lymphoma (MYC+ LBCL) have an inferior prognosis following standard first-line therapy with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) compared to patients without MYC rearrangement. Although intensive chemotherapy regimens yield higher remission rates, toxicity remains a concern. Lenalidomide is an oral immunomodulatory drug which downregulates MYC and its target genes thereby providing support using lenalidomide as additional therapeutic option for MYC+ LBCL. A phase II trial was conducted evaluating the efficacy of lenalidomide (15 mg day 1-14) in combination with R-CHOP (R2CHOP) in newly diagnosed MYC+ LBCL patients identified through a nationwide MYCFISH screening program. The primary endpoint was complete metabolic response (CMR) on centrally reviewed 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)-computer tomography (CT)-scan at end-of-treatment. Secondary endpoints were overall survival (OS), diseasehaematologica | 2020; 105(12)

Correspondence: MARTINE E.D. CHAMULEAU m.chamuleau@amsterdamumc.nl/ m.chamuleau@vumc.nl Received: September 25, 2019. Accepted: December 6, 2019. Pre-published: December 19, 2019. doi:10.3324/haematol.2019.238162 ©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.

2805

M.E.D. Chamuleau et al.

free survival (DFS) and event-free survival (EFS). Eighty-two patients with stage II-IV MYC+ LBCL were treated with six cycles of R2CHOP. At end of treatment, 67% (95% Confidence interval [CI]: 58-75) of the patients reached CMR. With a median follow-up of 25.4 months, 2-year estimates for OS, DFS, EFS were 73% (95% CI: 62-82), 75% (95% CI: 63-84) and 63% change to: (95% CI: 52-73) respectively. In this prospective trial for newly diagnosed MYC+ LBCL patients, we found that administering R2CHOP was safe, and yields comparable CMR and survival rates as in studies applying more intensive chemotherapy regimens. Hence, these findings offer new prospects for MYC+ LBCL patients and warrant comparison in prospective randomized clinical trials. This trial was registered at www.clinicaltrialsregister.eu (#2014-002654-39).

Introduction Diffuse large B-cell lymphoma (DLBCL) comprises about 35% of all non-Hodgkin lymphomas (NHL) and is the most common lymphoma subtype.1 The outcome of patients treated with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) is heterogeneous for which the IPI score and cell-of-origin (COO) are the most well-known denominators.2,3 MYC rearrangement status is an independent prognostic factor, and is reported in 10-15% of DLBCL patients (hereafter MYC+ LBCL).4-7 In about 30% of these patients, only a single MYC rearrangement is found (single hit [SH]), while in 70% MYC rearrangement is detected together with either a BCL2 or BCL6 rearrangement (double hit [DH]: MYC+/BCL2+ or MYC+/BCL6+) or with both (triple hit [TH]: MYC+/BCL2+/BCL6+).4 It has been shown that in patients with MYC+ LBCL, standard first-line therapy with R-CHOP results in an inferior prognosis compared to those without MYC rearrangement (2-year OS 35% vs. 61%8 and 5-year OS 31% vs. 66%6). Moreover, patients with MYC+ LBCL have an increased risk of central nervous system (CNS) relapse.5,6 Recently, Rosenwald et al. demonstrated that the inferior prognosis of MYC rearranged patients is however largely observed in patients with DH/TH lymphoma.7 In the revised World Helath Organisation (WHO) 2017 classification, SH is not recognized as a separate entity in contrast to DH/TH lymphoma.1 In search for improvement, intensified chemotherapy regimens, such as hyper-CVAD and R-CODOX-M/RIVAC, have been investigated. Data mainly come from subanalyses of MYC rearrangement positive patients in trials designed for unselected DLBCL patients. These studies indicate that intensified treatment results in improvement of progression free survival (PFS), but not OS.9-11 Only recently, a prospective, multicenter, single arm phase II study specifically designed for MYC+ LBCL patients showed that DA-EPOCH-R resulted in a promising CMR rate at end of treatment (EOT) of 74% and 4 year EFS and OS of 71% and 77%, respectively.12 Lenalidomide is an oral immunomodulatory drug with direct antitumor effects and indirect effects on the tumor microenvironment.13 In vitro studies have demonstrated that lenalidomide exposure results in down-regulation of MYC and its target genes via cereblon and IRF4 in lymphoid cells, thereby providing the rationale for introdu-cing lenalidomide as a therapeutic option in MYC+ LBCL.14 Two phase II studies in ABC/non-GCB-subtype DLBCL have demonstrated that the addition of lenalidomide to R-CHOP (R2CHOP) is indeed feasible and may contribute to a favorable outcome by decreasing CNS relapse.15,16 Against expectation, R2CHOP did not result in a survival advantage in ABC-subtype DLBCL, as has recently been shown in a phase III study (ROBUST).17 2806

The present study reports the results of a prospective single-arm phase II trial for MYC+ LBCL patients treated with R2CHOP. Patients were identified through a nationwide molecular biomarker diagnostics program. We report outcome based on the primary endpoint, which was CMR by centrally reviewed a 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)-computer tomography (CT) scan at EOT, as well as 2-year OS, DFS and EFS rates.

Methods Screening program and patient eligibility To support timely diagnosis of MYC+ LBCL and optimal enrolment in the present clinical trial, a nationwide diagnostic support program for MYC rearrangement assessment by fluorescence in situ hybridization (FISH) was implemented.18 Patients ≥18 years with newly diagnosed DLBCL or with B-cell lymphoma, unclassifiable, with features intermediate between DLBCL and Burkitt lymphoma (BLC-U) according to the WHO 2008 classification with a proven MYC rearrangement by FISH analysis including SH (not Burkitt lymphoma), DH or TH DLBCL were eligible. During the screening period one cycle of R-CHOP, a short course of steroids, or irradiation to control local symptoms was allowed. Patients with Ann Arbor stage II-IV, a WHO performance status (PS) of 0-3, ≥ one lesion of ≥1.5 cm on a contrast-enhanced CT scan and ≥one positive lesion on PET-CT scan were eligible. Patients diagnosed with any other subtype of aggressive B-cell lymphoma, a history of follicular lymphoma, proven CNS localization or HIV positivity were excluded.

Treatment Treatment consisted of six cycles of standard R-CHOP every 3 weeks plus lenalidomide 15 mg orally on day 1-14 (R2CHOP; Online Supplementary Table S1), followed by two additional rituximab administrations. Prophylactic intrathecal methotrexate or cytarabine (≥4 administrations), pegfilgrastim, venous thromboembolism prophylaxis (with aspirin or low-molecular-weight-heparin), and Pneumocystis prophylaxis were mandatory.

Safety assessments Adverse events were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events (AE), version 4.03. AE grade 1 were not reported.

Study overview This multicenter, phase II study was designed by investigators of HOVON and was approved by the medical ethics committee of the Amsterdam UMC. All patients provided written informed consent. The trial was conducted in accordance with the haematologica | 2020; 105(12)

MYC+ lymphoma treated with R-CHOP + lenalidomide

Declaration of Helsinki and Good Clinical Practice guidelines. An independent data and safety monitoring board conducted a review during the planned interim analysis.

Endpoints The primary endpoint of the study was CMR on EOT PET-CT scan as determined by central review plus EOT bone marrow (BM) examination in case of BM localization at diagnosis. In case a BM examination was not repeated at EOT in patients with baseline BM localization and the EOT-PET scan showed no BM uptake localization, the response was classified as CMR (based on recent findings that CMR on PET-CT has a high negative predictive value for BM localization).23,24 Secondary endpoints were: OS defined as time from registration to death; DFS defined as time from achievement of first CMR on protocol until relapse or death whichever comes first; EFS defined as the time from registration to lack of CMR on EOT-PETCT, relapse or death; and positive predictive value (PPV) and negative predictive value (NPV) of iPET-CT for EOT result.

Statistical analyses An optimal Simon two-stage design was used with a response rate of 45% as the null hypothesis, and 60% as the alternative hypothesis. With a statistical significance level of 5% and a power of 80%, the required number of patients was 77, with an interim analysis for futility involving the first 26 included patients. In order to overcome dropouts due to ineligibility, 85 patients were enrolled. All efficacy analyses were restricted to eligible patients, while safety analyses included all enrolled patients. Data cut-off was June 28, 2019. For the clinical protocol, central pathology review, central PETCT review and additional statistical information, see the Online Supplementary Data.

Results Clinical characteristics From April 2015 to February 2018, 85 patients were included from 20 hospitals in the Netherlands and Belgium. Three patients were declared ineligible (two because the MYC+ status was based on immunohistochemistry and not on FISH and one because of a transformed lymphoma), leaving 82 patients for efficacy and 85 patients for safety analyses. Baseline patient and disease characteristics are shown in Table 1. The median age was 63 years (range: 28-82 years). 49 of 81 patients (60%) had a WHO performance status (PS) of 0; 58 of 71 patients (71%) had stage IV disease, and 42 of 82 patients (51%) had â&#x2030;Ľ2 extranodal localizations. The IPI score was high-intermediate and high in 65% of patients. During treatment 12 of 82 patients went off protocol before completion (progressive disease [n=7], toxicity [n=2]; pulmonary embolism and diarrhea, other reasons [n=3]; new diagnosis of colon cancer, patient refusal, and vertebral fracture), see Figure 1.

Pathology review Diagnostic biopsy samples of all 85 patients were available for pathology review. Results of all 82 eligible patients are summarized in Table 1 and the Online Supplementary Table S2. A diagnosis of DLBCL according to the WHO 2008 classification was confirmed in 65 of 82 patients (79%) and BCL-U in 12 of 82 patients (15%) and haematologica | 2020; 105(12)

morphology was indecisive between DLBCL and BCL-U in 5 of 82 patients (6%). For classification according to the WHO classification 2017 see the Online Supplementary Table S2. In 81 of 82 patients MYC rearrangement was confirmed at central review. Based on the intention to treat principle, the one patient in whom MYC rearrangement could not be confirmed was included in all analyses. In 9 of 82 cases, insufficient material was available to perform additional BCL2 and BCL6 rearrangement. In 73 of 82 cases, data on BCL2 and BCL6 rearrangement were available: 20 of 82 (26%) had a single MYC rearrangement (SH); 44 of 82 (54%) had DH lymphoma (31 patients had MYC/BCL2 rearrangements and 13 patients MYC/BCL6 rearrangements), and 9 of 82 (11%) had all three rearrangements (TH). COO classification using a standard Hans algorithm showed GCB phenotype in 63 of 71 (89%) and non-GCB phenotype in 8 of 71 (11%). Lymph2Cx classification was performed in 38 cases showing GCB-subtype in 29 of 38 patients (76%), ABC-subtype in 7 of 38 patients (18%), and intermediate subtype in 2 of 38 patients (5%). Out of the 24 DH of TH patients, 21 showed GCB-subtype and 3 ABC-subtype. Out of the 12 SH patients, 8 showed GCB-subtype and 4 ABC-subtype.

Treatment Most patients (n=68) started with lenalidomide in the second cycle and continued lenalidomide for 14 days after the sixth cycle of R-CHOP. When MYC FISH results were available at diagnosis, R2CHOP was started in the first cycle (n=14) (Figure 1). Patients received a median (interquartile range [IQR]) dose of the planned drugs in the R-CHOP regimen as follows: cyclophosphamide 99.9% (99.0-101); vincristine 100% (72.5-100); doxorubicin 99.5% (97.7-101); prednisone 100% (100-100); rituximab 98.1 (95.1-100); pegfilgrastim 100% (100-100). Lenalidomide was given at a median dose intensity of 100% (range: 85.7-100). 57 of 82 patients (70%) received the planned â&#x2030;Ľ4 intrathecal prophylactic administrations.

Primary endpoint: CMR at EOT At EOT PET-CT, 55 of 82 patients (67%) reached the primary endpoint of CMR (95% CI: 58-75, P<0.001), 5 of 82 patients (6%) reached a partial metabolic response (PMR), and 21of 82 patients (26%) had progressive metabolic disease (PMD) (Table 2). One patient went off protocol due to toxicity after cycle 5 without EOT PET-CT (response unknown). Univariate logistic regression analysis of baseline characteristics (BM localization, WHO performance status, stage, B symptoms, IPI, number of extranodal sites and age) did not reveal any significant predictors for reaching CMR. Exploratory descriptive subgroup analyses revealed no differences between SH and DH/TH patients regarding achievement of the primary endpoint: CMR rate in both groups was 70% and 66% respectively (nine patients with unknown BCL2 and BCL6 rearrangement not included).

Secondary endpoints: survival analyses With a median follow-up of 25.4 months (IQR 18.330.3), 1-year OS was 85% (95% CI: 76-91), DFS 77% (95% CI: 65-85) and EFS 66% (95% CI: 54-75). 2-year estimates for OS, DFS, EFS were and 73% (95% CI: 62-82), 2807

M.E.D. Chamuleau et al.

75% (95% CI: 63-84) and 63% (95% CI: 52-73) respectively (Figure 2A-C). Baseline patient characteristics (BM localization, WHO performance status, stage, B symptoms, IPI, number of extranodal sites and age) were not significantly predictive for prolonged OS in a univariate analysis at the 5% significance level. Univariate regression analyses indicated that SH and DH/TH patients had comparable EFS and DFS, however DH/TH patients had a tendency for a higher risk of death compared to SH patients (Hazard ratio [HR]4.18, P=0.055; 95% CI: 0.97-18.02) (Online Supplentary Figure S1A-C). Separate analyses of DH MYC/BCL2 and DH MYC/BCL6 and TH in comparison to SH revealed no significant differences in OS (Online Supplementary Figure S2A-B). In univariate analyses with response as time dependent covariate we found that patients who had achieved CMR at EOT PET experienced a reduced risk of death compared to patients who had not achieved CMR (HR 0.1, 95% CI: 0.03–0.33, P<0.001), (Figure 3). EOT PET-CT predicted relapse within 12 months, with a positive predictive value (PPV) of 81% and a negative predictive value (NPV) of 93% (Online Supplementary Table S3A). In total, 29 patients showed progressive disease (11 without achieving CMR, 18 after achieving CMR [(at interim or EOT PET-CT)] including one patient with a CNS relapse.

Safety Grade 2, 3 and 4 AE were seen in 27 (32%), 33 (39%) and 14 (16%) of all 85 registered patients respectively (Table 3). The most common grade 3–4 AE were neutropenia (18%), infections (14%) and gastrointestinal disorders (14%). Four patients experienced deep venous thrombosis (grade 2), and two patients pulmonary embolism (grade 3). Two of these patients (one with deep venous thrombosis and one with pulmonary embolism) had not received the mandatory thrombosis prophylaxis (protocol violation). One patient went off protocol due to grade 3 diarrhea. 71 serious AE were reported in 36 patients; 66 were due to hospitalization (42% infections, 26% gastrointestinal disorders), four to other conditions [two second primary malignancies, two recurrence of previously diagnosed (>5 year) malignancies]. One patient died during treatment due to progression. There were no treatment related deaths.

Observational analysis: predictive value of iPET-CT At iPET-CT after three cycles of R2CHOP, 57 of 82 patients (70%) were in CMR; of these 45 of 57 (79%) were still in CMR and 11 of 57 (19%) showed PMD at EOT PETCT, and one missed EOT evaluation (Table 2). 23 of 82 patients (28%) were in PMR at iPET-CT; 10 of 23 (43%) of these converted to CMR, 4 of 23 (17%) remained in PMR, 9 of 23 (39%) showed PMD at EOT. The PPV of iPET-CT for predicting EOT PET-CT result was 60% (15 of 25), the NPV 79% (45 of 57) (Online Supplementary Table S3B).

Discussion From retrospective series it is clear that first-line R-CHOP therapy is not sufficiently effective for patients with MYC+ LBCL with CR rates of 40-50% and 3-year OS rates of 35% only.6,8 Intensified chemotherapy regi2808

mens such as R-CODOX-M/R-IVAC or autologous stem cell transplantation have not improved OS, and result in increased toxicity.5,9-11 We designed a prospective clinical trial for MYC+ LBCL patients, in which a time window of one cycle of R-CHOP was allowed to perform molecular diagnostics. This approach permitted high risk patients to start treatment immediately and overcame the bias of inclusion of mainly

Table 1. Patient demographics and disease characteristics. Patients completed treatment

Median age (range) in years Sex male female WHO performance status 0 1 2 3 Prior treatment no 1 course of R-CHOP only corticosteroids Ann Arbor stage II III IV Extranodal localisations 0 1 ≥2 LDH> ULN yes no unknown Bone Marrow involvement yes no not done IPI Low Low-intermediate High-intermediate High Morphology (WHO2008) DLBCL BCL-U indecisive between DLBCL or BCL-U COO IHC (Hans classification) GCB subtype Non-GCB subtype Not evaluable COO GEP (Nanostring) n=38 GCB subtype ABC subtype Intermediate FISH analysis single hit double hit MYC+/BCL2+ MYC+/BCL6+ triple hit MYC+ (BCL2 and BCL6 status unknown)

82 63 (28-82)

100

56 26

68 32

49 26 5 2

60 32 6 2

13 68 1

16 83 1

12 12 58

15 15 71

31 9 42

38 11 51

57 20 5

70 24 6

16 44 22

20 54 27

12 17 32 21

15 21 39 26

65 12 5

79 15 6

63 8 11

77 10 13

29 7 2

76 18 5

20 44 31* 13** 9 9

24 54 11 11

Demographics and disease characteristics of 82 MYC+ LBCL patients treated with R2CHOP. LBLC: large B-cell lymphoma; DLBCL: diffuse large B-cell lymphoma; WHO: World Health Organisation; R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone; FISH: fluorescence in situ hybridization, LDH: lactate dehydrogenase, ULN: upper limit of normal; GCB: germinal center B-cell subtype ; COO: cellof-origin; IHC: immune-histochemistry; GEP: gene expression profiling; BCL-U; B-cell lymphoma, unclassifiable, with features intermediate between DLBCL and Burkitt lymphoma. *from 4 of these patients BCL6 status is unknown, **from 1 of these patients BCL2 status is unknown.

haematologica | 2020; 105(12)

MYC+ lymphoma treated with R-CHOP + lenalidomide

lower risk patients due to enrolment delays.25 In this trial we show that the addition of lenalidomide to R-CHOP resulted in EOT CMR rate of 67% CMR and 2-year survival rates of 73%, 75%, and 63% for OS, DFS and EFS respectively. To our knowledge, this is the second prospective trial especially designed for MYC+ LBCL

patients. Recently, Dunleavy and colleagues reported a single arm phase II study in which the efficacy of DAEPOCH-R for MYC+ LBCL patients was explored.12 Results for EOT CMR and survival rates are largely comparable between both approaches with EOT CMR rate of 74%, 4-year OS rates of 77% and EFS of 71% in the study

Figure 1. Disposition of the patients. Eighty-two patients were included. In 14 patients, MYC fluorescence in situ hybridization (FISH) was performed immediately at diagnosis, these patients started with R2CHOP (lenalidomide in combination with rituximab cyclophosphamide, doxorubicin, vincristine, and prednisolone) in cycle 1. In 68 patients, MYC results became available during the first cycle of R-CHOP; these patients were registered after the first cycle of R-CHOP and started with R2CHOP in the second cycle and continued lenalidomide for 14 days after the sixth cycle of R-CHOP. During treatment 13 patients went off protocol (progressive disease [n=7], toxicity [n=2; pulmonary embolism and diarrhea], other reasons [n=3; new diagnosis of colon cancer, patient refusal, and vertebral fracture]). R: rituximab, R2: rituximab + lenalidomide

haematologica | 2020; 105(12)

2809

M.E.D. Chamuleau et al. A

Figure 2. Survival analyses. (A) Overall survival (OS; time from registration to death, n=82); (B) disease-free survival (DFS; time from achievement of first complete metabolic response [CMR] on protocol until relapse or death whichever comes first, n=69); (C) event-free survival (EFS; defined as the time from registration to lack of CMR on end of treatment [EOT] positron emission tomography [PET]-computer tomography [CT] scan, relapse or death, n=82) of MYC+ LBCL patients.

Table 2. Response rates on interim and EOT PET-CT scan. Interim CMR PMR PMD total

EOT

CMR

PMR

PMD

unknown§

total

n n n n

n 45 10 0 55

n 0 4 1 5

n 11 9 1 21

n 1 0 0 1

n 57 23 2 82

Response rates on interim and EOT PET-CT scan. Correlation of interim and end of treatment (EOT) response rates by centrally reviewed positron emission tomography (PET)computer tomography (CT)scan. §One patient was in complete metabolic response (CMR) at interim scan but went off protocol due to toxicity without an EOT scan. PMR: partial metabolic response, PMD: progressive metabolic disease.

of Dunleavy. When compared to the trial of Dunleavy, our patient population was larger (82 vs. 53 patients), comparable in age (median 63 vs. 61 years) but included more patients with IPI ≥3 (65% vs. 49%) and more patients with DH/TH (65% vs. 45%). Regarding safety, grade 3/4 infections were seen in 24% of cycles with DA-EPOCH-R versus grade 3 (and no grade 4 infections) in only 2,8 % of cycles (18 episodes) with R2CHOP. DA-EPOCH-R resulted in three treatment related deaths vs. none with R2CHOP. Lenalidomide penetrates the CNS, and thereby may aid to prevent CNS relapses as has been suggested for nongerminal center B-cell (GCB) subtype lymphoma patients treated with R2CHOP.15 Indeed, in this study, which combined lenalidomide and intrathecal prophylaxis, a remarkably low rate of CNS relapse at a median follow-up of 2810

25.4 months was seen (n=1). Several remarks regarding our study can be made. First, although correlation CMR at EOT PET-CT with survival in MYC+ LBCL has been described in a retrospective study,26 one might argue that it is not an ideal primary endpoint. Given the high FDG-avidity of MYC+ LBCL and the fact that CMR at EOT PET-CT in our study was highly predictive for DFS (NPV of 93%), we feel that using CMR at EOT PET-CT as a surrogate endpoint for highly aggressive B-cell lymphomas such as MYC+ LBCL is justified. Second, clinical prognostic markers, including age, stage, IPI score, as well as COO were not significantly correlated to CMR on EOT PET-CT and survival, which might be explained by the inclusion of high risk patients; 65% of our patients have an IPI score of ≥3, versus only 27% in Ziepert’s meta-analysis of the value of IPI in the rituximab era.2 haematologica | 2020; 105(12)

MYC+ lymphoma treated with R-CHOP + lenalidomide

Figure 3. Survival according to end-oftreatment PET-CT scan result. Patients who have achieved complete metabolic response (CMR) at the end of treatment (EOT) positron emission tomography (PET)-computer tomography (CT) scan experienced a reduced risk of death compared to patients who have not yet achieved CMR (Hazard ratio [HR] 0.1, 95% Confidence interval [CI]: 0.03– 0.33, P<0.001). Response was simplified to “CMR” versus “no-CMR”.

Table 3. Adverse events.

grade2 Hematologic neutropenia febrile neutropenia thrombocytopenia anemia Infectious Vascular disorders pulmonary embolism deep venous trombosis superfical thromboflebitis Nervous system disorders (PNP) Gastrointestinal disorders Hepatobiliary disorders (ALT, AST increased) General disorder Any*

grade 3 %

2 6 15

2 7 18

5 6 2 5 12

6 7 2 6 14

4 4 25 17 5 11 27

5 5 29 20 6 13 32

9 12 1 5 33

11 14 1 6 39

grade 4 n

2 1 14

2 1 16

Adverse events Adverse events (AE) were graded per patient (maximum grade per cycle) according to National Cancer Institute Common Terminology Criteria for AE. AE events grade 1 were not reported. *In this row each patient is counted only once with the highest grade AE experienced. PNP: polyneuropathy, ALT: alanine transaminase; AST: aspartate transaminase.

Furthermore, our patient population included patients with SH lymphoma (24%) based on previous reports demonstrating poor prognosis of these patients following R-CHOP.5,6,8,27 However, in the revised WHO 2017 classification, SH is not recognized as a separate entity in contrast to DH/TH lymphoma. Recently, Rosenwald et al. demonstrated that the inferior prognosis of MYC rearranged patients is largely observed in patients with DH lymphoma.7 However, SH patients still have a worse prognosis compared to patients without a MYC rearrangement, although this is not statistically significant when regarding OS (P=0.077). Our trial was not powered to study the prognostic impact of SH versus DH/TH in the R2CHOP setting. haematologica | 2020; 105(12)

Finally, we explored the role of iPET-CT scanning as a tool for early identification of refractory MYC+ LBCL. In non-selected cases of DLBCL, CMR on iPET-CT after two to four R-CHOP cycles has a high NPV for 2-year PFS, but the PPV varies widely.28 In our study, the PPV of iPET-CT for achievement of CMR on EOT PET-CT was only 60% and therefore does not support the use of an interim PETCT scan as interpreted with the current standard criteria to identify primary refractory cases treated with R2CHOP. In this prospective trial for newly diagnosed MYC+ LBCL patients, we found that administering R2CHOP was safe, and yielded comparable CMR and survival rates as in studies applying intensive chemotherapy regimens. Moreover, R2CHOP can be delivered on an outpatient 2811

M.E.D. Chamuleau et al.

basis in contrast to Burkitt schemes and is easier to deliver than DA-EPOCH-R, since it does not require placement of a central line. These findings offer new prospects for MYC+ LBCL patients and warrant comparison in prospective randomized trials. Acknowledgments The authors wish to acknowledge: the trial managers and central data managers of the study (N. Lamers and A. Elsinghorst, R. Sewsaran and H. Hofwegen), HOVON data Center in Rotterdam; Dr Staudtâ&#x20AC;&#x2122;s Laboratory at NCI/NIH Bethesda, MD, USA, for online analysis of Lymph2Cx raw data for COO char-

References 1. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-2390. 2. Ziepert M, Hasenclever D, Kuhnt E, et al. Standard International prognostic index remains a valid predictor of outcome for patients with aggressive CD20+ B-cell lymphoma in the rituximab era. J Clin Oncol. 2010;28(14):2373-2380. 3. Lenz G, Wright GW, Emre NC, et al. Molecular subtypes of diffuse large B-cell lymphoma arise by distinct genetic pathways. Proc Natl Acad Sci U S A. 2008;105(36):13520-13525. 4. Aukema SM, Kreuz M, Kohler CW, et al. Biologic characterization of adult MYCtranslocation positive mature B-cell lymphomas other than molecular Burkitt lymphoma. Haematologica. 2014;99(4):726-735. 5. Oki Y, Noorani M, Lin P, et al. Double hit lymphoma: the MD Anderson Cancer Center clinical experience. Br J Haematol. 2014;166(6):891-901. 6. Savage KJ, Johnson NA, Ben-Neriah S, et al. MYC gene rearrangements are associated with a poor prognosis in diffuse large B-cell lymphoma patients treated with R-CHOP chemotherapy. Blood. 2009;114(17):35333537. 7. Rosenwald A, Bens S, Advani R, et al. Prognostic significance of MYC rearrangement and translocation partner in diffuse large B-cell lymphoma: a study by the Lunenburg Lymphoma Biomarker Consortium. J Clin Oncol. 2019;37(35):33593368. 8. 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):3360-3365. 9. Howlett C, Snedecor SJ, Landsburg DJ, et al. Front-line,dose-escalated immunochemotherapy is associated with a significant progression-free survival advantage in patients with double-hit lymphomas: a systematic review and meta-analysis. Br J Haematol. 2015; 170(4):504-514. 10. Petrich AM, Gandhi M, Jovanovic B, et al. Impact of induction regimen and stem cell transplantation on outcomes in double-hit

2812

11.

12.

13.

14.

15.

16.

17.

18.

19.

acterization; and the DSMB members M. Jerkeman, K. Dunleavy and E. van Werkhoven for their advice following the interim analysis. Funding This work was supported by Celgene who funded the research, provided lenalidomide free of charge, approved the original protocol and all amendments as written, but had no role in either the design of the study, the collection, analysis, and interpretation of the data, or the preparation of the manuscript. The Dutch Cancer Society provided funding for the screening program and the clinical trial.

lymphoma: a multicenter retrospective analysis. Blood. 2014;124(15):2354-2361. Landsburg DJ, Falkiewicz MK, Maly J, et al. Outcomes of patients with double-hit lymphoma who achieve first complete remission. J Clin Oncol. 2017;35(20):2260-2267. 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. Witzig TE, Wiernik PH, Moore T, et al. Lenalidomide oral monotherapy produces durable responses in relapsed or refractory indolent non-Hodgkin's Lymphoma. J Clin Oncol. 2009;27(32):5404-09. Lopez-Girona A, Heintel D, Zhang LH, et al. Lenalidomide downregulates the cell survival factor, interferon regulatory factor-4, providing a potential mechanistic link for predicting response. Br J Haematol. 2011; 154(3):325-336. Ayed AO, Chiappella A, Pederson L, et al. CNS relapse in patients with DLBCL treated with lenalidomide plus R-CHOP (R2CHOP): analysis from two phase 2 studies. Blood Cancer J. 2018;8(7):63. Castellino A, Chiappella A, LaPlant BR, et al. Lenalidomide plus R-CHOP21 in newly diagnosed diffuse large B-cell lymphoma (DLBCL): long-term follow-up results from a combined analysis from two phase 2 trials. Blood Cancer J. 2018;8(11):108. Vitolo U, Witzig TE, Gascoyne RD, et al. ROBUST: first report of phase III randomized study of lenalidomide/R-CHOP (R2CHOP) vs placebo/R-CHOP in previously untreated ABC-type diffuse large B-cell lymphoma. Haematol Oncol. 2019;37(52):36-37. Chamuleau M, Nijland M, Lamers N, et al. First Report on a successful screening program for MYC rearrangements and a prospective clinical trial based on MYC rearrangement in newly diagnosed DLBCL patients in the Netherlands. Blood. 2017; 130(Supplement 1):4144. 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):12142117.

20. Barrington SF, Mikhaeel NG, Kostakoglu L, et al. Role of imaging in the staging and response assessment of lymphoma: consensus of the International Conference on Malignant Lymphomas Imaging Working Group. J Clin Oncol. 2014;32(27):3048-3058. 21. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014; 32(27):3059-3068. 22. Boellaard R, Delgado-Bolton R, Oyen WJ, et al. FDG PET/CT: EANM procedure guidelines for tumour imaging: version 2.0. Eur J Nucl Med Mol Imaging. 2015;42(2):328-354. 23. Teagle AR, Barton H, Charles-Edwards E, Dizdarevic S, Chevassut T. Use of FDG PET/CT in identification of bone marrow involvement in diffuse large B cell lymphoma and follicular lymphoma: comparison with iliac crest bone marrow biopsy. Acta Radiol. 2017;58(12):1476-1484. 24. Berthet L, Cochet A, Kanoun S, et al. In newly diagnosed diffuse large B-cell lymphoma, determination of bone marrow involvement with 18F-FDG PET/CT provides better diagnostic performance and prognostic stratification than does biopsy. J Nucl Med. 2013; 54(8):1244-1250. 25. Maurer MJ, Ghesquieres H, Link BK, et al. Diagnosis-to-treatment interval is an important clinical factor in newly diagnosed diffuse large B-cell lymphoma and has implication for bias in clinical trials. J Clin Oncol. 2018;36(16):1603-1610. 26. Cohen JB, Geyer SM, Lozanski G, et al. Complete response to induction therapy in patients with Myc-positive and double-hit non-Hodgkin lymphoma is associated with prolonged progression-free survival. Cancer. 2014;120(11):1677-1685. 27. Landsburg DJ, Falkiewicz MK, Petrich AM, et al. Sole rearrangement but not amplification of MYC is associated with a poor prognosis in patients with diffuse large B cell lymphoma and B cell lymphoma unclassifiable. Br J Haematol. 2016;175(4): 631-640. 28. Burggraaff CN, de Jong A, Hoekstra OS, et al. Predictive value of interim positron emission tomography in diffuse large B-cell lymphoma: a systematic review and metaanalysis. Eur J Nucl Med Mol Imaging. 2019; 46(1):65-79.

haematologica | 2020; 105(12)

ARTICLE

Plasma Cell Disorders

Targeting CD47/TNFAIP8 by miR-155 overcomes drug resistance and inhibits tumor growth through induction of phagocytosis and apoptosis in multiple myeloma

Ferrata Storti Foundation

Nasrin Rastgoo,1 Jian Wu,1 Aijun Liu,2 Maryam Pourabdollah,1 Eshetu G. Atenafu,3 Donna Reece,4 Wenming Chen2 and Hong Chang1,2

Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada; 2Department of Hematology, Beijing Chaoyang Hospital, Capital University Beijing, Beijing, China; 3Department of Biostatistics, University Health Network, Toronto, Ontario, Canada and 4Department of Hematology and Medical Oncology, University Health Network, Toronto, Ontario, Canada 1

Haematologica 2020 Volume 105(12):2813-2823

ABSTRACT

he mechanisms of drug resistance in multiple myeloma (MM) are poorly understood. Here we show that CD47, an integrin-associated receptor, is significantly up-regulated in drug resistant myeloma cells in comparison with parental cells, and that high expression of CD47 detected by immunohistochemistry is associated with shorter progression-free and overall survivals in MM patients. We show that miR-155 is expressed at low levels in drug resistant myeloma cells and is a direct regulator of CD47 through its 3´UTR. Furthermore, low miR-155 levels are associated with advanced stages of disease. MiR-155 overexpression suppressed CD47 expression on myeloma cell surface, leading to induction of phagocytosis of myeloma cells by macrophages and inhibition of tumor growth. MiR-155 overexpression also re-sensitized drug-resistant myeloma cells to bortezomib leading to cell death through targeting TNFAIP8, a negative mediator of apoptosis in vitro and in vivo. Thus, miR-155 mimics may serve as a promising new therapeutic modality by promoting phagocytosis and inducing apoptosis in patients with drug-refractory/relapsed MM.

Introduction Multiple myeloma (MM) is a plasma cell malignancy characterized by abnormal proliferation of clonal plasma cells in the bone marrow (BM).1 Current therapies, such as the proteasome inhibitor bortezomib (BTZ), have improved the outcome of patients. Nevertheless, MM remains an incurable disease with a high rate of relapse and development of drug resistance.2 So far, the pathogenic mechanisms underlying drug resistance of MM have not been fully elucidated. Ubiquitously expressed CD47 is known to block phagocytosis by binding to signal regulatory protein α (SIRPα) on macrophages. This receptor-ligand binding can also inhibit initiation of an innate immune response.3,4 High expression of CD47 is associated with tumor growth, metastasis, recurrence, or drug resistance in hematologic malignancies.5-7 In plasma cell neoplasms, transition from monoclonal gammopathy of undetermined significance (MGUS) to multiple myeloma (MM) has been associated with a significant increase in CD47 expression.8 It has also been shown that CD47 is expressed in different sub-populations of peripheral blood mononuclear cells (PBMC); however, its highest expression is observed in MM cells and it is significantly increased in the presence of a tumor microenvironment.9 In addition, CD47 plays a role in bone resorption in MM by inducing macrophage fusion and osteoclast formation.10 However, the clinical relevance of CD47 expression in MM patients and the role of CD47 upregulation in MM drug resistance have not yet been established. In the current study, we examined the expression of CD47 in MM patients’ cells and assessed its correlation with clinical outcomes. Moreover, to gain an insight into regulatory mechanisms underlying CD47 high expression in MM, we sought to explore if miRNAs could be involved. This is supported by the fact that miRNAs play an important role in the pathogenesis of MM. Particularly, deregulated miRNAs can haematologica | 2020; 105(12)

Correspondence: HONG CHANG hong.chang@uhn.on.ca Received: May 21, 2019. Accepted: November 27, 2019. Pre-published: November 28, 2019. doi:10.3324/haematol.2019.227579 ©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.

2813

N. Rastgoo et al.

influence gene-expression and functional responses of MM cells.11 Importantly, miRNAs are now considered as a new class of agents for therapeutic intervention.12 Here, we found that the tumor suppressor miRNA, miR-155, directly targeted CD47 and was under-expressed in drug resistant MM cells. Restoration of miR-155 in drug resistant MM cells induced phagocytosis of MM cells by macrophages. Moreover, miR-155 overexpression sensitized MM resistant cells to bortezomib by targeting TNF-Îą-induced protein 8 (TNFAIP8). TNFAIP8 acts as a negative mediator of apoptosis and may play a role in tumor progression. It suppresses the TNF-mediated apoptosis by inhibiting caspase-8 activity but not the processing of procaspase-8, subsequently resulting in inhibition of BID cleavage and caspase-3 activation.13-15 The negative correlation between miR-155 and its targets, CD47 and TNFAIP8, is significantly associated with disease progression in MM patients. Thus, the dysregulation of miR-155/CD47/TNFAIP8 axis contributes to drug resistance and represents a new therapeutic target for MM.

Methods Patients A total of 74 cases diagnosed with MM were included in this study. The selection criteria were availability of the clinical/laboratory data and pathologic specimens as well as receiving the same treatment protocol: 4-5 cycles of vincristine, Adriamycin and dexamethasone as induction and one course of melphalan 200 mg/m2 followed by autologous stem cell transplant (ASCT). The BM specimens used for staining were obtained before starting any treatment. There were 42 males and 32 females with median age of 55 (range: 34-73) years. Median follow-up time after transplant was 4.91 (range: 0.20-15.94) years (Online Supplementary Table S1). This study was approved by the research ethics committee of University Health Network, Toronto, in accordance with the Declaration of Helsinki.

Myeloma cell lines and primary multiple myeloma samples The MM parental cell lines (RPMI-8226 and MM.1S) and MM.1R, which is resistant to dexamethasone, were obtained from ATCC. RPMI-8226-R5, a multidrug-resistant MM cell line that is cross-resistant to BTZ, was kindly provided by Dr. R Buzzeo.16 Resistance of RPMI-8226-R5 and MM.1R to the proteasome inhibitors BTZ and MG132 had been shown in our previous study.17 All cell lines were cultured in complete RPMI-1640 medium supplemented with 10% FBS, as described previously.18 CD138+ cells were freshly isolated and purified from the bone marrow of MM patients and normal healthy donors. Details of immunohistochemical (IHC) staining, cell culture and generation of stable cell lines, cytotoxicity and apoptosis assay, immunostaining cell surface targets and indirect flow cytometry, luciferase reporter assay, miRNA mimics transfection, real-time polymerase chain reaction (RT-PCR), phagocytosis assay, animal xenograft model studies and statistical analysis are all available in the Online Supplementary Methods.

Results CD47 is increased in multiple myeloma and its overexpression correlated with disease progression and poor survival of multiple myeloma patients To investigate the potential clinical significance of 2814

CD47 expression in MM, we evaluated CD47 protein expression in a cohort of 74 newly diagnosed MM patients by immunohistochemical analysis on consecutive tissue sections of CD138 positive myeloma cells (Figure 1A). The results showed that low (30-155) and high (160-240) H-score (Online Supplementary Figure S1A) were present in 58% and 42% of the cases, respectively. High score was associated with shorter median progression-free survival (PFS) and overall survival (OS) in comparison to low score (PFS: 11.4 vs. 25.46 months, P=0.0005; OS: 32.1 vs. >90 months, P=0.0001, respectively) (Figure 1B and C). High CD47 expression was associated with 17p (p53) deletions (P=0.0407) and elevated b2 microglobulin level (P=0.0323) (Online Supplementary Table S1), two well-known poor risk factors in MM; it was also correlated with higher percentage of myeloma cells in the BM specimens (P=0.0157) (Online Supplementary Figure S1B). Multivariate analysis adjusting above three variates confirmed that high CD47 expression was an independent poor risk factor for PFS (HR=0.465, 95% confidence 0 interval [CI]: 0.258-0.839, P=0.0110) and OS (HR=0.149, 95%CI: 0.060-0.373, P<0.0001). There was no significant association between CD47 protein expression and other clinical or biological factors such as age, sex, hemoglobin, creatinine, calcium, albumin, and other cytogenetic risk factors including 13q deletion, t(4;14), 1p21 deletion, and 1q21 (CKS1B) amplification (Online Supplementary Table S1). Moreover, in support of our experimental finding, analyses of publicly available data in the CoMMpass database for 767 MM patients showed that OS and PFS were significantly shorter in patients with high CD47 mRNA level (n=383) in comparison to patients with low CD47 mRNA level (n=384) (Online Supplementary Figure S1C). In addition, the CD47 mRNA level was significantly higher in the MM cells from relapsed patients compared to newly diagnosed MM patients and normal plasma cells (Figure 1D). We have also identified correlation between higher CD47 expression level and progression of the disease in MM patients in bortezomib clinical trials (Online Supplementary Figure S1D). In addition, we evaluated the endogenous CD47 expression in two drug-resistant MM cell lines (8226-R5 and MM.1R) and their parental lines (8226 and MM.1S). By Western blot the CD47 expression level was significantly higher in 8226-R5 and MM.1R cells in comparison to 8226 and MM.1S, respectively (Figure 1E), indicating that CD47 level is associated with drug response in MM cells.

MiR-155 is down-regulated in multiple myeloma cells and directly targets CD47 To elucidate a molecular mechanism underlying CD47 overexpression in drug resistant MM cells, several miRNA-target prediction algorithms were exploited19-21 to identify miRNAs which can potentially target CD47; 82 CD47 targeting miRNA candidates were found. Comprehensive bioinformatics analysis on miRNA/mRNA MM patient datasets on CD47 targeting miRNA candidates was carried out to identify the miRNAs that had a negative correlation with CD47 expression level in MM patients. We found four miRNAs (miR-425, miR-135b, and miR-326 and miR-155) were negatively correlated with CD47 expression (Figure 2A and Online Supplementary Figure S2A-C). Furthermore, analyzing the patient dataset revealed that among these haematologica | 2020; 105(12)

Novel mechanism of miR-155/CD47/TNFAIP8 axis in MM drug resistance

CD47 (P=0.0005)

Overall survival (%)

Progression-free survival (%)

High expression Low expression

ed ps la re

M M w ne

CD47 expression level

CD47 (P=0.0001) High expression Low expression

Figure 1. Upregulation of CD47 in multiple myeloma (MM) is correlated with poor patient survival. (A) Representative immunohistochemistry (IHC) staining with CD47 antibody in two MM patients with low and high expression of CD47. (B and C) Kaplan-Meier plots indicate the progression free survival (PFS) (B) and overall survival (OS) (C) for 67 MM patients categorized by CD47 expression level. P-value is determined by log-rank test. (D) The box whisker plot shows CD47 expression in different stages of MM as shown in the graph, GSE6477 dataset. Normal donor (ND) n=15, newly diagnosed MM n =73, and Relapsed MM n=28. One-way analysis of variance and Student t-test was used. P<0.05, **P<0.01 were considered as significant at 95% confidence interval. (E) CD47 expression was determined by Western blot analysis in 8226 versus 8226-R5 and MM.1S versus MM.1R cells, respectively. ASCT: autologous stem cell transplant.

miRNAs, miR-155 was significantly lower in stage III of the MM disease in comparison to stage I and II and healthy donor samples (Figure 2B), confirming the clinical relevance between miR-155 downregulation and the advanced stage of MM disease. In addition, the expression level of miR-155 was assessed in 8226-R5 and MM.1R resistant cell lines by quantitative PCR and compared to 8226 and MM.1S parental cells. MiR-155 showed a significant downregulation in 8226-R5 and MM.1R relative to 8226 and MM.1S, respectively (Figure 2C). To determine whether CD47 expression was selectively regulated by miR-155, we transfected two drug resistant MM cell lines with synthetic miR-155 mimics and identified that CD47 protein level was suppressed in haematologica | 2020; 105(12)

miR-155 transfected cell lines, suggesting miR-155 is a specific regulator of CD47 in myeloma cells (Figure 2D and E). The target scan analysis revealed a conserved domain within the 3′ UTR of CD47 with a potential miR155 binding site (Figure 2F). To validate CD47 as a direct target of miRNA-155, the 3ÚTR sequence of human CD47 was cloned into the luciferase-expressing vector pEZX-MT01 downstream of the firefly luciferase gene. In order to further substantiate the site-specific repression of miR-155 on CD47, we constructed a mutant 3ÚTR reporter clone and miR-155 binding site in which the 3ÚTR of CD47 was inactivated by several mutations. Co-transfection of cells with the mutant or wild type of CD47-UTR luciferase reporter vectors together with miR2815

N. Rastgoo et al. B

P value: 0.0178 R= -0.3824

Figure 2. CD47 is a direct target of miR-155 in multiple myeloma (MM) cells. (A) Correlation analysis of endogenous miR-155 with CD47 expression in patient dataset (GSE17498, n=38 MM patients) presented as scatter plots. Linear regression with Pearson’s correlation coefficients (r) and P-value were presented in the graph. (B) Box whisker plot showing miR-155 expression in GSE17498 dataset. Normal donor (ND): n=3, stage I (Durie-Salmon): n=9, stage II (Durie-Salmon): n=9 and stage III (Durie-Salmon): n=16. (C) The total RNA including miRNA was isolated from MM resistant and sensitive cell lines and the level of miR-155 was assessed by qualitative polymerase chain reaction (qPCR). The values were normalized to SNORD72. (D) The 8226-R5 and MM.1R cells were transfected with 20 mM of either miR-155 mimics or scramble control using Hiperfect transfection reagent for 48 hours (h), and the level of miR-155 was assessed by qPCR. (E) The 8226-R5 and MM.1R cells were transfected with 20 mM of either miR-155 mimics or scramble control using Hiperfect transfection reagent for 48 h, and the whole cell lysate was subjected to Western blot with indicated antibodies. (F) The miR-155 binding site in the CD47 3′-UTR. Putative conserved target sites in the CD47 3′-UTR were identified using the TargetScan algorithm. Matched nucleic acid–base pairs were linked as “-” (G) The 3’ UTR sequence of human CD47 was cloned into the luciferaseexpressing vector pEZX-MT01 to the downstream of the firefly luciferase gene. For construction a mutant 3’UTR reporter clone, the miR-155 binding site in the 3’UTR of CD47 was inactivated by several mutations. The cells were transiently co-transfected with reporter plasmids (pEZX-MT-Control, pEZX-wt-CD47-3′UTR or pEZX-mutCD47-3′UTR) and miR-155 mimics or control miRNA. Cells were harvested 48 h after transfection and luciferase activities were analyzed as the relative activity of firefly to Renilla. Readings from the empty plasmid (pEZX-MT-control) were used for normalization. *P<0.05, **P<0.01, ***P<0.001 were considered as significant at 95% confidence interval. NS: not significant.

155 mimics or scramble control showed that miR-155 mimics significantly reduced wild type CD47-UTR reporter luciferase activity, but not that of the mutantUTR reporter (Figure 2G), indicating that miR-155 can directly target the CD47 3´UTR.

MiR-155-mediated loss of cell surface CD47 promotes the phagocytosis of multiple myeloma cells Previous studies suggested that cells may lose surface CD47 during apoptosis to enable phagocytic clearance.22,23 To demonstrate the effect of miR-155 on cell surface CD47 level, we transfected two drug resistant MM cell lines with synthetic miR-155 mimics. The FACS analysis revealed that the level of cell surface CD47 was significantly decreased in miR-155 over-expressing resistant MM cells (Figure 3A). Blocking CD47 with anti-CD47 monoclonal antibodies has enabled the phagocytosis of myeloma cells.24 We investigated whether targeting of CD47 by miR-155 overexpression promotes the phagocytosis of myeloma cells. Our results showed that macrophage-like cells differentiated from THP-1 mono2816

cytes phagocytosed human myeloma cells at a low frequency when treated with scramble control. However, miR-155 overexpression in drug resistant 8226-R5 and MM.1R MM cells significantly increased the phagocytosis of the cells (Figure 3B and C). Additionally, we explored whether overexpression of CD47 can rescue the effect of miR-155 on phagocytosis of myeloma cells by macrophages. To this end, we performed functional rescue assays by the ectopic expression of CD47 in 8226-R5 and MM.1R resistant cells and co-transfecting with synthetic miR-155 mimics or miRNA negative control. The CD47 overexpression partially abolished the phagocytosis of MM cells by macrophages confirming that targeting of CD47 by miR-155 increased the phagocytosis of resistant MM cells (Figure 3D and E). These data indicate that increased phagocytosis of myeloma cells by high level miR-155 expression results as a consequence of CD47 targeting by miR-155 which leads to reduction of cell surface CD47 on MM cells and consequently reduction of the 'do not eat me signal', followed by phagocytosis of MM cells by macrophages. haematologica | 2020; 105(12)

Novel mechanism of miR-155/CD47/TNFAIP8 axis in MM drug resistance

Figure 3. Targeting of CD47 by miR-155 increased the phagocytosis of myeloma resistant cells. (A) The 8226-R5 and MM.1R cell lines were transfected with synthetic miR-155 or scramble control for 48 hours (h) and stained with CD47 antibody and analyzed by flow cytometry to determine the level of CD47 in the cell surface of multiple myeloma (MM) cells. The percentage of CD47 stained cells is indicated in each histogram. (B) The GFP-tagged 8226-R5 and MM.1R cell lines were transfected with synthetic miR-155 or scramble control for 48 h and co-cultured with THP-1 like macrophages for 2 h at 37Â°. Macrophages were repeatedly washed four times with PBS 1X and subsequently imaged with fluorescent microscope. A representative result of each condition is shown. Scale bars are 100 mm. (C) The phagocytic index was calculated as the number of phagocytosed GFP+ cells per 100 macrophages. (D) The GFP-tagged 8226-R5 and MM.1R cell lines were transduced with CD47 lentiviral or empty vector. The CD47 transduced or control cells were transfected with synthetic miR-155 or scramble control for 48 h and co-cultured with THP-1 like macrophages for 2 h at 37Â°. Macrophages were repeatedly washed and subsequently imaged with fluorescent microscope. A representative result of each condition is shown. Scale bars are 100 mm. (E) The phagocytic index was calculated as the number of phagocytosed GFP+ cells per 100 macrophages. *P<0.05, **P<0.01, were considered as significant based on the 95% of confidence intervals. The result is the sum of three triplicate experiments. NS: not significant.

haematologica | 2020; 105(12)

2817

N. Rastgoo et al. A

P value: 0.0264 r: -0.3078

Figure 4. miR-155 restoration inhibits cell proliferation and induces apoptosis in multiple myeloma (MM) cells. (A) The 8226-R5 and MM.1R cell lines were transfected with synthetic miR-155 or scramble control and treated with different concentrations of BTZ for 48 hours (h) and cell viability was measured using MTT assay. The result is the sum of three triplicate experiments. (B) The 8226-R5 and MM.1R cells were transfected with 20 mM of either miR-155 mimics or scramble control using Hiperfect transfection reagent for 24 h, and then treated with 10 nM BTZ or vehicle for 24 h. Then the cells were stained with annexin-V/propidium iodide and analyzed by flow cytometry to determine the percentage of apoptotic cells. The percentage of double positive cells indicated in each dot plot as a representation of apoptosis. (C) Correlation analysis of endogenous miR-155 with TNFAIP8 expression in patient dataset (GSE17306) presented as scatter plots. Linear regression with Pearson’s correlation coefficients (r) and P-value were presented in the graph. (D) TNFAIP8 expression was determined by qualitative-polymerase chain reaction (q-PCR) a in 8226 versus 8226-R5 and MM.1S versus MM.1R cells, respectively. (E and F) Cell proliferation was assessed in RPMI-8226 and MM.1S cells with or without stable TNFAIP8 overexpression alone or in combination with 5 nM BTZ. (G) RPMI-8226 and MM.1S parental cell and TNFAIP8 over-expressing cells were subjected to Western blot with indicated antibodies. (H and I) The 8226-R5 and MM.1R cells were transfected with 20 mM of either miR-155 mimics or scramble control using Hiperfect transfection reagent for 48 h, and the TNFAIP8 level was assessed with q-PCR and Western blot. (J) The cells were transiently co-transfected with reporter plasmids (pEZX-MT-Control, pEZX-wt-TNFAIP8-3′UTR or pEZX-mut- TNFAIP8-3′UTR) and miR-155 mimics or control miRNA. Cells were harvested 48 h after transfection and luciferase activities were analyzed as the relative activity of firefly to Renilla. Readings from the empty plasmid (pEZX-MT-control) were used for normalization. (K) The 8226-R5 cells with or without stable TNFAIP8 overexpression were transfected with synthetic miR-155 or scramble control. Cell proliferation was assessed in the absence or presence of 5 nM BTZ 48 hrs after treatment (L). The 8226-R5 cells with or without stable TNFAIP8 overexpression were transfected with synthetic miR-155 or scramble control. The cell lysate was prepared 48 h after transfection and subjected to western blot with indicated antibodies. *P<0.05, **P<0.01, ***P<0.001,****P<0.0001 and significant difference based on the 95% of confidence intervals. NS: not significant.

2818

haematologica | 2020; 105(12)

Novel mechanism of miR-155/CD47/TNFAIP8 axis in MM drug resistance

MiR-155 inhibits multiple myeloma cell proliferation and induces apoptosis in resistant myeloma cells by targeting TNFAIP8 To examine the functional effect of miR-155 on cell proliferation and apoptosis in drug resistant MM cells, two resistant cell lines (8226-R5 and MM.1R) were transfected with miR-155 mimics or scramble control. We found that transfection of synthetic miR-155 in drug resistant cells could re-sensitize the resistant cells to BTZ in a dose and time dependent manner (Figure 4A and Online Supplementary Figure S3A). Furthermore, reduction of cell proliferation upon overexpression of miR-155 was associated with apoptosis of the resistant cells as indicated by an increase in the percentage of annexin-V positive cells compared to the control (Figure 4B). Collectively, these results support the concept that miR-155 functions as a tumor suppressor miRNA and its dysregulation contributes to MM cell growth and drug resistance. To understand the molecular mechanism of action of miR-155 on apoptosis, several MicroRNA target prediction databases were analyzed and the targets with highest target scores in at least four different databases were selected. Moreover, we assessed the oncogenic potential of the selected genes using OncoScore Internet-based tool (Online Supplementary Figure S3B). Furthermore, analyzing the MM patient datasets on the targets with the most oncogenic scores revealed that among these targets, TNFAIP8, a negative mediator of apoptosis, showed more clinical relevance with MM disease. Target scan analysis showed miR-155 can potentially target TNFAIP8 3’UTR (Online Supplementary Figure S3C). We identified a negative correlation (with the low/moderate coefficient value -0.24, -0.26 and -0.3) between miR-155 and TNFAIP8 mRNA expression level in MM patient samples in comparison to other potential targets (Figure 4C and Online Supplementary Figure S3D and E). We also found that the mRNA expression level of TNFAIP8 was higher in MM patients at relapsed stage compared to those at diagnosis (Online Supplementary Figure S3F). Moreover, in support of our finding, we evaluated the TNFAIP8 expression in two drugresistant MM cell lines (8226-R5 and MM.1R) and their parental lines (8226 and MM.1S). We found that the TNFAIP8 mRNA and protein levels were significantly higher in 8226-R5 and MM.1R cells in comparison to 8226 and MM.1S, respectively (Figure 4D and Online Supplementary Figure S3G). Notably, ectopic expression of TNFAIP8 in the parental cell lines promoted cell proliferation and increased resistance to BTZ by inhibiting caspase-8 activity and subsequently caspase-3 activation (Figure 4E-G), indicating that TNFAIP8 can modulate drug response in MM cells. In addition, overexpression of miR-155 in drug resistant MM cells significantly reduced mRNA and protein level of TNFAIP8 (Figure 4H and I). To validate TNFAIP8 as a direct target of miR-155, the 3ÚTR of human TNFAIP8 was cloned into the luciferaseexpressing vector pEZX-MT01 downstream of the firefly luciferase gene. We also constructed a mutant 3ÚTR reporter clone. To this end, the miR-155 binding site in the 3ÚTR of TNFAIP8 was inactivated by several insertions in miRNA binding site. Co-transfection of cells with the mutant or wild type of TNFAIP8-UTR luciferase reporter vectors together with miR-155 mimics or scramble control showed that miR-155 mimics significantly reduced wild type TNFAIP8-UTR reporter luciferase activity, but not that of the mutant-UTR reporter (Figure 4J), indicating that miRhaematologica | 2020; 105(12)

155 can directly target the TNFAIP8 3′ UTR. Reversely, overexpression of TNFAIP8 in MM.1S cells could significantly decrease the miR-155 expression level, indicating that miR-155 is also regulated by TNFAIP8 through a negative feedback loop (Online Supplementary Figure S3H). Additionally, we explored whether TNFAIP8 can rescue the effect of miR-155 on MM cell growth. To this end, we performed functional rescue assay by over-expressing TNFAIP8 in 8226-R5 and MM.1R resistant cells and cotransfecting with synthetic miR-155 mimics or miRNA scramble control. The MTT results revealed that TNFAIP8 overexpression partially abolished the anti-tumor effect induced by miR-155 plus BTZ treatment confirming that combining miR-155 and BTZ induces a growth inhibitory effect in MM cells by targeting TNFAIP8 (Figure 4K and Online Supplementary Figure S3I). In addition, Western blot results also confirmed that TNFAIP8 suppressed the TNFmediated apoptosis in MM resistant cells by inhibiting caspase-8 activity, subsequently resulting in inhibition of BID cleavage and caspase-3 activation (Figure 4L). Collectively, these results support the concept that miR-155 functions as a tumor suppressor miRNA and contributes to the effect of CD47 on MM cell growth and drug resistance.

Restoration of miR-155 by miRNA mimics suppresses tumorigenesis in multiple myeloma xenograft model To translate our findings to a therapeutic model, we next investigated the effect of miR-155 overexpression on tumorigenesis of MM resistant cells in vivo. We established a mouse xenograft model with 8226-R5 resistant MM cells in severe combined immunodeficient (SCID) mice and treated them with miR-155 mimics alone or in combination with BTZ. Intratumoral injection of miR-155 significantly suppressed tumorigenesis in combination with BTZ (Figure 5A) without showing any untoward toxicity as indicated by the body weight (Figure 5B). The combination treatment of miR-155 mimics and BTZ significantly extended the OS in comparison to all other three treatments (Figure 5C). To assess in vivo targeting of TNFAIP8 by miR-155, we evaluated the level of TNFAIP8 and CD47 in mice tumors. Consistent with our in vitro data, the protein levels of TNFAIP8 were dramatically decreased in miR-155-treated groups compared with control (Figure 5D). We also verified the efficiency of miR-155 mimics delivery into tumor cells by quantitative (q)-PCR (Figure 5E). IHC analysis of tumor sections showed that treatment with miR-155 mimics and BTZ resulted in a decrease in the proliferation index (Ki67) and an increase in the apoptotic index (Tunnel), compared to either BTZ or miR-155 mimics alone (Figure 5F). These findings indicate that targeting of TNFAIP8 by miR-155 sensitizes myeloma cells to BTZ treatment and contributes to the marked induction of apoptosis of MM cells, as well as suppression of MM tumor growth in vivo. The proposed model for mechanism of MM drug resistance is shown in Figure 6.

Discussion Increased CD47 expression has been reported in various patient tumor cells and in some cases, CD47 high level expression was correlated with a worse prognosis.5,6,25 Examination of global gene expression profiling of MM patients revealed that CD47 expression is up-regulated in MM patients in comparison to MGUS and normal donor 2819

N. Rastgoo et al. plasma cell samples.26 In addition; flow cytometric results demonstrated that CD47 protein expression was higher on MM cells in comparison to non-myeloma cells.24 To validate clinical relevance of CD47 overexpression on patient’s outcome, several public datasets for MM patients were analyzed and revealed that CD47 expression was significantly higher in relapsed patients in comparison to newly diagnosed MM patients. Moreover, we performed

IHC in a cohort of MM patients and to the best of our knowledge, this is the first study to demonstrate that high CD47 protein expression was associated with unfavorable patient’s outcome. Therefore, therapeutic intervention with CD47 targeting could be used as a promising strategy to treat MM. Consistent with our results, a recent IHC study on the clinical specimens of gastric cancer also indicated that CD47 positivity was an independent adverse

Figure 5. Synthetic miR-155 retards tumor growth and prolongs survival in human multiple myeloma (MM) mouse xenograft model. Intratumoral injection with either miR-155 mimics or control miRNA (1 mg/kg), with or without intraperitoneal injection of 0.5 mg/kg BTZ (n=5 mice per group) was carried out at an interval of 3 days for 15 days. (A) Overexpression of miR-155 enhanced BTZ-induced retardation of tumor growth in vivo. (B) Body weight was measured from first day of drug injection (day 20) every 3 days till day 42 and presented as mean ± standard error mean (SEM). (C) Survival was evaluated using Kaplan-Meier curves and log-rank analysis from the first day of tumor cells injection until death or occurrence of an event. (D) Mice tumors from in vivo experiment were analyzed by immunoblotting for CD47 and TNFAIP8 protein expression. (E) The total RNA including miRNA was isolated from the mice of four groups and the level of miR-155 was measured by qualitative-polymerase chain reaction (q-PCR) to evaluate the delivery efficiency of miRNA mimics into tumor cells after intratumoral injection of miR-155 mimics using the novel formulation of neutral lipid emulsion (NLE; MaxSuppressor in vivo RNA Lancer II, BIOO Scientific). Fold change was expressed as log2-fold induction over control group (mean ±SEM). (F) Representative microscopic images of immunohistochemical analysis of tumor sections from four treated groups with hematoxylin & eosin, the proliferation index (Ki-67) and the apoptotic index, TUNEL staining. *P<0.05, **P<0.01, ***P<0.001, and significant difference based on the 95% of confidence intervals.

2820

haematologica | 2020; 105(12)

Novel mechanism of miR-155/CD47/TNFAIP8 axis in MM drug resistance

prognostic factor.27 Our finding also showed that high CD47 protein expression was associated with higher BM myeloma cell infiltration. A possible explanation is that high CD47 expression could cause the escape of MM cells from immune surveillance and result in progression of the disease. To understand the mechanism of increased expression of CD47, we found an inverse correlation between the expression of CD47 and its potential negative regulator, miR-155. We provided strong evidence that miR-155 negatively regulated CD47 expression in MM. miR-155 was down-regulated in drug resistant MM cell lines compared to parental lines and inversely correlated with the progression of MM disease. We also showed that there was a significant negative correlation between miR-155 and CD47 expression levels in MM patients. The luciferase assay also revealed that 3â&#x20AC;&#x2122;UTR of CD47 was directly targeted by miR-155. In line with these findings, miRNA related CD47 overexpression has been reported in solid tumors, with high expression of CD47 being correlated with poor clinical outcome and tumor progression.28 Moreover, considering our observation that miR-155 mimics can downregulate CD47 protein in MM cells, and CD47 is reportedly a direct target of miR-155 in other contexts such as multiple sclerosis and adipogenesis,29,30 it indicates that low level of miR-155 in myeloma cells may contribute to upregulation of CD47. Furthermore, an important finding here is that overexpression of miR-155 in drug resistant MM cells could decrease the CD47 on the cell surface and

induce phagocytosis by macrophages through activation of an 'eat me' signal in resistant cells, whereas overexpression of CD47 antagonized this effect of miR-155. Consistent with our results, another study demonstrated that blocking CD47 increased phagocytosis of myeloma cells (in vitro) induced tumor regression and alleviated bone resorption in animal models.31 We demonstrated that restoration of miR-155 level in drug resistant MM cells by miRNA mimics could efficiently reduce cell proliferation and induce apoptosis in resistant cells by targeting of TNFAIP8. It should be mentioned that TNFAIP8 is known to counteract apoptosis by inhibiting caspase-8 activity, subsequently resulting in inhibition of BID cleavage and caspase-3 activation.32,33 It has also been shown that TNFAIP8 might serve as a predominant pro-tumor factor by modulating different modulators and molecular targets such as growth factor receptors (EGFR and VEGFR), cell cycle protein (Cyclin D1, phospho-Rb), cell surface transmembrane receptor (Integrin).34 We found that TNFAIP8 was significantly upregulated in drug resistant MM cells compared to parental cells. Analysis of the MM patient dataset also revealed that it was elevated in MM patients at relapsed stage in comparison to presentation time point, suggesting that TNFAIP8 is an oncogene and could contribute to drug resistance in MM. Along this line, other studies found evidence linking TNFAIP8 oncogene to the survival of several cancer types.35,36 Moreover, we have demonstrated that TNFAIP8 overexpression increased cell proliferation of

Figure 6. Proposed model for the role of CD47 overexpression in drug resistance of multiple myeloma (MM). Low level of miR-155 in drug resistant cells leads to upregulation of CD47 expression, which results in activation of 'do not eat me' signal and inhibition of phagocytosis by macrophages. miR-155 can also target anti-apoptotic gene, TNFAIP8 which inhibits apoptosis through caspase 8 inactivation. On the other hand, TNFAIP8 can regulate miR-155 expression level through a negative feedback loop.

haematologica | 2020; 105(12)

2821

N. Rastgoo et al.

parental sensitive MM cells and conferred resistance to BTZ in parental MM cells. Of note, it has also been shown that TNFAIP8 overexpression was correlated with drug resistance in several cancers including MM.37-39 Collectively, these data suggest that TNFAIP8 has a significant role in the oncogenesis and progression of human malignancies. Interestingly, we demonstrated that targeting of TNFAIP8 by miR-155 could re-sensitize the drug resistant MM cells to anti-myeloma drugs, and TNFAIP8 overexpression partially rescued the cytotoxic effect of miR-155 in drug resistant MM cells. These results suggest that, besides directly targeting the cell surface protein CD47 and promoting phagocytosis, ectopic expression of miR-155 could target the oncogene TNFAIP8 and induce apoptosis at the same time (Figure 6), whereas the previous studies found that anti-CD47 therapy could just increase phagocytosis of MM cells but did not induce antibody-dependent cell-mediated cytotoxicity (ADCC) or complement-dependent cytotoxicity (CDC) on myeloma cells.24 Our pre-clinical study on xenograft MM mouse model illustrated that synthetic miR-155 mimics can restrain the tumor growth and prolong survival of the resistant MM mice model. In line with these findings, a very recent study also showed that miR-155 has tumor suppressor activity in MM cells and antagonized bortezamib resistanse by targeting proteasome subunit b5.40 Several studies have demonstrated that miRNA may be applied for the targeted delivery of personalized medicine to improve the outcome of MM patients and the number of studies focusing on pre-clinical applications of miRNAs in MM is increasing.41-44 Importantly, formulated NLE-miR-34a was safely administered to mice bearing MM tumors by intratumor injection, suggesting a favorable therapeutic index for synthetic miRNAs.45 In particular, miRNA-based therapeutics can be relevant both for safety issues and to abrogate late onset of resistance because of the complexity of miRNA-targeted pathways and the consequent low chance of developing individual 'escape' mutations in the treated cells. In addition, although current therapeutic methods such as small molecule inhibitors or monoclonal antibodies have shown promise, there are many genes that are not druggable using these methods. An advantage of miRNA-based therapy is the ability to rapidly develop new therapies with broad adaptability. In addition, miRNA offers the opportunity to target multiple mRNA

References 1. Rollig C, Knop S, Bornhauser M. Multiple myeloma. Lancet. 2015;385(9983):21972208. 2. Sonneveld P and Broij A. Treatment of relapsed and refractory multiple myeloma. Haematologica. 2016;101(4):396-406. 3. Barclay AN, Berg TK. The interaction between signal regulatory protein alpha (SIRPα) and CD47: structure, function, and therapeutic target. Annu Rev Immunol. 2014;32:25-50. 4. Oldenborg PA, Gresham HD, Lindberg FP. CD47-signal regulatory protein alpha (SIRPalpha) regulates Fc gamma and complement receptor-mediated phagocytosis. J Exp Med. 2001;193(7):855-862.

2822

targets with a single miRNA in a given pathway.46,47 It is important to mention that, although miR-based therapeutics have demonstrated great promise for the treatment of different diseases, this is still an evolving field. The main obstacle for clinical application of RNAbased therapeutics is determining how to best deliver the agent to targeted cells. Recent progress in miRNA delivery such as nanoparticle-based technology shows great promise as it may reduce doses, which will be beneficial for treatment of cancer.48 Other factors such as safety, efficacy, and target selection will also require optimization to produce successful drugs. Engineered nanoparticles are especially used for delivery to specific cells, which will help to achieve this goal. RNA-based therapeutics combined with conventional chemotherapy agents might be a new approach to use for cancer treatment and will ultimately help to bring the RNA-based therapeutics strategy to the clinic. Nevertheless, therapeutic miRNAs definitely have the potential to contribute significantly to the future of medicine.42,49,50 In the current study, local administration of miRNA has been applied, which is the most commonly used model for miRNA-based therapeutic. Importantly, our study provides a proof of principle that miR-155 can effectively eradicate tumor growth in vivo by targeting CD47 and TNFIP8. In future studies, a xenograft model would be required to confirm the efficacy of the treatment, with an alternative route of administration. In conclusion, we show that CD47 could serve as an adverse prognostic factor in MM and demonstrate a novel mechanism of miR-155/CD47/TNFAIP8 axis in MM drug resistance. We illustrate a tumor suppressor role for miR155 in MM, which contributes to deregulation of CD47 and TNFAIP8 oncogene. Therefore, targeting CD47 by miR-155 mimics implies a novel therapeutic strategy for relapsed/refractory MM, particularly with high CD47 expression. Acknowledgments We thank Drs. Mark Minden and Eldad Zacksenhaus for helpful discussions. Funding The study was supported in part by the grants from Leukemia and Lymphoma Research Society of Canada (LLSC), and Cancer Research Society (CRS).

5. Majeti R, Chao MP, Alizadeh AA, et al. CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells. Cell. 2009;138(2):286-299. 6. Chao MP, Alizadeh AA, Tang C, et al. AntiCD47 antibody synergizes with rituximab to promote phagocytosis and eradicate nonHodgkin lymphoma. Cell. 2010;142(5):699713. 7. Galli S, Zlobec I, Schürch C, et al. CD47 protein expression in acute myeloid leukemia: a tissue microarray-based analysis. Leuk Res. 2015;39(7):749-756. 8. Rendtlew Danielsen JM, Knudsen LM, Dahl IM, Lodahl M, Rasmussen T. Dysregulation of CD47 and the ligands thrombospondin 1 and 2 in multiple myeloma. Br J Haematol. 2007;138(6):756-760.

9. Muz B, Kusdono H D, King J, et. al. Targeting CD47 As a novel therapeutic strategy in multiple myeloma. Blood. 2017; 130(Suppl 1):3099. 10. Tucci M, Stucci S, Strippoli S, Dammacco F, Silvestris F. Dendritic cells and malignant plasma cells: an alliance in multiple myeloma tumor progression? Oncologist. 2011; 16(7):1040-1048. 11. Seckinger A, Meißner T, Moreaux J, et al. miRNAs in multiple myeloma-a survival relevant complex regulator of gene expression. Oncotarget. 2015;6(36):39165-39183. 12. Zhang YK, Wang H, Leng Y, et al. Overexpression of microRNA-29b induces apoptosis of multiple myeloma cells through down regulating Mcl-1. Biochem Biophys Res Commun. 2011;414(1):233-239. 13. Kumar D, Whiteside TL, Kasid U.

haematologica | 2020; 105(12)

Novel mechanism of miR-155/CD47/TNFAIP8 axis in MM drug resistance

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Identification of a novel tumor necrosis factor-alpha-inducible gene, SCC-S2, containing the consensus sequence of a death effector domain of fas-associated death domainlike interleukin-1beta-converting enzymeinhibitory protein. J Biol Chem. 2000;275(4):2973-2978. You Z, Ouyang H, Lopatin D, Polver PJ, Wang CY. Nuclear factor-kappa B-inducible death effector domain-containing protein suppresses tumor necrosis factor-mediated apoptosis by inhibiting caspase-8 activity. J Biol Chem. 2001;276(28):26398-26404. Kumar D, Gokhale P, Broustas C, Chakravarty D, Ahmad I, Kasid U. Expression of SCC-S2, an antiapoptotic molecule, correlates with enhanced proliferation and tumorigenicity of MDA-MB 435 cells. Oncogene. 2004;23(2):612-616. Buzzeo R, Enkemann S, Nimmanapalli R, et al. Characterization of a R115777-resistant human multiple myeloma cell line with cross-resistance to PS-341. Clin Cancer Res. 2005;11(16):6057-6064. Yang Y, Chen Y, Saha MN, et al. Targeting phospho-MARCKS overcomes drug-resistance and induces antitumor activity in preclinical models of multiple myeloma. Leukemia. 2015;29(3):715-726. Saha MN, Chen Y, Chen MH, Chen G, Chang H. Small molecule MIRA-1 induces in vitro and in vivo anti-myeloma activity and synergizes with current anti-myeloma agents. Br J Cancer. 2014;110(9):22242231. Wang X. miRDB: a microRNA target prediction and functional annotation database with a wiki interface. RNA. 2008; 14(6):1012-1017. Wang X, El Naqa IM. Prediction of both conserved and non-conserved microRNA targets in animals. Bioinformatics. 2008; 24(3):325-332. Betel D, Wilson M, Gabow A, Marks DS, Sander C. The microRNA.org resource: targets and expression. Nucleic Acids Res. 2008;36(Database issue):D149-153. Burger P, Hilarius-Stokman P, de Korte D, van den Berg TK, van Bruggen R. CD47 functions as a molecular switch for erythrocyte phagocytosis. Blood. 2012; 119(23):5512-5521. Khandelwal S, van Rooijen N, Saxena RK. Reduced expression of CD47 during murine red blood cell (RBC) senescence and its role in RBC clearance from the circulation. Transfusion. 2007;47(9):1725-1732. Kim D, Wang J, Willingham SB, Martin R, Wernig G, Weissman IL. Anti-CD47 antibodies promote phagocytosis and inhibit the growth of human myeloma cells.

haematologica | 2020; 105(12)

Leukemia. 2012;26(12):2538-2545. 25. Chao MP, Alizadeh AA, Tang C, et al. Therapeutic antibody targeting of CD47 eliminates human acute lymphoblastic leukemia. Cancer Res. 2011;71(4):13741384. 26. Zhan F, Hardin J, Kordsmeier B, et al. Global gene expression profiling of multiple myeloma, monoclonal gammopathy of undetermined significance, and normal bone marrow plasma cells. Blood. 2002;99(5):17451757. 27. Yoshida K, Tsujimoto, H, Matsumura K, et al. CD47 is an adverse prognostic factor and a therapeutic target in gastric cancer. Cancer Med. 2015;4(9):1322-1333. 28. Suzuki S, Yokobori T, Tanaka N, et al. CD47 expression regulated by the miR-133a tumor suppressor is a novel prognostic marker in esophageal squamous cell carcinoma. Oncol Rep. 2012;28(2):465-472. 29. Junker A, Krumbholz M, Eisele S, et al. MicroRNA profiling of multiple sclerosis lesions identifies modulators of the regulatory protein CD47. Brain. 2009;132(Pt 12):3342-3352. 30. Peshdary V and Atlas E. Dexamethasone induced miR-155 up-regulation in differentiating 3T3-L1 preadipocytes does not affect adipogenesis. Sci Rep. 2018;19;8(1):1264. 31. Kikuchi Y, Uno S, Kinoshita Y, et al. Apoptosis inducing bivalent single-chain antibody fragments against CD47 showed antitumor potency for multiple myeloma. Leuk Res. 2005;29(4):445-450. 32. Patel S, Wang FH, Whiteside T, Kasid U. Identification of seven differentially displayed transcripts in human primary and matched metastatic head and neck squamous cell carcinoma cell lines: implications in metastasis and/or radiation response. Oral Oncol. 1997;33(3):197-203. 33. Laliberté B, Wilson AM, Nafisi H, et al. TNFAIP8: a new effector for Galpha (i) coupling to reduce cell death and induce cell transformation. J Cell Physiol. 2010; 225(3):865-874. 34. Zhang L, Liu R, Luan Y and Yao Y. Tumor necrosis factor-α induced protein 8: pathophysiology, clinical significance, and regulatory mechanism. Int J Biol Sci. 2018;14(4): 398-405. 35. Liu T, Gao H, Chen X, et al. TNFAIP8 as a predictor of metastasis and a novel prognostic biomarker in patients with epithelial ovarian cancer. Br J Cancer. 2013; 109(6):1685-1692. 36. Shi TY, Cheng X, Yu KD, et al. Functional variants in TNFAIP8 associated with cervical cancer susceptibility and clinical outcomes. Carcinogenesis. 2013;34(4):770-778.

37. Liu T, Xia B, Lu Y, Xu Y, Lou G. TNFAIP8 overexpression is associated with platinum resistance in epithelial ovarian cancers with optimal cytoreduction. Hum Pathol. 2014;45(6):1251-1257. 38. Eisele L, Klein-Hitpass L, Chatzimanolis N, et al. Differential expression of drug-resistance-related genes between sensitive and resistant blasts in acute myeloid leukemia. Acta Haematol. 2007;117(1):8-15. 39. Mutlu P, Ural AU, Gündüz U. Different types of cell cycle and apoptosis related gene expressions alter in corticosteroid, vincristine and melphalan resistant U-266 multiple myeloma cell lines. Turk J Hematol. 2014;31(3):231-238. 40. Amodio N, Gallo Cantafio ME, Botta C, et al. Replacement of miR-155 elicits tumor suppressive activity and antagonizes bortezomib resistance in multiple myeloma. Cancers (Basel). 2019;11:236. 41. Caracciolo D, Montesano M, Altomare E, et al. The potential role of miRNAs in multiple myeloma therapy. Expert Rev Hematol. 2018;11(10):793-803. 42. Chakraborty C, Sharma AR, Sharma G, Doss CGP, Lee SS. Therapeutic miRNA and siRNA: moving from bench to clinic as next generation medicine. Mol Ther Nucleic Acids. 2017;8:132-143. 43. Ahmad N, Haider S, Jagannathan S, Anaissie E, Driscoll JJ. MicroRNA theragnostics for the clinical management of multiple myeloma. Leukemia. 2014;28(4):732-738. 44. Nana-Sinkam SP, Croce CM. Clinical applications for microRNAs in cancer. Clin Pharmacol Ther. 2013;93(1):98-104. 45. Di Martino MT, Leone E, Amodio N, et al. Synthetic miR-34a mimics as a novel therapeutic agent for multiple myeloma: in vitro and in vivo evidence. Clin Cancer Res. 2012;18(22):6260-6270. 46. Rupaimoole R, Han HD, Lopez-Berestein G, and Sood AK. MicroRNA therapeutics: principles, expectations, and challenges. Chin J Cancer. 2011;30(6):368-370. 47. Kwok G, Zhao TJ, Weiss J, et al. Translational applications of microRNAs in cancer, and therapeutic implications. Noncoding RNA Res. 2017;2(3-4):143-150. 48. Denizli M, Aslan B, Mangala LS, et al. Chitosan nanoparticles for miRNA delivery. Methods Mol Biol. 2017;1632:219-230. 49. Shah MY, Ferrajoli A, Sood AK, LopezBerestein G, Calin GA. microRNA Therapeutics in cancer – an emerging concept. EBioMedicine. 2016;12:34-42. 50. Chen Y, Gao DY, Huang L. In vivo delivery of miRNAs for cancer therapy: challenges and strategies. Adv Drug Deliv Rev. 2015; 81:128-141.

2823

ARTICLE Ferrata Storti Foundation

Hemostasis

Functional plasminogen activator inhibitor 1 is retained on the activated platelet membrane following platelet activation Gael B. Morrow,° Claire S. Whyte and Nicola J. Mutch

Institute of Medical Sciences, University of Aberdeen, Aberdeen, UK °Current address: Radcliffe Department of Medicine, University of Oxford, Oxford, UK

Haematologica 2020 Volume 105(12):2824-2833

ABSTRACT

Correspondence: NICOLA J. MUTCH n.j.mutch@abdn.ac.uk

latelets harbor the primary reservoir of circulating plasminogen activator inhibitor 1 (PAI-1), but the reportedly low functional activity of this pool of inhibitor has led to debate over its contribution to thrombus stability. Here we analyze the fate of PAI-1 secreted from activated platelets and examine its role in maintaining thrombus integrity. Activation of platelets results in translocation of PAI-1 to the outer leaflet of the membrane, with maximal exposure in response to strong dual agonist stimulation. PAI-1 is found to co-localize in the 'cap' of phosphatidylserine-exposing platelets with its co-factor, vitronectin, and fibrinogen. Inclusion of tirofiban or Gly-Pro-Arg-Pro significantly attenuated exposure of PAI-1, indicating a crucial role for integrin αIIbb3 and fibrin in delivery of PAI-1 to the activated membrane. Separation of platelets post stimulation into soluble and cellular components revealed the presence of PAI-1 antigen and activity in both fractions, with approximately 40% of total platelet-derived PAI-1 remaining associated with the cellular fraction. Using a variety of fibrinolytic models, we found that platelets produce a strong stabilizing effect against tissue plasminogen activator (tPA)-mediated clot lysis. Platelet lysate, as well as soluble and cellular fractions, stabilize thrombi against premature degradation in a PAI-1-dependent manner. Our data show for the first time that a functional pool of PAI-1 is anchored to the membrane of stimulated platelets and regulates local fibrinolysis. We reveal a key role for integrin αIIbb3 and fibrin in delivery of PAI-1 from platelet α-granules to the activated membrane. These data suggest that targeting platelet-associated PAI-1 may represent a viable target for novel profibrinolytic agents.

Introduction Received: June 26, 2019. Accepted: November 21, 2019. Pre-published: November 28, 2019. doi:10.3324/haematol.2019.230367 ©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.

2824

The fibrinolytic system is primarily responsible for thrombus resolution in vivo thus maintaining vessel patency. The principal enzyme, is formed via cleavage of the inactive circulating zymogen plasminogen. The main plasminogen activators are tissue plasminogen activator (tPA) derived largely from endothelial cells1-3 and urokinase (uPA), which is synthesized by cells of fibroblast morphology,4 epithelial cells, monocytes and macrophages.5 The activity of tPA is primarily regulated by one-to-one complex formation with the serpin inhibitor, plasminogen activator inhibitor-1 (PAI-1).6,7 PAI-1 is unusual amongst the family of serpin inhibitors, as in its free form it can exist in an active or latent state.8-10 The active form of secreted cellular PAI-1 has a relatively short half-life of around 30 minutes (min) in plasma8,11-14 but is stabilized by binding to the adhesive glycoprotein vitronectin (Vn), thereby prolonging its half-life 2-3-fold in vivo.15-17 Vn is crucial for PAI-1 function in fibrinolysis, acting as an intermolecular bridge between PAI-1 and fibrin,18 localizing PAI-1 within the fibrin clot.19 The primary reservoir of circulating PAI1 resides within platelet α-granules;20 however, it has been suggested that only 510% of platelet PAI-1 exists in an active configuration.20-22 Platelets play a crucial role in hemostasis and are the first to respond to vessel injury. Activation of platelets gives rise to multiple platelet subpopulations with diverse phenotypes and differential functions.23,24 Aggregating, or spread, platelets mediate clot retraction and are defined by expression of the active integrin αIIbb3 and a lack of phosphatidylserine (PS) exposure.25,26 In contrast, PS-exposing haematologica | 2020; 105(12)

Functional PAI-1 on activated platelets

platelets demonstrate a characteristic balloon shape, increased cytosolic Ca2+ and enhanced ability to bind coagulation factors27,28 and promote thrombin generation.29,30 'Coated' platelets are a subset of PS-exposing platelets, which harbor several procoagulant α-granule proteins on their surface, such as fibrinogen, factor V and von Willebrand factor.31-34 It has been demonstrated that proteins are anchored via a transglutaminase-dependent mechanism and require integrin αIIbb3 activation to permit anchoring of fibrin at the platelet surface.35 PS-exposing platelets possess a protruding 'cap' on their membrane, also described as the platelet body,36,37 that is rich in aminophospholipids and harbors a number of plateletderived and plasma proteins. Our laboratory has identified platelet FXIII-A and plasma-derived plasminogen within the 'cap' of PS-exposing platelets,38,39 that have the potential to direct fibrinolysis in platelet-rich areas of thrombi. The susceptibility of a thrombus to fibrinolysis is influenced by platelet content and fibrin structure.40 Platelets anchor to fibrinogen via the integrin αIIbb3; this binding interaction stabilizes the forming thrombus and initiates the process of clot retraction.41 Outside-in signaling, initiated through engagement of αIIbb3 by fibrin(ogen), stimulates contraction of the platelet intracellular cytoskeleton.42 This process reels in the fibrin network to create a tightly compacted clot with increased resistance to fibrinolysis.43,44 We have previously shown that the fibrin immediately adjacent to platelet aggregates is markedly more resistant to degradation under flow,39 in agreement with observations under static conditions.39,45 In this study, we examine the fate of PAI-1 released from platelet α-granules. We provide the first evidence that a pool of platelet-derived PAI-1 is retained on the activated platelet membrane via a fibrin and integrin αIIbb3 mechanism. Importantly, this pool of PAI-1 retains functional activity and directly participates in thrombus stability against fibrinolytic degradation.

bin (100 nM). Slides were blocked with 5% BSA before addition of washed platelets (0.5x108 plt/mL). In some cases, platelets were pre-treated with 5 mM GPRP or 1 mg/mL tirofiban prior to activation. Fluorescently labeled antibodies to either PAI-1 (5.8 mg/mL), fibrin(ogen) (37 mg/mL) or Vn (13 mg/mL), P-selectin (1/20) or CD41 (1/20) were included during stimulation. After 30 min Annexin-A5 FITC or AF647 (1/20) was added in the presence of 2 mM CaCl2. At 45 min platelets were visualized using a x63 1.40 oil immersion objective and Zeiss 710 laser scanning confocal microscope.

Fluorescence imaging of platelet-rich plasma clots Clots were formed from 30% platelet rich plasma (PRP) with 0.25 mM fibrinogen-Alex fluor 546 (AF546) (Thermo Fisher Scientific) ± a neutralizing antibody to PAI-1 (400 mg/mL). Clotting was initiated using 0.125 U/mL thrombin and 10 mM CaCl2. Annexin A5-AF647 and fluorescently-labeled rabbit polyclonal antibody to PAI-1 or Vn were incorporated. Clots were polymerized in Ibidi m-slide VI0.4 chambers at 37°C for 2 hours (h) in a moist box. In some cases, 75 nM tPA (Genetech) was added to the edge of clot and lysis monitored by taking images every 10 seconds (s). Clots were imaged using a x63 1.40 oil immersion objective and Zeiss 710 laser scanning confocal microscope.

Chandler model thrombi Thrombi were formed using the Chandler model.46 Pooled normal plasma (PNP) thrombi containing 45 mg/mL FITC-labeled fibrinogen and 10.9 mM CaCl2 ± a neutralizing antibody to PAI-1 (400 mg/mL) were rotated at 30 rpm for 90 min. Thrombi were removed and lysed in 1 mg/mL tPA at 37°C and samples taken every 30 min for 4 h. The plate was read at excitation 485 nm and emission 525 nm using a BioTek FLx800 fluorescence reader and Gen5 software. Fluorescence release is directly proportional to the rate of fibrinolysis in the sample.

Ethical consent Ethical approval was obtained from the University of Aberdeen College Ethics Review Board. Further details of the methods used can be found in the Online Supplementary Appendix.

Methods Isolation of soluble and cellular fraction

Platelets were activated with 1 mg/mL convulxin (CVX; Enzo Life Sciences) and 100 nM thrombin (Sigma-Aldrich). The soluble fraction was collected by centrifugation at 13,000xgr for 4 minutes (min). The pellet, containing the cellular components, was re-suspended in HEPES buffer.

Flow cytometry analysis of platelets

Washed platelets (2x108 plt/mL) were stimulated with 1 mg/mL CVX ± 0.2 mM TRAP-6 (Sigma-Aldrich) or 100 nM thrombin in the presence of 2 mM CaCl2. In some cases, platelets were pretreated for 30 min with 5 mM Gly-Pro-Arg-Pro (GPRP) (SigmaAldrich) or 1 mg/mL tirofiban (Sigma-Aldrich). Fluorescentlylabeled antibodies to either PAI-1 (5.8 mg/mL), fibrin(ogen) (37 mg/mL) or Vn (13 mg/mL) were added during stimulation. After 40 min Annexin A5-Alexa fluor 647 (AF647) (1/20) (BD Biosciences) was added in the presence of 2 mM CaCl2. Exposure of PAI-1 and PS were analyzed using a BD LSRII cytometer with FACS DIVA 6.1.3 software.

Fluorescence imaging of platelets

Ibidi m-slide VI0.4 chambers were coated with collagen (20 mg/mL) (American Biochemical Pharmaceuticals) and thromhaematologica | 2020; 105(12)

Results Plasminogen activator inhibitor-1 is retained on the activated platelet membrane Plasminogen activator inhibitor-1 is abundant in platelet α-granules and is known to be a constituent of the platelet secretome. Here we address whether platelet-derived PAI-1 is retained on the surface of platelets. Using confocal microscopy we analyzed PAI-1 on the membrane of platelets stimulated on a collagenand thrombin-coated surface for 45 min. The majority (78.8±1.7%) of collagen- and thrombin-stimulated platelets were shown to be positive for PAI-1 (Figure 1). The serpin was found to be located within the ‘cap’ of PS-exposing platelets, which are characterized by Annexin V-AF647 staining and a characteristic balloon shape.27,39,47,48 P-selectin was included as a marker of platelet degranulation post stimulation and was found to co-localize with PAI-1 on the activated membrane of PSpositive platelets (Online Supplementary Figure S1). PAI-1 was found to co-localize with fibrin(ogen) and αIIbb3 in the 'cap' of PS-exposing platelets (Figure 1A and B, arrows), with co-efficient (r) values of 0.92 and 0.57, 2825

G.B. Morrow et al.

respectively. Platelet-derived Vn, a co-factor of PAI-1, was also found within the 'cap' region of PS-exposing platelets (r=0.67) (Figure 1C). Flow cytometry analysis revealed a negligible amount of PAI-1 on the membrane of unstimulated platelets (Figure 2A and Online Supplementary Figure S2A). Following activation of platelets with CVX ± TRAP-6 or thrombin there was a significant increase in the presence of PAI-1 compared to unstimulated platelets (P<0.0001). Maximum PAI-

1 exposure occurred following stimulation of platelets with CVX and thrombin, with a 35-fold increase in mean fluorescence intensity (MFI) compared to unstimulated platelets. Annexin V-AF647 staining revealed that the majority (93%) of PAI-1 was associated with PS-exposing platelets (data not shown). Similarly, maximal exposure of platelet-derived Vn and fibrinogen was observed in response to CVX and thrombin (Figure 2B and C, and Online Supplementary Figure S2B and C).

2826

Figure 1. Platelet-derived plasminogen activator inhibitor 1 (PAI-1) co-localizes with fibrin(ogen), integrin αIIbb3 and vitronectin (Vn) in the “cap” of phosphatidylserine-exposing platelets. Platelets (0.5x108 plt/mL), were adhered to a slide coated with collagen (20 mg/mL) and thrombin (100 nM) for 30 minutes (min) at ambient temperature. Annexin V (red) was added to stain phosphatidylserine and left for a further 15 min before imaging by confocal microscopy. (A) PAI-1 was detected using a rabbit polyclonal antibody labeled with DL550 (yellow) and integrin αIIbb3 using a FITC-conjugated antibody to the CD41 subunit (green). (B) Fibrin(ogen) was analyzed using a rabbit polyclonal antibody labeled with DL405 (blue). (C) Vitronectin (Vn) was detected using a rabbit polyclonal antibody labeled with DL488 (green). Arrows highlight examples of co-localization. Images shown are representative of n ≥3. Scale bars represent 2 mm.

haematologica | 2020; 105(12)

Functional PAI-1 on activated platelets

Plasminogen activator inhibitor-1 retention on platelets is dependent on αIIbb3 and fibrin

We next analyzed the potential mechanism of retention of PAI-1 on the platelet membrane by blocking fibrin polymerization and the integrin αIIbb3 with GPRP and tirofiban, respectively. A significant reduction in PAI-1 was observed on incorporation of tirofiban (2.3-fold; P<0.01) or GPRP (2-fold; P<0.0001) (Figure 3A). Interestingly, there was no change in association of Vn with the activated platelet membrane upon incorporation of tirofiban or GPRP (Figure 3B). Consistent with flow cytometry data confocal microscopy revealed significantly less membrane-associated PAI-1 upon inclusion of tirofiban or GPRP (Figure

3C). These data indicate that despite the clear co-localization of PAI-1 and Vn on the activated platelet membrane, the mechanism of retention on the activated platelet surface is different.

Distribution of platelet-derived plasminogen activator inhibitor-1 antigen and activity Our data show for the first time that a pool of PAI-1 can be retained on the activated platelet membrane. The distribution of PAI-1 antigen and activity between the secretome and membrane fractions was then analyzed. Platelets were subjected to dual agonist stimulation to induce complete degranulation. The soluble fraction and

Figure 2. Exposure of platelet-derived plasminogen activator inhibitor 1 (PAI-1) on the activated membrane is maximized by strong dual agonist stimulation. Platelets (2x108 plt/mL), were unstimulated (US) or activated with convulxin (CVX) (1 mg/mL) ± TRAP-6 (200 nM) or thrombin (TH; 100 nM) and analyzed using flow cytometry. (A) Platelet-derived PAI-1 was detected with a rabbit polyclonal antibody labeled with DL488. (B) Vitronectin (Vn) was detected with DL488 labeled rabbit polyclonal antibody. (C) Fibrin(ogen) was analyzed using a rabbit polyclonal antibody labeled with DL405. Percentage of platelets positive for PAI-1, Vn and fibrinogen and mean fluorescence intensity data are expressed as mean±standard deviation. n≥ 4. ***P<0.001 versus **P<0.01, *P<0.05 when comparing agonist treatment versus unstimulated.

haematologica | 2020; 105(12)

2827

G.B. Morrow et al. A

Figure 3. Plasminogen activator inhibitor 1 (PAI-1) is dependent on αIIbb3 and fibrin for maximal exposure. Platelets (2x108 plt/mL), were pre-treated with tirofiban (1 mg/mL) or GPRP (5 mM) before activation with CVX (1 mg/mL) and thrombin (100 nM) and analyzed using flow cytometry. Platelet-derived (A) PAI-1 or (B) vitronectin (Vn) were detected using rabbit polyclonal antibodies labeled with DL488. Data are expressed as mean fluorescence intensity (mean±standard deviation). ***P<0.001 versus untreated. n≥3. (C) Platelets (0.5x108 plt/mL), were pre-treated with tirofiban (1 mg/mL) or GPRP (5 mM) before adhering to a collagen (20 mg/mL) and thrombin (100 nM) coated slide. AnnexinV-AF647 (red) was used to stain phosphatidylserine and PAI-1 was detected using a rabbit polyclonal antibody labeled with DL550 (yellow). Images shown are representative of n=3. Scale bars represent 10 mm and magnified image of a single platelet. ns: not significant.

the remaining cellular fraction, consisting of the platelet internal and external membranes, were analyzed for PAI1 antigen by ELISA and activity assay. PAI-1 antigen was more abundant in the soluble fraction (19.2 ng/108 plt), but almost a third of the total PAI-1 (33.8 ng/108 plt) remained associated with the cellular fraction (10.5 ng/108 plt) (Figure 4A). PAI-1 activity, determined by complex formation with tPA, revealed a similar distribution within the soluble and membrane fractions as the antigen (Figure 4B). These data indicate that a significant proportion of functional PAI-1 (~40%) is retained on the activated platelet surface where it can potentially regulate fibrinolysis. Significant attenuation of PAI-1 antigen and activity in the soluble and cellular fraction was observed when αIIbb3 or fibrin polymerization were inhibited (Figure 5A and B). These data indicate an essential role for functional αIIbb3 and fibrin in translocation of PAI-1 from the platelet α-granules to both the activated platelet membrane and the secretome.

Platelet-derived PAI-1 localizes in platelet-dense areas and stabilizes thrombi Platelet-derived PAI-1 staining in clots was localized to platelet-dense regions and emanated into the surrounding fibrin network (Figure 6). These data indicate that following platelet activation, the pool of PAI-1 is translocated from α-granules to the activated membrane and distally to 2828

Table 1. Thrombodynamic analysis of clot formation and lysis.

Thrombodynamic parameter

Control

Lag time (min) 1.33 ± 0.41 Rate of clot growth (µm/min) 15.80 ± 14.40 Clot density (a.u.) 14716 ± 906.5 Full lysis time (min) 51.64 ± 2.01

+ PAI-1 Ab

0.60 ± 0.12 65.60 ± 6.55 7927 ± 123.4 24.19 ± 2.83

0.139 0.0346 0.0003 0.0002

Platelet-rich clots were formed ± antibody to PAI-1 (400 mg/mL) and lysed with tPA (5 nM). Clot formation and lysis was monitored using a Hemacore thrombodynamic a.u. = arbitrary units, data represent mean ± SEM, n=4. Statistical significance was determined by an unpaired Student’s t-test; PAI-1: plasminogen activator inhibitor 1; tPA: tissue plasminogen activator.

platelet-associated fibrin. We also found evidence of colocalization of Vn with the fibrin network in the clot (Online Supplementary Figure S3). The role of the platelet reservoir of PAI-1 in stabilization of thrombi has been a subject of debate as it reportedly chiefly exists in a latent inactive form.21 Here we analyze tPA-mediated lysis of platelet-rich clots using multiple static and flow-based models. Lysis of clots in real-time was visualized by confocal microscopy in the absence and presence of a neutralizing antibody to PAI-1. Inhibition of PAI-1 resulted in significantly faster lysis of clots (5.7±0.8 min, P<0.001 vs. 24±1.5 min) (Figure 6B and C, and Online Supplementary Video S1). Similarly, tPA-mediated lysis of PRP clots, monitored by change in absorbance, revealed sighaematologica | 2020; 105(12)

Functional PAI-1 on activated platelets

nificantly faster 50% lysis times on inclusion of the neutralizing antibody to PAI-1 (96±3.2 vs. 119±3.5 min, respectively; P<0.001. n=3). A control polyclonal rabbit IgG had no effect (data not shown). Thrombodynamic analysis of PRP clots revealed a faster rate of clot formation and a reduction in clot density (Table 1). A significant enhancement of lysis was observed when PAI-1 was inhibited (Table 1 and Online Supplementary Video S2). Lysates of activated platelets stabilized thrombi formed under arterial flow rates against premature lysis (Figure 7A).

Our activity data (Figure 4) revealed that there were two pools of functional PAI-1, therefore the contribution of the cellular and soluble fractions of platelets to thrombus stability were analyzed. Inclusion of soluble and cellular platelet fractions during thrombus formation resulted in a 2-fold and 2.7-fold reduction in lysis, respectively, compared to a 3-fold reduction on inclusion of whole platelet lysate (Figure 7A). Incorporation of a neutralizing antibody to PAI1 completely attenuated the stabilizing effect of the platelet lysate and soluble fraction on thrombus lysis (Figure 7B and

Figure 4. Active platelet-derived plasminogen activator inhibitor 1 (PAI-1) is retained within the cellular fraction of stimulated platelets. Platelets (2.5x108 plt/mL) were left unstimulated (US) or activated with 1 μg/mL CVX and 100 nM thrombin for 30 minutes at 37°C. Total platelet lysate or soluble and cellular fractions separated by centrifugation post-stimulation were analyzed for (A) PAI-1 protein (n=4) or (B) PAI 1 activity (n=5). Data are expressed as mean±standard deviation. ***P<0.001, **P<0.01 versus whole platelet lysate; ns: not significant.

Figure 5. Plasminogen activator inhibitor 1 (PAI-1) release and retention on the activated platelet membrane is dependent on functional αIIbb3 and fibrin. Platelets were activated with 1 mg/mL CVX and 100 nM thrombin for 30 minutes at 37°C ± pre-treatment with tirofiban (1 mg/mL) or GPRP (5 mM). Soluble and cellular fractions were isolated by centrifugation. PAI-1 antigen and activity levels were quantified by ELISA and activity assays. (A and B) Antigen and activity in the soluble fraction. (C and D) Antigen and activity in the cellular fraction. Data are expressed as mean±standard deviation. n≥4. ***P<0.001.

haematologica | 2020; 105(12)

2829

G.B. Morrow et al.

C). In contrast, the antibody only partially abrogated the stabilizing effect of the cellular fraction, suggesting that additional factors on the platelet membrane contribute to thrombus stabilization (Figure 7D).

Discussion Platelets are well known to be the primary circulating source of the fibrinolytic inhibitor PAI-1. Despite this the fate of the inhibitor following stimulation and degranulation of platelets is poorly defined. To our knowledge this is the first study to show that functional PAI-1 is retained on the activated platelet membrane following stimulation where it functions to regulate local fibrinolysis. Strong dual agonist stimulation of platelets maximizes PAI-1 exposure on the activated platelet membrane. PAI-1 was localized in the aminophosphoplipid-rich ‘cap’ of PSexposing platelets, and over the granulomere of spread platelets. There was evident co-localization of PAI-1 with its co-factor Vn and fibrinogen. Our data are also the first to show that the retention and release of platelet PAI-1 is dependent on integrin αIIbb3 and fibrin, alluding to the importance of this inhibitor in fibrin stabilization. In accordance with this we have utilized several functional models of fibrinolysis to reveal a crucial role for plateletderived PAI-1 in stabilizing thrombi against premature degradation. Our lab and others have previously reported the accumulation of hemostatic and adhesive proteins within a small (~1 mm) concave cap area on PS-exposing platelets,

Annexin V

Fibrin(ogen)

PAI-1

Merged

2830

these include platelet-derived factor XIII-A,38 39 49 plasminogen, fibrinogen, thrombospondin and coagulation factors such as prothrombin, factor V and factor X.50 We have shown that PAI-1 was localized within the “cap”, alongside its co-factor Vn. Pre-treating platelets with tirofiban or GPRP to inhibit αIIbb3 and fibrin polymerization, down-regulated PAI-1 but not Vn exposure on the activated platelet membrane. Interestingly, PAI-1 and Vn are reportedly not in complex within α-granules, instead PAI-1 is stabilized by calcium which is thought to mask the Vn binding site.51 These data imply that the PAI-1/Vn complex must form subsequent to platelet activation to permit PAI 1 interaction with fibrin.18 An important observation in this study is that only approximately 60% of total platelet-derived PAI-1 was released into the soluble fraction, often termed platelet releasate, while the remaining 40% was associated with the cellular fraction and was found to be functionally active. These data are consistent with findings that platelet-derived PAI-1 is more active than previously described.52 The discrepancy between our study and older literature20,21 is most likely accounted for by variations in experimental set-up, in particular the activation status of the platelets following strong dual agonist stimulation. Platelets harbor mRNA for PAI-1 and are thought to be capable of de novo synthesis of the inhibitor.53 Interestingly, the rate of synthesis of platelet PAI-1 increases 25% over 24 h, post-stimulation with thrombin, and the serpin is found within an active conformation.53 Inclusion of tirofiban and GPRP prior to platelet activation essentially abolished PAI-1 antigen and activity in the sol-

Figure 6. Platelet-derived plasminogen activator inhibitor 1 (PAI-1) is localized within platelet aggregates and attenuates tissue plasminogen activator (tPA)-mediated lysis of platelet-rich plasma (PRP) clots. PRP clots (30%) were formed in the presence of fibrinogen-AF546 (red) for 2 hours at 37°C ± rabbit polyclonal neutralizing antibody to PAI-1 (400 mg/mL) by addition of 0.125 U/mL thrombin. (A) Phosphatidylserine-exposing platelets were detected using AnnexinVAF647 (green) and platelet PAI-1 was visualized using a DL-488-rabbit polyclonal (yellow). Arrows highlight PAI-1 localized in the platelet “cap”. Representative images of n=5. (B) Platelets were labeled with DIOC-6 (0.5 mg/mL) and clots lysed ± neutralizing antibody to PAI-1 by addition of tPA (75 nM). Images were recorded at 0 minutes (min) before addition of tPA and at 4 and 17 min. Representative images of n=4. Scale bars represents 10 mm. (C) Average lysis time (min) of PRP clots ± a neutralizing antibody to PAI-1. ***P<0.001.

haematologica | 2020; 105(12)

Functional PAI-1 on activated platelets

uble and cellular fraction, suggesting that release of active PAI-1 from α-granules and its retention on the platelet surface is dependent on αIIbb3 and polymerized fibrin. This could arise due to an outside-in signaling mechanism whereby binding of extracellular fibrin(ogen) to αIIbb3 mediates intracellular signaling events that trigger granule secretion and translocation of PAI-1 to the outer leaflet of the membrane.54 We have clearly shown that addition of whole platelet lysate or soluble and cellular fractions, derived post-stimulation, stabilize thrombi formed under flow against lytic degradation. Neutralizing PAI-1 completely abolished the stabilizing effect of the soluble fraction, attributing it to PAI1 inhibitory activity. The cellular fraction had a stronger stabilizing effect on thrombi which could not be completely alleviated by inhibition of PAI-1, indicating that additional factors on the platelet membrane contribute to thrombus resistance. Our work has previously shown that plateletderived FXIII-A is retained on the activated platelet-membrane and stabilizes thrombi against premature degradation in an α2AP-dependent manner.38 Consistent with our results a recent study using a novel inhibitor, PAItrap, in a laser-induced vascular injury mouse model showed a significant reduction in platelet accumulation and thrombus formation but did not impact on global hemostasis.55 Interestingly, in addition to the significant impact that neutralization of PAI-1 has on fibrinolysis we show using thrombodynamic analysis a trend toward altered clot growth. Studies in PAI-1 deficient mice reveal markedly prolonged time to occlusion in arterial and venous mouse models of injury.56 A significant, but less pronounced effect, on occlusion was observed in Vn deficient mice. This sug-

gests that neutralization of PAI-1 during clot formation tilts the hemostatic balance toward fibrin lysis rather than fibrin formation. Collectively, these data highlight the huge potential of targeting fibrinolytic inhibitors in terms of modulating thrombus formation, propagation and stability. It is now well documented that thrombi formed in vivo exhibit a hierarchical structure, with two distinct regions of platelet activation.26,57-60 The inner core is rich in fibrin(ogen) and thrombin and is comprised of tightly packed degranulated platelets. This is encapsulated by an outer shell of loosely packed platelets with minimal α-granule release.57 A role for αIIbb3 outside-in signalling has been described in consolidation of the platelet mass, indicating the key role of these signaling events in platelet packing, interplatelet molecular transport, agonist distribution, and subsequent platelet activation.61 Our studies reveal that PAI 1 exposure on platelets is highly dependent on αIIbb3 and fibrin, suggesting that these signaling mechanisms may mediate solute transport of PAI-1 within the micro-environment of the thrombus. There are currently no drugs in clinical trials that target fibrinolytic inhibitors, including PAI-1. Several approaches have been reported in the literature, including the use of a diabody directed against PAI-1 and TAFIa,62 monoclonal antibodies to PAI-1 and TAFIa,63 PAItrap, an antagonist based on a variant of uPA,55 and an inhibitory hexapeptide that corresponds to amino acids 350-355 of PAI-1.64 These compounds demonstrate strong profibrinolytic capacity in various mouse models of ischemic stroke and thromboembolism without an increase in global bleeding. However, none have progressed further into phase II clinical trials. Our novel data

Figure 7. Platelet-(plt)-derived plasminogen activator inhibitor 1 (PAI-1) stabilizes thrombi against premature fibrinolytic degradation. Platelets (2.5x108 plt/mL), were activated with thrombin (100 nM) and CVX (1 μg/mL) for 30 minutes at 37°C. Soluble and cellular fractions were isolated from stimulated platelets by centrifugation. Model thrombi were formed with pooled normal plasma (PNP) and FITC-labeled fibrinogen under arterial shear rates and lysis was subsequently induced by tissue plasminogen activator (tPA) (1 mg/mL). Fluorescence release is directly proportional to the degree of fibrinolysis. (A) Pooled normal plasma (PNP) thrombi formed ± whole platelet lysate (blue), soluble (green) or cellular (orange) fractions. (B-D) Thrombi were formed in the absence (closed symbols) and presence (open symbols) of a neutralizing antibody to PAI-1 (400 mg/mL) with (B) whole platelet lysate (C) platelet soluble fraction, and (D) platelet cellular fraction. Data represent mean±standard error of mean. **P<0.01, ***P<0.001; , ****P<0.0001 versus plasma thrombi. n=5.

haematologica | 2020; 105(12)

2831

G.B. Morrow et al.

are the first to show that platelet-associated PAI-1 is functionally active and functions to maintain thrombus integrity. These results underscore the potential of PAI-1 as a target for novel profibrinolytic drugs to augment thrombus dissolution in vivo. Funding This work was supported by a PhD studentship from the British Society of Haematology, British Society of Thrombosis &

References 1. van den Eijnden-Schrauwen Y, Kooistra T, de Vries RE, et al. Studies on the acute release of tissue-type plasminogen activator from human endothelial cells in vitro and in rats in vivo: evidence for a dynamic storage pool. Blood. 1995;85(12):3510-3517. 2. Levin EG, Loskutoff DJ. Cultured bovine endothelial cells produce both urokinase and tissue-type plasminogen activators. J Cell Biol. 1982;94(3):63163-63166. 3. Levin EG, Loskutoff DJ. Regulation of plasminogen activator production by cultured endothelial cells. Ann N Y Acad Sci. 1982; 401:184-194. 4. Larsson LI, Skriver L, Nielsen LS, et al. Distribution of urokinase-type plasminogen activator immunoreactivity in the mouse. J Cell Biol. 1984;98(3):894-903. 5. Manchanda N, Schwartz BS. Interaction of single-chain urokinase and plasminogen activator inhibitor type 1. J Biol Chem. 1995; 270(34):20032-20035. 6. Mutch NJ, Thomas L, Moore NR, et al. TAFIa, PAI-1 and alpha-antiplasmin: complementary roles in regulating lysis of thrombi and plasma clots. J Thromb Haemost. 2007;5(4):812-817. 7. Booth NA, Anderson JA, Bennett B. Plasminogen activators in alcoholic cirrhosis: demonstration of increased tissue type and urokinase type activator. J Clin Pathol. 1984; 37(7):772-777. 8. Gils A, Declerck PJ. Plasminogen activator inhibitor-1. Curr Med Chem. 2004; 11(17):2323-2334. 9. Hekman CM, Loskutoff DJ. Endothelial cells produce a latent inhibitor of plasminogen activators that can be activated by denaturants. J Biol Chem. 1985;260(21):1158111587. 10. Sprengers ED, Kluft C. Plasminogen activator inhibitors. Blood. 1987;69(2):381-387. 11. Loskutoff DJ, Curriden SA. The fibrinolytic system of the vessel wall and its role in the control of thrombosis. Ann N Y Acad Sci. 1990;598:238-247. 12. Chandler WL, Alessi MC, Aillaud MF, et al. Clearance of tissue plasminogen activator (TPA) and TPA/plasminogen activator inhibitor type 1 (PAI-1) complex: relationship to elevated TPA antigen in patients with high PAI-1 activity levels. Circulation. 1997;96(3):761-768. 13. Kooistra T, Sprengers ED, van Hinsbergh VW. Rapid inactivation of the plasminogenactivator inhibitor upon secretion from cultured human endothelial cells. Biochem J. 1986;239(3):497-503. 14. MacGregor IR, Booth NA. An enzymelinked immunosorbent assay (ELISA) used to study the cellular secretion of endothelial plasminogen activator inhibitor (PAI-1).

2832

Haemostasis and Thrombosis UK (formally Lifeblood) and the British Heart Foundation (PG/15/82/31721). Acknowledgments We acknowledge the Microscopy and Histology Core Facility and the Iain Fraser Cytometry Centre at the University of Aberdeen, UK, for excellent advice and use of the facilities. We thank Michela Donnarumma for technical assistance and Prof. Nuala Booth for critical appraisal of the manuscript.

Thromb Haemost. 1988;59(1):68-72. 15. Mimuro J, Loskutoff DJ. Binding of type 1 plasminogen activator inhibitor to the extracellular matrix of cultured bovine endothelial cells. J Biol Chem. 1989;264(9):5058-5063. 16. Seiffert D, Loskutoff DJ. Evidence that type 1 plasminogen activator inhibitor binds to the somatomedin B domain of vitronectin. J Biol Chem. 1991;266(5):2824-2830. 17. Declerck PJ, De Mol M, Alessi MC, et al. Purification and characterization of a plasminogen activator inhibitor 1 binding protein from human plasma. Identification as a multimeric form of S protein (vitronectin). J Biol Chem. 1988;263(30):15454-15461. 18. Podor TJ, Peterson CB, Lawrence DA, et al. Type 1 plasminogen activator inhibitor binds to fibrin via vitronectin. J Biol Chem. 2000;275(26):19788-19794. 19. Blouse GE, Dupont DM, Schar CR, et al. Interactions of plasminogen activator inhibitor-1 with vitronectin involve an extensive binding surface and induce mutual conformational rearrangements. Biochemistry. 2009;48(8):1723-1735. 20. Booth NA, Simpson AJ, Croll A, et al. Plasminogen activator inhibitor (PAI-1) in plasma and platelets. Br J Haematol. 1988;70(3):327-333. 21. Declerck PJ, Alessi MC, Verstreken M, et al. Measurement of plasminogen activator inhibitor 1 in biologic fluids with a murine monoclonal antibody-based enzyme-linked immunosorbent assay. Blood. 1988; 71(1):220-225. 22. Schleef RR, Sinha M, Loskutoff DJ. Immunoradiometric assay to measure the binding of a specific inhibitor to tissue-type plasminogen activator. J Lab Clin Med. 1985;106(4):408-415. 23. Heemskerk JW, Mattheij NJ, Cosemans JM. Platelet-based coagulation: different populations, different functions. J Thromb Haemost. 2013;11(1):2-16. 24. Kempton CL, Hoffman M, Roberts HR, et al. Platelet heterogeneity: variation in coagulation complexes on platelet subpopulations. Arterioscler Thromb Vasc Biol. 2005; 25(4):861-866. 25. Munnix IC, Cosemans JM, Auger JM, et al. Platelet response heterogeneity in thrombus formation. Thromb Haemost. 2009;102(6):1149-1156. 26. Nesbitt WS, Westein E, Tovar-Lopez FJ, et al. A shear gradient-dependent platelet aggregation mechanism drives thrombus formation. Nat Med. 2009;15(6):665-673. 27. Munnix IC, Kuijpers MJ, Auger J, et al. Segregation of platelet aggregatory and procoagulant microdomains in thrombus formation: regulation by transient integrin activation. Arterioscler Thromb Vasc Biol. 2007;27(11):2484-2490. 28. Berny MA, Munnix IC, Auger JM, et al. Spatial distribution of factor Xa, thrombin,

29.

30. 31. 32.

33.

34.

35.

36.

37.

38.

39.

40.

41. 42.

and fibrin(ogen) on thrombi at venous shear. PloS One. 2010;5(4):e10415. Bevers EM, Comfurius P, van Rijn JL, et al. Generation of prothrombin-converting activity and the exposure of phosphatidylserine at the outer surface of platelets. Eur J Biochem. 1982;122(2):429436. Heemskerk JW, Bevers EM, Lindhout T. Platelet activation and blood coagulation. Thromb Haemost. 2002;88(2):186-193. Dale GL. Coated-platelets: an emerging component of the procoagulant response. J Thromb Haemost. 2005;3(10):2185-2192. Alberio L, Safa O, Clemetson KJ, et al. Surface expression and functional characterization of alpha-granule factor V in human platelets: effects of ionophore A23187, thrombin, collagen, and convulxin. Blood. 2000;95(5):1694-1702. Dale GL, Friese P, Batar P, et al. Stimulated platelets use serotonin to enhance their retention of procoagulant proteins on the cell surface. Nature. 2002;415(6868):175179. Szasz R, Dale GL. Thrombospondin and fibrinogen bind serotonin-derivatized proteins on COAT-platelets. Blood. 2002; 100(8):2827-2831. Mattheij NJ, Swieringa F, Mastenbroek TG, et al. Coated platelets function in plateletdependent fibrin formation via integrin alphaIIbbeta3 and transglutaminase factor XIII. Haematologica. 2016;101(4):427-436. Agbani EO, Hers I, Poole AW. Temporal contribution of the platelet body and balloon to thrombin generation. Haematologica. 2017;102(10):e379-e381. Agbani EO, Williams CM, Hers I, et al. Membrane ballooning in aggregated platelets is synchronised and mediates a surge in microvesiculation. Sci Rep. 2017; 7(1):2770. Mitchell JL, Lionikiene AS, Fraser SR, et al. Functional factor XIII-A is exposed on the stimulated platelet surface. Blood. 2014; 124(26):3982-3990. Whyte CS, Swieringa F, Mastenbroek TG, et al. Plasminogen associates with phosphatidylserine-exposing platelets and contributes to thrombus lysis under flow. Blood. 2015;125(16):2568-2578. Collet JP, Park D, Lesty C, et al. Influence of fibrin network conformation and fibrin fiber diameter on fibrinolysis speed: dynamic and structural approaches by confocal microscopy. Arterioscler Thromb Vasc Biol. 2000;20(5):1354-1361. Ginsberg MH, Du X, Plow EF. Inside-out integrin signalling. Curr Opin Cell Biol. 1992;4(5):766-771. Knezevic I, Leisner TM, Lam SC. Direct binding of the platelet integrin alphaIIbbeta3 (GPIIb-IIIa) to talin. Evidence that interaction is mediated through the cytoplasmic

haematologica | 2020; 105(12)

Functional PAI-1 on activated platelets

43.

44.

45.

46. 47.

48.

49.

domains of both alphaIIb and beta3. J Biol Chem. 1996;271(27):16416-16421. Blinc A, Keber D, Lahajnar G, et al. Magnetic resonance imaging of retracted and nonretracted blood clots during fibrinolysis in vitro. Haemostasis. 1992;22(4):195-201. Sabovic M, Lijnen HR, Keber D, et al. Effect of retraction on the lysis of human clots with fibrin specific and non-fibrin specific plasminogen activators. Thromb Haemost. 1989;62(4):1083-1087. Collet JP, Montalescot G, Lesty C, et al. A structural and dynamic investigation of the facilitating effect of glycoprotein IIb/IIIa inhibitors in dissolving platelet-rich clots. Circ Res. 2002;90(4):428-434. Chandler AB. In vitro thrombotic coagulation of the blood; a method for producing a thrombus. Lab Invest. 1958;7(2):110-114. Siljander P, Farndale RW, Feijge MA, et al. Platelet adhesion enhances the glycoprotein VI-dependent procoagulant response: involvement of p38 MAP kinase and calpain. Arterioscler Thromb Vasc Biol. 2001; 21(4):618-627. Agbani EO, van den Bosch MT, Brown E, et al. Coordinated membrane ballooning and procoagulant spreading in human platelets. Circulation. 2015;132(15):1414-1424. Abaeva AA, Canault M, Kotova YN, et al. Procoagulant platelets form an alpha-granule protein-covered "cap" on their surface that promotes their attachment to aggregates. J Biol Chem. 2013;288(41):29621-29632.

haematologica | 2020; 105(12)

50. Podoplelova NA, Sveshnikova AN, Kotova YN, et al. Coagulation factors bound to procoagulant platelets concentrate in cap structures to promote clotting. Blood. 2016; 128(13):1745-1755. 51. Lang IM, Schleef RR. Calcium-dependent stabilization of type I plasminogen activator inhibitor within platelet alpha-granules. J Biol Chem. 1996;271(5):2754-2761. 52. Brogren H, Wallmark K, Deinum J, et al. Platelets retain high levels of active plasminogen activator inhibitor 1. PLoS One. 2011;6(11):e26762. 53. Brogren H, Karlsson L, Andersson M, et al. Platelets synthesize large amounts of active plasminogen activator inhibitor 1. Blood. 2004;104(13):3943-3948. 54. Li Z, Delaney MK, O'Brien KA, et al. Signaling during platelet adhesion and activation. Arterioscler Thromb Vasc Biol. 2010; 30(12):2341-2349. 55. Peng S, Xue G, Gong L, et al. A long-acting PAI-1 inhibitor reduces thrombus formation. Thromb Haemost. 2017;117(7):1338-1347. 56. Eitzman DT, Westrick RJ, Nabel EG, et al. Plasminogen activator inhibitor-1 and vitronectin promote vascular thrombosis in mice. Blood. 2000;95(2):577-580. 57. Stalker TJ, Traxler EA, Wu J, et al. Hierarchical organization in the hemostatic response and its relationship to the plateletsignaling network. Blood. 2013;121(10): 1875-1885. 58. Welsh JD, Stalker TJ, Voronov R, et al. A sys-

59.

60.

61.

62.

63.

64.

tems approach to hemostasis: 1. The interdependence of thrombus architecture and agonist movements in the gaps between platelets. Blood. 2014;124(11):1808-1815. Welsh JD, Muthard RW, Stalker TJ, et al. A systems approach to hemostasis: 4. How hemostatic thrombi limit the loss of plasmaborne molecules from the microvasculature. Blood. 2016;127(12):1598-1605. Welsh JD, Poventud-Fuentes I, Sampietro S, et al. Hierarchical organization of the hemostatic response to penetrating injuries in the mouse macrovasculature. J Thromb Haemost. 2017;15(3):526-537. Stalker TJ, Welsh JD, Tomaiuolo M, et al. A systems approach to hemostasis: 3. Thrombus consolidation regulates intrathrombus solute transport and local thrombin activity. Blood. 2014; 124(11):1824-1831. Wyseure T, Rubio M, Denorme F, et al. Innovative thrombolytic strategy using a heterodimer diabody against TAFI and PAI1 in mouse models of thrombosis and stroke. Blood. 2015;125(8):1325-1332. Denorme F, Wyseure T, Peeters M, et al. Inhibition of thrombin-activatable fibrinolysis inhibitor and plasminogen activator inhibitor-1 reduces ischemic brain damage in mice. Stroke. 2016;47(9):2419-2422. Armstead WM, Nassar T, Akkawi S, et al. Neutralizing the neurotoxic effects of exogenous and endogenous tPA. Nat Neurosci. 2006;9(9):1150-1155.

2833

ARTICLE Ferrata Storti Foundation

Haematologica 2020 Volume 105(12):2834-2840

Tranfusion Medicine

Mortality reduction in 46 patients with severe COVID-19 treated with hyperimmune plasma. A proof-of-concept, single-arm, multicenter trial

Cesare Perotti,1* Fausto Baldanti,1,2* Raffaele Bruno,1,2 Claudia Del Fante,1 Elena Seminari,1 Salvatore Casari,3 Elena Percivalle,1 Claudia Glingani,3 Valeria Musella,1 Mirko Belliato,1 Martina Garuti,3 Federica Meloni,1,2 Marilena Frigato,3 Antonio Di Sabatino,1,2 Catherine Klersy,1 Giuseppe De Donno3 and Massimo Franchini3 on behalf of The collaborative COVID-19 Plasma Task Force Departments of Immunohematology and Transfusion, Infectious Diseases, Respiratory Diseases, Intensive Care, Virology and Clinical Epidemiology & Biometry, Fondazione IRCCS Policlinico San Matteo, Pavia 2University of Pavia, Pavia and 3Departments of Immunohematology and Transfusion, Infectious Diseases, Respiratory Diseases, Carlo Poma Hospital, ASST Mantova, Mantova, Italy. 1

*CP and FB contributed equally as co-first authors.

ABSTRACT

Correspondence: CATHERINE KLERSY klersy@smatteo.pv.it Received: June 5, 2020. Accepted: July 15, 2020. Pre-published: July 23, 2020. doi:10.3324/haematol.2020.261784 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.

2834

yperimmune plasma from patients convalescing from COVID-19 is a potential treatment for severe COVID-19 in other patients. We conducted a multicenter, one-arm, proof-of-concept interventional study. Patients with COVID-19 with moderate-to-severe acute respiratory distress syndrome, elevated C-reactive protein level and need for mechanical ventilation and/or continuous positive airway pressure were enrolled. One to three units (each of 250-300 mL) of hyperimmune plasma (neutralizing antibody titer â&#x2030;Ľ1:160) were administered. The primary outcome measure was 7-day hospital mortality. Secondary study outcomes were PaO2/FiO2, changes in laboratory and radiological parameters, weaning from mechanical ventilation and safety of the intervention. The study enrolled 46 patients between March 25 and April 21, 2020. The mean age of the patients was 63 years and 61% were male. Thirty of the patients were on continuous positive airway pressure and seven were intubated. The mean PaO2/FiO2 was 128 (standard deviation [SD] 47). Bilateral infiltrates on chest X-ray were present in 36 patients (84%). The mean duration of symptoms and ARDS was 14 (SD 7) and 6 (SD 3) days, respectively. Three patients (6.5%) died within 7 days as compared to an expected 15% according to national statistics and 30% in a small concurrent cohort of 23 patients. The upper one-sided 90% confidence interval (CI) was 13.9%, allowing rejection of the null hypothesis of a 15% mortality. The PaO2/FiO2 increased by 112 units (95% CI: 82-142) in survivors and the severity of the chest X-ray findings decreased in 23% (95% CI: 5%-42%). C-reactive protein, ferritin and lactate dehydrogenase levels decreased by 60%, 36% and 20%, respectively. Weaning from continuous positive airway pressure was achieved in 26/30 patients and it was possible to extubate three of the seven patients who had been intubated. Five serious adverse events occurred in four patients (2 likely and 2 possibly treatment-related). In conclusion, hyperimmune plasma showed promising benefits in COVID-19. Although these benefits need to be confirmed in a randomized controlled trial, this proof-of-concept study could open the way to future developments including hyperimmune plasma banking, standardized pharmaceutical products and monoclonal antibodies. haematologica | 2020; 105(12)

Hyperimmune plasma in severe COVID-19

Introduction At the end of 2019, a new coronavirus strain was reported in the Chinese province of Wuhan and was named 2019-nCoV or SARS-CoV-2.1-3 The rapid spread of infection by this virus and its resultant disease, COVID-19, in western countries almost overcame the capacity of health systems to respond, leading to high numbers of hospitalized people and deaths. There has been an inevitable lag between the onset of the pandemic and the availability of an effective therapy, and, as of today, no treatment has been convincingly shown to be effective.4-7 Previous data on the use of convalescent plasma during the SARS and MERS epidemics suggest that antibodies could be passively transferred to patients by administering specific antibodies contained in the plasma from recovered/convalescent subjects.8-16 A meta-analysis on the use of hyperimmune immunoglobulins in severe acute respiratory infections of viral etiology, published in 2014, concluded that the technique is effective and safe, although well-designed clinical trials were advocated.17 At the time of designing this study, there were very few studies in the literature demonstrating the feasibility and efficacy of hyperimmune plasma in the SARS-CoV-2 pandemic, and all of them reported small case series. Shen et al. described five severely ill patients who showed an improvement in a variety of signs and symptoms of COVID-19 after the infusion of hyperimmune plasma.18 In the same journal, Roback and Guarner discussed the need for larger studies.19 Duan and colleagues presented a study of ten severely ill COVID-19 patients;20 the primary endpoint was safety. They demonstrated that all patients tolerated plasma transfusion without severe adverse events and had improvements in clinical symptoms and laboratory values from day 3 after infusion. On this background, we designed and conducted a proof-of-concept, interventional, multicenter study to determine the potential efficacy and safety of infusions of hyperimmune plasma, obtained from convalescent donors, in COVID-19 patients with respiratory failure and lung infiltration at chest radiogram, hospitalized in the participating Centers.

Methods Design This was a proof-of-concept, one-arm, multicenter interventional study on the short-term (7 days) efficacy and safety of the infusion of hyperimmune plasma in COVID-19 patients with moderately to severely compromised respiratory function, according to the Berlin score. It was hoped that the evidence collected would help either to plan an informed large clinical trial or to dismiss the proposed treatment if irrelevant. The primary endpoint of the study was 7-day mortality; secondary endpoints, all evaluated at 7 days, were changes in respiratory function (PaO2/FiO2 ratio), laboratory values (C-reactive protein, ferritin, lactate dehydrogenase, viral load) and radiological signs, as well as weaning from mechanical ventilation (continuous positive airway pressure [CPAP] and/or naso-tracheal intubation).

Setting and population The study was conducted in two university hospitals and one general hospital in northern Italy and was registered at clinicaltrials.gov as NCT 04321421. It was approved by the local ethical committee on March 17, 2020 (n. 20200027967). Patients were haematologica | 2020; 105(12)

enrolled between March 25, 2020 and April 21, 2020. Follow-up was closed on April 28, 2020. Eligibility criteria are summarized in Online Supplementary Table S1. Data were entered into a database in REDCap hosted at the Fondazione IRCCS Policlinico San Matteo (Pavia, Italy) and monitored remotely for missing data. The schedule of assessments is summarized in Online Supplementary Table S2. Mortality data from a control cohort of 23 consecutive patients from the Pavia COVID Registry, observed between March 10, 2020 and March 24, 2020 and satisfying the same entry criteria, were retrieved for comparison. These patients were observed for 7 days for comparison with the trial cohort.

Selection of convalescent donors and hyperimmune plasma Male adults or females with no previous pregnancy who had recovered from COVID-19 and had two consecutive negative naso-pharyngeal swabs performed in the 7 to 30 days before potential recruitment as donors were identified from the hospital records; their suitability was assessed according to current Italian guidelines and transfusion law.21,22 Their plasma was collected using latest-generation cell separators (Trima Accel –Terumo BCT and Amicus –Fresenius Kabi). A plasma volume of about 660 mL was collected during each procedure and immediately divided equally into two bags using a sterile tubing welder. Plasma pathogen reduction was performed with the INTERCEPT processing system (Cerus Europe BV) or the Mirasol PRT System (Terumo BCT, Lakewood, CO, USA). The collected units were stored at a controlled temperature ranging from -40°C to -25°C.23 (see the Online Supplementary Material for further details).

Plasma infusion Plasma was delivered ready-for-use by the Immunohematology Service to the COVID Units and was administered to the patients over 30 to 60 min, under supervision of the treating physician.

SARS-CoV2 RNA detection

Total nucleic acids (DNA/RNA) were extracted from 200 mL of respiratory specimens. Clinical samples were pretreated with 1:1 ATL lysis buffer and the nucleic acids were extracted using the QIAsymphony® instrument with a QIAsymphony® DSP Virus/Pathogen Midi Kit (Complex 400 protocol) according to the manufacturer’s instructions (QIAGEN, Hilden, Germany). Specific real-time reverse transcriptase-polymerase chain reactions targeting RNA-dependent RNA polymerase and E genes were used to detect the presence of SARS-CoV-2, in accordance with World Health Organization guidelines24 and the protocol published by Corman et al.25

SARS-CoV-2 microneutralization assay The titer of neutralizing antibodies against SARS-CoV-2 was determined using the following protocol.26,27 Briefly, 50 mL of a serum sample from each patient, starting from 1:10 in a serial 4fold dilution series, were added to two wells of a flat-bottomed tissue culture microtiter plate (COSTAR, Corning Incorporated, NY, USA), mixed with an equal volume of 50 TCID50 of a SARSCoV-2 strain isolated from a symptomatic patient. The plates were incubated at 33°C in 5% CO2. The SARS-CoV-2 strain was previously titrated to calculate the 50TCID50 to be used in the test. All dilutions were made in Eagle minimum essential medium with addition of 1% penicillin, streptomycin and glutamine and 5 mL/mL of trypsin. After 1 h of incubation at 33°C in 5% CO2, VERO E6 cells (VERO C1008 [Vero 76, clone E6, Vero E6]; ATCC® CRL-1586™] were added to each well. After another 48 h of incubation at 33°C in 5% CO2, the wells were stained with Gram crys2835

C. Perotti et al. Table 1. Description of medical history, baseline symptoms and laboratory findings.

Variable

Surviving patients

Age (years), mean (SD) 62 (11) Male, n (%) 28 (65) BMI (kg/m2), mean (SD) 27.5 (6.6) Days of symptoms, mean (SD) 14 (7) Days of ARDS, mean (SD) 6 (3) Temperature (°C), mean (SD) 38.4 (0.5) Comorbidities ≥2, n (%) 17 (39) Hypertension, n. (%) 20 (46) Diabetes, n. (%) 8 (17) Cardiovascular, n. (%) 6 (14) Cerbrovascular, n. (%) 1 (2) Neoplastic, n. (%) 7 (16) COPD, n. (%) 2 (5) Chronic kidney disease, n. (%) 4 (9) Dyslipidemia, n (%) 9 (21) Symptoms (prevalence ≥10%), n. (%) Fever 43 (100) Cough 29 (64) Fatigue 20 (46) Diarrhea 5 (12) Myalgia 6 (14) Dyspnea (on mild exertion) 25 (58) Dyspnea (at rest) 16 (37) Concomitant treatment, n. (%) Antivirala 18 (42) Antibiotics 36 (84) Hydroxychloroquine 37 (86) Anticoagulant 42 (98) Oxygen support, n. (%) 43 (100) Intubation 7 (16) CPAP 30 (70) High flow 5 (12) Low flow 1 (2) Oxygen saturation (%), 94% (3) mean (SD) PaO2/FiO2, mean (SD) 131 (47) Berlin score, severe, n (%) 12 (30) CRP (mg/dL), median (IQR) 10.7 (2.7-17.2) Ferritin (mg/L), 1102 (669-2231) median (IQR) LDH (m/L), median (IQR) 470 (382-691) Creatinine (mg/dL), 0.64 (0.57-0.97) median (IQR) Hs-troponin (TnI) (ng/L), 8.3 (4.3-23.4) median (IQR) Chest radiogram bilateral 36 (84%) multilobe infiltrates, n (%) Nasal swab viral load (U), 3.4 (2.1- 5.9) median (IQR) Number of plasma units administered, n (%) 1 21 (49) 2 21 (49) 3 1 (2)

Deceased Deceased Deceased Patient 1 Patient 2 Patient 3 (1 day) (5 days) (6 days) 68 F 22.7 29 8 37.1 3 Yes Yes No No Yes No No No

86 F 31.2 2 1 38.2 4 Yes Yes No No Yes No Yes No

80 F 21.3 11 7 37.5 0 No No No No No No No No

Yes Yes Yes Yes No Yes No

No No No No No Yes No

Yes No Yes No No No Yes

No Yes Yes No CPAP

No No Yes Yes CPAP

No Yes Yes Yes High flow

97.2

92.5

92.4

95 Yes 12.9 4919

116 No 18.6 1149

67 Yes 3.5 -

580 0.89

572 2.07

480 0.62

158

Yes

5.3

4.3

3.7

Neutralization titer (first unit), n (%) 1:80 1:160 1:320 1:640

1:640

1:320

1:160

1 (2) 27 (61) 11 (24) 6 (13)

Lopinavir/ritonavir (n=1); darunivir/ritonavir (n=16), darunavir/cobicistat (n=1). SD: standard deviation; BMI: body mass index; ARDS: acute respiratory distress syndrome; COPD: chronic obstructive pulmonary disease; CPAP: continuous positive airway pressure; PaO2: partial pressure of arterial oxygen; FiO2: fraction of inspired oxygen; CRP: C-reactive protein; IQR: interquartile range; LDH: lactate dehydrogenase. a

tal violet solution (Merck KGaA, 64271 Damstadt, Germany) plus 5% formaldehyde 40% m/v (Carlo ErbaSpA, Arese, MIlan, Italy) for 30 min. The microtiter plates were then washed in running water. Wells were scored to evaluate the degree of cytopathic effect compared to that of the virus control. Blue staining of wells indicates the presence of neutralizing antibodies. The neutralizing titer was the maximum dilution with reduction of 90% of the cytopathic effect. A positive titer was defined as one equal or greater than 1/10. Positive and negative controls were included in all test runs.

Sample size The first case of COVID-19 was diagnosed in Italy on February 20, 2020. Considering the hospitalization and mortality data retrieved from the Italian National Institute of Health on March 16,28,29 for benchmarking purposes, we used a conservative estimate of mortality of about 15% (i.e., a null hypothesis, H0, of survival of 85%) in patients treated according to the standard of care, corresponding to the general COVID-related mortality in Italy. We expected the mortality to decrease to 5% (survival 95%; H1) with the proposed hyperimmune plasma infusion. This being a proof-of-concept study, we used a one-sided type I error of 10%. According to the one-stage Fleming design, 43 patients would give a power of more than 80% to reject H0. If we were to observe at least 40 successes, the H0 hypothesis could be rejected and we would consider proceeding with a future, larger trial. Three additional patients were enrolled to allow for possible drop-outs.

Statistical analysis All continuous variables are summarized using the mean and standard deviation (SD) or the median and interquartile range (IQR). Frequencies and percentages are reported for all categorical measures. All enrolled patients who received a plasma infusion constitute the analysis population.

Primary endpoint The observed mortality rate was computed as the number of deaths over the full analysis population. The one-sided exact binomial confidence interval (CI), at the 90% level (by design) is presented. The clinical and laboratory findings at baseline are described in aggregate for patients surviving 7 days and listed individually for patients who died. No formal tests were performed.

Secondary endpoints

continued in the next column

2836

continued from the previous column

To assess changes in PaO2/FiO2 ratio, lactate dehydrogenase, Creactive protein and ferritin levels, and viral load over time we fitted repeated measures linear models (with Huber-White clustered robust standard errors to account for intra-patient correlation) or bootstrapped median regression models (depending on the distribution). The coefficients comparing day 7 to day 1 together with haematologica | 2020; 105(12)

Hyperimmune plasma in severe COVID-19 Table 2. List of adverse events and relation to treatment for each patient.

Patient

Type of adverse event

Relation to treatment

Resolution

1408-7

Chills and fever during transfusion Subsegmental pulmonary embolism Anaphylaxis/hypersensitivity Transfusion-related acute lung injury Urticaria

Likely, probable Excluded or unlikely Possible Possible Likely, probable

No (transfusion interrupted) No Yes No Yes

1408-18 1408-19 1408-21

their 95% CI are presented to describe the changes at the end of the study.

Results Study cohort We enrolled 46 patients from the three centers participating in this proof-of-concept study. The patientsâ&#x20AC;&#x2122; mean age was 63 years (SD 12) and 28 were male (61%). Their mean oxygen saturation was 94% (SD 3) and their mean PaO2/FiO2 ratio was 128 (SD 47). Fourteen (33%) had a severe Berlin score, 30 patients (70%) were on CPAP and seven (16%) were intubated. Nineteen (41%) had two or more comorbidities and 36 (84%) had bilateral multilobe infiltrates as shown by chest X-ray. Patients had been symptomatic for a mean of 14 days (SD 7) and had had acute respiratory distress syndrome for a mean of 6 days (SD 3). More than 80% of patients were treated with antibiotics, hydroxychloroquine and anticoagulants (Table 1).

Plasma infusion and safety Twenty-four patients received one unit of plasma, 21 received two units and one patient received three units. The plasma units administered at the first infusion had a neutralizing antibody titer of 1:160 or 1:320 in 85% of patients and of 1:320 in 12%; one patient only received plasma with a 1:80 titer (Table 1). At the second infusion, the titers were 1:80 in two patients, 1:160 in 11 patients, 1:320 in seven patients and 1:640 in one patient. The third infusion performed in a single patient had a titer of 1:320. The plasma infusion was well tolerated in 42/46 patients. It was interrupted in one case. Five serious adverse events occurred in four patients. In one case the transfusion had to be interrupted. In two cases the relation to treatment was considered as likely and in two as possible (Table 2). Three adverse events did not resolve spontaneously and were treated accordingly.

Primary endpoint Three patients of the 46 (6.5%) died within 7 days (on days 1, 4 and 6); the upper one-sided 90% CI was 13.9% and 40 of the first 43 patients enrolled survived, allowing us to reject the null hypothesis of a 15% mortality. The main characteristics of the three patients who died are listed in Table 2. Two had important comorbidities, including diabetes, hypertension and cancer, while the third had an extremely low PaO2/FiO2 ratio of 67 at the time of the plasma infusion. Among survivors, the severity of the condition at baseline was confirmed by the low oxygen saturation (mean 94%) and PaO2/FiO2 (mean 131). More than 89% of patients showed bilateral multilobe infiltrates on chest X-rays and all had markedly elevated laboratory biomarkers (Table 1). haematologica | 2020; 105(12)

In a concurrent cohort of 23 patients from the Pavia COVID Registry, observed between March 10, 2020 and March 24, 2020 who met the same entry criteria as used for this study and who were followed up for 7 days (Online Supplementary Table S3), the observed mortality was 30% (two-sided 80% CI: 18%-46%, or equivalently, the lower 90% one-sided limit was 18%, which is higher than the upper limit, reported above, of 13.9% for the treated cohort)

Secondary endpoints At 7 days after plasma infusion the PaO2/FiO2 increased by 112 units (95% CI: 82-142) in survivors and bilateral multilobe infiltrates on the chest X-ray had disappeared in 23% of patients (95% CI: 5%-42%). C-reactive protein, ferritin and lactate dehydrogenase levels all decreased, by 90%, 36% and 20%, respectively (Table 3, Figure 1). Conversely, no or little improvement was documented in the three patients who died (Figure 1). Overall, 30 patients were on CPAP and seven were intubated. Weaning from CPAP was achieved in 26 patients over a median time of 2 days (IQR 0-3) and three of the intubated patients were extubated after a median of 2 days (IQR 1-5). Two of 16 patients who were in the Intensive Care Unit were discharged from the Unit within the 7 days following the infusion (both on day 3). Two patients were put on extracorporeal membrane oxygenation 1 and 6 days after the plasma infusion. No patient was discharged from the hospital within the 7day study observation period.

Discussion This proof-of-concept study showed that the infusion of highly specific hyperimmune plasma in COVID-19 patients with severe respiratory failure reduces short-term mortality by 2.5 times, from an expected 15% (i.e., about 1 in 6 patients) at the time of study design to 6% (i.e., about 1 in 15 patients). Compared with the mortality rate in a concurrent series of 23 patients, satisfying the same entry criteria and observed in the period March 10 to March 24, 2020, the decrease in deaths was even more dramatic (5fold, from 30% to 6%). Only three patients died during the 7-day study period. Of these three patients, two had important comorbidities while the third had an extremely low PaO2/FiO2 ratio at the time of the plasma infusion. Patients had been symptomatic for 2 weeks at the time of plasma infusion. Most had been on treatment with antibiotics, hydroxychloroquine and anticoagulants. Among survivors, the PaO2/FiO2 ratio increased 2-fold, mirrored by decreases in the levels of biomarkers: C-reactive protein by 90%, ferritin by 36%, and lacate dehydrogenase by 20%. The bilateral multilobe infiltrates on chest radio2837

C. Perotti et al.

grams disappeared in one-third of the study population. The viral load was reduced to null. Serious adverse events occurred in four patients; in two cases, they were likely related to the transfusion. Of note, one adverse event was potentially transfusion-related acute lung injury (TRALI), the features of which are similar to those of COVID-19 and thus may be underdiagnosed. Importantly TRALI may be triggered by transfused anti-

bodies.30 Indeed TRALI was reported in 11 cases by Joyner et al. in a large, safety study of 5,000 COVID-19 patients treated with convalescent plasma.31 In an attempt to minimize the risk of TRALI and other transfusion-related serious adverse events, we excluded women who had had previous pregnancies from donating plasma for this study. Overall, both our data and, on a larger scale, those from Joyner et al.31 and Duan et al.20 confirm that transfusion of

Table 3. Changes from baseline in functional, laboratory and radiological parameters in survivors.

Variable Oxygen saturation (%), mean (SD) PaO2/FiO2, mean (SD) Chest radiogram bilateral multilobe infiltrates, n (%) CRP (mg/dL), medan (IQR) Ferritin (mg/L), median (IQR) LDH (U/L), median (IQR) Viral load (U) Â°

Day 1 N=43

Day 7 N=43

Difference (95% CI)

94 (3) 131 (47)

97 (2) 243 (40)

3 (1 to 4) 112 (82 to 142)

36 (84) 10.3 (2.5-17.2) 1102 (669-2231) 470 (382-691 3.4 (2.1- 5.9)

26 (61) 0.9 (0.5-8.0) 668 (406-1315) 378 (323-422) 0 (0-3.2)

-23% (-42% to -5%) -9.5( -13.3 to -5.4) -397 (-955 to 161) -92 (-157 to -27) -3.5 (-6.2 to -0.8)

Â°Pavia cases only (assessed in 17 patients ). 95% CI: 95% confidence interval; SD: standard deviation; PaO2: partial pressure of arterial oxygen; FiO2: fraction of inspired oxygen; CRP: C-reactive protein; IQR: interquartile range; LDH: lactate dehydrogenase.

Figure 1. Changes of respiratory function and laboratory parameters over time from day 1 to day 7 for survivors and patients who died. (A) Whisker plots of the mean and 95% confidence interval (95% CI) of the PaO2/FiO2 values. (B-D) Whisker plots of the median and 95% CI values for C-reactive protein (B), lactate dehydrogenase (C) and ferritin (D). Estimates and 95% CI values were obtained from linear (A) and quantile (B-D) regression models for repeated measures. PaO2: partial pressure of arterial oxygen; FiO2: fraction of inspired oxygen; CRP: C-reactive protein; LDH: lactate dehydrogenase

2838

haematologica | 2020; 105(12)

Hyperimmune plasma in severe COVID-19

convalescent plasma appears to be safe in hospitalized patients with COVID-19. Although hyperimmune plasma was used for the treatment of severe cases in the 2002-2004 SARS outbreak,8-14,17 few data are available from the COVID-19 epidemics. Our results are consistent with those of preliminary experiences from China. As recently described in JAMA,18,19 five critically ill patients at Shenzhen Third People’s Hospital (Shenzhen, China) were treated with convalescent plasma with a neutralizing titer of 1:80 to 1:480, 10 to 20 days after admission. The patients’ clinical conditions and laboratory findings improved. As a result, three patients were discharged; the other two patients were still hospitalized after 1 month. In a second study, performed in three hospitals in Wuhan (China),20 the outcomes of ten patients with severe disease, who were treated with convalescent plasma with high neutralizing titers ( ≥1:640) at a median of 16 days after the onset of symptoms, were determined. Following improvements in clinical and laboratory parameters, three patients were discharged, and the other seven were ready for discharge. In contrast, among a group of historical controls, similar for baseline characteristics, only one patient improved, six were stable and three died. A third report of compassionate use of hyperimmune plasma gave encouraging results as well.32 In contrast, in a fourth retrospective study, administration of convalescent plasma to six patients led to suboptimal results.33 Prior determination of neutralizing antibody response had not been performed. Indeed, both the Food and Drug Administration34 and the European Commission35 strongly recommend that SARSCoV-2 neutralizing antibody titers be measured in the donated plasma. While the Food and Drug Administration recommends a minimum neutralizing antibody titer of 1:160, indicating that a titer of 1:80 might be acceptable in some cases, the European Commission considers titers of 1:320 or more to be optimal, although lower thresholds could be considered. In our study, of the 68 units of hyperimmune plasma administered, only three (4%) had a titer below 1:160; 58 (84%) had a titer between 1:160 and 1:320 and seven had a titer of 1:640, largely consistent with the international recommendations. Of note, the promising efficacy of using neutralizing antibodies was described in a recent study that reported on a human monoclonal antibody neutralizing SARS-CoV-2 (and SARS-CoV) in cell cultures.36 A strong and novel point of our study is that we titrated the plasma from patients who had recovered from COVID-19 by quantifying COVID-19-specific neutralizing antibodies. For this purpose, we developed a new rapid microneutralization test based on evaluation of a 90% reduction of cytopathic effect in 48 h with Gram crystal violet staining. We introduced this staining to increase the readout of our neutralization test compared to that of other microneutralization test assays, such as the plaque reduction neutralization test, which require the overlay of cells and a longer time for the result.37 The plaque reduction neutralization test is a gold standard for the detection of neutralizing antibodies and has a high sensitivity and specificity. Recently, Li and colleagues presented the results of the first randomized clinical trial of convalescent plasma therapy for patients with COVID-19 conducted in China.38 The study, which enrolled 103 patients, showed more favorable haematologica | 2020; 105(12)

outcomes (measured as time to clinical improvement within 28 days, 28-day mortality and time to discharge) for patients who received convalescent plasma compared to the outcomes of patients in the control group. However, the differences were not statistically significant, the trial having been terminated early and as such, was underpowered.38 Our study has some limitations. First of all, it lacked a randomized control arm; it was, however, designed as a proof-of-concept study to verify the potential efficacy and safety of the administration of hyperimmune plasma in severely compromised COVID-19 patients and to inform the design of a rigorous, randomized controlled trial. Despite the lack of a randomized control arm, mortality was shown to be decreased by the treatment, both when compared to the mortality of hospitalized patients in Italy at the time of designing the study and when compared to the mortality in our concurrent, similar observational cohort. Interestingly, the 8.5% absolute mortality risk reduction observed in our study was very similar to that in the recent randomized trial from China.38 Indeed, in a subgroup analysis including 23 cases and 22 controls with severe COVID-19, the authors observed a 9.1% reduction in the mortality rate in the group of patients treated with convalescent plasma. A second limitation is that the study was designed at the very beginning of the pandemic in Italy. The patients were included under the pressure of a medical emergency in order to provide them with a potentially effective treatment in the very short term, given the high mortality.39 For this reason, some information was not planned to be collected, such as, but not only, the levels of D-dimer or other markers of inflammation, and long-term outcome. In conclusion, we were able to show a promising benefit of hyperimmune plasma in COVID-19 patients, both through a reduction of mortality, an improvement in respiratory function and decreases in inflammatory indices. This was a proof-of-concept study, thus these findings should not be over-interpreted and efficacy cannot be advocated yet. Nevertheless, the results pave the way for future developments including the rigorous demonstration of hyperimmune plasma efficacy in a randomized clinical trial, and possibly, the need for hyperimmune plasma banking to anticipate a potential second wave of the pandemic, the development of standardized pharmaceutical products made from the purified antibody fraction (concentrated COVID-19 H-Ig) and last, but not least, the production of monoclonal antibodies on a large scale. Acknowledgments The authors would like to thank Valeria Scotti, librarian at Fondazione IRCCS Policlinico San Matteo for her help with the references. We also thank all the donors who, after experiencing a difficult time being victims of the COVID-19 disease, are freely giving their convalescent plasma for the benefit of all. COVID-19 Plasma Task Force Angelo Corsico,1 Federica Melazzini,1,2 Marco Lenti,1,2 Cristina Mortellaro,1 Edoardo Vecchio Nepita,1 Gianluca Viarengo,1 Giorgio Iotti,1 Luciano Perotti,1 Marco Maurelli,1 Margherita Sambo,1 Mariangela Delliponti,1 Raffaella Di Martino,1 Roberta Maserati,1 Valentina Zuccaro1 and Gennaro Mascaro.3 1 Fondazione IRCCS Policlinico San Matteo, Pavia; 2University of Pavia, Pavia and 3Ospedale Maggiore della Carità, Novara, Italy. 2839

C. Perotti et al.

References 1. Zhai P, Ding Y, Wu X, Long J, Zhong Y, Li Y. The epidemiology, diagnosis and treatment of COVID-19. Int J Antimicrob Agents. 2020;55(5):105955. 2. Peeri NC, Shrestha N, Rahman MS, et al. The SARS, MERS and novel coronavirus (COVID-19) epidemics, the newest and biggest global health threats: what lessons have we learned? Int J Epidemiol. 2020:49 (3);717-726 3. Xie M, Chen Q. Insight into 2019 novel coronavirus - an updated interim review and lessons from SARS-CoV and MERS-CoV. Int J Infect Dis. 2020;94:119-124. 4. Zhang L, Liu Y. Potential interventions for novel coronavirus in China: a systematic review. J Med Virol. 2020;92(5):479-490. 5. Zhou M, Zhang X, Qu J. Coronavirus disease 2019 (COVID-19): a clinical update. Front Med. 2020;14(2):126-135. 6. Rome BN, Avorn J. Drug evaluation during the Covid-19 pandemic. N Engl J Med. 2020;382(24):2282-2284. 7. World Health Organization. Clinical management of severe acute respiratory infection when novel coronavirus (nCoV) infection is suspected. Published March 2020. https://www.who.int/publicationsdetail/clinical-management-of-severe-acuterespiratory-infection-when-novel-coronavirus-(ncov)-infection-is-suspected [Accessed April 24, 2020]. 8. Wong HK, Lee CK. Pivotal role of convalescent plasma in managing emerging infectious diseases. Vox Sang. 2020;115(7):545547. 9. Luke TC, Kilbane EM, Jackson JL, Hoffman SL. Meta-analysis: convalescent blood products for Spanish influenza pneumonia: a future H5N1 treatment? Ann Intern Med. 2006;145(8):599-609. 10. Zhou B, Zhong N, Guan Y. Treatment with convalescent plasma for influenza A (H5N1) infection. N Engl J Med. 2007;357(14):14501451. 11. World Health Organization. Use of convalescent whole blood or plasma collected from patients recovered from Ebola virus disease for transfusion, as an empirical treatment during outbreaks. Interim guidance for national health authorities and blood transfusion services. Published September 2004. http://apps.who.int/iris/bitstream/10665/1 35591/1/WHO_HIS_SDS_2014.8_eng.pdf? ua=1 [Accessed April 24, 2020]. 12. Marano G, Vaglio S, Pupella S, et al. Convalescent plasma: new evidence for an old therapeutic tool? Blood Transfus. 2016;14(2):152-157. 13. Hung IF, To KK, Lee CK, et al. Convalescent plasma treatment reduced mortality in patients with severe pandemic influenza A

2840

(H1N1) 2009 virus infection. Clin Infect Dis. 2011;52(4):447-456. 14. Keller MA, Stiehm ER. Passive immunity in prevention and treatment of infectious diseases. Clin Microbiol Rev. 2000;13(4):602614. 15. Bloch EM, Shoham S, Casadevall A, et al. Deployment of convalescent plasma for the prevention and treatment of COVID-19. J Clin Invest. 2020;130(6):2757-2765. 16. Casadevall A, Pirofski LA. The convalescent sera option for containing COVID-19. J Clin Invest. 2020;130(4):1545-1548. 17. Mair-Jenkins J, Saavedra-Campos M, Baillie JK, et al. The effectiveness of convalescent plasma and hyperimmune immunoglobulin for the treatment of severe acute respiratory infections of viral etiology: a systematic review and exploratory meta-analysis. J Infect Dis. 2015;211(1):80-90. 18. Shen C, Wang Z, Zhao F, et al. Treatment of 5 critically ill patients with COVID-19 with convalescent plasma. JAMA. 2020;323(16): 1582-1589. 19. Roback JD, Guarner J. Convalescent plasma to treat COVID-19: possibilities and challenges. JAMA. 2020;323(16):1561-1562. 20. Duan K, Liu B, Li C, et al. Effectiveness of convalescent plasma therapy in severe COVID-19 patients. Proc Natl Acad Sci U S A. 2020;117(17):9490-9496. 21. Istituto Superiore di SanitĂ . Centro Nazionale Sangue. https://www.centronazionalesangue.it/sites/default/files/GU%2 0SG%20n.300%20del%2028-122015_SO_069.pdf [Accessed May 11, 2020]. 22. Franchini M, Marano G, Velati C, Pati I, Pupella S, Liumbruno GM. Operational protocol for donation of anti-COVID-19 convalescent plasma in Italy. Vox Sang. 2020 Apr 23. [Epub ahead of print]. 23. Perotti C, Del Fante C, Baldanti F, et al. Plasma from donors recovered from new Corona virus 2019 as therapy for critical patients with COVID-19 (COVID-19 PLASMA study). A multicentre study protocol. Int Emerg Med. 2020 May 28. [Epub ahead of print] 24. World Health Organization. Diagnostic detection of 2019-nCoV by real-time RTPCR. Published January 2020. https://www.who.int/docs/defaultsource/coronaviruse/protocol-v2-1.pdf [Accessed April 24, 2020]. 25. Corman VM, Landt O, Kaiser M, et al. Detection of 2019 novel coronavirus (2019nCoV) by real-time RT-PCR. Euro Surveill. 2020;25(3):2000045. 26. Percivalle E, Cassaniti I, Sarasini A, et al. West Nile or Usutu virus? A three-year follow-up of humoral and cellular response in a group of asymptomatic blood donors. Viruses. 2020;12(2):157. 27. Percivalle E, CambiĂ¨ G, Cassaniti I, et al. Prevalence of SARS-CoV-2 specific neutralis-

ing antibodies in blood donors from the Lodi Red Zone in Lombardy, Italy, as at 06 April 2020. Euro Surveill. 2020;25(24): 2001031. 28. Torresi M. Statistiche coronavirus Lombardia. https://statistichecoronavirus.it/ regioni-coronavirus-italia/lombardia/ [Accessed May 11, 2020]. 29. Rosini U. COVID-19 Italia - Monitoraggio situazioneDati Regionali. https://github.com/pcm-dpc/COVID19/blob/master/dati-regioni/dpc-covid19ita-regioni-20200316.csv [Accessed May 11, 2020]. 30. Semple JW, Rebetz J, Kapur R. Transfusionassociated circulatory overload and transfusion-related acute lung injury. Blood. 2019;133(17):1840-1853. 31. Joyner M, Wright RS, Fairweather DL, et al. Early safety indicators of COVID-19 convalescent plasma in 5,000 patients. medRxiv 2020; doi: https://doi.org/10.1101/ 2020.05.12.20099879 32. Zhang B, Liu S, Tan T, et al. Treatment with convalescent plasma for critically ill patients with SARS-CoV-2 infection. Chest. 2020;158(1):e9-e13. 33. Zeng QL, Yu ZJ, Gou JJ, et al. Effect of convalescent plasma therapy on viral shedding and survival in COVID-19 patients. J Infect Dis. 2020;222(1):38-43. 34. Food and Drug Administration. Center for Biologics Evaluation and Research. Investigational COVID-19 convalescent plasma. Guidance for industry. Published May 2020. https://www.fda.gov/media/ 136798/download [Accessed May 11, 2020]. 35. European Commission Directorate-General For Health And Food Safety. An EU programme of COVID-19 convalescent plasma collection and transfusion. Guidance on collection, testing, processing, storage, distribution and monitored use. Published April 2020. https://ec.europa.eu/health/sites/ health/files/blood_tissues_organs/docs/guid ance_plasma_covid19_en.pdf [Accessed May 11, 2020]. 36. Wang C, Li W, Drabek D, et al. A human monoclonal antibody blocking SARS-CoV-2 infection. Nat Commun. 2020;11(1):2251. 37. Perera RA, Mok CK, Tsang OT, et al. Serological assays for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), March 2020. Euro Surveill 2020;25(16): 2000421. 38. Li L, Zhang W, Hu Y, et al. Effect of convalescent plasma therapy on time to clinical improvement in patients with severe and life-threatening COVID-19: a randomized clinical trial. JAMA. 2020;324(5)1-11. 39. Characteristics of SARS-CoV-2 patients dying in Italy. Report based on available data on June 18th, 2020. https://www.epicentro.iss.it/en/coronavirus/bollettino/ReportCOVID-2019_18_june_2020.pdf [Accessed June 26, 2020].

haematologica | 2020; 105(12)

LETTERS TO THE EDITOR ABO blood groups are not associated with the risk of acquiring SARS-CoV-2 infection in young adults Four studies have previously implicated ABO blood groups in susceptibility to COVID-19 (symptomatic SARS-CoV-2 infection).1–4 This study assessed a large population of 1,769 crewmembers confined to an aircraft carrier who, therefore, constituted an excellent epidemiological model. Our results show that ABO and Rh(D) blood groups are not associated with increased or decreased risk of infection by SARS-CoV-2. These data contrast with those of previous studies. Unlike the other studies, we did not compare the distribution of blood groups in hospitalized patients with SARS-CoV-2 infection and in the general population, but rather the distribution of SARS-CoV-2 infection according to blood groups in people exposed at the same time and in the same place to SARS-CoV-2. Our study confirms that in this period of the SARS-CoV-2 epidemic, no young adults can consider themselves more or less at risk in relation to their blood type. Numerous studies indicated that people with blood group A had a significantly higher risk of SARS-CoV-2 infection, whereas people with blood group O had a significantly lower risk.1-4 These studies could be criticized for the lack of data about patients’ exposure to the virus before developing COVID-19 and for probably not including all forms of the disease. On this background, it is important to further explore the relationships between ABO blood groups and SARS-CoV-2 infection. It is also necessary to determine whether different rates of infection reflect the effect of infection prevention strategies or a sharp increase of asymptomatic or mild forms (not

seeking medical assistance), which would affect collective immunity. A COVID-19 outbreak occurred in April 2020 on the French Navy nuclear aircraft carrier “Charles de Gaulle”, which had 1,769 crewmembers. The aircraft carrier was deployed at sea from January 22 to April 13, 2020 as part of an operational mission. A SARS-CoV-2 epidemic broke out on-ship, requiring its early (2 weeks) return to Toulon (the ship’s main harbor). The first confirmed case occurred on February 28, 2020. During the mission, the crew was contained on board and the members were, therefore, exposed at the same time and in the same place to SARS-CoV-2. When the ship returned to France, the entire crew was confined at different military bases and benefited from daily medical monitoring for 14 days after landing. This study was undertaken with the main objective of investigating possible relationships between ABO blood groups and SARS-CoV-2 infection in these well-defined conditions of exposure. To investigate the effect of ABO blood group on developing SARS-CoV-2 infection, we conducted this observational study on a retrospective cohort comprised of the crew of the aircraft carrier. The Ethics Committee of Sainte Anne Military Hospital (Toulon, France) approved the study (Institutional Review Board n. 0011873-202009). All data were collected, in the context of care, from completely anonymized files, in accordance with French and European laws, including the General Data Protection Regulation. All crewmembers were informed and provided free written consent to the use of their data. All 1,769 persons on board underwent a physical examination and reverse transcriptase -polymerase chain reaction testing (RT-PCR) of two nasopharyngeal samples obtained immediately upon arrival in Toulon and at

Table 1. Characteristics of the study participants.

Characteristics

All crewmembers

Confirmed/suspected SARS-CoV-2

No SARS-CoV-2

Number (%) Median age (IQR), years Male, n (%) Blood group, n (%)* A A+ A– B B+ B– AB AB+ AB– O O+ O–

1,688 (100.0) 28 (23–35) 1466 (87.0)

1279 (76.0) 28 (23–36) 1112 (87.0)

409 (24.0) 27 (23–33) 354 (87.0)

674 (39.9) 581 (34.4) 93 (5.5) 183 (10.8) 154 (9.1) 29 (1.7) 70 (4.1) 59 (3.5) 11 (0.7) 742 (44.0) 645 (38.2) 97 (5.7)

521 (40.7) 450 (35.1) 71 (5.6) 135 (10.6) 115 (9.0) 20 (1.6) 54 (4.2) 47 (3.6) 7 (0.5) 553 (43.2) 480 (37.5) 73 (5.7)

153 (37.4) 131 (32.0) 22 (5.4) 48 (11.7) 39 (9.5) 9 (2.2) 16 (3.9) 12 (2.9) 4 (1.0) 189 (46.2) 165 (40.3) 24 (5.9)

19 (1.1) 3 (0.2) 0 (0)

19 (1.5) 3 (0.2) 0 (0)

– – 0 (0)

Oxygen therapy, n (%) ICU, n (%) Mortality

IQR: interquartile interval (25%-75%): RT-PCR: reverse transcriptase polymerase chain reaction; ICU: Intensive Care Unit. *19 crewmembers lacked blood group information in their medical records or in the French database.

haematologica | 2020; 105(12)

2841

Letters to the Editor

Table 2. Univariable analyses of factors associated with SARS-CoV-2 infection.

Factor

ABO blood group (N=1669) A (n=674) B (n=183) AB (n=70) O (n=742) Rh(D) negative (N=230)

Positive RT-PCR and/or clinical signs (n=1279) OR (95% CI) P 0.89 (0.68-1.08) 1.12 (0.78-1.59) 0.91 (0.51-1.62) 1.11 (0.89-1.40) 1.08 (0.78-1.49)

0.22 0.52 0.88 0.32 0.61

least 7 days later at the end of a period of containment. Total RNA was extracted from the nasopharyngeal swab samples, and 200 mL of the sample were subjected to thermal and chemical inactivation and RNA extraction with Magnapure (Roche Diagnostic®) or Genoxtract (Biocentric®). The RT-PCR assay was run according to the French National Center protocol, targeting the IP2 and IP4 regions of the SARS-CoV-2-specific RNA-dependent RNA polymerase (RdRp) gene. Individually housed during confinement (14 days and at least 2 days after symptoms had disappeared), crewmembers with clinical signs and/or who were PCRpositive were isolated in a given area and asymptomatic PCR-negative subjects in another. The vital signs and clinical symptoms of all subjects were monitored daily by the medical teams of the French Army Health Service (Service de Santé des Armées). Subjects were considered to be infected by SARS-CoV-2 if they had at least one positive RT-PCR (confirmed cases) and/or clinical symptoms highly suggestive of COVID-19 in this epidemiological context (fever, myalgias, arthralgia, dyspnea, cough, headache, anosmia, ageusia, rhinitis, diarrhea, fatigue, cutaneous signs). Subjects were considered not to have SARS-CoV-2 infection if they had two negative RT-PCR tests and no clinical signs. The following medical information was extracted from the medical records: age, sex, blood group, COVID-19related symptoms and their duration, indication for hospitalization and whether or not a computed tomography scan was done. The impact of ABO and Rh(D) types on susceptibility to SARS-CoV-2 infection was evaluated using logistic-regression univariable analyses (risks assessed by odds ratios [OR]). All analyses were computed with R software, version 3.6.3. Given our population we determined that the study would have a power of 99.9% to detect a 5% difference between blood groups. Of the 1,769 crewmembers on the ship, 1,688 (95%) were included in this study (59 did not give consent and data were missing for 22, most of whom had left the ship because of specific duties). While on board, 1,279 crewmembers developed SARS-CoV-2 infection, accounting for a 76% (1,279/1,688) infection attack rate. The SARS-CoV-2 infection was confirmed by RT-PCR in 1,038 (62%) patients, whereas in 241 (14%) cases it was medically diagnosed (no positive RT-PCR but typical symptoms of COVID-19). After returning to France, 19 crewmembers were hospitalized with oxygen therapy, including three requiring admission to an Intensive Care Unit. The SARS-CoV-2 infection was asymptomatic in 172 (14%) cases. Among symptomatic cases, the mean duration of symptoms was 9 days (range, 1-40; 95% confidence interval, 9.08-9.88). Demographic and clinical 2842

Positive RT-PCR (n=1038) OR (95% CI) 0.83 (0.68-1.02) 1.16 (0.85-1.59) 1.12 (0.69-1.82) 1.10 (0.90-1.34) 0.91 (0.68-1.22)

P 0.08 0.33 0.70 0.36 0.56

Asymptomatic with positive RT-PCR (n=172) OR (95% CI) P 1.06 (0.76-1.49) 0.55 (0.29-1.07) 3.12 (1.6-6.03) 0.88 (0.63-1.23) 0.84 (0.51-1.38)

0.73 0.09 0.0017 0.49 0.54

characteristics of all participants and SARS-CoV-2 status are reported in Table 1. The median age of the crewmembers was 28 years and 13% were women. The population appeared to be healthy, with no significant comorbidities. Among the whole crew, the ABO blood group showed a distribution of 39.9%, 10.8%, 4.1% and 44.0% for A, B, AB and O types, respectively. The corresponding proportions for crewmembers infected with SARS-CoV-2 were 40.7%, 10.6%, 4.2% and 43.2% for A, B, AB and O types, respectively. In univariate analysis no significant relationship was found between SARS-CoV-2 infection and ABO or Rh(D) types (Table 2). Excluding the minority of cases that were diagnosed clinically without laboratory support (14%), the results did not differ: ABO or Rh(D) types were also not related to RT-PCR-positive SARS-CoV-2 infections. Concerning asymptomatic cases, AB group seemed to be related to the absence of symptom among crewmembers with RT-PCR positivity for SARS-CoV-2 (odds ratio 3.12 [95% confidence interval: 1.6-6.03], P=0.0017). The median age of the 19 patients requiring oxygen therapy was 45 years. Five patients were blood group A, 5 group B and 10 group O. All patients admitted to the Intensive Care Unit were older than 50 years: one had group B and 2 had group O blood. In this study, which assessed a large population confined to an aircraft carrier, thus constituting an excellent epidemiological model, ABO and Rh(D) blood groups were not associated with increased or decreased risk of SARS-CoV-2 infection. These data are opposed to those of previous studies which had a certain impact in the press. Unlike those studies, we did not compare the distribution of blood groups in hospitalized patients with SARS-CoV-2 and in the general population, but rather compared the distribution of SARS-CoV-2 infection according to blood groups in people exposed at the same time and in the same place to SARS-CoV-2. The contradictory results can also be explained in other ways. Our population appeared to be young and healthy, with no significant comorbidities. ABO blood types are potentially related only to severe forms of COVID-19. Another previous study hypothesized that although ABO blood type and/or cardiovascular diseases are prognostic of the severity of COVID-19 in patients, they are not factors predisposing to the risk of getting SARS-CoV-2 infection.5 There is a pathophysiological mechanism to support this hypothesis: subjects with A blood type are at risk of the development of cardiovascular diseases and severe COVID-19 because of the positive association of this blood type with angiotensin-converting enzyme activity, and the attachment of adhesion molecules on the vascular wall, which increases inflammation and decreases haematologica | 2020; 105(12)

Letters to the Editor

blood circulation.6,7 In our population, the lack of comorbidities and cardiovascular risk factors could therefore be considered a bias. However, a recent study observed no association between ABO blood group polymorphisms and deaths from COVID-19. Nineteen crewmembers had no blood group recorded in their medical records or in the French blood-bank database: 16 cases infected by SARS-CoV-2 and three not infected. This is potential confounder, although probably a limited one given that it represents such a small fraction of the total group. Otherwise the reported rate of infection (76%) is remarkably high. Close contact in an aircraft carrier does not reflect normal social conditions and this extreme situation may have possibly downgraded any protective effect of O blood group. The population with AB blood type is small. The relationship between an asymptomatic form of SARS-CoV-2 infection and this blood type needs to be confirmed on a larger scale. If confirmed, the underlying molecular mechanism of our findings will need further study.8 Regarding viral strains, the viral genomes of samples from the first 60 cases of RT-PCR-confirmed COVID-19 were sequenced. The diversity of the mutational signatures within these samples indicates the presence of several viral variants. These different variants can result from several contaminations by different strains or from evolution of the initial strain.9 In conclusion our study confirms that in this period of the SARS-CoV-2 epidemic, no young adults can consider themselves more or less at risk in relation to their blood type. Laurys Boudin,1 Frédéric Janvier,2 Olivier Bylicki3 and Fabien Dutasta4 1 Department of Oncology and Hematology; 2Biology Unit;

haematologica | 2020; 105(12)

Respiratory Disease Unit and 4Department of Internal Medicine, HIA Sainte-Anne, Toulon, France Correspondence: LAURYS BOUDIN - laurys.boudin@intradef.gouv.fr doi:10.3324/haematol.2020.265066

References 1. Li J, Wang X, Chen J, Cai Y, Deng A, Yang M. Association between ABO blood groups and risk of SARS-CoV-2 pneumonia. Br J Haematol. 2020;190(1):24-27. 2. Zhao J, Yang Y, Huang H, et al. Relationship between the ABO blood group and the COVID-19 susceptibility. medRxiv. 2020 March 27 [Epub ahead of print]. 3. Zietz M, Tatonetti NP. Testing the association between blood type and COVID-19 infection, intubation, and death. medRxiv 2020 Apr 11 [Epub ahead of print]. 4. Ellinghaus D, Degenhardt F, Bujanda L, et al. Genomewide association study of severe Covid-19 with respiratory failure. N Engl J Med. 2020;383(16):1522-1534. 5. Dai X. ABO blood group predisposes to COVID-19 severity and cardiovascular diseases. Eur J Prev Cardiol. 2020;27(13):1436-1437. 6. Wu O, Bayoumi N, Vickers MA, Clark P. ABO(H) blood groups and vascular disease: a systematic review and meta-analysis. J Thromb Haemost. 2008;6(1):62-69. 7. Paré G, Chasman DI, Kellogg M, et al. Novel association of ABO histo-blood group antigen with soluble ICAM-1: results of a genome-wide association study of 6,578 somen. PLoS Genet. 2008; 4(7):e1000118. 8. Dzik S, Eliason K, Morris EB, Kaufman RM, North CM. COVID-19 and ABO blood groups. Transfusion. 2020 Jun 19 [Epub ahead of print]. 9. Communiqué_Publication des conclusions des enquêtes sur la contamination au COVID-19 au sein du groupe aéronaval. https://www.defense.gouv.fr/salle-de-presse/communiques/communique_publication-des-conclusions-des-enquetes-sur-la-contamination-au-covid-19-au-sein-du-groupe-aeronaval (Accessed July 13, 2020).

2843

Letters to the Editor

Perforin gene variant A91V in young patients with severe COVID-19 Since early 2020 SARS-CoV-2 infectious disease (COVID-19) has been responsible for more than 300.000 deaths across the globe.1 Advanced age and previous comorbidities are clearly related to the development of severe forms of the disease (sCOVID) and an increased mortality risk.2,3 However young healthy subjects with sCOVID are also admitted to intensive care units (ICU) and die, suggesting that individual variations and/or genetic predisposing factors might play a role in modifying the clinical course and severity of the disease.4 sCOVID-19 is characterized by fever, bilateral pneumonia, lymphopenia, hyperferritinemia, elevated acute phase reagents and cytokine storm, altogether conforming a hyperinflammation scenario similar to that in secondary hemophagocytic lymphochystiocytosis (sHLH), also known as macrophage activation syndrome.5 In adults, sHLH is mostly triggered by viral infections and approximately 50% of patients experience pulmonary

disease.6 In contrast, familial HLH (fHLH) is genetically determined by mutations in genes coding for proteins related to lymphocyte cytotoxicity such as perforin (PRF1 gene). Studies in juvenile idiopathic arthritis or systemic lupus erithematosus patients show that up to 40% of individuals suffering sHLH carry heterozygous mutations in fHLH genes. In a fatal influenza A (H1N1) series, 36% of patients carried one or several mutations in fHLHrelated genes.7,8 These findings suggest an important, not yet totally recognized overlap between primary and secondary forms of HLH. It has been previously observed that the highly prevalent, fHLH-associated c.272C>T variant (p.A91V; rs35947132) in the PRF1 gene impairs the processing to the active form of perforin protein.9 Published reports associate this variant with immune diseases but it has not been validated as pathological in larger cohorts.10 The A91V PRF1 gene translates into a protein with reduced stability and abnormal trafficking which associates with a significant decrease of NK-cell cytotoxicity.11,12 Previous studies reported higher prevalence of the A91V variant in

Table 1. Description of the main clinical and laboratory characteristics of patients positive for the c.272C>T (p.A91V; rs35947132) change in the perforin PRF1 gene. Patient 1 Patient 2 Demographic characteristics Age (years) Sex Initial findings Medical history Symptoms at disease onset Imaging features

Treatment before admission to ICU

Days from disease onset to death Findings at ICU admission Days from disease onset to ICU admission Disease severity Laboratory findings at ICU admission Albumin (g/deciliter) [3.5 - 5.0] Alanine aminotransferase (U/liter) [5 - 45] Aspartate aminotransferase (U/liter) [5 - 33 Creatinine (mg/deciliter) [0.70 - 1.20] Lactate dehydrogenase (U/liter) [135 - 225] Triglycerides (mg/deciliter) Creatinine Kinase U/liter [34 - 171] Troponin T (ng/liter) [< 14] Procalcitonin (ng/liter) [â&#x2030;¤ 0.50] Prothrombin time (sec) D-dimer (ng/mililiter) [0 - 500] Serum ferritin ng/mililiter [30 - 400] Fibrinogen mg/deciliter [200 - 560] C-reactive protein mg/deciliter [0.10 - 0.50] Hemoglobin (g/deciliter) White-cell count (x103 per mm3) [4.0 - 11.3] Lymphocytes (x13 per mm3) [1.2 - 4.0] Platelets (x103 per mm3) [140 - 450] Neutrophils (x103 per mm3) [1.8 - 7.4] c.272C>T (p.A91V; rs35947132) variant

45 Female

46 Male

Hypertriglyceridemia Confusion, tachypnea and dyspnea Multi-lobar/bilateral patchy consolidations (100% involvement of both lungs). Worst radiologic findings of the series. No previous treatment

Grade 1 obesity Dry cough and fever Multi-lobar/bilateral patchy consolidations (75% involvement of both lungs) Azithromycin 500 mg/24h, Lopinavir/Ritonavir 400/100 mg/12h, hydroxychloroquine 200 mg/12h 14

0 Severe

6 Severe

3.6 25 29 0.78 663 483 110 21.3 1.15 12.5 1020 1107 870 22.93 10.9 16 1.2 380.000 13.5 Heterozygosis

3.1 49 81 1.01 757 Not tested 297 8.9 0,52 12.8 672 3032 765 23.96 12.1 4.9 1.1 175.000 3.4 Heterozygosis continued on the nex page

2844

haematologica | 2020; 105(12)

Letters to the Editor continued from the previous page

Patient 1

Patient 2

Evolution of laboratory parameters

Initial lymphopenia resolved (from 0.5 to 2.7 per mm3), hemoglobin decreased from 9.9 to 7.9 g/dL, requiring transfusion of 2 erythrocytes concentrates. D-dimer stabilized at 2240 ng/mL. CRP values increased up to 46.7 mg/dL and fell to 4,5 mg/dl after antibiotics.

Ventilation management

Mechanical ventilation was initiated soon after admission. Tracheostomy was performed on day 11. After 3 days without improvement, extracorporeal membrane oxygenation was implemented. Hydroxychloroquine 200 mg/12h, Azithromycin 500 mg/24h, Lopinavir/Ritonavir 400/100 mg/12h, Interferon b 250 mcg/48h, Methylprednisolone 250 mg/12h and antibiotics. 188 points

Sustained lymphopenia (0.5 to 1.1 per mm3), hemoglobin decreased, platelets recovered to normal levels. D-Dimers abruptly rose up to 8706 ng/ml. CRP increased up to 35.5 mg/dL. LDH stabilized. Creatinine increased to 2.81 mg/dl and Troponin T to 20 ng/L. Oxygen therapy with face mask with reservoir bag was initiated at admission. Due to progressive deterioration, mechanical ventilation was indicated on day 1. Hydroxychloroquine 200 mg/12h, Azithromycin 500 mg/24h, Lopinavir/Ritonavir 400/100 mg/12h, Methylprednisolone 50 mg/12h, Tocilizumab 20 mg/mL, and antibiotics.

Treatment received

HScore for HLH (without bone marrow aspirate and organomegaly evaluation)1

175 points

HLH: hemophagocytic lymphochystiocytosis; CPR: c-reactive protein; ICU: intensive care unit; LDH: lactate dehydrogenase.

HLH patients.13,14 It is reasonable to think that perforin bearing the A91V change could be related to suboptimal activation and effector capacities of CD8 and/or natural killer (NK) cells. In the context of a viral infection, the correct function of these cells is required to contain the viral replication, clear the virus and overcome the infection. Ineffective killing of SARS-CoV-2 infected cells might lead to a sustained activation of lymphocytes and macrophages contributing to the cytokine storm and hyperinflammation that characterizes sCOVID-19. Based on the above premises, we hypothesized that the fHLH-associated A91V PRF1 variant is prevalent in patients suffering severe forms of COVID-19. We therefore tested for the A91V PRF1 variant in all sCOVID-19 patients in the ICU of our hospital on a random day (March 27). Exon 2 of the PRF1 gene coding region was amplified using PCR. PCR products were purified and sequenced as previously reported.15 Elderly and patients with comorbidities were excluded. Twenty-two previously healthy patients between the age of 24-52 years were identified: 17 of 22 males; 14 of 22 Latin- American, 7of 22 Spanish and 1 of 22 Polish. Among the studied patients, 2 of 22 showed A91V PRF1 in heterozygosis (allele frequency of 0.045). According to the Genome Aggregation Database (gnomAD gnomad.broadinstitute.org), the calculated A91V PRF1 variant frequency in European plus Latino population is 0.031. Considering that no A91V-positive patients were detected among the Latin-American patients in intive care, the allele frequency found in our Europeans COVID-19 cohort was 0.125, almost 3-times higher than that described for Europeans in gnomAD (0.046). After 6 weeks, 17of 20 A91V-negative patients had been discharged, 2 of 20 continued hospitalization with significant clinical improvement without ventilator requirement and 1 of 20 had died. Remarkably, both A91V-positive patients died. In these patients we calculated the value of the HScore, a previously validated score which includes the most important variables independently associated with sHLH and helps to form an haematologica | 2020; 105(12)

accurate diagnosis of HLH. HScore values higher than 169 are considered positive for HLH, with a sensitivity of 93% and specificity of 86%.16 Both patients showed a high HScore for HLH (188 and 175),5 a shorter time from the disease onset to ICU admission (0 and 6 vs. 9.36 days on average) and more severe initial radiological findings (Table 1). Clinically, our A91V-positive patients had high fever associated with the respiratory symptoms. The HLH-related laboratory parameters triglycerides, fibrinogen, ferritin and aspartate aminotransferase were markedly elevated in both subjects, even while receiving immunosuppressive therapy. Unfortunately, because of the pandemic situation and rapid death of both patients, functional studies with cell samples could not be performed. In conclusion, in our young sCOVID-19 patient cohort, A91V PRF1 was prevalent. A defective A91V PRF1 may translate into suboptimal lymphocyte cytotoxicity and ineffective SARS-CoV-2 clearance, favoring the progress to sCOVID-19 with an HLH-like clinical phenotype and high mortality. Our observation merits further investigations to assess the specific influence of this variant in COVID-19 clinical course. International collaborative efforts are needed to elucidate the role of genetics in COVID-19. Oscar Cabrera-Marante, Edgard Rodríguez de Frías, Daniel E. Pleguezuelo, Luis M. Allende, Antonio Serrano, Rocío Laguna-Goya, María Esther Mancebo, Paloma Talayero, Luis Álvarez-Vallina, Pablo Morales, María José Castro-Panete and Estela Paz-Artal Servicio de Inmunología, Instituto de Investigaciones Sanitarias, Hospital 12 de Octubre, Madrid, Spain Correspondence: OSCAR CABRERA-MARANTE - oscar.cabrera@salud.madrid.org doi:10.3324/haematol.2020.260307 Acknowledgments: we thank the patients and their families, lab technicians and other health care providers of the Hospital 12 de Octubre for their participation in this research. 2845

Letters to the Editor

Funding: this work was supported by FIS-Instituto de Salud Carlos III, Grant Number: COV20/00181. This study was approved by the Institutional Ethical Board (20/167).

References 1. Novel Coronavirus (2019-nCoV) situation reports-117 (May 2020). World Health Organization (WHO). https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200516-covid-19sitrep-117.pdf?sfvrsn=8f562cc_2. 2. Zhou F, Yu T, Du R, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study [published correction appears in Lancet. 2020;395(10229):1038. 3. Yang J, Zheng Y, Gou X, et al. Prevalence of comorbidities and its effects in patients infected with SARS-CoV-2: a systematic review and meta-analysis. Int J Infect Dis. 2020;94:91‐95. 4. DeBiasi RL, Song X, Delaney M, et al. Severe COVID-19 in children and young adults in the Washington, DC Metropolitan Region. J Pediatr. 2020;223:199-203. 5. Mehta P, McAuley DF, Brown M, et al. COVID-19: consider cytokine storm syndromes and immunosuppression. Lancet. 2020; 395(10229):1033-1034. 6. Seguin A, Galicier L, Boutboul D, Lemiale V, Azoulay E. Pulmonary involvement in patients with hemophagocytic lymphohistiocytosis. Chest. 2016;149(5):1294-1301. 7. Schulert GS, Zhang M, Fall N, et al. Whole-exome sequencing reveals mutations in genes linked to hemophagocytic lymphohistiocytosis and macrophage activation syndrome in fatal cases of H1N1 influenza. J Infect Dis. 2016;213(7):1180-1188. 8. Crayne CB, Albeituni S, Nichols KE, Cron RQ. The immunology of

2846

macrophage activation syndrome. Front Immunol. 2019;10:119. 9. Trambas C, Gallo F, Pende D, et al. A single amino acid change, A91V, leads to conformational changes that can impair processing to the active form of perforin. Blood. 2005;106(3):932‐937. 10. Voskoboinik I, Lacaze P, Jang HS, et al. Prevalence and disease predisposition of p.A91V perforin in an aged population of European ancestry. Blood. 2020;135(8):582-584 11. Voskoboinik I, Sutton VR, Ciccone A, et al. Perforin activity and immune homeostasis: the common A91V polymorphism in perforin results in both presynaptic and postsynaptic defects in function. Blood. 2007;110(4):1184‐1190. 12. House IG, Thia K, Brennan AJ, et al. Heterozygosity for the common perforin mutation, p.A91V, impairs the cytotoxicity of primary natural killer cells from healthy individuals. Immunol Cell Biol. 2015;93(6):575‐580. 13. Busiello R, Fimiani G, Miano MG, et al. A91V perforin variation in healthy subjects and FHLH patients. Int J Immunogenet. 2006; 33(2):123‐125. 14. Carvelli J, Piperoglou C, Farnarier CC, et al. Functional and genetic testing in adults with hlh do not reveal a cytotoxicity defect but rather an inflammatory profile. Blood. 2020;136(5):542-552. 15. Mancebo E, Allende LM, Guzman M, et al. Familial hemophagocytic lymphohistiocytosis in an adult patient homozygous for A91V in the perforin gene, with tuberculosis infection. Haematologica 2006; 91(9):1257-1260. 16. Fardet L, Galicier L, Lambotte O, Marzac C, Aumont C, Chahwan D, et al. Development and validation of the HScore, a score for the diagnosis of reactive hemophagocytic syndrome. Arthritis Rheumatol. 2014;66:2613–2620.

haematologica | 2020; 105(12)

Letters to the Editor

Complement C5 inhibition in patients with COVID19 - a promising target? Since the report of the first cases in December 2019, infection with the severe acute respiratory coronavirus 2 (SARS-CoV-2) commonly referred as COVID-19 has become a worldwide pandemic.1 In patients with COVID-19 infection, respiratory deterioration has been associated with not only increased viral loads in the lung but also with an inadequate and exaggerated immune response.2 Preclinical data have demonstrated a role for complement activation in CoV-mediated disease. Gralinski et al., found activation of the complement system in a mouse model of CoV.3 In some patients with COVID-19, significant deposits of terminal complement components C5b-9 (membrane attack complex), C4d, and mannose binding lectin (MBL)-associated serine protease 2 (MASP-2) have been found in the microvasculature of different organs, consistent with sustained, systemic activation of the lectin- complement pathway.5 MASP-2 mediated complement overactivation has also been reported in some patients by a Chinese group.6 We assessed complement activity in 103 patients with COVID-19 followed in the Saint-Louis hospital (Pneumology Unit, Infectious Disease Unit and Intensive Care Unit [ICU]). We used a validated routine complement hemolytic activity assay (reported as CH50) which consists of testing the capacity of patient plasma to lyse sheep erythrocytes coated with antibodies, C3, C4 and sC5b-9 circulating levels by nephelometry (Siemens) and ELISA (Quidel, San Diego, CA, USA), respectively, according to the manufacturersâ&#x20AC;&#x2122; instructions. We found that the levels of C3 and C4 were increased in 57.2 % (59 of 103) and 36.9 % (38 of 103) of patients, respectively (data not shown). Moreover, the level of circulating sC5b-9 was increased in 64% of the patients (66 of 103), highlighting the systemic C5 cleavage during Covid-19 infection (healthy controls median 160 ng/mL, range: 49-362 vs. Covid-19 patients median 344 ng/mL, range:

71-883, P<0.0001). We then classified the severity of COVID-19 patients as moderate (dyspnea requiring a maximum of 3 L/min of oxygen and no other organ failure), severe (respiratory distress requiring more than 3 L/min of oxygen and no other organ failure) and critical cases (respiratory failure requiring mechanical ventilation or high flow oxygen support, shock and/or other organ failure necessitating intensive care unit [ICU] care). The plasma levels of sC5b9 were significantly higher in the three groups of patients than in the heathy donors (P<0.001 in the three groups) and higher in the patients with critical disease than in moderate disease (P=0.01) (Figure 1). Altogether, these findings suggest that C5 activation might be associated with disease severity, even if we agree causality still needs to be investigated. Nevertheless, this result has to be taken with caution since sC5b-9 may not properly reflect the situation on a cellular level. Moreover, tissue biopsies were not performed in any of our patients and the presence or absence of complement deposition in the lungs or microvasculature could, thus not be assessed. Taken together, however, these findings suggest that complement might be targeted for specific intervention. In China, two deteriorating patients were rescued using an anti-C5a monoclonal antibody.6 In Italy, four patients with severe pneumoniae successfully recovered after treatment with eculizumab.8 In the absence of proven effective therapy, we decided to treat COVID-19 patients with severe pneumonia with eculizumab on an emergency compassionate-use basis. Five patients with severe pneumonia requiring â&#x2030;Ľ 5L/min of oxygen to maintain SpO2 >97% (but not requiring ICU) and three patients with respiratory failure requiring mechanical ventilation and suffering from renal injury (defined by AKI â&#x2030;Ľ2 or requiring dialysis) and vasopressive drugs support thus received eculizumab off label on a compassionate-use basis. All patients had confirmed severe COVID19 using specific RT-PCR (positive PCR on nasal swabs). Characteristics of the patients are detailed in Table 1. This report is based on data from patients who received

Figure 1. The level of C5b9 according to clinical severity. Plasma levels of sC5b-9 in healthy controls (n=68) and in patients with moderate (n=17), severe (n=18), critical COVID-19 (n=60), as well as patients with COVID-19 sampled at least 2 months after hospital discharge (n=40) are represented. Patients with COVID-19 were sampled during hospitalization up to 5 days after the admission. up to 5 days after the admission. The normal values of sC5b-9 are below 300 ng/mL. The plasma level of sC5b-9 was increased in 41 % (7 of 17), 50 % (9 of 18) and 68 % (41 of 60) of the patients with moderate, severe and critical disease. The median plasma levels of sC5b9 (Q1- Q3) in the patients with moderate, severe and critical disease were 281 ng/mL (range: 168-348), 314 ng/mL (range: 235-501) and 367 ng/mL (range: 262-467) respectively. The plasma levels of sC5b9 returned back to the normal range for the patients sampled 2 months after their discharge from hospital.

haematologica | 2020; 105(12)

2847

Letters to the Editor

Table 1. Patients baseline characteristics, at time of Eculizumab initiation and during treatment.

Patient#1

Patient#3

Patient#4

Patient#5

Patient#6

Patient#7

Patient#8

Age, years Sex Coexisting conditions

65 58 31 M F M Multiple 7 months High Myeloma post Blood Amyloidosis AlloBMT (AML) Pressure High Blood Takayasu Pressure disease Obesity Hypothyroidism

73 M High Blood Pressure

58 M High Blood Pressure Diabetes Obesity

56 F High Blood Pressure Diabetes Obesity

64 F High Blood Pressure Diabetes Obesity Asthma

47 M None

Symptoms

Fever, cough, dysgeusia, anosmia 5

Fever, dyspnea

Time from first symptoms (days) Oxygen supplementation to maintain SpO2 ≥97% (L/mn) Creatinine clearance (mL/mn) C-reactive protein (mg/L)

Patient#2

Baseline characteristics at admission Fever Dyspnea, Fever, Fever, cough, throat pain cough, cough dyspnea, dyspnea nausea, vomiting 5 3 12 5

24 96

49 195

115 90

79 187

On dialysis 137

20 221

40 NA

109 95

Time from first symptoms 6 to Eculizumab first injection (days) Oxygen supplementation 5 to maintain SpO2 ≥97% (L/minute) Additional treatments Enoxaparin 60 mg/day -

Characteristics at time of Eculizumab initiation 6 3 13 8 10 6 6 6 Mechanical ventilation Enoxaparin Enoxaparin Enoxaparin Unfractioned 40 mg/day 40 mg/day 60 mg/day heparin Dexamethasone Dexamethasone

Evolution under Eculizumab treatment sC5b9 (ng/mL) 593 315 492 550 435 (at time Eculizumab first injection) sC5b9 (ng/mL) (at day +1 post 182 147 182 422 169 Eculizumab first injection)* sC5b9 (ng/mL) at time of last 388 329 390 305 415 follow-up (days from the (27) (46) (12) (41) (9) first Eculizumab injection) CH50 and trough eculizumab 0 (49)/0 0 (30)/ 22 (116)/ 36(174)/ 21 (118)/ concentration at day 1 and day 4 (271) 0 (<24) 28 (48) 41 (65) 99 (<24) after the first Ecu injection Time from first injection to 39 33 13 13 NA ambient air of low flow supplemental oxygen (≤2 L/minute) (days) Therapeutic schedule 900 mg 900 mg 900 mg 1,200 mg 900 mg and total number of every 4 days every 4 days every 4 days every 4 days every 7 days Eculizumab injection (4) (5) (3) (3) (2) Status at last follow-up, Alive Alive Alive Alive Death location and length of stay discharged discharged Home Home (MOF) (13) (14) (13) in the hospital (days) from hospital from hospital# (40) (34)

Mechanical ventilation Enoxaparin 60 mg/day Dexamethasone

Mechanical High ventilation flow Enoxaparin Enoxaparin 60 mg/day 40 mg/day Dexamethasone -

354

505

393

223

429

271 (4)

252 (48)

290 (6)

22 (208)/ 26 (70)

18 (163)/ 0 (88)

NA (215)

1,200 mg every 4 days (1) Death (PE) (6)

1,200 mg 1,200 mg every 4 days every 4 days (4) (2) Alive Alive discharged Home from (5) hospital (25)

Thrombosis history: none of the eight patients have previously presented thrombosis prior to the episode of COVID; patients 1 was diagnosed 10 days after Eculizumab injection with deep vein thrombosis and pulmonary embolism; the evolution was favorable under Eculizumab continuation and anticoagulation; patients 6 died 4 days after Eculizumab first injection (massive PE). Complement pathway activity monitoring: CH50 (screening hemolytic assay using sheep Erythocytes) is routinely used to monitor patients under Eculizumab; at day 1, all patients showed drastic diminution of CH50 below 20% of the normal value. AML: acute myeloid leukemia; alloBMT: allogeneic bone marrow transplantation; MOF: multi organ failure; NA: not available; PE: pulmonary Embolism. #BMT related bronchiolitis flaired 15 days after Eculizumab injection with a concomitant diagnosis of Parainfluenzae Virus infection; Dexamethasone was given intravenously at dose of 20 mg once daily from day 1 to day 5, and then 10 mg once daily from day 6 to day 10. *At day +1 after injection, seven patients with available samples showed a decreased of circulating levels of sC5b9. The median sC5b9 levels pre (463 ng/mL) and post (317 mg/mL) Eculizumab was significantly different (P=0.01) using non-parametric statistical analysis (Wilcoxon test). However, no correlation between the percentage reduction and response to Eculizumab was found.

2848

haematologica | 2020; 105(12)

Letters to the Editor

eculizumab during the period from March 17, 2020 through May 19, 2020. The first ICU patient (patient 1) was treated according to SOLIRIS® SmPC – dosing regimen of atypical Hemolytic and Uremic Syndrome (aHUS, induction period with 900 mg every week). He continues to be monitored closely with regard to the complement activity during follow-up according to our usual practice.9 The level of free eculizumab in the plasma was assessed using in house ELISA as previously described.9 At day 7, we observed a lack of complete inhibition of C5 with normal CH50 activity and undetectable free eculizumab circulating levels in patient 1. These findings suggest a much higher clearance of eculizumab in patients with COVID19 than usually observed after a single injection in other diseases like aHUS and paroxysmal nocturnal hemoglobinuria.9 Patients 2, 3 and 5 thus received 900 mg every 4 days allowing better but not optimal and prolonged complement blockade. We observed low eculizumab levels at day 4 (below 50 mg/mL in 2 of the 3 patients) with an efficient complement inhibition since day 1. The next four patients thus received three induction doses of 1,200 mg at day 1, 4 and 8, which appears satisfactory on a PK/PD standpoint (CH50 blocked before reinjection and residual free eculizumab upper to 50 mg/mL in all evaluable patients). Because of the complement blockade the patients received prophylactic antibiotics against meningococcal infection prior to initiating eculizumab treatment and were vaccinated when possible. At time of eculizumab initiation, three patients were intubated, one received high flow oxygen and four patients were treated with standard oxygen support only. They all had elevated sC5b-9 circulating levels. Over a median follow-up of 19.5 days (range: 13-31 days) after receiving the first dose of eculizumab, six patients showed an improvement in the category of oxygen support, including one patient receiving mechanical ventilation who was subsequently extubated. At last follow-up, six patients had been discharged (day +5, +13, +13, +23, +34, +40 days after first eculizumab injection, respectively). Two patients who received invasive ventilation died (day +4 and +10 after the first eculizumab injection, respectively). Patient 5 presented a septic shock and multi organ failure while patient 6 was diagnosed with massive pulmonary embolism and cardiac arrest. In addition, patient 1 also presented severe thrombotic complications during evolution (deep venous thrombosis and pulmonary embolism). Those latter observations are in line with the recent findings suggesting that up to 30% of patients with severe COVID-19 infection develop life-threatening thrombotic complications.10 Those considerations reinforce the recommendation to strictly apply thrombosis prophylaxis in COVID19 patients,11 also in the context of complement inhibition. While the relationship between thrombosis occurrence and complement activation is still unclear in patients with COVID 19, evidence is emerging that innate immunity contributes to the inflammatory storm that leads to respiratory failure in many patients with COVID-19. In particular, neutrophils-derived neutrophil extracellular traps (NET) which are extracellular webs of DNA, histones, microbicidal proteins, and oxidant enzymes that are released by neutrophils to corral infections have the potential to propagate inflammation and thrombosis if not properly regulated12,13 as is the case in patients with COVID 19.14 Excessive inflammation together with massive complement activation might thus participate in the predisposition of patients to a higher thrombotic risk. haematologica | 2020; 105(12)

Overall, we showed that the terminal pathway of the complement is overactivated in 64% of COVID-19 patients on admission to hospital. This may contribute to the severity of the disease. Naturally, it is difficult to draw any robust conclusion on the efficacy of complement blocking, firstly because of the small number and heterogeneity of our patients, and secondly in the absence of a control group. However, all of our eight patients were particularly severe at the time of eculizumab initiation and six improved significantly. Even if further investigation is needed, our results suggest that the inhibition of one mechanism of COVID-induced organ damage may be an add-on treatment for this condition. Our experience also highlights that eculizumab pharmacokinetics in COVID patients differ from reports in other complement-mediated diseases. In this particular disease, higher eculizumab doses and/or shorter intervals to ensure an efficient and sustained blockade seem required. The degree of complement activation at the start of therapy, possibly including C5b9 deposition in the tissue may be a significant determinant of eculizumab clearance and disease response. Controlled trials are timely to confirm the value of eculizumab in patients with COVID-19. Regis Peffault de Latour,1,2 Anne Bergeron,2,3,4 Etienne Lengline,5,6 Thibault Dupont,6 Armance Marchal,7 Lionel Galicier,8 Nathalie de Castro,9 Louise Bondeelle,3 Michael Darmon,2,6 Clairelyne Dupin,3 Guillaume Dumas,6 Pierre Leguen,3 Isabelle Madelaine,3 Sylvie Chevret,2,4,10 Jean-Michel Molina,2,9 Elie Azoulay2,6 and Veronique Fremeaux-Bacchi7,11,12 on behalf of the CORE group 1 Bone Marrow Transplantation (BMT) Unit, Saint-Louis Hospital, Paris; 2Université de Paris, Paris; 3Pneumology Unit, Saint-Louis Hospital, Assistance Publique Hôpitaux de Paris (APHP), Paris; 4UMR 1153 CRESS, Biostatistics and Clinical Epidemiology Research Team, Paris; 5Hematology Unit, SaintLouis Hospital, AP-HP, Paris; 6Intensive Care Unit, Saint-Louis Hospital, AP-HP, Paris; 7Laboratory of Immunology, GeorgesPompidou European Hospital, AP-HP, Paris; 8Clinical Immunology Unit, Saint-Louis Hospital, AP-HP, Paris; 9Infectious Disease Unit, Saint-Louis Hospital, Paris; 10Biostatistic Unit, Saint Louis Hospital, AP-HP, Paris; 11Immunology, HEGP Hospital, AP-HP, Paris and 12Centre de Recherche des Cordeliers and Pasteur Institute, Paris, France The CORE group is detailed in Online Supplementary Appendix 1. Correspondence: REGIS PEFFAULT DE LA TOUR regis.peffaultdelatour@aphp.fr doi:10.3324/haematol.2020.260117 Acknowledgments: the authors would like to thank the colleagues who actively participated in the study and the laboratory team in HEGP who performed the complement analysis. We would particularly like to thank the patients for kindly accepting to take part in this research.

References 1. Cucinotta D, Vanelli M. WHO declares COVID-19 a pandemic. Acta Biomed. 2020;91(1):157-160. 2. Risitano AM, Mastellos DC, Hubert-Lang M, et al. Complement as a target in COVID-19? Nat Immunol. 2020;20(6):343-344. 3. Gralinski LE, Sheahan TP, Morrison TE, et al. Complement activation contributes to severe acute respiratory syndrome Coronavirus pathogenesis. mBio. 2018;9(5):e01753- 18. 4. Jiang Y, Zhao G, Song N, et al. Blockade of the C5a-C5aR axis alleviates lung damage in hDPP4-transgenic mice infected with

2849

Letters to the Editor

MERS-CoV. Emerg. Microbes Infect. 2018;7(1):77. 5. Magro C, Mulvey JJ, Berlin D et al. Complement associated microvascular injury and thrombosis in the pathogenesis of severe COVID-19 infection: A report of five cases. Transl Res. 2020;220:1-13. 6. Gao T, Hu M, Zhang X, et al. Highly pathogenic coronavirus N protein aggravates lung injury by MASP-2- mediated complement over-activation. Medrxiv. 2020. 2020.03.29.20041962 7. Kabat EA, Mayer MM. Experimental immunochemistry. Vol. 4 (ed Second): Springfield, Illinois; 1961. 8. Diurno F, Numis FG, Porta G, et al. Eculizumab treatment in patients with COVID-19: preliminary results from real life ASL Napoli 2 Nord experience. Eur Rev Med Pharmacol Sci. 2020;24(7):4040-4047. 9. Peffault de Latour R, Fremeaux-Bacchi V, Porcher R, et al.

2850

10.

11. 12.

13. 14.

Assessing complement blockade in patients with paroxysmal nocturnal hemoglobinuria receiving eculizumab. Blood. 2015;125(5): 775-783. Klok FA, Kruip MJHA, van der Meer NJM, et al. Incidence of thrombotic complications in critically ill ICU patients with COVID-19. Thromb Res. 2020;191:145-147. Connors JM and Levy JH. COVID-19 and its implications for thrombosis and anticoagulation. Blood. 2020;135(23):2033-2040. Twaddell SH, Baines KJ, Grainge C, et al. The emerging role of neutrophil extracellular traps in respiratory disease. Chest. 2019;156(4):774-782. Porto BN, Stein RT. Neutrophil extracellular traps in pulmonary diseases: too much of a good thing? Front Immunol. 2016;7:311. Zuo Y, Yalavarthi S, Shi H, et al. Neutrophil extracellular traps in COVID-19. JCI Insight. 2020; 5(11):e138999.

haematologica | 2020; 105(12)

Letters to the Editor

The proteome of neutrophils in sickle cell disease reveals an unexpected activation of the interferon alpha signaling pathway Polymorphonuclear neutrophils (PMN) are key actors in the pathophysiology of sickle cell disease (SCD), but signaling pathways underlying their activation and sustained inflammation are not well documented. We thus investigated the protein profile of neutrophils from SCD patients (SS genotype) using a proteomic approach. Unexpectedly, SCD neutrophils exhibit a high expression of interferon signaling proteins (ISP) belonging to the type 1 interferon (IFN-1) response pathway. We also showed that SCD patients at steady state displayed a higher level of plasmatic IFNα. Overall, we reported a dramatic high-level expression of ISP in neutrophils from SS patients suggesting an abnormal activation that could be important in developing new anti-inflammatory therapies. SCD is a hemoglobinopathy leading to major red blood cell (RBC) dysfunction, but other cell types (vascular endothelium, leukocytes, platelets)1-3 also represent key

actors in the pathophysiology of the disease. Important studies have highlighted the role of PMN, both during the vaso-occlusive crisis (VOC) and the associated long-term morbidity and mortality.4 In SCD, patients have an increased leukocyte count at steady state, and exhibit neutrophil activation, rendering them more susceptible to inflammatory stimuli.5 Moreover recent data have demonstrated the presence of different sub-phenotypes of PMN especially in cancer and inflammation6,7 as well as in a preclinical model of SCD.8 Despite these advances, signaling pathways underlying sustained inflammation in SCD remain elusive. In addition, a fine understanding of PMN activation profile is necessary to better decipher the inflammatory paradigm in SCD and develop tailored therapies. In the present study, we investigated for the first time the proteomic profile of PMN in SCD at basal state by a label-free global proteomic approach. We performed a proteomic comparative study of purified neutrophils from four SS patients (SS1-4) at basal state and four AA blood type healthy donors (AA1-4). All patients included in this study were homozygous (SS genotype), aged 2 to 18 years (mean age 9.7 years), and

Figure 1. Proteomic analysis of neutrophils from healthy donors (AA) and SS genotype patients (SS) at steady state. (A) Overall proteomics results: number of proteins identified and quantified in at least one sample (A) or in at least three samples and in at least one group (B). (B) 2D enrichment analysis of the proteins expressed in at least 75% of the samples and in at least one group. Protein annotation databases from GOBP. Annotations out of the diagonal corresponding to differential expression in SS or AA samples are indicated. (C) Cluster analysis of proteins differentially expressed in SS and AA neutrophils. Proteins with significantly different expression values (P<0.05 and fold change>0.3) were selected, their LFQ values were z score transformed and analyzed by Euclidian clustering. Clusters 1 and 2 correspond to proteins upregulated and downregulated in SS neutrophils, respectively. (D) Cluster analysis of differentially expressed proteins with the “cellular response to type I interferon” GOBP annotation. (E) Characterization of neutrophils from SS patients by flow cytometry analysis, typical result: over-expression of the Fc fragment of IgG/CD64 and under-expression of L-selectin/CD62L. Data are presented as mean ± standard devaition (SD). Mann-Whitney test was used to compare SS and AA; *P<0.05 compared with AA; **P<0.01 compared with AA.

haematologica | 2020; 105(12)

2851

Letters to the Editor

IFNb (pg/mL)

IFNα (fg/mL)

Nucleosome (ng/mL)

PMN elastase (ng/mL)

Figure 2 Activation of IFN-1 pathway in SS gentype neutrophils. (A) Representative images of Western blots for MX1, IFIT-1, ISG15, STAT 1, Phospho-STAT 1, STAT 2 and Phospho-STAT 2 expressed in neutrophils from healthy volunteers (AA, n=10) and SS patients (SS, n=10). The Western blots are represented as the ratio of the density of the specific band on the total protein in each sample. (B) Level of IFNα and IFNb in plasma from healthy donors (AA, n=34) and SS patients at basal state (SS, n=37) using a digital-ELISA assay (SIMOA). (C) Level of neutrophil elastase and nucleosome in plasma from healthy donors (AA, n=12) and SS patients at basal state (SS, n=28) using ELISA assay. Data are presented as mean ± standard devaitaion (SD). Mann-Whitney test was used to compare SS and AA; *P<0.05 compared with AA; **P<0.01 compared with AA; ***P<0.001.

without any associated co-morbidity. They were free from any infections and exhibited C-reactive protein level < 10 mg/L at the day of the inclusion. The controls were voluntary blood donors, all healthy and of AA genotype. After mass spectrometry analysis, 4,634 proteins were identified and 4,487 of them could be reliably quantified. Restricting the analysis to proteins quantified in at least 75% of the samples and in at least one group (AA and/or SS) led to the comparison of 3,069 proteins (Figure 1A). To identify biological pathways modified in neutrophils from SS patients, we first performed a 2D annotation enrichment test9 using GO, KEGG and keywords annotation databases. This analysis revealed the presence of many neutrophil membrane and secreted proteins involved in immune response (Figure 1B). Next, we analyzed the SS and the AA proteomes and found 101 pro2852

teins significantly differentially expressed using a SS/AA ratio >1.3 or <0.7. Sixty-eight proteins were overexpressed (cluster 1, Figure 1C), and 33 proteins were down-regulated in the SS group compared to the AA group (cluster 2, Figure 1C). Prior to further investigating a new biological pathway, we aimed to confirm the already known surface markers of sickle cell neutrophils. Indeed, several studies in mouse models or in patients have shown an activated and aged phenotype of PMN in SCD.8,10 By using our proteomics data, we found two proteins described as markers of activation (Fc fragment of IgG, high affinity Ia, receptor/CD64) and ageing (L-Selectin/CD62L) of neutrophils respectively overexpressed (5.8-fold) and underexpressed (0.6-fold) in SS patients compared to the AA group (Online Supplementary Table S1 and Online Supplementary Proteomic File). These data were confirmed by cytometry haematologica | 2020; 105(12)

Letters to the Editor

analysis of freshly isolated neutrophils from five SS patients and five controls (Figure 1E). We therefore concluded that the proteomic analysis could be a relevant tool for exploring neutrophil abnormalities in SCD. A Fisher exact test performed using proteins down-regulated in SS neutrophils did not evidence any common biological pathway (data not shown). In contrast, analyses of the upregulated proteins revealed a major involvement of the type1 interferon (IFN-I) response (Online Supplementary Table S1 and Online Supplementary Proteomic File). Importantly, major ISP including IFIT1, IFIT2, IFIT3, ISG15, ISG20, GBP2, IFI35, MX1 and MX2 were increased 3- to 84-fold in the proteome of SS neutrophils compared to the one of AA neutrophils (Figure 1D). Moreover, we found a significant overexpression of STAT1 and STAT2 in the neutrophil proteome of the SS group consistent with an activation of the IFN-related JAK/STAT signaling pathways. In order to confirm the proteomic data, we assessed the overexpression of the main ISP using Western blotting experiments of purified neutrophils from 10 other SS patients at steady state and 10 other AA controls. In agreement with the proteomic data, we found a significant increase of ISP expression including MX1, ISG15 and IFIT1 as well as the STAT1 and STAT2 proteins, in neutrophil lysates from SS patients compared to controls (Figure 2A). The nuclear translocation of STAT1 and STAT2 are activated by JAK and TYK2-mediated phosphorylation of the Y701 and Y689, respectively, that stimulates the IFN-1 responses.11,12 We showed that both Y701 of STAT1 and Y689 of STAT2 were highly phosphorylated in SS compared to AA neutrophils (Figure 2A). These findings confirmed the strong activation of the IFN-1 signaling pathway in SS neutrophils via the JAK/STAT1/2 pathway. In order to investigate further and to assess whether the IFN-I response was due to either IFNα or IFNb, we measured the level of both cytokines in the plasma of 34 healthy AA donors and 37 SS patients at steady state, using the novel digital-ELISA technology. Interestingly, we found a significantly increased level of IFNα in the plasma from half of the SS patients compared to AA controls (Mann-Whitney test, P<0.001), even though no difference was observed for IFNb (Figure 2B). Although the specific role of the different types of IFN-1 is not fully understood, it appears that IFNα, in contrast to IFNb is mainly involved in autoimmune and auto-inflammatory diseases.13 It is noteworthy that 20 SS patients exhibited an increase of IFNα from 10- to 1,000-fold compared to healthy individuals although 17 of the 37 plasma samples had normal levels of IFNα. Clinical and biological investigations of these 37 SS patients did not show any correlation between the plasmatic level of IFNα and biologic markers including leucocyte, neutrophils, reticulocytes and platelets counts, hemoglobin (Hb) level, Hb haplotypes or age (Online Supplementary Table S2), and plasmatic cytokine concentration (including CX3CL-1, Rantes, MCP-1, MCP-3, TNFα, IL1b, IL10, IL18, and IL6). Since it is known that neutrophil extracellular trap formation (NETosis) plays a role in the pathogenesis of SCD, we next investigated the NETosis by measuring the neutrophil elastase and nucleosome, in the plasma from 28 patients investigated for IFNα and IFNb. As expected, we found a significant high level of both markers of NETosis (Figure 2C) in SS patients compared to AA controls, but no correlation with the IFNα level (data not shown). Finally, we analyzed the clinical data from the patients, and found no significant difference between haematologica | 2020; 105(12)

the “high IFN” and “low IFN” group of patients and the number of acute events (including number of VOC per year, acute chest syndrome, stroke, cerebral vasculopathy, acute splenic sequestration nor splenectomy). It is noteworthy that no patient has been treated with hydroxycarbamide and none of them has followed a transfusion program. Moreover, it is interesting to note that of the four SS plasma samples used for the neutrophil proteomic analysis one had low IFNα level, while the other three exhibited 7- to 60-fold increased levels compared to controls, although all four neutrophil samples expressed high level of ISP. Therefore, it is highly probable that plasma IFNα has a transient secretion while the downstream activation of the signaling pathway is persistent. Altogether, our data indicated that SS patients may have inappropriate transient high IFNα secretions (i.e., outside of any acute and infectious events), responsible for the activation of the IFN-1 signaling pathway in neutrophils. Although the mechanism of this activation remains to be elucidated, some recent data described a clear relationship between INF-1 responses and red blood cell alloimmunization in murine models.14,15 Since alloimmunization represents a detrimental issue in SCD, our data highlight the importance of testing the link between ISP and alloimmunization in SS patients. In conclusion, we showed for the first time by quantitative proteomic analyses of purified neutrophils a particular immune and inflammatory signature in SCD. Our findings provide evidence of a dysfunction of the IFNα signaling pathway that could play an important role in the pathogenesis of SCD. Future studies using a cohort of patients are needed to determine the relationship between IFNα activation and clinical complications and to establish if ISP may represent therapeutic targets to decrease inflammation in SCD. Patricia Hermand,1,2 Slim Azouzi,1,2 Emilie-Fleur Gautier,2,3 François Guillonneau,2,3 Vincent Bondet,4 Darragh Duffy,4 Sebastien Dechavanne,1,2 Pierre-Louis Tharaux,2,5 Patrick Mayeux,2,3 Caroline Le Van Kim1,2# and Berengere Koehl1,2,6# 1 Université de Paris, UMR_S1134, BIGR, Inserm, Institut National de la Transfusion Sanguine; 2Laboratoire d’Excellence GR-Ex; 3Université de Paris, UMR_S1016, UMR 8104, Plateforme de Protéomique (3P5), Institut Cochin, Inserm, CNRS; 4 Immunobiology of Dendritic Cells, Institut Pasteur, Inserm UMR 1223; 5Université de Paris, Paris Cardiovascular Centre, PARCC, INSERM and 6Sickle Cell Disease Center, Hematology Unit, Hôpital Robert Debré, Assistance Publique – Hôpitaux de Paris, Paris, France #

CLVK and BK contributed equally as co-senior authors.

Correspondence: BERENGERE KOEHL - berengere.koehl@inserm.fr CAROLINE LE VAN KIM - caroline.le-van-kim@inserm.fr doi:10.3324/haematol.2019.238295 Acknowledgments: we thank Dr Marie-Helène Odièvre for her contribution to patient’s recruitment. We are indebted to Wassim El Nemer for helpful comments and for reading the manuscript. Funding: We thank the patients and their families for their participation in the study and all members of the Sickle Cell Disease Center from the Robert Debré Hospital for the management of blood samples. This work was supported by a grant from l’Association Recherche et Transfusion, the Institut National de la Transfusion Sanguine, and the Laboratory of Excellence GR-Ex, reference ANR-11-LABX-0051; GR-Ex is funded by the program “Investissements d’avenir” of the French National Research Agency, reference ANR-11-IDEX-0005-02. 2853

Letters to the Editor

The Orbitrap Fusion mass spectrometer was acquired with funds from the FEDER through the "Operational Programme for Competitiveness and Employment 2007-2013" and from the "Canceropole Ile de France".

References 1. Kaul DK, Finnegan E, Barabino GA. Sickle red cell-endothelium interactions. Microcirculation. 2009;16(1):97-111. 2. Proenca-Ferreira R, Brugnerotto AF, Garrido VT, et al. Endothelial activation by platelets from sickle cell anemia patients. PLoS One. 2014;9(2):e89012. 3. Koehl B, Nivoit P, El Nemer W, et al. The endothelin B receptor plays a crucial role in the adhesion of neutrophils to the endothelium in sickle cell disease. Haematologica. 2017;102(7):1161-1172. 4. Platt OS, Brambilla DJ, Rosse WF, et al. Mortality in sickle cell disease. Life expectancy and risk factors for early death. N Engl J Med. 1994;330(23):1639-1644. 5. Lum AF, Wun T, Staunton D, Simon SI. Inflammatory potential of neutrophils detected in sickle cell disease. Am J Hematol. 2004;76(2):126-133. 6. Yang P, Li Y, Xie Y, Liu Y. Different faces for different places: heterogeneity of neutrophil phenotype and function. J Immunol Res. 2019;2019:8016254. 7. Wang X, Qiu L, Li Z, Wang XY, Yi H. Understanding the multifaceted role of neutrophils in cancer and autoimmune diseases. Front Immunol. 2018;9:2456.

2854

8. Zhang D, Xu C, Manwani D, Frenette PS. Neutrophils, platelets, and inflammatory pathways at the nexus of sickle cell disease pathophysiology. Blood. 2016;127(7):801-809. 9. Geiger T, Wehner A, Schaab C, Cox J, Mann M. Comparative proteomic analysis of eleven common cell lines reveals ubiquitous but varying expression of most proteins. Mol Cell Proteomics. 2012;11(3):M111.014050. 10. Fadlon E, Vordermeier S, Pearson TC, et al. Blood polymorphonuclear leukocytes from the majority of sickle cell patients in the crisis phase of the disease show enhanced adhesion to vascular endothelium and increased expression of CD64. Blood. 1998;91(1):266-274. 11 . Bancerek J, Poss ZC, Steinparzer I, et al. CDK8 kinase phosphorylates transcription factor STAT1 to selectively regulate the interferon response. Immunity. 2013;38(2):250-262. 12. Wiesauer I, Gaumannmuller C, Steinparzer I, Strobl B, Kovarik P. Promoter occupancy of STAT1 in interferon responses is regulated by processive transcription. Mol Cell Biol. 2015;35(4):716-727. 13. Crow MK, Ronnblom L. Type I interferons in host defence and inflammatory diseases. Lupus Sci Med. 2019;6(1):e000336. 14. Liu D, Gibb DR, Escamilla-Rivera V, et al. Type 1 IFN signaling critically regulates influenza-induced alloimmunization to transfused KEL RBCs in a murine model. Transfusion. 2019;59(10):3243-3252. 15. Gibb DR, Liu J, Natarajan P, et al. Type I IFN is necessary and sufficient for inflammation-induced red blood cell alloimmunization in mice. J Immunol. 2017;199(3):1041-1050.

haematologica | 2020; 105(12)

Letters to the Editor

Plasticity in growth behavior of patients’ acute myeloid leukemia stem cells growing in mice Resistance against chemotherapy remains a major obstacle in treating patients with acute myeloid leukemia (AML).1 Novel therapeutic concepts are especially desired to target and eliminate resistant AML stem cells. Here we show that AML stem cells harbor plasticity, a changing pattern of biological behavior, by demonstrating that AML stem cells reversibly switch from a low-cycling, chemotherapy resistant state into an actively proliferating state associated with a response to standard chemotherapy. We used patient-derived xenograft (PDX) cells from patients with high risk or relapsed AML that were lentivirally transduced for marker expression. We stained these cells with the proliferation-sensitive dye carboxyfluorescein succinimidyl ester (CFSE), and re-transplanted them into next-recipient mice. A rare subpopulation of AML cells displayed reduced proliferation in vivo, associated with resistance against standard chemotherapy. The proportion of AML cells with stem cell potential was identical in both, the high and low proliferative sub-fractions. In re-transplantation experiments, proliferation behavior proved reversible, and AML stem cells were able to switch between a high and low proliferation state. Our data indicate that AML stem cells display functional plasticity in vivo, which might be exploited for therapeutic purposes, to prevent AML relapse and ultimately improve the prognosis of patients with AML. AML patients are at risk to suffer disease relapse associated with dismal prognosis. The rare subpopulation of AML stem cells (or leukemia initiating cells [LIC]) might be responsible for relapse by combining self-renewal capacity with dormancy and resistance against standard chemotherapy.2 AML LIC features, including the growth phenotype, have long been considered mainly constant and persistent;2-5 in contrast, recent data suggest unsteady features under therapeutic pressure,6,7 while data without experimental treatment pressure remain elusive. Putative functional plasticity of AML LIC is of major clinical importance as it might enable novel therapeutic options. We previously reported functional plasticity in acute lymphoblastic leukemia (ALL), where we showed in vivo that long-term dormant, treatment-resistant ALL cells were able to convert into highly proliferative, treatmentsensitive cells and vice versa.8 Nevertheless, AML and ALL differ widely regarding stem cell biology and a defined stem cell hierarchy - characteristic for AML – has not been proven in ALL. Based on diverse stem cell characteristics, we considered the functional plasticity of LIC conceivable in ALL, but hypothesized its absence in AML. To test our hypothesis, we studied cells from ten patients with high-risk or relapsed AML of different karyotypes, genotypes and clinical histories (Online Supplementary Table S1). As a clinic-close model system, primary cells were transplanted into immunocompromised mice, and AML PDX models were established.9 PDX models were selected to allow for serial transplantation; as this ability is restricted to highly aggressive disease, our study is biased towards high risk AML. AML PDX models were genetically engineered to express luciferase for bioluminescence in vivo imaging and mCherry for cell enrichment by flow cytometry. Marker expression remained stable over serial re-transplantation and allowed enrichment of minute numbers of PDX AML cells from murine bone marrow (Figure 1A, for details see the Online Supplementary Materials and Methods). As conhaematologica | 2020; 105(12)

trols, three samples (AML-356, AML-358 and AML-538) were studied without prior genetic engineering. AML PDX samples showed more than 3-fold differences in doubling times in vivo, resulting in variable time to overt disease in mice (Online Supplementary Figure S1AB). When PDX cells were re-isolated from murine bone marrow, mCherry expression enabled unbiased enrichment of AML PDX cells, independent of other, putatively subpopulation-restricted, surface markers on AML cells (Figure 1B).8,10 Re-isolation of PDX cells revealed that homing was heterogeneous between samples, as 0.011% of PDX cells could be re-isolated from mice early after transplantation (Online Supplementary Figure S1C). The frequency of LIC, as determined in limiting dilution transplantation assays, varied by a factor of 10 between samples (Online Supplementary Figure S1D and Online Supplementary Table S2). Thus, our AML PDX cohort of aggressive samples displayed major functional inter-sample heterogeneity in vivo, reflecting the known phenotypic heterogeneity of AML.11 In order to track in vivo proliferation of AML cells from individual samples, PDX cells were stained with CFSE, a dye that is not metabolized in eukaryotic cells, but decreases upon cell divisions, indicating proliferation.12 CFSE records a cell’s proliferative history rather than providing a snapshot of the cell´s proliferative state at a given moment. CFSE content was measured by flow cytometry at different time points following injection into groups of mice. In accordance with an increase in leukemic burden and numbers of re-isolated cells (Figures 1CD and Online Supplementary Figure S2A), most AML PDX cells entirely lost CFSE within days of in vivo growth, indicating high proliferative activity in the majority of cells (Figures 1E-F and Online Supplementary Figure S2A). However, a minor subpopulation of cells retained CFSE over several weeks, indicating a low-cycling, putatively dormant phenotype (Figures 1E-F and Online Supplementary Figure S2A). We called these cells label-retaining cells (LRC) according to the literature.8 LRC were found in 9 of 10 samples tested (Figures 1E-F and Online Supplementary Figure S2A-B). Only a single sample originating from a child with a fatal AML relapse had entirely lost the LRC population between day 7 to 15 (Online Supplementary Figure S2C), again highlighting the known heterogeneity of AML.11 Cell cycle analysis confirmed that LRC divide less compared to non-LRC (Online Supplementary Figure S3A). Together, our data reveal, in the majority of cases, heterogeneity of in vivo growth behavior within individual AML PDX samples, including a subpopulation of low-cycling LRC. Hence, our results add an important level of phenotypic heterogeneity to AML on top of the known heterogeneity of e.g., immunophenotypes, or gene expression profiles. As a large range of AML subtypes were studied (Online Supplementary Table S1), the novel characteristic is not limited to a specific cytogenetic or genetic subgroup. To further characterize attributes of LRC, gene expression analysis of 24 LRC and non-LRC samples isolated from AML-393 and AML-491 was performed.13 Among the top down regulated gene sets in LRC were cell cycle regulators, confirming the reduced proliferative state of these cells (Online Supplementary Figure S3B-C); among the top upregulated gene sets were cell adhesion molecules (Online Supplementary Figure S3C). Notably, LRC of AML393 were more similar to LRC of AML-491 than to their own non-LRC (Figure 1G), despite the substantial differences in the mutational profile of AML-393 and AML-491 (Online Supplementary Table S1). Even more striking, geneset enrichment analysis identified a high gene-set enrich2855

Letters to the Editor

NES=1.88; NOM P-val<0.0001; FDR <0.0001

Figure 1. Acute myeloid leukemia patient-derived xenograft cells contain a rare subpopulation of low-cycling cells. (A) Experimental procedure; primary patients’ acute myeloid leukemia (AML) cells were transplanted into NSG mice, resulting patient-derived xenograft (PDX) cells were genetically engineered, sorted, and amplified. At advanced disease stage, mCherry+ AML PDX cells were isolated, stained with carboxyfluorescein succinimidyl ester (CFSE), and re-transplanted. At different time points, AML cells were re-isolated from mouse bone marrow, enriched, and CFSE content measured by flow cytometry, to detect CFSEpositive, low-cycling label-retaining cells (LRC), and CFSE-negative, proliferating non-LRC (nLRC). NSG: NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ; EF1α: elongation factor 1-α promoter; Luc: enhanced firefly luciferase. (B) Gating strategy: bone marrow cells depleted of murine cells by MACS were gated on (i) leukocytes, (ii) DAPImCherry+ AML PDX cells, and (iii) separated into LRC and non-LRC according to their CFSE content. Maximum CFSE mean fluorescence intensity (MFI) was measured at day 2 after cell injection or in in vitro cultivation, and divided by factor 2 to model cell divisions (dotted lines); upon less than three divisions, cells were considered as low-cycling LRC, upon more than seven divisions as proliferating non-LRC; days indicate time after cell injection. (C-D) Growth of AML-393 cells monitored by in vivo imaging (C) or by quantifying PDX cells re-isolated from mouse bone marrow using flow cytometry (n=21) (D); each square represents data from one mouse. (E-F) A rare subpopulation of AML PDX cells retains CFSE upon prolonged in vivo growth. AML-393 cells from different time points in (D) were analyzed by flow cytometry for CFSE using the gating strategy described in (B); representative dot plots (E) and percentage of LRC cells among all isolated PDX cells are shown (F); each square represents data of one mouse. (G-H) Gene expression analysis of LRC and non-LRC. LRC and non-LRC were isolated from mice carrying AML-393 (n=4) or AML-491 (n=4) 10 or 14 days after cell injection, respectively and subjected to RNA sequencing. Technical replicates were analyzed in 6 of 8 samples, resulting in a total of 24 samples analyzed. (G) Heatmap of top differentially regulated genes (false discovery rate [FDR] ≤0.05) between LRC (green) and nLRC (black) of AML-393 and AML-491. (H) LRC of AML-393 and AML-491 show significant overlap with the previously published LRC signature of acute lymphoblastic leukemia (ALL). See the Online Supplementary Figure S1-S2 for additional data.

2856

haematologica | 2020; 105(12)

Letters to the Editor

ment analysis identified a high overlap of significantly deregulated genes between AML LRC and our previously defined LRC signature in ALL8 (Figure 1H and Online Supplementary Figure S3C), suggesting comparable biologic processes activated in LRC of both, AML and ALL.

Given the long-known link between dormancy and chemo-resistance,14 we compared the drug response between low-cycling LRC and high-cycling non-LRC. Groups of mice engrafted with CFSE-labeled cells were treated with a chemotherapeutic regimen mimicking

Figure 2. Low-cycling acute myeloid leukemia patient-derived xenograft cells are treatment resistant in vivo. (A) Experimental procedure; groups of mice were injected with carboxyfluorescein succinimidyl ester (CFSE) labeled acute myeloid leukemia patient-derived xenograft (AML-PDX) cells and treated with PBS (control [ctrl]) or a combination of 20 mg/kg DaunoXomeÂŽ [DNXl] on day 7 and 150 mg/kg cytarabine (Ara-C) on days 7 to 9; PDX cells were re-isolated from murine bone marrow on day 10 and analyzed as described in Figure 1B. (B) Tumor load was monitored by in vivo imaging in AML-393. (C) Total number of isolated PDX cells is shown of control and treated mice as mean+/- standard deviation (SD) of AML-393 (n=8), AML-491 (n=6), AML-372 (n=10) and AML-388 (n=7) (C); each dot/square represents one mouse. (D) Representative dot plots (AML-393). (E-F) Absolute number (E) and percentage (F) of non-label-retaining cells (non-LRC) and LRC among all isolated PDX cells are shown from the same mice as in (C); Log2 fold reduction for each subpopulation is displayed. *P<0.05

haematologica | 2020; 105(12)

2857

Letters to the Editor

2858

Figure 3. Acute myeloid leukemia patient-derived xenograft cells display reversible growth behavior, independently from stemness potential. (A) Experimental procedure: acute myeloid leukemia (AML)-393 cells were isolated from advanced disease donor mice (n=5 in three independent experiments), labeled with carboxyfluorescein succinimidyl ester (CFSE), and re-transplanted into first recipient mice. Ten days after injection, cells were re-isolated and sorted into label-retaining cells (LRC) and nonLRC (nLRC) using the gates as described in Figure 1B and re-injected into secondary recipient mice. (B) Secondary recipient mice receiving either 300 LRC or 300 non-LRC (n=5) were monitored by in vivo imaging. (C) LRC and non-LRC were re-injected into secondary recipient mice (n=38) in limiting dilutions at numbers indicated in the Online Supplementary Table S3. Positive engraftment of patient-derived xenograft (PDX) cells was determined by in vivo imaging and/or flow cytometry. Leukemia-initiating cell (LIC) frequency was calculated using the ELDA software and is depicted +/- 95% confidence interval. No statistically significant difference between LIC frequency of LRC and non-LRC was found according to c2 test (P=0.0638). (E) Experimental procedure; from first recipient mice (n=2 in two independent experiments) harboring CFSE stained cells, non-LRC were isolated at day 21, re-stained with CFSE and 3.6x106 cells were injected into secondary recipients (n=8); cells were re-isolated 10, 14 and 20 days later, and LRC were quantified using gates as described in Figure 1B. The experiment is technically unfeasible for LRC as the high number of cells needed cannot be generated. (E-F) Representative dot plots (E) and quantification (F) of the percentage of LRC among all PDX cells isolated from secondary recipients is displayed (dark green squares). LRC of first recipient mice as determined in Figure 1E are shown for comparison (light green dots). See the Online Supplementary Figure S4 and Online Supplementary Table S3 for additional data.

haematologica | 2020; 105(12)

Letters to the Editor

“7+3” induction therapy,1 consisting of cytarabine and liposomal daunorubicin (DaunoXome) (Figure 2A). In vivo treatment diminished tumor burden as monitored by in vivo imaging (Figure 2B), resulting in a decrease of the total isolated PDX cells by at least one order of magnitude (Figure 2C). Interestingly, while non-LRC numbers were strongly reduced by treatment, even to undetectable levels in some mice (Figures 2DE), low-cycling LRC revealed decreased sensitivity towards systemic treatment in all samples tested. As a net effect, the relative proportion of LRC was significantly enriched among cells surviving after treatment in 3 of 4 samples (Figures 2DF). Thus, low-cycling LRC show increased resistance against conventional chemotherapy in vivo compared to high-cycling non-LRC. We next asked whether LRC and non-LRC differ in their ability to form tumors and performed re-transplantation experiments. Low numbers of sorted LRC and non-LRC were re-injected into secondary recipient mice in limiting dilutions close to sample-specific LIC frequency (Figures 3A and Online Supplementary Figure S4A). Interestingly, both, LRC and non-LRC gave rise to leukemia upon re-transplantation, indicating both subpopulations contained LIC (Figures 3B and Online Supplementary Figure S4B). As leukemia development was highly similar in mice transplanted with either LRC or non-LRC, low-cycling LRC must have converted into an actively proliferative state. Furthermore, we found similar LIC frequencies in LRC and non-LRC (Figure 3C, Online Supplementary Figure S4C and Online Supplementary Table S3), and no difference in CD34+CD38– cells between the two groups (Online Supplementary Figure S5), strengthening previous findings.15 Notably, CD34+CD38– cells were barely detectable in the aggressive AML-393 sample, despite the high LIC frequency (Online Supplementary Figure S5 and Online Supplementary Table S2). These data indicate that LIC reside not only in the low-cycling LRC, but also in the high-cycling non-LRC compartment, indicating heterogeneity in proliferation dynamics within the AML LIC pool. As low-cycling cells were able to convert to active proliferation, we asked whether the switch could also occur vice versa. To test whether LRC could be replenished from non-LRC, we re-transplanted high cell numbers of non-LRC restained with CFSE (Figure 3D and Online Supplementary Figure S4D). Upon secondary transplantation, non-LRC gave rise to a clear LRC fraction, comparable to the one from bulk cells at first transplantation, even at late time points (Figures 3E-F and Online Supplementary Figure S4E-F), indicating that high-cycling cells converted to a low-cycling phenotype. These experiments revealed major functional plasticity of AML LIC phenotypes, and the ability to change their proliferation rate upon changes in external stimuli, such as re-transplantation. Taken together, our data shows that low proliferation or dormancy characterizes a temporary, reversible cell state rather than a defined subpopulation of cells. AML contains a rare fraction of low-cycling, chemo-resistant LIC which are functionally plastic; AML LIC might temporarily adopt a low-cycling LRC phenotype or switch to a rapidly proliferating non-LRC phenotype, triggered by external stimuli such as re-transplantation. Even the highly aggressive AML samples used in this study harbor the potential to adopt a proliferative phenotype associated with response to standard chemotherapy. Unexpectedly, we detected similar functional plasticity in AML as previously observed in ALL.8 This was haematologica | 2020; 105(12)

accompanied by similar changes in gene expression profiles, although both diseases differ substantially with regards to their stem cell biology as ALL has never revealed a stem cell hierarchy as proven in AML. In contrast to ALL, AML plasticity comes as a major surprise, as we show here that high-cycling cells harbor the potential to convert into low-cycling cells, while both populations retain stem cell capacities. In our experiments, neither functionally nor immunophenotypically defined LIC were enriched in the LRC fraction, suggesting that dormancy and stemness are not consistently linked in AML, but that dormancy characterizes a temporary cell state rather than a defined subpopulation of cells. In addition to the known constant, presumably deterministic factors defining stemness, AML stem cells appear to be regulated by additional, transient and putatively stochastic factors.16 Our data indicates that stemness and resistance to anti-leukemic therapy is not strictly linked in AML. This opens up an exciting therapeutic potential to prevent relapse and strongly supports the concept that recruiting AML LIC from their low-cycling phenotype into proliferation might sensitize them towards conventional chemotherapy.3,4,7 Taking advantage of the discovered heterogeneity and reversibility of the low- and highcycling phenotypes implicates the need to identify factors responsible for AML plasticity, in addition to known microenvironment-derived regulators such as G-CSF.2 The detected similarity in the transcriptome signature between LRC of AML and ALL might aid in the identification of factors that regulate these processes in both diseases. As an attractive therapeutic concept, inhibition of the reversible low-cycling state might enable to overcome treatment resistance, remove AML LIC, prevent relapse, and ultimately increase patients’ prognosis. Sarah Ebinger,1 Christina Zeller,1 Michela Carlet,1 Daniela Senft,1 Johannes W. Bagnoli,2 Wen-Hsin Liu,1 Maja Rothenberg-Thurley,3 Wolfgang Enard,2 Klaus H. Metzeler,3-5 Tobias Herold,1,3,4 Karsten Spiekermann,3-5 Binje Vick1,4 and Irmela Jeremias1,4,6 1 Research Unit Apoptosis in Hematopoietic Stem Cells, Helmholtz Zentrum München, German Research Center for Environmental Health (HMGU), Munich; 2Anthropology & Human Genomics, Department of Biology II, LudwigMaximilians-University, Martinsried, Germany; 3Laboratory for Leukemia Diagnostics, Department of Medicine III, University Hospital, Ludwig-Maximilians-University Munich, Munich; 4 German Cancer Consortium (DKTK), partner site Munich; 5 German Cancer Research Center (DKFZ), Heidelberg and 6 Department of Pediatrics, Dr. von Hauner Childrens Hospital, Ludwig Maximilian University, Munich, Germany Correspondence: IRMELA JEREMIAS - Irmela.Jeremias@helmholtz-muenchen.de doi:10.3324/haematol.2019.226282 Acknowledgments: we thank Liliana Mura, Fabian Klein, Maike Fritschle, Annette Frank and Miriam Krekel for excellent technical assistance; Markus Brielmeier and team (Research Unit Comparative Medicine, Helmholtz Zentrum München) for animal care services; Andreas Beyerlein (Core Facility Statistical Consulting, Institute of Computational Biology, Helmholtz Zentrum München) for assisting with statistical analysis of treatment studies; Helmut Blum and Stefan Krebs (Laboratory for Functional Genome Analysis, Gene Center, LMU, Munich) for sequencing, Claudia Baldus and Lorenz Bastian (Divison of Hematology and Oncology, Charité Universitätsmedizin Berlin, Germany) for kindly providing primary cells of AML-538, and Maya C. André and Martin Ebinger (Department of Pediatric 2859

Letters to the Editor

Hematology/Oncology, University Children’s Hospital Tübingen) for kindly providing pediatric AML PDX samples. Funding: the work was supported by grants from the European Research Council Consolidator Grant 681524; a Mildred Scheel Professorship by German Cancer Aid; German Research Foundation (DFG) Collaborative Research Center 1243 “Genetic and Epigenetic Evolution of Hematopoietic Neoplasms”, projects A05, A06 (to KHM), A07 (to KS) and A14 (to JWB and WE), and associate member (TH); DFG proposal MA 1876/13-1; Bettina Bräu Stiftung and Dr. Helmut Legerlotz Stiftung (all to IJ, if not indicated differently). This work was further supported by the Physician Scientists Grant (G-509200-004) from the Helmholtz Zentrum München to TH.

References 1. Dohner H, Weisdorf DJ, Bloomfield CD. Acute myeloid leukemia. N Engl J Med. 2015;373(12):1136-1152. 2. Thomas D, Majeti R. Biology and relevance of human acute myeloid leukemia stem cells. Blood. 2017;129(12):1577-1585. 3. Ishikawa F, Yoshida S, Saito Y, et al. Chemotherapy-resistant human AML stem cells home to and engraft within the bone-marrow endosteal region. Nat Biotechnol. 2007;25(11):1315-1321. 4. Saito Y, Uchida N, Tanaka S, et al. Induction of cell cycle entry eliminates human leukemia stem cells in a mouse model of AML. Nat Biotechnol. 2010;28(3):275-280. 5. Hope KJ, Jin L, Dick JE. Acute myeloid leukemia originates from a hierarchy of leukemic stem cell classes that differ in self-renewal capacity. Nat Immunol. 2004;5(7):738-743. 6. Farge T, Saland E, de Toni F, et al. Chemotherapy-resistant human

2860

7. 8. 9.

10. 11. 12. 13. 14. 15. 16.

acute myeloid leukemia cells are not enriched for leukemic stem cells but require oxidative metabolism. Cancer Discov. 2017;7(7):716-735. Boyd AL, Aslostovar L, Reid J, et al. Identification of chemotherapyinduced leukemic-regenerating cells reveals a transient vulnerability of human AML recurrence. Cancer Cell. 2018; 34(3):483-498.e5. Ebinger S, Ozdemir EZ, Ziegenhain C, et al. Characterization of rare, dormant, and therapy-resistant cells in acute lymphoblastic leukemia. Cancer Cell. 2016;30(6):849-862. Vick B, Rothenberg M, Sandhofer N, et al. An advanced preclinical mouse model for acute myeloid leukemia using patients' cells of various genetic subgroups and in vivo bioluminescence imaging. PLoS One. 2015;10(3):e0120925. de Boer B, Prick J, Pruis MG, et al. Prospective isolation and characterization of genetically and functionally distinct AML subclones. Cancer Cell. 2018;34(4):674-689. 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. Takizawa H, Regoes RR, Boddupalli CS, Bonhoeffer S, Manz MG. Dynamic variation in cycling of hematopoietic stem cells in steady state and inflammation. J Exp Med. 2011;208(2):273-284. Bagnoli JW, Ziegenhain C, Janjic A, et al. Sensitive and powerful single-cell RNA sequencing using mcSCRB-seq. Nat Commun. 2018; 9(1):2937. Cheung WH, Rai KR, Sawitsky A. Characteristics of cell proliferation in acute leukemia. Cancer Res. 1972;32(5):939-942. Griessinger E, Vargaftig J, Horswell S, Taussig DC, Gribben J, Bonnet D. Acute myeloid leukemia xenograft success prediction: Saving time. Exp Hematol. 2018;59:66-71. Gupta PB, Fillmore CM, Jiang G, et al. Stochastic state transitions give rise to phenotypic equilibrium in populations of cancer cells. Cell. 2011;146(4):633-644.

haematologica | 2020; 105(12)

Letters to the Editor

Donor cell leukemia: is reappearance of gene mutations in donor cells more than an incidental phenomenon? Acute myeloid leukemia (AML) is one of the extreme outcomes of age-related clonal hematopoiesis (ARCH).1 With aging, mutations accumulate in hematopoietic stem and progenitor cells (HSPC).2,3 Based on the estimated number of HSPC (~50,000) in the human body and the number of somatic mutations in single adult cells (~1000),4 it is predicted that every ~100 nucleotides, a somatic mutation will occur at a low variant allele frequency (VAF). Indeed, recent studies have demonstrated that virtually all elderly healthy individuals carry proteindamaging mutations in DNMT3A and TET2 at a low VAF.5 Although mutations in preleukemic genes are inevitable, only 10-30% of older adults develop ARCH (large clone VAF >0.02).6 Moreover, only one out of 1,000 carriers of somatic mutations will develop a myeloid malignancy. Recent studies have revealed that individuals with ARCH who carry larger clones and/or more than one mutation and specific mutations, are at a higher risk of developing AML. However, the underlying mechanisms resulting in clonal evolution that eventually leads to full-blown AML remain unclear. HSPC reside in the bone marrow (BM), where they are supported by mesenchymal stromal cells that maintain their quiescence and/or promote their proliferation and differentiation. The BM microenvironment can determine whether HSPC carrying mutations will outgrow their wild-type counterparts. The role of the BM microenvironment in the induction of leukemia has been described in a rodent model,7 but studies in humans are scarce. We present here two rare cases of late-onset donor cell leukemia (DCL) that developed in allogeneic stem cell transplant (SCT) recipients and not in their sex-mismatched donors (Table 1). These findings might imply the involvement of the BM microenvironment in the initiation and evolution of myeloid malignancies. The study was approved by the Institutional Review Board of Rambam Health Care Campus (approval n. 0016-15). The molecular profile of the patients was assessed using deep, error-corrected sequencing (average coverage ~5000X) as described in the Online Supplementary Methods. The presence of mutations (somatic and germline) at diagnosis and at the occurrence of DCL was evaluated (variants with VAF >0.1) in the patients' peripheral blood and saliva as well as in the donors' peripheral blood. Karyotype and chimerism analyses were performed in DCL samples. To confirm the presence of donor sex chromosomes and full donor chimerism, polymerase chain reaction (PCR) for short tandem repeats and fluorescence in situ hybridization

(FISH) analyses were done (Online Supplementary Table S1). To explore the possibility that DCL originated from ARCH mutations present in the donor we applied ARCH variant calling algorithms aimed at accurately detecting recurrent AML mutations at a VAF >0.005. Additionally, we longitudinally followed complete blood count (CBC) parameters in 71 long-term survivors post-allogeneic SCT. These individuals were identified in the electronic database of a health maintenance organization encompassing 3.45 million people. The CBC of these patients were compared to those of 500,000 age- and gendermatched controls. The current report describes the analysis of two patients with myelodysplastic syndrome (MDS)/AML who underwent allogeneic SCT from HLA-matched, sexmismatched family donors and relapsed with DCL years after the procedure. In the first case, a 54-year old male, diagnosed with AML in 2006, exhibited a mutation in the U2AF1 S34F gene (Online Supplementary Tables S1 and S2). The patient achieved complete remission and was transplanted from his HLA-matched sister (52 years old) who carried a mutation in the DNMT3A F752Y gene at a low VAF (2.96%) at the time of donation (Online Supplementary Table S3). Nine years later (2015) the patient relapsed, presenting with different mutations, i.e., U2AF1 (S34Y instead of the original S34F), IDH1 R132C and DNMT3A F752Y at higher VAF (43.3%). At the time of DCL onset he had full chimerism and 100% XX donor cells, as confirmed by PCR for short tandem repeates and FISH. Notably, the mutations in the U2AF1 and IDH1 genes were not detected in the donor's peripheral blood either at the time of cell donation or 10 years later (Figure 1: case 1). To date, the donor remains in good general condition with normal CBC. In the second case, a 26-year old female was diagnosed with high-risk MDS in 2000 and transplanted from her haplo-matched father (73 years old) (Online Supplementary Tables S1 and S2). At the time the MDS was diagnosed a mutation in the U2AF1 S34F gene was found. Seventeen years later (2017) the woman relapsed with high-risk MDS, presenting with mutations in TET2 S1059X and ASXL1 W1411X, not observed at diagnosis. At the onset of DCL the patient had full donor chimerism and 100% XY donor cells, as confirmed by PCR for short tandem repeats and FISH. These mutations were not detected in donor's peripheral blood during donation. (Figure 1, case 2; Online Supplementary Table S3). Notably, the donor died from a cerebrovascular accident. In both analyzed DCL patients, the red blood cell distribution width (RDW) remained elevated even 9 and 17 years after the transplant (Figure 1). As increased RDW has been reported to be associated with dyserythropoiesis and a poor prognosis in patients with hematologic

Table 1. Potential confounders responsible for the clinical course after allogeneic stem cell transplantation. Conditioning regimens GvHD prophylaxis Acute GvHD grade Chronic GvHD grade Pharmacological immunosuppression after SCT

Case 1

Case 2

Reduced-intensity: fludarabine/melphalan CSA + MTX 2 Mild 1.5 years

Myeloablative: total body irradiation/ fludarabine/thiotepa/ATG T-cell depletion 3 Mild-moderate 3 years

ATG: antithymocyte globulin; CSA: cyclosporine; GvHD: graft-versus-host disease; MTX: methotrexate; SCT: stem cell transplantation; .

haematologica | 2020; 105(12)

2861

%VAF

Letters to the Editor

Figure 1. The variant allele frequency profile and red cell distribution width of two patients with donor cell leukemia that developed long after transplantation for acute myeloid leukemia. RDW: red cell distribution width; Dx: diagnosis; VAF: variant allele frequency; DCL: donor cell leukemia (DCL).

malignancies, the present study analyzed longitudinal CBC of another 71 allogeneic SCT recipients, recorded in the dataset of Clalit, a health maintenance organization (Online Supplementary Methods). While most CBC parameters, including RDW, gradually normalized in the majority of patients in this cohort over the first 4 years following allogeneic SCT, ~25% of these long-term survivors still exhibited high RDW. This evidence could be suggestive of underlying MDS, or other hematologic pathologies associated with increased RDW,8 DCL being one of them. Moreover, in an Italian study including 94 adults with acute leukemia, unfavorable cytogenetics were found in 26% of patients with high RDW compared to 8% of patients with RDW within the normal range (P=0.10).8 Additionally, increased RDW was reported among individuals with ARCH6 and specifically in those at a greater risk of AML development. This change in erythropoiesis might be mediated, at least in part, by an impaired BM microenvironment, although an abnormal crosstalk between cells in the BM and clonal HSPC (ARCH) could also be a possible contributor.9 While a damaged microenvironment might be involved in both DCL evolution and RDW dynamics, the mechanisms underlying these effects may differ. The long period between allogeneic SCT and DCL development observed in our patients could imply either that the evolution occurred in donor HSPC, originating from mutations that were scanty (below the detection limit) in the donor and then expanded in the recipient, or 2862

that new mutations were acquired and the mutated cells proliferated in the recipient only. In case 1, a preleukemic mutation (DNMT3A) was found in the donor at the time of donation.10-12 While inherited predisposition to MDS/AML is known, it is less likely to be the case in the current patient, as no other leukemias have been reported among family members. Hence, the initiation of DCL in the cases reported here could have been triggered by involvement of the BM microenvironment in abnormal hematopoiesis.13-15 This interpretation of ours is supported by the fact that in patient #1 the same gene has been affected twice in the same position, as revealed at diagnosis and in the transplanted donor’s cells 9 years after SCT. Of note, the chance of the same patient developing AML twice in the same position within the gene (U2AF1 S34) is less than 1:500,000.2 Overall, our findings suggest that at least in some MDS/AML cases, following allogeneic SCT, certain mutations in the donors’ cells might have a selective advantage in specific conditions of the BM microenvironment. Hematopoiesis is not fully normalized as a result of abnormalities in the recipient’s BM microenvironment, which trigger leukemogenesis in the donor’s cells that may eventually lead to DCL. Better understanding of this interplay could shed light on the mechanisms of AML evolution and contribute to advances in prevention and treatment of this disease. Tal Shahar Gabay,1,2 Noa Chapal-Ilani,3 Yoni Moskovitz,3 Tamir Biezuner,3 Barak Oron,3 Yardena Brilon,3 haematologica | 2020; 105(12)

Letters to the Editor

Anna Fridman-Dror,1 Rawan Sabah,1,2 Ran Balicer,4 Amos Tanay,3,5 Netta Mendelson-Cohen,5 Eldad J. Dann,6 Riva Fineman,6 Nathali Kaushansky,3 Shlomit Yehudai-Reshef,1*# Tsila Zuckerman2,6# and Liran I. Shlush3,6# 1 Hematology Research Center, Rambam Health Care Campus, Haifa; 2Bruce Rappaport Faculty of Medicine, Technion, Haifa; 3 Department of Immunology, Weizmann Institute of Science, Rehovot; 4 Clalit Research Institute, Tel Aviv; 5Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot and 6Department of Hematology and BMT, Rambam Health Care Campus, Haifa, Israel #

SY-R, TZ and LIS contributed equally as co-senior authors.

Correspondence: TSILA ZUCKERMAN - t_zuckerman@rambam.health.gov.il doi:10.3324/haematol.2019.242347 Funding: this work was supported by the Israel Cancer Association through grants n. 20170179 and 20181136.

References 1. Shlush LI. Age-related clonal hematopoiesis. Blood. 2018;131(5):496504. 2. Shlush LI, Zandi S, Mitchell A, et al. Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature. 2014; 506(7488):328-333. 3. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264-278. 4. Lee-Six H, Obro NF, Shepherd MS, et al. Population dynamics of normal human blood inferred from somatic mutations. Nature. 2018; 561(7724):473-478.

haematologica | 2020; 105(12)

5. Young AL, Challen GA, Birmann BM, Druley TE. Clonal haematopoiesis harbouring AML-associated mutations is ubiquitous in healthy adults. Nat Commun. 2016;7:12484. 6. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371(26):2488-2498. 7. Raaijmakers MH, Mukherjee S, Guo S, et al. Bone progenitor dysfunction induces myelodysplasia and secondary leukaemia. Nature. 2010;464(7290):852-857. 8. Milone G, DiMarco A, Sapienza G, et al. Red cell distribution width (RDW) as prognostic factor in acute leukemia patients undergoing allogeneic hematopoietic transplantation. Blood. 2017; 130 (Supp_1):5098. 9. Leimkuhler NB, Schneider RK. Inflammatory bone marrow microenvironment. Hematology Am Soc Hematol Educ Program. 2019;2019(1):294-302. 10. Hahn CN, Ross DM, Feng J, et al. A tale of two siblings: two cases of AML arising from a single pre-leukemic DNMT3A mutant clone. Leukemia. 2015;29(10):2101-2104. 11. Gondek LP, Zheng G, Ghiaur G, et al. Donor cell leukemia arising from clonal hematopoiesis after bone marrow transplantation. Leukemia. 2016;30(9):1916-1920. 12. Yasuda T, Ueno T, Fukumura K, et al. Leukemic evolution of donorderived cells harboring IDH2 and DNMT3A mutations after allogeneic stem cell transplantation. Leukemia. 2014;28(2):426-428. 13. Wong TN, Miller CA, Jotte MRM, et al. Cellular stressors contribute to the expansion of hematopoietic clones of varying leukemic potential. Nat Commun. 2018;9(1):455. 14. Suarez-Gonzalez J, Martinez-Laperche C, Martinez N, et al. Wholeexome sequencing reveals acquisition of mutations leading to the onset of donor cell leukemia after hematopoietic transplantation: a model of leukemogenesis. Leukemia. 2018;32(8):1822-1826. 15. Engel N, Rovo A, Badoglio M, et al. European experience and risk factor analysis of donor cell-derived leukaemias/MDS following haematopoietic cell transplantation. Leukemia. 2019;33(2):508-517.

2863

Letters to the Editor

Chronic lymphocytic leukemias with trisomy 12 show a distinct DNA methylation profile linked to altered chromatin activation Chronic lymphocytic leukemia (CLL) is a neoplasm derived from mature B cells showing a broad spectrum of clinico-biological features.1 The landscape of genetic alterations of CLL is well characterized2 and found to be extremely heterogeneous, with multiple chromosomal aberrations and dozens of driver genes mutated in relatively small proportions of the cases.3,4 In spite of this heterogeneity, four cytogenetic alterations, i.e., del(13q) (>50% of the patients), del(11q) (18%), +12 (16%), and less frequently del(17p) (7%), are collectively detected in at least 80% of patients.1 These copy number changes are part of the routine risk assessment of CLL, as they are robustly associated with treatment choices and the clinical course of the patients. At one end of the prognostic spectrum, the isolated del(13q) is related to favorable prognosis, +12 with intermediate prognosis, del(11q) with poor prognosis and del(17p) with the worst prognosis of all groups. This latter subgroup identifies patients with particular resistance to chemoimmunotherapy who, instead, benefit considerably from biological agents.1 Although the targets of del(11q), del(13q) and del(17p) have been narrowed down to ATM/mir-34b-c, the DLEU2â&#x20AC;&#x201C;mir-15-16 cluster, and TP53, respectively, how +12 contributes to CLL pathogenesis remains elusive.1,5 This group of CLL is characterized by high rates of cell proliferation as well as clinical and biological heterogeneity, since it is linked to additional genomic aberrations such us trisomy 19, high frequency of NOTCH1 mutations6-8 and enrichment in the aggressive stereotyped subset 89 and IgG-switched heavy chains.10 Moreover, +12 has recently been identified as an important modulator of response to kinase inhibitors in CLL.11 In addition to genetic changes, various layers of the CLL epigenome have also been characterized and these analyses have revealed novel insights into the cellular origin and molecular mechanisms underlying disease pathophysiology.12-14 However, the relationship between the genetic and epigenetic changes in CLL is still unclear. Therefore, the goal of the present study is to identify the epigenetic correlates of the three most frequent cytogenetic subgroups of CLL and to shed light onto the molecular pathogenesis of +12 cases. We used 450k arrays to analyze the DNA methylomes from 255 CLL cases in whom del(13q), del(11q) and +12 were mutually exclusive. In particular, we investigated 29 CLL cases with del(11q) (3 IGHV-mutated CLL [M-CLL] and 26 IGHV-unmutated CLL [U-CLL]), 45 with +12 (17 M-CLL and 28 U-CLL) and 181 with del(13q) (131 MCLL and 50 U-CLL), as well as five biological replicates each of normal naive (NBC) and memory (MBC) B cells sorted from healthy donors (Online Supplementary Table S1). We initially applied a series of unsupervised methods to characterize the DNA methylome of the cases. We found that the mean methylation levels of the three CLL subgroups was lower than in normal B cells, and that +12 cases globally had a significantly hypomethylated genome compared to the other CLL cases (+12 vs. del(11q), P=0.006; +12 vs. del(13q), P=7.2e-08; del(13q) vs. del(11q), P=0.01) (Figure 1A). Next, we analyzed the CpG with most variable methylation levels (Standard deviation [SD]â&#x2030;Ľ0.1, n=185,936) by Principal Component Analysis (PCA) and studied the information provided by each of the five principal components (Online 2864

Supplementary Figure S1). As previously reported,13 the first two components revealed different fractions of the variability to reflect different concepts: (i) overall CLL are epigenetically more similar to memory B cells (PC1, 7.9%); (ii) U-CLL and M-CLL are epigenetically different (PC1 and PC2); and, (iii) CLL as a whole are different from naive and memory B cells (PC2, 3.7%). The third and fourth components (1.9% and 1.3%) were less clear and showed subtle differences between naive and memory B cells as well as between U-CLL and M-CLL. Remarkably, the fifth component (1%) was unrelated to the IGHV gene somatic hypermutation status and revealed clear differences between +12 cases and CLL cases from the other cytogenetic subgroups (Figure 1B and Online Supplementary Figure S1), suggesting that +12 cases may indeed show a specific epigenetic configuration. We then aimed at identifying the +12-specific CpG sites by two complementary approaches. First, we correlated the eigenvalues of the PC5 and the methylation b-values and found that 1,760 CpG were significantly associated with this source of variability and confirmed to be related with +12 by hierarchical clustering (Online Supplementary Figure S2). In order to more specifically detect +12-related CpG, we performed a supervised differential methylation (DM) analysis in cases with and without +12 considering U-CLL and M-CLL separately (absolute mean b-value difference of at least 0.25 and a false discovery rate [FDR]<0.05). In +12 U-CLL (n=28), we observed a signature of 646 DMCpG, which was mostly composed of hypomethylated sites in the +12 cases (80.1%) (Figure 1C-D and Online Supplementary Table S2). This hypomethylation signature was present in +12 CLL lacking and showing NOTCH1 mutations, and therefore, was specifically associated with +12. The +12 M-CLL (n=17) were more heterogeneous and a lower number of DMCpG were detected (Figure 1C). However, the 646 DMCpG in +12 U-CLL showed the same trend in M-CLL cases (Online Supplementary Figure S3). Furthermore, analyzing B cells spanning the entire B-ell maturation program15 revealed that the hypomethylation signature was mostly acquired de novo in the +12 cases, as for the great majority of the CpG it was not present in any B-cell subpopulation (Figure 1D and Online Supplementary Figure S3). Based on these findings, downstream analyses in the +12 cases were focused on the UCLL subgroup. We also performed a similar analysis for del(13q) and del(11q) cases, which showed few DMCpG (Figure 1C). The predominantly de novo acquired hypomethylation signature in +12 U-CLL was related to gene bodies and untranslated (UTR) regions outside CpG islands, and targeted both transcribed regions and enhancer elements (Figure 1E). The link between +12-specific hypomethylation and enhancer elements prompted us to evaluate the association between DNA methylation and histone 3 lysine 27 acetylation (H3K27ac), a chromatin mark related to active regulatory elements. In a previous study,12 we generated H3K27ac ChIP-seq data in CLL patients, which included 21 of the 255 cases whose DNA methylome is studied in the present report. We initially observed that 48.3% of the DMCpG (312 of 646) specific for +12 UCLL were located within 246 H3K27ac peaks (Online Supplementary Table S3). Remarkably, studying the H3K27ac signal within the overlapping peaks, we noticed that the regions losing methylation showed a trend towards higher H3K27ac levels, i.e., more activation, and those regions gaining methylation showed haematologica | 2020; 105(12)

Letters to the Editor

Log2 FC Figure 1. DNA methylation analysis of del(13q), del(11q) and +12 chronic lymphocytic leukemia. (A) Barplot showing the overall DNA methylation levels in the examined cytogenetic subgroups and normal B cells (+12 vs. del(11q), P=0.006; +12 vs. del(13q), P=7.166e-08; del(13q) vs. del(11q), P= 0.01). (B) Principal component analysis showing components 1 and 5 in the cytogenetic subgroups and normal B cells. (C) Number of differential methylation CpG (DMCpG) resulting from the comparison of each cytogenetic subgroup versus all the others chronic lymphocytic leukemia (CLL) cases. The white bar represents the IGHVunmutated CLL (U-CLL) while the grey bar represents the IGHV-mutated CLL (M-CLL) cases. (D) Hierarchical clustering based on 646 DMCpG of +12 U-CLL cases as compared to other U-CLL and normal B cells. (E) Enrichment analysis of DMCpG hypomethylated in +12 U-CLL on gene location, location related to islands and chromatin states of memory B cells. The density represents the log2FC of the enriched CpG compared to the background in each condition (hypo- and hypermethylation). Each color is linked to a particular subgroup: orange for del(11q), blue for del(13q), red for +12, green for memory B cells and yellow for naĂŻve B cells. PC: principal component, FC: fold change.

lower H3K27ac levels and, therefore, less active chromatin (Figure 2A-B). A differential analysis of the 246 H3K27ac peaks in U-CLL with and without +12 (adjusted P<0.05) revealed 35 regions with significantly increased H3K27ac levels (Figure 2C). Although a previous report indicated that the overall H3K27ac pattern of +12 cases was similar to normal B cells,12 the increased statistical power of our targeted analysis of de novo hypomethylated sites focused on +12 U-CLL revealed the presence of chromatin activation. As transcription factor (TF) binding has been described to be related to DNA methylation and chromatin activity,12,16 we postulated that TF could be implicated in inducing the +12-specific epigenomic signature. As TF bind to DNA lacking nucleosomes, we used previously reported chromatin accessibility data generated by ATAC-seq (available data for 20 of 255 examined cases)12 (Figure 2D). We found that the 35 regions of interest contained 52 sites of accessible chromatin where TF potentially bind (Online Supplementary Table S4). The DNA sequences within these 52 accessible sites were sighaematologica | 2020; 105(12)

nificantly enriched (P<0.05) in binding sites of the NFIBC, MYCN, TFCP2 and XBP1 TF (Figure 2E). Interestingly, the TFCP2 gene is located in chromosomal band 12q13 and show moderate but significant overexpression cases with +12 compared to CLL lacking this genetic change (log2FC= 0.284, FDR=0.024) (Online Supplementary Figure S4). Although detailed functional studies are needed to establish the role of TFCP2 in CLL, our data suggest that its overexpression may be one of the potential mechanisms through which +12 exerts its pathogenic effect in CLL. In order to identify the potential target genes of the regulatory elements targeted by hypomethylation in +12 cases, we analyzed the gene expression arrays of U-CLL with del(11q), del(13q) or +12. As regulatory elements can affect the expression of distant genes within topologically associating domains (TAD),12 the +12-specific hypomethylated regions associated with increased acetylation (n=35) were integrated with the TAD from GM12878, a lymphoblastoid B-cell line widely used to 2865

Letters to the Editor

characterize epigenetic patterns in CLL.12-14 Using gene expression arrays from an extended series of 20 U-CLL with +12 and 54 U-CLL lacking this genetic change, we identified 25 genes with increased expression in +12 compared to del(11q) and del(13q) U-CLLs (Figure 2F and Online Supplementary Table S5). Amongst them, RUNX3

was a remarkable example. This gene is a master regulator of gene expression during development that has been reported to act as tumor suppressor or oncogene in cancer.17 We identified that +12 U-CLL show a cluster of seven hypomethylated CpG plus increased H3K27ac levels in a regulatory region 35 Kb upstream the RUNX3

NFIB (P=5.57e-5)

MYCN (P=1.24e-4) TFCP2 (P=2.60e-4)

D XBP1 (P=5.17e-4) NFIC (P=6.14e-4)

Figure 2. Characterization of the chromatin and transcriptional features of regions de novo hypomethylated in +12 cases. (A) DNA methylation levels of 312 CpG that overlap with H3K27ac-containing regions. (B) Heatmap of signal intensities of H3K27ac regions associated with CpG from panel (A). (C) Heatmap of the 35 hypomethylated regions showing significantly higher H3K27ac levels in IGHV-unmutated CLL (U-CLL) compared with cases without +12. (D) Scheme of the strategy used to detect accessible sites within regions with differential methylation CpG (DMCpG) and increased chromatin activity. (E) Transcription factors whose binding sites are significantly enriched in accessible regions within hypomethylated and active regulatory elements. (F) Heatmap showing the relative expression levels of the 25 genes epigenetically-upregulated in +12 U-CLL compared with cases lacking +12. (G) On the left, a genome browser display of the RUNX3 region including H3K27ac and gene expression of two representative +12 and non +12 U-CLL, as well as DNA methylation data from all U-CLL studied. The red square points to the distant regulatory region that becomes hypomethylated and active in +12 U-CLL. On the right, a zoom-in panel of the H3K27ac levels of the distant enhancer region as well as box plots of H3K27ac and gene expression in all U-CLL studied. Each color is linked to a particular subgroup: orange for del(11q), blue for del(13q), red for +12.

2866

haematologica | 2020; 105(12)

Letters to the Editor

promoter, a gene that was found to be significantly overexpressed (Figure 2G). No other region across the RUNX3 gene changed DNA methylation levels and the gene promoter showed similar H3K27ac levels in cases showing and lacking +12 (Figure 2G). These results suggest that epigenetic activation of dozens of genes, and in particular the activation of a distant RUNX3 regulatory element leading to gene overexpression, may account for the distinct biological background of this CLL subtype. Taken together, our findings further support the unique biological features of +12 CLL from the epigenetic perspective. This group of CLL is associated with a subset of epigenetically-upregulated genes that may account for its distinct biological background. These novel insights into +12 CLL may provide a biological rationale to identify specific therapies to treat this unique subtype of CLL. Maria Tsagiopoulou,1,2 Vicente Chapaprieta,3 Martí Duran-Ferrer,3 Theodoros Moysiadi,1,4 Fotis Psomopoulos,1,4 Panagoula Kollia,2 Nikos Papakonstantinou,1 Elias Campo,3,5,6 Kiostas Stamatopoulos1,4 and Jose I. Martin-Subero3,5,6,7 1 Institute of Applied Biosciences, Center for Research and Technology Hellas, Thessaloniki, Greece; 2Department of Biology, National and Kapodistrian University of Athens, Athens, Greece; 3Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Universitat de Barcelona, Barcelona, Spain; 4Department of Molecular Medicine and Surgery, Karolinska Institute, Stockholm, Sweden; 5 Departamento de Fundamentos Clínicos, Universitat de Barcelona, Barcelona, Spain; 6Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), Madrid, Spain and 7Institució Catalana de Recerca I Estudis Avançats (ICREA), Barcelona, Spain. Correspondence: JOSE I. MARTIN-SUBERO - imartins@clinic.cat doi:10.3324/haematol.2019.240721 Acknowledgements: this work was partially developed at the Centro Esther Koplowitz (CEK, Barcelona, Spain). Funding: data used in the present study have been generated through funding from the European Union’s Seventh Framework Programme through the Blueprint Consortium (grant agreement 282510), the International Cancer Genome Consortium (Chronic Lymphocytic Leukemia Genome consortium to EC), the World Wide Cancer Research Foundation Grant No. 16–1285 (to JIM-S), the Generalitat de Catalunya Suport Grups de Recerca AGAUR 2017SGR-736 (to JIM-S) and 2017-SGR-1142 (to EC), and CIBERONC (CB16/12/00225). EC is an Academia Researcher of the Institució Catalana de Recerca I Estudis Avançats (ICREA) of the Generalitat de Catalunya. MT is a recipient of a fellowship from the State Scholarships Foundation of Greece, and a short term collaboration award winner by the European Hematology Association.

haematologica | 2020; 105(12)

References 1. Kipps TJ, Stevenson FK, Wu CJ, et al. Chronic lymphocytic leukaemia. Nat Rev Dis Primers. 2017;3:16096. 2. Bosch F, Dalla-Favera R. Chronic lymphocytic leukaemia: from genetics to treatment. Nat Rev Clin Oncol. 2019;16(11):684-701. 3. 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. 4. Puente XS, Bea S, Valdes-Mas R, et al. Non-coding recurrent mutations in chronic lymphocytic leukaemia. Nature. 2015; 526(7574):519-524. 5. Deneberg S, Kanduri M, Ali D, et al. microRNA-34b/c on chromosome 11q23 is aberrantly methylated in chronic lymphocytic leukemia. Epigenetics. 2014;9(6):910-917. 6. 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. 7. Del Giudice I, Rossi D, Chiaretti S, et al. NOTCH1 mutations in +12 chronic lymphocytic leukemia (CLL) confer an unfavorable prognosis, induce a distinctive transcriptional profiling and refine the intermediate prognosis of +12 CLL. Haematologica. 2012;97(3):437-441. 8. Ibbotson R, Athanasiadou A, Sutton LA, et al. Coexistence of trisomies of chromosomes 12 and 19 in chronic lymphocytic leukemia occurs exclusively in the rare IgG-positive variant. Leukemia. 2012; 26(1):170-172. 9. Baliakas P, Hadzidimitriou A, Sutton LA, et al. Clinical effect of stereotyped B-cell receptor immunoglobulins in chronic lymphocytic leukaemia: a retrospective multicentre study. Lancet Haematol. 2014; 1(2):e74-84. 10. Vardi A, Agathangelidis A, Sutton LA, et al. IgG-switched CLL has a distinct immunogenetic signature from the common MD variant: ontogenetic implications. Clin Cnacer Res. 2014;20(2):323-330. 11. Dietrich S, Oles M, Lu J, et al. Drug-perturbation-based stratification of blood cancer. J Clin Invest. 2018;128(1):427-445. 12. Beekman R, Chapaprieta V, Russinol N, et al. The reference epigenome and regulatory chromatin landscape of chronic lymphocytic leukemia. Nat Med. 2018;24(6):868-880. 13. 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. 14. 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. 15. Kulis M, Merkel A, Heath S, et al. Whole-genome fingerprint of the DNA methylome during human B cell differentiation. Nat Genet. 2015;47(7):746-756. 16. Stadler MB, Murr R, Burger L, et al. DNA-binding factors shape the mouse methylome at distal regulatory regions. Nature. 2011;480(7378):490-495. 17. Selvarajan V, Osato M, Nah GSS, et al. RUNX3 is oncogenic in natural killer/T-cell lymphoma and is transcriptionally regulated by MYC. Leukemia. 2017;31(10):2219-2227.

2867

Letters to the Editor Table 1. Pre-treatment characteristics.

Lenalidomide, dexamethasone and alemtuzumab or ofatumumab in high-risk chronic lymphocytic leukemia: final results of the NCRI CLL210 trial Therapeutic response in chronic lymphocytic leukemia (CLL) is variable, with deletion or inactivating mutation of the TP53 gene on chromosome 17p13 being strongly associated with chemotherapy resistance and short survival. The UK CLL206 and German/French CLL2O trials demonstrated the effectiveness of combining the antiCD52 monoclonal antibody alemtuzumab with highdose methylprednisolone (HDMP) or dexamethasone in high-risk CLL,1,2 and these p53-independent drug combinations became the standard of care for such patients in many centers prior to the advent of novel agents such as ibrutinib, idelalisib and venetoclax.3 The CLL210 trial was developed to evaluate the potential benefit of adding the cereblon-targeting drug lenalidomide to the alemtuzumab/glucocorticoid backbone. Lenalidomide was of interest owing to its established activity in 17p-deleted CLL coupled with its potential to act in synergy with the other two drugs in a p53-independent manner.4,5 During the course of the study, alemtuzumab became unavailable and was replaced by the anti-CD20 monoclonal antibody ofatumumab, which has a reported efficacy similar to that of alemtuzumab.6 Although the study showed that both regimens had therapeutic activity, the predefined co-primary endpoints for efficacy and toxicity were not met. CLL210 was designed as a single-arm phase II trial with a randomisation to lenalidomide maintenance versus placebo for patients who responded to the induction phase. Patients were eligible if they had CLL requiring therapy by International Workshop on Chronic Lymphocytic Leukemia (iwCLL) criteria and were highrisk defined by a previously documented 17p deletion or TP53 mutation affecting at least 20% of CLL cells, or a history of not responding to or relapsing within 12 months of responding to fludarabine-containing combination therapy irrespective of TP53 status. The study treatment consisted of dexamethasone (40 mg on day 1-4 of alternate weeks from week 1-15), lenalidomide (5 mg daily during weeks 3 and 4 and then 10 mg daily during weeks 5-24) and alemtuzumab (30 mg by subcutaneous injection thrice weekly during weeks 722). Supportive care included aciclovir, pneumocytistis jiroveci prophylaxis, cytomegalovirus (CMV) PCR surveillance and granulocyte colony-stimulating factor (GCSF) support. In the amended protocol, alemtuzumab was replaced by 12 doses of intravenous ofatumumab (300 mg on day 1 of week 7, then 1,000 mg weekly on day 1 of weeks 8-15, then fortnightly on day 1 of weeks 17-21). Patients who achieved a complete response (CR) or partial response were allowed to proceed to allogeneic haemopoietic stem-cell transplantation (HSCT) or were randomised to stopping treatment or continuing lenalidomide as maintenance therapy (10 mg daily until disease progression). The efficacy and toxicity of induction therapy were evaluated using co-primary endpoints comprising CR rate and tolerability defined as absence of treatmentrelated grade 5 serious adverse events (SAE) and grade â&#x2030;Ľ3 SAE due to infection. The criteria for considering the study treatment to be of potential or definite interest were set at a CR rate of more than 10% or 20%, respectively, and an intolerance rate of less than 50% or 30%, respectively. Secondary outcomes included overall response (OR) rate, progression-free survival (PFS), overall survival (OS) and toxicity. Minimal residual disease 2868

Ofatumumab Alemtuzumab Total (N=48) (N=16) (N=64) Age, median (IQR) 66 (59-70) Sex, n (%) Female 15 (31%) Male 33 (69%) Binet stage, n (%) A 10 (21%) B 12 (25%) C 25 (52%) Unknown 1 (2%) IGHV Status* Mutated 13 (27%) Unmutated 29 (60%) Other** 6 (13%) WHO performance status, n (%) 0 25 (52%) 1 17 (35%) 2 6 (13%) CIRS Total Score*** 2 (0-4) median (IQR) CIRS Severity Index 1 (0-2) median (IQR) Previous Treatment, n (IQR) No 21 (44%) Yes 27 (56%) TP53 defect****, n (%) No 8 (17%) Yes 40 (83%)

68 (57-74) 66 (59-70) 3 (19%) 13 (81%)

18 (28%) 46 (72%)

7 (44%) 4 (25%) 5 (31%) 0 (0%)

17 (27%) 16 (25%) 30 (47%) 1 (1%)

2 (12%) 11 (69%) 3 (19%)

15 (23%) 40 (63%) 9 (14%)

9 (56%) 7 (44%) 0 (0%) 2 (1-4)

34 (53%) 24 (38%) 6 (9%) 2 (1-4)

1 (1-2)

8 (50%) 8 (50%)

29 (45%) 35 (55%)

3 (19%) 13 (81%)

11 (17%) 53 (83%)

*IGHV genes showing >98% homology to the germline DNA were classed as unmutated and the remainder as mutated. **Six patients had no clonal heavy-chain variable region identified and three patients had insufficient sample to assess for IGHV status. ***CIRS score did not include points for having CLL. ****Previously documented TP53 defects were confirmed in pre-treatment blood samples from 47 of 53 (89%) patients and consisted of 17p deletion and TP53 mutation (33 patients), 17p deletion only (eight patients) or TP53 mutation only (six patients). IGHV: immunoglobulin heavy-chain variable region; IQR: interquartile range; WHO: World health organisation.

(MRD) was assessed centrally by 4-color flow cytometry with a sensitivity of 10-4. Efficacy data were assessed by an independent endpoint review committee using the 2008 National Cancer Institute/iwCLL (NCI/iwCLL) criteria.7 Patients without progressive disease (PD) were deemed evaluable for response assessment if at least 10 weeks of study treatment had been administered. Toxicity assessment was in accordance with common terminology criteria for adverse events (CTCAE) v4.0 with the exception of hematological toxicity which was assessed using the 2008 NCI/iwCLL criteria. Sixty-four patients were registered from 21 UK sites between February, 6 2012 and October, 8 2015. Sixteen patients were recruited to the original alemtuzumab protocol until September, 4 2012, after which 48 additional patients were recruited to the revised ofatumumab protocol from September, 13 2013. Baseline features of registered patients are summarized in Table 1 and were broadly as expected. Twenty-nine (45%) patients were treatment-naĂŻve, while the other 35 (55%) had received haematologica | 2020; 105(12)

Letters to the Editor

Figure 1. Kaplan-Meier plots showing progression-free and overall survival for the different induction and post-induction treatments. (A) Progression-free survival of the alemtuzumab and ofatumumab cohorts from study registration; (B) overall survival of the alemtuzumab and ofatumumab cohorts from study registration; (C) progression-free survival of patients who were randomised to lenalidomide maintenance or no further treatment or received a hematopoietic stemcell transplant; (D) overall survival of patients who were randomised to lenalidomide maintenance or no further treatment or received a hematopoietic stemcell transplant.

between one and three lines of prior therapy. Fifty-three (83%) patients had a previously recorded 17p deletion including all 29 treatment-naĂŻve patients and 24 of 35 (69%) previously treated patients. Patient characteristics were generally well balanced between the alemtuzumab and ofatumumab cohorts. Within the alemtuzumab cohort, 9 of 16 patients received all of the planned induction therapy, whereas treatment was terminated prematurely in 7 patients who received a median of 29 (interquartile range [IQR]: 12-54) percent of the planned treatment. Within the ofatumumab cohort (excluding one untreated patient who did not start trial treatment due to acute immune thrombocytopenia [ITP]), 24 of 47 patients received all the planned induction therapy, whereas treatment was terminated prematurely in 23 patients who received a median of 29 (IQR: 10-38) percent of the planned treatment. Among the 16 patients in the alemtuzumab cohort, the CR/CRi, PR, SD and PD rates were 6%, 69%, 0% and 6%, respectively, while 19% were non-evaluable due to missing data and/or receiving less than 10 weeks of study treatment in the absence of disease progression. Among the 47 patients in the ofatumumab cohort, the CR/CRi, PR, SD and PD rates were 2%, 51%, 9% and 11%, respectively, with 28% being non-evaluable. Consequently, neither regimen met the predefined boundary for being of interest from an efficacy perspective. Of note, the 6% CR rate in the alemtuzumab cohort was substantially lower than the 36% CR rate observed haematologica | 2020; 105(12)

in the CLL206 trial1 which employed an 8-fold higher relative glucocorticoid dose. Kaplan-Meier curves for PFS and OS are shown in Figure 1. Despite the lower-than-expected CR rate in the alemtuzumab cohort, the 2-year PFS rate was surprisingly good at 58% (95% confidence interval [CI]: 27-91%). This compares with ~17% in the CLL206 trial, 12% in the previously treated cohort of CLL2O and 56% in the treatment-naĂŻve cohort of CLL2O (two thirds of whom received alemtuzumab maintenance or HSCT)1,2 and suggests that adding lenalidomide to alemtuzumab and dexamethasone may prolong PFS without increasing the CR rate. In contrast, the 2-year PFS rate in the ofatumumab cohort of CLL210 was only 30% (95% CI: 18-49%) with a striking difference between previously treated versus treatment-naĂŻve patients (9% and 52%, respectively). Two-year OS rates were higher for the alemtuzumab cohort compared to the ofatumumab one (79% vs. 57%). Our findings revealed interesting differences between the responses induced by the alemtuzumab and ofatumumab regimens. In addition to being more effective in terms of OR rate (75% vs. 53%), CR rate (6% vs. 2%), 2-year PFS (58% vs. 30%), and 2-year OS (79% vs. 57%), the alemtuzumab regimen produced much higher rates of blood MRD negativity (37% vs. 0%) and morphological bone marrow clearance (50% vs. 8% of responders). In contrast, the two regimens were comparably effective at clearing nodal and splenic enlargement (25% vs. 20% of 2869

Letters to the Editor Table 2. Summary of all grade ≥3 adverse events (AE) (reported as either serious AE [SAE] or non-serious AE) occurring with a frequency of >1%.

Toxicity

Lung infection Neutropenia Sepsis Infection, other Febrile neutropenia Neoplasms, other Anemia Hyperglycemia Hypophosphatemia Thrombocytopenia Upper respiratory infection Vomiting General, other Infusion related reaction Bronchial infection Infective enterocolitis Hyponatremia Hypercalcemia Hypokalemia Maculopapular rash Thromboembolic event Localized edema Laryngitis

Induction phase Alemtuzumab Ofatumumab group (n=16) group (n=47) 8 4 13 2 1 2 3 2 1 3 1 1 1

1 1 2

13 15 1 5 4 2 3 3 4 3 4 3 1 4 1 1 4 1 1 4

Post-induction phase Control arm Not randomized (n=9) (n=18)

3 3

1 1 6

3 2 3 2 3 1 2 1

2 1

1 1 1

patients, respectively). Twenty patients (5 from the alemtuzumab cohort and 15 from the ofatumumab cohort) were randomized to lenalidomide maintenance (11) versus placebo (9). The median duration of lenalidomide maintenance was 6 (IQR: 2-10) months. There was a non-significant trend for superior PFS in the lenalidomide arm compared to the control arm and HSCT group (Figure 1). However, these results should be interpreted with caution owing to the small number of patients in each group and the high post-induction drop-out rate. A total of 252 grade ≥3 adverse events (AE) were identified from SAE and non-serious AE reports, among which infections (83), hematological alterations (61) and metabolic disturbances (30) were the most common (Table 2). Grade ≥3 SAE were reported in 13 of 16 (81%) patients in the alemtuzumab cohort and 28 of 47 (60%) patients in the ofatumumab cohort. These included eight treatment-related grade 5 SAE, of which two were in the alemtuzumab cohort (one infection and one neoplasm) and six in the ofatumumab cohort (four infections, one hematoma and one visceral arterial ischemia). The intolerance rate was 0.67 (95% CI: 0.51-0.80) for the alemtuzumab cohort and 0.38 (95% CI: 0.30-0.46) for the ofatumumab cohort. Consequently, neither regimen met the predefined boundary for being of interest from a tolerability perspective. Neither of the two regimens evaluated in CLL210 compare favorably with newer drugs such as ibrutinib, idelalisib and venetoclax when applied as monotherapy to a 2870

Lenalidomide arm (n=11)

Total events 30 24 17 11 9 9 8 7 6 5 5 4 4 4 4 4 4 4 4 4 4 3 3

similar patient population. For example, ibrutinib produced a 2-year PFS rate of 85% in a retrospective study of 108 patients with treatment-naïve 17p-deleted CLL8 and 65% in a combined analysis of 230 patients with a 17p deletion who were recruited into three prospective clinical trials of relapsed/refractory CLL.9 Similarly, the 2-year PFS rate among 46 patients with a 17p deletion or TP53 mutation who were recruited into a prospective clinical trial of idelalisib in relapsed/refractory CLL was ~43%,10 while the 2-year PFS for venetoclax in the pivotal study of 158 patients with predominantly relapsed/refractory 17p-deleted CLL was 54%.11 In summary, although the NCRI CLL210 trial showed that lenalidomide and dexamethasone combined with either alemtuzumab of ofatumumab is feasible and active in high-risk CLL, the study did not meet the pre-specified dual primary endpoints. Furthermore, interest in glucocorticoid/ antibody combinations has now been eclipsed by the emergence of highly effective and well-tolerated novel agents that target BCL-2 or components of the B-cell receptor signalling pathway. Andrew R. Pettitt,1,2 Richard Jackson,1 Silvia Cicconi,1 Fotis Polydoros,1 Christina Yap,3 James Dodd,1 Matthew Bickerstaff,1 Michael Stackpoole,1 Umair T. Khan,1,2 Stacey Carruthers,1 Melanie Oates,1 Ke Lin,1,4 Sarah E. Coupland,1,4 Geetha Menon,1,4 Nagesh Kalakonda,1,4 Helen McCarthy,5 Adrian Bloor,6 Anna Schuh,7 Andrew Duncombe,8 Claire Dearden,9 Christopher Fegan,10 Ben Kennedy,11 Renata Walewska,5 Scott Marshall,12 haematologica | 2020; 105(12)

Letters to the Editor

Christopher P. Fox13 and Peter Hillmen14 1 University of Liverpool, Liverpool; 2Clatterbridge Cancer Center NHS Foundation Trust, Liverpool; 3University of Birmingham, Birmingham; 4Royal Liverpool & Broadgreen University Hospitals NHS Trust, Liverpool; 5The Royal Bournemouth Hospital, Bournemouth; 6The Christie NHS Foundation Trust, Manchester; 7 Churchill Hospital, Oxford; 8Southampton General Hospital, Southampton; 9Royal Marsden NHS Foundation Trust, Sutton; 10 University Hospital of Wales, Cardiff; 11Leicester Royal Infirmary, Leicester; 12City Hospitals Sunderland NHS Trust, Sunderland; 13 Nottingham University Hospitals NHS Trust, Nottingham and 14 University of Leeds, Leeds, UK Correspondence: ANDREW R. PETTITT - arp@liv.ac.uk doi:10.3324/haematol.2019.230805 Acknowledgments: we are grateful to Cancer Research UK for endorsing the study, to the NIHR for supporting local trial delivery, and to industry partners for funding trial co-ordination (Celgene, Chugai) and providing (Celgene, GSK/Novartis) or subsidising (Baxter, Chugai) investigational medical products.

3. 4.

5. 6. 7.

8. 9.

References 10. 1. Pettitt AR, Jackson R, Carruthers S, et al. Alemtuzumab in combination with methylprednisolone is a highly effective induction regimen for patients with chronic lymphocytic leukemia and deletion of TP53: final results of the national cancer research institute CLL206 trial. J Clin Oncol. 2012;30(14):1647-1655. 2. Stilgenbauer S, Cymbalista F, Leblond V, et al. Alemtuzumab com-

haematologica | 2020; 105(12)

11.

bined with dexamethasone, followed by alemtuzumab maintenance or allo-SCT in â&#x20AC;&#x153;ultra High-riskâ&#x20AC;? CLL: final results from the CLL2O phase II study. Blood. 2014;124(21):1991-1991. Oscier D, Dearden C, Eren E, et al. Guidelines on the diagnosis, investigation and management of chronic lymphocytic leukaemia. Br J Haematol. 2012;159(5):541-564. Arumainathan A, Kalakonda N, Pettitt AR. Lenalidomide can be highly effective in chronic lymphocytic leukaemia despite T-cell depletion and deletion of chromosome 17p. Eur J Haematol. 2011; 87(4):372-375. Riches JC, Gribben JG. Mechanistic and clinical apects of lenalidomide treatment for chronic lymphocytic leukemia. Curr Cancer Drug Targets. 2016;16(8):689-700. Wierda WG, Kipps TJ, Mayer J, et al. Ofatumumab as single-agent CD20 immunotherapy in fludarabine-refractory chronic lymphocytic leukemia. J Clin Oncol. 2010;28(10):1749-1755. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008;111(12):5446-5456. Mato AR, Roeker LE, Allan JN, et al. Outcomes of front-line ibrutinib treated CLL patients excluded from landmark clinical trial. Am J Hematol. 2018;93(11):1394-1401. Jones J, Mato A, Coutre S, et al. Evaluation of 230 patients with relapsed/refractory deletion 17p chronic lymphocytic leukaemia treated with ibrutinib from 3 clinical trials. Br J Haematol. 2018; 182(4):504-512. 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):1391-1402. Stilgenbauer S, Eichhorst B, Schetelig J, et al. Venetoclax for patients with chronic lymphocytic leukemia with 17p deletion: results from the full population of a phase II pivotal trial. J Clin Oncol. 2018;

2871

Letters to the Editor

Clinical characteristics and outcome of multiple myeloma patients with concomitant COVID-19 at Comprehensive Cancer Centers in Germany The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was discovered in Wuhan in December 2019 where it quickly led to a severe outbreak. In just a few weeks it evolved into a pandemic, with >3 million confirmed cases and >220,000 deaths attributed to COVID-19. Around 15% of infected people develop severe symptoms requiring hospitalization, and 3-10% of patients subsequently succumb to COVID-19, often due to acute respiratory distress syndrome (ARDS).1–3 Currently, neither specific treatment for COVID-19 nor a vaccine against SARS-CoV-2 is available, while trials evaluating pharmacological interventions are ongoing and specific risk groups are being defined. Whether cancer in general and hematological malignancies in particular bear substantial risks, is of eminent interest, as these patients receive immune-modulatory treatment. Of additional concern is the risk for severe COVID-19 disease in hematological malignancies given cytotoxic chemotherapy, including novel agents such as immunomodulatory drugs (IMiD) and immunotherapies (i.e., monoclonal antibodies such as daratumumab, elotuzumab and others).2,3 It is also unclear, whether the immunosuppressant state of cancer patients predisposes them to a severe COVID-19 disease or, if a diminished host immune response may decrease the risk of multiorgan complications. Early reports from China and others indicate that cancer patients with COVID-19 may have a higher risk of a severe disease course and less favorable outcome compared to non-cancer patients, albeit not shared by others and if diligently compared with non-cancer cohorts.4–10 Most of these studies are, as yet, hampered by sample sizes, and patient cohorts consist mainly of patients with solid tumors.2–7 Other risk factors associated with adverse outcome seem to be nearly the same as for non-cancerpatients: advanced age, male sex, presence of substantial comorbidities (neurological: advanced Alzheimer, Parkinson, cardiovascular disease, diabetes mellitus, respiratory disease, chronic kidney disease, liver disease and immunosuppression).11,12 In this German Multiple Myeloma (MM) Study Group Consortium (DSMM and GMMG), we aimed to characterize a population of MM patients registered from 10 institutions who developed COVID-19 at hotspot areas in Germany. All MM patients with concomitant SARSCoV-2 infection were treated at secondary and tertiary Comprehensive Cancer Centers (CCC). Our goal was to determine whether COVID-19 in MM patients resulted in a greater morbidity and mortality compared to prior COVID-19 reports in cancer and specifically in MM patients. We performed a retrospective multicenter DSMM/GMMG cohort study involving 10 secondary and tertiary CCC in German pandemic epicenters. From March 1 to May 31, 2020, all MM in- and outpatients with concomitant SARS-CoV-2 infection were included (registered via DSMM/GMMG-performed incentive). SARS-CoV-2 infection was confirmed by reverse-transcriptase-PCR (RT-PCR) assay (Table 1A-B). Laboratory findings and radiological data were retrieved from the electronic medical records of each center. Comparative analysis of prior reports in cancer and MM patients was additionally performed as summarized in Table 2. The study was conducted in accordance and in compliance with the Declaration of Helsinki and International 2872

Conference on Harmonization Guidelines for Good Clinical Practice. The data cutoff date was May 31, 2020. Descriptive data with median and ranges are presented. Our cohort of 21 MM patients encompassed 81% males and a median age of 59 years (range: 46-83). Most were Caucasians (Table 1A-B). Their median Karnofsky Performance Status (KPS) was 80% and the median number of comorbidities was 1 (0-3), the most prevalent being cardiovascular (hypertension), renal impairment and others (polyneuropathy [PNP], diabetes; Table 1A). Smokers versus non-smokers comprised 4 (19%) versus 17 (81%) of patients, respectively. In line with the patients' KPS and comorbidities, the median revised myeloma comorbidity index (R-MCI)13,14 and International Myeloma Working Group (IMWG)frailty scores14 were 3 (=fit) and 1 (=intermediate-fit), respectively. Nonetheless, 48% (via R-MCI) and 62% (via IMWG-frailty score) were in the intermediate-fit/frail group (Table 1A). The median time from MM diagnosis to SARS-CoV-2 infection was 20 months (0-142). MM patients had a median of 1 (0-4) prior line of therapy. Fifteen (71%) patients had a prior autologous stem cell transplantation (ASCT). Preceding anti-myeloma treatments were proteasome inhibitors (PI) in 19 of 21 (90%; in all except both patients with newly diagnosed [IDMM]), IMiD in 12 of 21 (57%) and antibodies (daratumumab, elotuzumab, isatuximab) in 10 of 21 (48%) patients (Table 1A). The most common MM subtype was IgG (67%), followed by IgA (24%) and light-chain (LC)-only MM in 9% of the patients. High-risk cytogenetics (del17p, t(4;14), t(14;16)) were present in six (29%) patients. Seven (33%), 19 (48%) and four (19%) patients had an international staging system (ISS) of 1, 2 and 3, respectively. A median of two CRAB criteria had led to anti-myeloma treatment, in line with prior reports in MM without SARS-CoV-2 infection.15 The disease status at the time of SARS-CoV-2 infection included six patients in complete remission (CR), three in very good partial remission (VGPR), 10 in partial remission (PR) and two patients with IDMM. At the time of SARS-CoV-2 infection, 12 (57%) were being treated, either with daratumumab (5), elotuzumab (1), VCd (2), KRd (1) or lenalidomide-maintenance (3). Nine patients were not on active anti-MM therapy. At the time of SARS-CoV-2 infection, anti-MM treatment was transiently stopped for ~4 weeks in all of them (Table 1A). The most common reported symptoms among all patients were a cough (81%) and fever (76%). Notably, two patients were almost asymptomatic. Seventeen (81%) patients were admitted to the hospital for inpatient care. The median time between self-reported symptom onset and admission was 3 days. Three patients required intensive care unit (ICU) support, all were treated with high-flow oxygen, and two were eventually intubated. These two developed ARDS, with one already fully recovered. The median time to recovery in all patients from symptom onset was 17 days and from test positivity 14 days. There were no deaths in the total cohort. Pulmonary infiltrates via computer tomography (CT) scans were present in 18 (86%) patients. Blood counts at SARS-CoV-2 infection showed absolute neutrophil counts (ANC) of 2.9x109/L, whereas median absolute lymphocyte counts (ALC) were suppressed (0.8x109/L). On initial presentation, platelets, lactate dehydrogenase (LDH), creatinine, PTT and D-dimer were normal or less compromised, whereas C-reactive protein (CRP) and ferritin were elevated (Table 1B). The number of patients haematologica | 2020; 105(12)

Letters to the Editor

with 1, 2 or all 3 paraprotein subclass suppression (immunoparesis) was substantial with six, nine and three patients, thus immunoparesis of >1 subclass was broadly present in 86% of patients. Our COVID-19 management involved most frequently intravenous (iv) antibiotics in 15 patients and orally in two patients. Additionally, four patients received azithromycin and seven patients hydroxychloroquine. Remdesivir, tocilizumab and anakinra were given in one patient each who had a more severe and/or prolonged SARS-CoV-2 infection. In four patients with mild SARSCoV-2 infection, only supportive care was required. The

median time of inpatient treatment was 14 days (range: 3-52). Approximately 4 weeks after full COVID-19 recovery, anti-MM treatment was restarted in eight patients: five patients with lenalidomide, one patient with KRd continuation with relapsed/refractory MM (RRMM), two IDMM patients received VCD and Dara-VRd induction after full SARS-CoV-2 recovery and were confirmed as SARS-CoV-2 negative by RT-PCR. We also performed a descriptive subgroup analysis of patients in need (n=3) versus not in need (n=18) of ICU support (Online Supplementary Table S1), which revealed,

Table 1A. Patient characteristics (n=21).

Parameters Patient characteristics Median age (years; range) Ethnicity, caucasian : asian Male : female Median Karnofsky Performance Status, % (range) Median number of comorbidities (range) none cardiac/hypertension renal impairment obesity other: PNP, diabetes, hypothyreosis Smoking : non-smoking Median R-MCI (range) # of fit patients # of intermediate patients # of frail patients Median IMWG-frailty score (range) # of fit patients # of intermediate patients # of frail patients MM disease characteristics MM type: IgG, IgA, LC only ISS stage: I : II : III Cytogenetic risk SR : HR Median # of CRAB criteria (range) MM disease status at SARS-CoV-2 infection ID CR VGPR / PR Number of prior anti-MM lines (range) Prior transplant, yes : no Prior anti-MM treatment with PI IMID antibody (daratumumab/isatuximab, elotuzumab) Receiving active anti-MM treatment, yes : no Anti-MM treatment at COVID-19 diagnosis daratumumab-combination elotuzumab-combination VCd / KRd R-maintenance no MM therapy Median time of SARS-CoV-2 infection since MM ID, ms (range) Anti-MM treatment discontinuation at SARS-CoV-2 infection

Median

Number of patients

59 (46-83) 20 : 1 17 : 4 80 (30-100) 1 (0-3) 4 11 3 1 4 4 : 17 3 (0-7) 11 7 3 1 (0-3) 8 6 7 14 : 5 : 2 7 : 10 : 4 15 : 6 2 (0-4) 2 6 3 / 10 1 (0-4) 15 : 6 19 (90%) 12 (57%) 10 (48%) 12 : 9 5 1 2/1 3 9 20 (0-142) all (100%)

PNP: polyneuropathy; R-MCI: revised myeloma comorbidity index; IMWG: International Myeloma Working Group; MM: multiple myeloma; LC: light-chain secreting MM; ISS: international staging system; SR: standard risk; HR: high risk del17p, t(4;14), t(14;16); SR: all others; CRAB: hypercalcemia, anemia, renal impairment, bone disease; CR: complete response; VGPR/PR: very good partial response/partial response; ASCT: autologous stem cell transplantation; PI: proteasome inhibitors; IMID: immunmodulatory drugs; VCd/KRd: bortezomib-cyclophosphamid-dexamethasone/carfilzomib-lenalidomide-dexamethasone; R: lenalidomide; ID: initial diagnosis of MM; ms: months.

haematologica | 2020; 105(12)

2873

Letters to the Editor

that the former patients were older (60 vs. 58 years), had more comorbidities (2 vs.1; KPS: 60% vs. 90%, R-MCI: 6 vs. 3, IMWG-frailty score: 2 vs. 1), had a longer time span from symptom onset to testing (5 vs. 2 days), lower ALC counts (0.5 vs. 0.9x109/L), highly elevated inflammatory parameters and needed substantially longer to recover (from symptom onset: 55 vs. 16 days; from COVID-19test: 52 vs. 14 days, respectively). Even though the ICUgroup predictably needed more intensified SARS-CoV-2

treatment support, the outcome in both groups was identical with 100% survival (OS; Online Supplementary Table S1). Table 2 summarizes seven selected MM/cancer-studies with SARS-CoV-2 infection. Remarkable from our data was the death rate of 0%, whilst MM disease features and characteristics were similar to those of real-life UKand US-MM-cohorts with different health systems.12,16 Even though the median patient age in both UK- and US-

Table 1B. COVID-19 disease characteristics (n=21).

Median (range)

Number of patients (%)

Symptoms & diagnostics Symptoms at SARS-CoV-2 infection cough fever myalgia GI asymptomatic Pulmonary infiltrates via CT, yes : no Laboratory parameters ANC (x109/L) ALC (x109/L) platelets (x109/L) LDH (U/L) CRP (mg/L) procalcitonin (ng/mL) ferritin (mg/L) creatinin clearance (ml/min) PTT (s) D-dimer (mg/L) Immunoparesis, 0 : 1 : 2 : 3 Immunoparesis â&#x2030;Ľ1 subclass (%)

17 16 4 2 2 18 : 3 2.9 (1.2-7.5) 0.8 (0.2-3.1) 150 (93-300) 200 (160-560) 50 (3-500) 0.1 (0.1-20) 100 (30-8,344) 75 (15-117) 30 (27-50) 0.1 (0.1-8.7)

Outcome Median time of symtoms to test (days; range) 3 (0-9) Median time to recovery (days; range) from symptom onset 17 (7-61) from test positivity 14 (6-52) Inpatient: outpatient treatment Median time of inpatient treatment (days; range) 14 (3-52) Requirement of ICU, yes : no ventilation, yes : no ARDS, yes : no Alive status COVID-19 treatment po antibiotics / iv antibiotics plus Azithromycin Hydroxychloroquin Remdesivir Tocilizumab Anakinra supportives only MM-therapy restart within ~4 weeks after COVID-19 recovery: yes : no MM-treatment restarted after SARS-CoV-2 infection with R-maintenance KRd continuation induction (VCD / Dara-VRD)

3:6:9:3 18 / 21 (86)

17 : 4

3 : 18 2 : 19 2 : 19 21 / 21 (100) 2 / 15 4 7 1 1 1 4 8 : 13 5 1 1/1

GI: gastrointestinal symptoms; CT: computer tomography scan; ANC: absolute neutrophile count; ALC: absolute lymphocyte count; LDH: lactat dehydrogenase; CRP: C-reactive protein; PTT: prothrombine time; ICU: intensive care unit; ARDS: acute respiratory distress syndrome; po: oral, iv: intravenous; R: Lenalidomide, VCD: Bortezomib/ Cyclophosphamide/Dexamethasone; Dara-VRD: Daratumumab/Bortezomib/ Lenalidomide/ Dexamethasone; KRd: Carfilzomib/Lenalidomide/Dexamethasone.

2874

haematologica | 2020; 105(12)

Letters to the Editor

cohorts was higher than ours and both had more nonwhites than typical in Germany, lines of prior therapy, ASCT-rates, median number of comorbidities and ALC/immunoparesis frequencies were comparable, suggesting that with maximum triage support and supportive care, MM/cancer patients per se may not have a dismal SARS-CoV-2 infection outcome.2,3,8â&#x20AC;&#x201C;10 Early data in cancer patients in general5,17,18 have suggested higher risks in those with systemic/chemotherapeutic treatment within 4 weeks prior to symptom onset, male gender, and poor constitution, whereas our experience in 39 cancer patients with SARS-CoV-2 infection compared with very well matched non-cancer controls suggests no difference in OS for both groups (K. Shoumariyeh et al., Cancer Medicine 2020; in print). This is in line with a prospective observational UK study of 800 cancer patients9 that was unable to identify evidence that cancer patients on cytotoxic chemotherapy or other anticancer treatment are at an increased risk of mortality from COVID-19 disease compared with those not on active treatment. Mortality from COVID-19 in these cancer patients appeared to be principally driven by age, sex and

comorbidities, as well as by cancer patients' access to intensive care support/ ICU.9 In conclusion, we here present the examined clinical and laboratory parameters of 21 MM patients who were treated in 10 CCC with confirmed SARS-CoV-2 infection. The features of these patients were comparable to others, except that our patients were younger.12,16 We have previously described that younger MM patients may come to German tertiary centers, even though "stage and age migration" is occurring,19,20 since elderly and frail patients have recognized the advantages of university centers'/CCC' support offers. Moreover, since all regional German hospitals have been enthusiastically dealing with COVID-19 patients, we were not the only centers, where COVID-19 patients were treated, which may be different to other more centralized medical systems. In all three MM analyses (Table 2),12,16 patient and MM disease characteristics were similar: notable was our male predominance (81%, Cook: 60%,16 Wang 52%12), the median prior lines of anti-MM treatment were in all 1, ALC and immunoparesis were present in substantial per-

Figure 1. Time point, when multiple myeloma (MM) patients aquired COVID-19 infection, key differences of patients with (w) versus without (w/o) intensive care (ICU) needs and learning experience of COVID-19 infection in MM. (A) Time point, when multiple myeloma (MM) patients acquired COVID-19 during their disease course, which was rather early than later and at a median of 20 months after the initial diagnosis (ID). (B) Key differences of patients w versus w/o ICU needs, who expectedly showed a worse Karnofsky performance status (KPS) in those ICU patients and who also had more comorbidities, higher R-MCI and IMWG-scores. These ICU patients also seemed to show a slightly longer time span from symptoms to COVID-19-testing and a lower absolute lymphocyte count (ALC). (C) Learning experience of possible risks for COVID-19 infection is displayed, who were in our series males in 81%, intermediate-fit or frail via R-MCI and IMWG scores in 48% and 62%, respectively and acquired COVID-19 infection fairly early during the MM course. This might be due to a "loss" of more vulnerable patients earlier (first 2 years after ID), so that later in the MM disease course, only more stoic or inured patients may survive who may be less prone to COVID-19 infection. Moreover, COVID-19 infection occurred in both patients on active MM treatment and in those without. We performed a MM therapy pause in all patients, both in milder or more substantial cases and observed no death. We recommend to stay vigilant on COVID-19, observe all precautions as described,2,3 avoid undertreatment and follow-up losses, restart anti-MM therapy with full COVID-19 recovery and encourage our MM patients and experts, that with maximum support, COVID-19 may not lead to a more dismal outcome compared to the general population.9,10

haematologica | 2020; 105(12)

2875

Letters to the Editor

centages and median comorbidities were 1 or 2.12,16 Nevertheless country-specific differences in death rates varied from 0% in ours to 54.8% in IDMM and 50% in RRMM in the UK cohort (non-MM UK mortality: 14.5%) and 24% in the US-cohort.12,16 Compared to non-cancer cohorts,4,9,21,22 MM patients seemed to show a longer duration to clinical improve-

ment (time from onset of symptoms to recovery) and longer hospitalization time of 17 and 14 days, respectively, even though intensive supportive care led to recovery. Therefore MM patients did not show a more dismal OS compared to non-cancer patients,2,3,8 which needs to be confirmed in larger analyses, with a cohort study from five New York academic centers already confirming

Table 2. Different COVID-19 outcome reports in multiple myeloma and other cancer patients.

Author

Engelhardt M et al.13

# pts Cancer Males Median type, time median from age diagnosis (range) (ms) MM, 59 y 81% (46-83)

0.8

86%

Cook G et al.16 75

MM, 73 y (47-88)

60%

0.6

90%

Wang B et al.12 58

MM, 67 y

52%

1.5

232

Cancer, 64 y (58-69)

51%

differed due to various cancers

n.g.

0.8

Yang K et al.18 205

Cancer, 63 y (56-70)

47%

differed due to various cancers

n.g.

0.99

n.g.

20%

various

Lee L et al.9

800

Cancer, 69 y (59-76)

56%

differed

n.g

n.g.

28%

various

Shoumariyeh K 39 et al. (in print)

Cancer, 73 y (54-95)

56%

differed

n.a.

0.6

n.a.

Tian J et al.17

Median Median Immuno- Death Median # Insights and risks prior ALC paresis rate comorbidities therapy count/mL lines 0%

IDMM: 54.8% RRMM: 50% (non-MM UK: 14.5%) 89% 24%

20% various (non-cancer: 11%)

21% 1 (age-matched COVID-controls: 36%)

COVID-19 in all disease stages, during active and non active treatment; leading to stop of SACT in all pts during SARS-CoV-2-infection and in 8/21 resumption after recovery Median pt age who died et al. significantly higher (78 vs. 66 ys) +greater level of comorbidities Risks for hospitalization: older age (>70y), males + cardiovascular risk race, severe

Highest risk for severity + death if last CTx within 2 wks of admission. Lower risk of severity when cancer diagnosis >1 year Hematological malignancies had poorer prognosis than solid tumors (but more hematological pts received CTx within 4 wks before symptom onset) No increased risk of COVID-19 mortality in cancer pts

Solid + hematological cancer vs. age-matched controls: no OS difference

Conclusions

With maximum supportive treatment: favorable outcome possible

Higher mortality than UK mortality rate for COVID-19

Risks for increased mortality: non-white hypogammaglobulinemia + elevated inflammatory markers associated with death. 17% pts with 5+ lines of therapy. Risks of COVID-19 severity: tumor stage↑, TNF-α↑, NTproBNP↑, CD4+ T cells↓, albumin-globulin ratio↓ Risks for death: CTx 4 wks before symptom onset and male sex

Mortality driven by age, gender, comorbidities, ICU admission low: 6% -> questions whether cancer decreases access to ICU support No per se risk of cancer for dismal COVID-19-outcome

MM: multiple myeloma; pts: patients; ms: months; y: years; ALC: absolute lymphocyte count; SACT: systemic anticancer therapy; CTx: chemotherapy; n.g.: not given; n.a.: not applicable; ICU: intensive care unit; OS: overall survival; wks: weeks; ms: months.

2876

haematologica | 2020; 105(12)

Letters to the Editor

this.10 In line, other centers have reported that MM patients have done remarkably well overall and that the type of chemotherapy being administered just prior to COVID-19 had no bearing on the outcome (personal communication: Dr. Sundar Jagannath, New York, International Myeloma Working Group [IMWG] and9,10). Although one has to be very careful with worldwide comparisons, one strategy in Germany had been to allocate all possible resources to medical centers during the COVID-19 pandemic, so that COVID-19 patients had utmost medical support, including sufficient intensive care facilities and no onrush of patients. Thus pandemicinduced overwhelming of capacities was effectively avoided, whereas if this cannot be ensured, mortality is likely to be higher (personal communication S.Jagannath and IMWG).9,10 Relevant learning experience is summarized in Figure 1. We are well aware, that our observations have to be interpreted with caution, because our experience is as yet limited to 10 CCC in Germany. We refrained from larger subgroup analysis, because all our patients survived and we had only 3 of 21 and 2 of 21 patients who required ICU support, ventilation/intubation or had ARDS, respectively. Nevertheless, our observation of a very reassuring outcome seems to be highly encouraging, as we are aiming to continue utmost cancer care and avoid "undertreatment", appointment stops and delay of follow-ups. Our results seem also soothing to those, who are greatly threatened by SARS-CoV-2 infections and who believe that cancer patients may fare far worse, whereas an opposite observation - as shown here - may be true and is worth to be spread to our patients and colleagues (Figure 1). Monika Engelhardt,1,2 Khalid Shoumariyeh,1,2 Amelie Rösner,1 Gabriele Ihorst,3 Francesca Biavasco,1,2 Katharina Meckel,4 Ivana von Metzler,5 Sebastian Theurich,6 Holger Hebart,7 Matthias Grube,8 Miriam Kull,9 Florian Bassermann,10 Kerstin Schäfer-Eckart,11 Anette Hoferer,12 Hermann Einsele,4 Leo Rasche4 and Ralph Wäsch1,2 1 Freiburg University Medical Center, Department of Internal Medicine I, Faculty of Medicine, University of Freiburg, Freiburg; 2 German Cancer Consortium (DKTK) Heidelberg, Heidelberg; 3 Clinical Trials Unit Freiburg, Faculty of Medicine, University of Freiburg, Freiburg; 4Medizinische Klinik und Poliklinik II, JuliusMaximilian-University of Würzburg, Würzburg; 5Universitätsklinik Frankfurt, Department of Hematology and Oncology, University of Franfurt, Frankfurt; 6LMU University Hospital München, Department of Internal Medicine III, Ludwig-Maximilian-University München, München; 7Kliniken Ostalb, Stauferklinikum, Mutlangen; 8 Universitätsklinik Regensburg, Department of Hematology and Oncology, University of Regensburg, Regensburg; 9Universitätsklinik Ulm, Department of Hematology and Oncology, Univeristy of Ulm, Ulm; 10Florian Bassermann Klinikum rechts der Isar, Department of Hematology and Oncology, München; 11Klinikum Nürnberg, Department of Hematology and Oncology, Nürnberg and 12Robert-Bosch Krankenhaus, Stuttgart, Germany Correspondence: MONIKA ENGELHARDT monika.engelhardt@uniklinik-freiburg.de doi:10.3324/haematol.2020.262758 Acknowledgments: ME, KS and RW specifically thank Prof. Dr. Winfried Kern (Division of Infectious Diseases, University of Freiburg (UKF)) and Prof. Dr. Hartmut Hengel (Institute of Virology, UKF) for their outstanding knowledge, support and cooperation in management of the UKF SARS-CoV-2 infected patients. ME was redeployed to general and acute medicine roles to support the COVID-19 admissions and in charge of a UKF COVID-19haematologica | 2020; 105(12)

ward. We thank our outstanding study coordinating and study nurse teams at all CCC for their exceptional support, great care and diligent documentation of study patients, including those with COVID-19 infection and to all our colleagues for the team spirit during the COVID-19 challenge, its pressure and during vulnerable times at pandemic hotspot areas of Germany. We also thank the UK, US and worldwide MM and cancer colleagues, including GMMG/DSMM members greatly, who we discussed our data with. Moreover, we are greatly indebted to Prof. Dr. Justus Duyster's prodigious support and the UKF-support in SARS-CoV-2 times in general, with specific gratitude to Prof. Dr. Frederik Wenz, Prof. Dr. Norbert Südkamp and Prof. Dr. Christoph Peters (all UKF).

References 1. Wu Z, McGoogan JM. Characteristics of and important lessons from the Coronavirus Disease 2019 (COVID-19) outbreak in China: summary of a report of 72 314 cases from the chinese Center for disease control and prevention. JAMA. 2020;323(13):1239-1242. 2. Terpos E, Engelhardt M, Cook G, et al. Management of patients with multiple myeloma in the era of COVID-19 pandemic: a consensus paper from the European Myeloma Network (EMN). Leukemia. 2020;34(8):2000-2011. 3. Mian H, Grant SJ, Engelhardt M, et al. Caring for older adults with multiple myeloma during the COVID-19 pandemic: perspective from the International Forum for optimizing care of older adults with Myeloma. J Geriatr Oncol. 2020;11(5):764-768. 4. Guan WJ, Ni ZY, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in China. N Engl J Med. 2020;382(18):1708-1720. 5. Liang W, Guan W, Chen R, et al. Cancer patients in SARS-CoV-2 infection: a nationwide analysis in China. Lancet Oncol. 2020; 21(3):335-337. 6. Zhang L, Zhu F, Xie L, et al. Clinical characteristics of COVID-19infected cancer patients: a retrospective case study in three hospitals within Wuhan, China. Ann Oncol. 2020;31(7):894-901. 7. Dai M, Liu D, Liu M, et al. Patients with cancer appear more Vulnerable to SARS-CoV-2: a Multicenter Study during the COVID19 outbreak. Cancer Discov. 2020;10(6):783-791. 8. Omarini C, Maur M, Luppi G, et al. Cancer treatment during the coronavirus disease 2019 pandemic: Do not postpone, do it! Eur J Cancer. 2020;133:29-32. 9. Lee LYW, Cazier JB, Starkey T, Turnbull CD, Kerr R, Middleton G. COVID-19 mortality in patients with cancer on chemotherapy or other anticancer treatments: a prospective cohort study. Lancet. 2020;395(10241):1919-1926. 10. Hultcrantz M, Richter J, Rosenbaum C, et al. COVID-19 infections and outcomes in patients with multiple myeloma in New York City: a cohort study from five academic centers. medRxiv. 2020 Jun 11. [Epub ahead of print] 11. von Lilienfeld-Toal M, Vehreschild JJ, Cornely O, et al. Frequently asked questions regarding SARS-CoV-2 in cancer patients-recommendations for clinicians caring for patients with malignant diseases. Leukemia. 2020;34(6):1487-1494. 12. Wang B, Van Oekelen O, Mouhieddine TH, et al. A tertiary center experience of multiple myeloma patients with COVID-19: lessons learned and the path forward. J Hematol Oncol. 2020;13(1):94. 13. Engelhardt M, Domm A-S, Dold SM, et al. A concise revised Myeloma Comorbidity Index as a valid prognostic instrument in a large cohort of 801 multiple myeloma patients. Haematologica 2017;102(5):910-921. 14. Dold SM, Möller M-D, Ihorst G, et al. Validation of the revised myeloma comorbidity index and other comorbidity scores in a multicenter German study group multiple myeloma trial. Haematologica. 2020 May 15. [Epub ahead of print] 15. Gengenbach L, Reinhardt H, Ihorst G, et al. Navigating the changing multiple myeloma treatment landscape: clinical practice patterns of MM patients treated in- and outside German DSMM study group trials. Leuk Lymphoma. 2018;59(11):2692-2699. 16. Cook G, Ashcroft AJ, Pratt G, et al. Real-world assessment of the clinical impact of symptomatic infection with severe acute respiratory syndrome coronavirus (COVID-19 disease) in patients with Multiple Myeloma receiving systemic anti-cancer therapy. Br J Haematol. 2020 May 21. [Epub ahead of print] 17. Tian J, Yuan X, Xiao J, et al. Clinical characteristics and risk factors associated with COVID-19 disease severity in patients with cancer in Wuhan, China: a multicentre, retrospective, cohort study. Lancet Oncol. 2020;21(7):893-903. 18. Yang K, Sheng Y, Huang C, et al. Clinical characteristics, outcomes, and risk factors for mortality in patients with cancer and COVID-19

2877

Letters to the Editor

in Hubei, China: a multicentre, retrospective, cohort study. Lancet Oncol. 2020;21(7):904-913. 19. Hieke S, Kleber M, KĂśnig C, Engelhardt M, Schumacher M. Conditional survival: a useful concept to provide information on how prognosis evolves over time. Clin Cancer Res. 2015; 21(7):1530-1536. 20. Schinke M, Ihorst G, Duyster J, WĂ¤sch R, Schumacher M, Engelhardt M. Risk of disease recurrence and survival in patients with multiple myeloma: a German Study Group analysis using a

2878

conditional survival approach with long-term follow-up of 815 patients. Cancer. 2020;126(15):3504-3515. 21. Huang C, Wang Y, Li X, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet. 2020;395(10223):497-506. 22. Lu H, Stratton CW, Tang Y-W. Outbreak of pneumonia of unknown etiology in Wuhan, China: The mystery and the miracle. J Med Virol. 2020;92(4):401-412.

haematologica | 2020; 105(12)

Letters to the Editor

The prognosis of older patients with newly diagnosed multiple myeloma (NDMM) who are not eligible for stem cell transplantation has greatly improved as a result of treatment with either a combination of bortezomib with lenalidomide, or the addition of daratumumab to bortezomib or to lenalidomide.1-3 Although not compared head-to-head, progression-free survival (PFS) was longer with lenalidomide used continuously as compared to bortezomib for a limited number of cycles only.4,5 Continuous treatment with the oral proteasome inhibitor, ixazomib, in combination with lenalidomide and dexamethasone (IRd) did not increase the incidence of grade â&#x2030;Ľ3 neuropathy as compared to lenalidomide and dexamethasone (Rd) only.6 A possible additional advantage of continuous therapy with this proteasome inhibitor is that it may overcome the negative impact of high-risk cytogenetic abnormalities.7 Furthermore, there was no need for discontinuation of ixazomib due to toxicity during the maintenance phase, whereas approxi-

mately 25% of patients discontinued lenalidomide treatment.8,9 Therefore, in this randomized phase II trial (registered at www.trialregister.nl as NTR4910), we investigated the efficacy and feasibility of ixazomib versus placebo maintenance in transplant-ineligible patients with NDMM, after nine cycles of induction with ixazomib, thalidomide and dexamethasone (ITd) (protocol details in the Online Supplementary Appendix). To improve knowledge about unfit and frail patients, we investigated the outcome in such patients, using a simplified frailty score.10 We did not observe an improvement in PFS with maintenance treatment with ixazomib compared to placebo. However, in the elderly population, including 44% of frail patients, only 55% of patients could be randomized after induction therapy. Importantly, for those patients who were randomized, ixazomib maintenance was very well tolerated and the PFS was comparable in patients >75 versus â&#x2030;¤75 years and in frail versus unfit or fit patients. The characteristics of the 143 eligible patients are presented in Table 1. According to the simplified frailty score, using World Health Organization (WHO) performance status as a replacement for (instrumental) activities of daily living, 33 (23%) of patients were classified as fit, 38 (27%) as unfit and 63 (44%) as frail; the frailty score was unknown for 6%. A total of 78/143 (55%) patients were randomized to maintenance treatment with either

P=0.92 (adjusted)

P=0.39 (adjusted)

Figure 1. Survival from randomization and registration. (A, B) Progression-free survival (A) and overall survival (B) after randomization (PFS-R and OS-R, respectively). With a median follow-up of 23.4 months after randomization (range, 6.9-35.5), the median PFS-R for patients treated with ixazomib was 9.5 months (95% confidence interval [95% CI]: 5.5-24.0) versus 8.4 months (95% CI: 3.0-13.8) for those given the placebo. The OS-R at 18 months for all patients was 96% (88-99%), with the value being comparable for patients treated with ixazomib (ixa, 100%) or placebo (92% [95% CI: 77-97%], P=1.00). (C, D) Progression-free survival (C) and overall survival (D) after registration (PFS and OS, respectively). With a median follow-up of 28.5 months (range, 0.9-44.1), the median PFS from registration for all patients was 14.3 months (95% CI: 11.5-16.8). The median OS from registration for all patients has not yet been reached.

haematologica | 2020; 105(12)

2879

Letters to the Editor

Table 1. Demographics of patients at registration and at randomization.

Total, n (%) Median age (range), years > 75 years n (%) > 80 years n (%) WHO performance months, n (%) 0 1 2 3 Frailty score, n (%) Fit Unfit Frail Unknown International Staging System, n (%) I II III Unknown LDH level, n (%) Normal Elevated Unknown FISH analysis done, n (%) t(4;14) del(17p) High-risk cytogenetics Standard-risk cytogenetics

Induction ITd

Placebo

Maintenance Ixazomib

% randomized

143 73 (64-90) 52 (36) 11 (8)

39 73 (67-82) 12 (31) 1 (3)

39 72 (66-80) 9 (23) -

78/143 (55) 73 (66-82) 21/52 (40) 1/11 (9)

51 (36) 59 (41) 32 (22) 1 (1)

20 (51) 13 (33) 6 (15) -

15 (38) 15 (38) 8 (21) 1 (3)

35/51 (68) 28/59 (47) 14/32 (44) 1/1 (100)

33 (23) 38 (27) 63 (44) 9 (6)

13 (33) 14 (36) 12 (31) -

12 (31) 9 (23) 14 (36) 4 (10)

25/33 (76) 23/38 (61) 26/63 (41) 4/9 (44)

28 (20) 72 (50) 42 (29) 1 (1)

9 (23) 19 (49) 11 (28) -

10 (26) 21 (54) 8 (21) -

19/28 (68) 40/72 (56) 19/42 (45) 0/1 (0)

121 (85) 18 (13) 4 (3) 132 (92) 12/128 (9) 17/127 (13) 27/123 (22) 96/123 (78)

35 (90) 3 (8) 1 (3) 34 (87) 5/34 (15) 2/34 (6) 6/34 (18) 28/34 (82)

31 (79) 6 (15) 2 (5) 36 (92) 1/3 (3) 5/35 (14) 6/33 (18) 27/34 (79)

66/121 (55) 9/18 (50) 3/4 (75) 70/132 (53) 6/12 (50) 7/17 (41) 12/27 (44) 55/96 (57)

ITd: ixazomib, thalidomide and dexamethasone; n: number; WHO: World Health Organization; LDH: lactate dehydrogenase; FISH: fluorescence in situ hybridization. *Frailty score: based on age, WHO performance status and comorbidities as defined by the Charlson Comorbidity Index;

ixazomib (n=39) or placebo (n=39). Patients who were randomized were younger and less frail at registration (Table 1). The median PFS from randomization (PFS-R) was 9.5 months (95% confidence interval [95% CI]: 5.524.0) for ixazomib-treated patients versus 8.4 months (95% CI: 3.0-13.8) for those given placebo (Figure 1A). The lack of difference in PFS-R between arms was independent of age, frailty, cytogenetics and best response on ITd induction (Online Supplementary Figure S1A). Importantly, although patients who were older, frail or had high-risk cytogenetics were less likely to reach randomization, those patients who were randomized for maintenance therapy experienced a similar outcome compared to younger, non-frail and standard-risk patients (Online Supplementary Figure S1B-D). The median overall survival (OS) from randomization has not been reached in either arm and was comparable in the two arms (Figure 1B). The median second progression-free survival (PFS2) from randomization has not been reached yet, since the PFS2 rate was 83% at 18 months, and was comparable in the two arms (Online Supplementary Figure S2). The median PFS from registration for all patients was 14.3 months (95% CI: 11.5-16.8) (Figure 1C). Subgroup analyses showed a comparable PFS in patients with high 2880

versus standard cytogenetic risk (median 12.0 vs. 14.6 months, respectively; P=0.11), patients aged â&#x2030;¤75 versus >75 years (14.3 vs. 13.9 months, respectively; P=0.96) and fit versus unfit versus frail patients (15.9 months vs. 13.6 months vs. 12.9 months, respectively; P=0.26). The median OS from registration has not yet been reached (Figure 1D). OS was independent of cytogenetic risk. Age >75 and frailty were associated with inferior OS rates at 2 years (73% vs. 90% in patients â&#x2030;¤75 years, P=0.002, and 74% vs. 89% in unfit and 90% in fit patients at 2 years, P=0.08). Response rates are presented in Table 2. The ITd induction regimen was effective with an overall response rate of 81%, including 47% of patients who achieved at least a very good partial remission, which is comparable to the rate obtained with bortezomib, lenalidomide, dexamethasone (VRd).1 Response was not affected by cytogenetic risk status, age or frailty. The flow of patients through the study is shown in CONSORT diagrams (Online Supplementary Figure S3A, B). Sixty-five of 143 patients (45%) discontinued the study prematurely, during or after induction treatment, and were not randomized. Reasons for discontinuation of induction treatment were toxicity (17%), progressive disease (15%), death (3%) and other reasons (10%) (Online haematologica | 2020; 105(12)

Letters to the Editor

Table 2. Response rates during induction and on protocol.

Response rate (%) Overall response (s)CR VGPR PR < PR ≥ VGPR Improvement in response during maintenance Median time to response (months) Median time to maximum response (months)

ITd induction N=143

On protocol placebo N=39

On protocol ixazomib N=39

81 9 38 34 19 47

97 28 44 26 3 72 13

100 23 38 38 62 13

1.1 2.9

ITd: ixazomib-thalidomide-dexamethasone; N: number; (s)CR: (stringent) complete response; VGPR: very good partial response; PR: partial response..

Supplementary Figure S4A). The main toxicity was neurotoxicity attributed to thalidomide (46%) (Online Supplementary Figure S4B). Patients >75 years had to discontinue induction treatment more often compared to patients ≤75 years (60% vs. 38%; P=0.023) (Online Supplementary Figure S4C). The early mortality was 8% in patients >75 years old compared to only 1% in patients ≤75 years of age. Similarly, more frail patients had to discontinue induction treatment than unfit and fit patients (59%, 39% and 27%, respectively; P=0.008). The main reasons for discontinuation were progressive disease (21% in frail, 13% in unfit and 9% in fit patients; P=0.34) and toxicity (17% in frail, 16% in unfit and 9% in fit patients; P=0.60) (Online Supplementary Figure S4D). Hematologic toxicity was limited (Online Supplementary Table S1). The main non-hematologic toxicities were infections and cardiac events. The incidence of grade 3 neuropathy was low (5%). Seventy-eight patients proceeded to randomization of maintenance treatment (Online Supplementary Figure S3B). During maintenance, 84% of patients discontinued treatment, mainly due to progressive disease (61%), which was comparable between the arms (59% with ixazomib, 63% with placebo). In both arms four patients had to discontinue therapy because of toxicity, of whom three because of neurotoxicity. This was ascribed to thalidomide use, as there was no new-onset neurotoxicity during maintenance therapy and the incidence of neurotoxicity was similar in the two arms. Older age did not negatively affect discontinuation of maintenance therapy, with the rates being 71% in patients >75 years versus 89% in patients ≤75 years, and toxicity accounting for 5% and 13% of the cases, respectively. The same was true for frailty, with the discontinuation rates being 73% in frail patients versus 87% in fit patients, with the discontinuation being due to toxicity in only one frail patient. The median relative dose intensity of both ixazomib and placebo maintenance was 100% (range, 58100% and 65-100%, respectively). The incidence of grade ≥3 adverse events with ixazomib maintenance therapy was comparable to that with placebo maintenance therapy. The lack of improvement in PFS with ixazomib was an unexpected finding as the TOURMALINE-MM3 study had shown an improvement in PFS of 5.2 months with ixazomib maintenance therapy compared to placebo following stem cell transplantation.11 A relevant, yet unexplained, observation is that a sub-analysis of the TOURMALINE-MM3 study showed that in the patients who were treated with induction therapy consisting of both a proteasome inhibitor and an immunomodulatory drug

(mainly thalidomide), as being used in our study, there was no improvement in PFS with ixazomib maintenance. However, most importantly, the small sample size in our phase II study, which was calculated hypothesizing a pronounced hazard ratio of 0.39 for PFS following randomization, might have caused a type II error. Therefore, the results of the TOURMALINE-MM4 study comparing ixazomib with placebo maintenance in patients with NDMM not eligible for transplantation, which is reported to have met its primary endpoint, will hopefully clarify the role of ixazomib maintenance in the non-transplant-eligible population. Importantly, based on preclinical data it may be that the standard maximum dose of ixazomib is suboptimal. We and others have shown sensitivity of the myeloma cell line RPMI 8226 to ixazomib in the nanomolar range; however, as compared to bortezomib, 10-fold higher concentrations of ixazomib were required for the inhibition of cell growth (Online Supplementary Figure S5).12 In this respect, current studies investigating higher doses of ixazomib are of great interest.13,14 Importantly, we found that the PFS in patients with high cytogenetic risk was comparable to that in patients with standard-risk cytogenetics, suggesting that ixazomib overcomes the negative impact of high-risk cytogenetics, which is in accordance with the results of the TOURMALINE-MM1 study.6 The shorter than expected PFS of 14.3 months might well be explained by different levels of frailty of the patients who were included in the different studies. That fitness level may indeed affect treatment efficacy was recently shown by Larocca et al., who observed a PFS of only 14 months with Rd in an unfit, not even a frail, population versus 25.5 months in the original FIRST trial.15 This is in accordance with the findings of a post-hoc analysis of the outcome of patients in the FIRST trial, showing a PFS of only 19.4 months with Rd in frail patients versus 31.3 months in the non-frail patients.10 In conclusion, in this phase II randomized trial we could not show an improvement in PFS with maintenance treatment with ixazomib as compared to placebo. However, the sample size was small, partly due to toxicity of the combination with thalidomide during induction therapy, only allowing randomization of 55% of all patients and 40% of the oldest and frail patients. Importantly, for those patients who were randomized, ixazomib maintenance was very well tolerated, irrespective of age and frailty. Therefore, the results of the randomized phase III trial comparing ixazomib versus placebo maintenance in transplant-ineligible patients are eagerly awaited, as the mild toxicity profile even in frail

Letters to the Editor

patients and the efficacy independent of high-risk disease would pave the way for a new therapy in those categories of patients with an unmet need for novel treatment options. Sonja Zweegman,1 Claudia A.M. Stege,1 Einar Haukas,2 Fredrik H. Schjesvold,3 Mark-David Levin,4 Anders Waage,5 Rineke B.L. Leys,6 Saskia K. Klein,7 Damian Szatkowski,8 Per Axelsson,9 Trung Hieu Do,10 Dorota Knut-Bojanowska,11 Ellen van der Spek,12 Asta Svirskaite,13 Anja Klostergaard,14 Morten Salomo,15 Celine Blimark,16 Paula F. Ypma,17 Ulf-Hendrik Mellqvist,18 Pino J. Poddighe,1 Marian Stevens-Kroef,19 Niels W.C.J. van de Donk,1 Pieter Sonneveld,20 Markus Hansson,9 Bronno van der Holt20 and Niels Abildgaard21 1 Department of Hematology, Amsterdam UMC, Vrije Universiteit Amsterdam, Cancer Center Amsterdam, Amsterdam, the Netherlands; 2 Stavanger University Hospital-Rogaland Hospital, Stavanger, Norway; 3Oslo Myeloma Center, Oslo, Norway; 4Albert Schweitzer Hospital, Dordrecht, the Netherlands; 5St Olavs Hospital and Norwegian University of Science, Trondheim, Norway; 6Maasstad Hospital, Rotterdam, the Netherlands; 7Meander Medical Center, Amersfoort, the Netherlands; 8FĂ¸rde Central Hospital, FĂ¸rde, Norway; 9 Skanes University Hospital Lund, Scania, Sweden; 10Herlev Hospital, Herlev, Denmark; 11NU-Hospital, Uddevalla Hospital, Uddevalla, Sweden; 12Rijnstate Hospital, Arnhem, the Netherlands; 13Aalborg Hospital, Aalborg, Denmark; 14Aarhus University Hospital, Aarhus, Denmark; 15Rigshospitalet Copenhagen, Copenhagen, Denmark; 16 Sahlgrenska University Hospital, Gothenburg, Sweden; 17Haga Hospital, Den Haag, the Netherlands; 18Sodra Alvsborgs Sjukhus Boras, Boras, Sweden; 19Radboud Medical Center, Nijmegen, the Netherlands; 20Erasmus Medical Center Cancer Institute, Rotterdam, the Netherlands and 21Hematology Research Unit and Academy of Geriatric Cancer Research, Odense University Hospital and University of Southern Denmark, Denmark Correspondence: SONJA ZWEEGMAN - s.zweegman@amsterdamumc.nl doi:10.3324/haematol.2019.240374

References 1. Durie BG, Hoering A, Abidi MH, et al. Bortezomib with lenalidomide and dexamethasone versus lenalidomide and dexamethasone alone in patients with newly diagnosed myeloma without intent for immediate autologous stem-cell transplant (SWOG S0777): a ran-

2882

domised, open-label, phase 3 trial. Lancet. 2017;389(10068):519-527. 2. Mateos MV, Dimopoulos MA, Cavo M, et al. Daratumumab plus bortezomib, melphalan, and prednisone for untreated myeloma. N Engl J Med. 2018;378(6):518-528. 3. Facon T, Kumar S, Plesner T, et al. Daratumumab plus lenalidomide and dexamethasone for untreated myeloma. N Engl J Med. 2019; 380(22):2104-2115. 4. San Miguel JF, Schlag R, Khuageva NK, et al. Bortezomib plus melphalan and prednisone for initial treatment of multiple myeloma. N Engl J Med. 2008;359(9):906-917. 5. Benboubker L, Dimopoulos MA, Dispenzieri A, et al. Lenalidomide and dexamethasone in transplant-ineligible patients with myeloma. N Engl J Med. 2014;371(10):906-917. 6. Moreau P, Masszi T, Grzasko N, et al. Oral ixazomib, lenalidomide, and dexamethasone for multiple myeloma. N Engl J Med. 2016; 374(17):1621-1634. 7. Sonneveld P, Avet-Loiseau H, Lonial S, et al. Treatment of multiple myeloma with high-risk cytogenetics: a consensus of the International Myeloma Working Group. Blood. 2016;127(24):29552962. 8. Kumar SK, Berdeja JG, Niesvizky R, et al. Ixazomib, lenalidomide, and dexamethasone in patients with newly diagnosed multiple myeloma: long-term follow-up including ixazomib maintenance. Leukemia. 2019;33(7):1736-1746. 9. Attal M, Lauwers-Cances V, Marit G, et al. Lenalidomide maintenance after stem-cell transplantation for multiple myeloma. N Engl J Med. 2012;366(19):1782-1791. 10. Facon T, Dimopoulos MA, Meuleman N, et al. A simplified frailty scale predicts outcomes in transplant-ineligible patients with newly diagnosed multiple myeloma treated in the FIRST (MM-020) trial. Leukemia. 2019;34(1):224-233. 11. Dimopoulos MA, Gay F, Schjesvold F, et al. Oral ixazomib maintenance following autologous stem cell transplantation (TOURMALINE-MM3): a double-blind, randomised, placebo-controlled phase 3 trial. Lancet. 2019;393(10168):253-264. 12. Brunnert D, Kraus M, Stuhmer T, et al. Novel cell line models to study mechanisms and overcoming strategies of proteasome inhibitor resistance in multiple myeloma. Biophys Acta Mol Basis Dis. 2019;1865(6):1666-1676. 13. Kumar SK, LaPlant BR, Reeder CB, et al. Randomized phase 2 trial of ixazomib and dexamethasone in relapsed multiple myeloma not refractory to bortezomib. Blood. 2016;128(20):2415-2422. 14. Richardson PG, Hofmeister CC, Rosenbaum CA, et al. Twice-weekly ixazomib in combination with lenalidomide-dexamethasone in patients with newly diagnosed multiple myeloma. Br J Haematol. 2018;182(2):231-244. 15. Larocca A, Salvini M, De Paoli L, et al. Efficacy and feasibility of dose/schedule-adjusted Rd-R vs. continuous Rd in elderly and intermediate-fit newly diagnosed multiple myeloma (NDMM) patients: RV-MM-PI-0752 phase III randomized study. Blood. 2018; 132(Suppl 1):305.

haematologica | 2020; 105(12)

CASE REPORTS Apparent recessive inheritance of sideroblastic anemia type 2 due to uniparental isodisomy at the SLC25A38 locus Sideroblastic anemias (SA) are a group of hematological disorders that are characterized by inadequate levels of hemoglobin and ringed sideroblasts in the bone marrow.1 They can be acquired or congenital. Acquired SA are part of the myelodysplastic syndromes with onset in adulthood.1 Congenital SA show both early onset in infancy or childhood and in adulthood, and they are the result of germline mutations (inherited or de novo).2 Several genes have been associated with SA, and these are involved in mitochondrial pathways, such as for heme synthesis, iron-sulfur cluster biogenesis, and mitochondrial metabolism.1 Point mutations in the ALAS2, SLC25A38, GLRX5, and HSPA9 genes have been reported for the nonsyndromic SA. Point mutations in the ABCB7, PUS1, YARS2, LARS2, TRNT1, SLC19A2, and MT-ATP6 genes, and deletion of mitochondrial DNA and the NDUFB11 gene are causal for the syndromic forms of SA. We report here on a case of an infant with microcytic anemia and paternal isodisomy at the SLC25A38 locus (Figure 1A). We revealed uniparental isodisomy (UPD) in the proband, which carried a homozygous recessive mutation, with only one parent heterozygous for the same variant (excluding a de novo mutation), and in the absence of parental consanguineity. UPD is characterized by the presence of two identical copies of a complete or partial chromosome from one of the parents. This results in homozygosity for an autosomal recessive gene, and thus the clinical expression of a recessive disease.3 To date, UPD has never been described as a pathogenic mechanism of SA.

The patient considered here was a newborn male of nonconsanguineous parents from Italy (Figure 1A). At 21 days old, he was admitted to the pediatric Emergency Department due to deep asthenia. He showed mucocutaneous pallor, hypospadia, and micropenis, without any jaundice, fever, or other clinical signs and symptoms. Peripheral blood tests showed severe hypochromic and microcytic anemia, normal white blood cells and platelet counts, and reticulocytopenia (Table 1). A peripheral blood smear showed marked anisopoichylocytosis, with some dacrocytes, microcytes, and spherocytes. His ferritin level was high (747 mg/L), with serum iron and transferrin in the normal range (Table 1); his bilirubin, haptoglobin, and lactate dehydrogenase levels were normal; and his serum erythropoietin (EPO) level was high (114 mUI/mL) (Table 1). Bone marrow smear evaluation was not performed. He received one blood transfusion (10 cm3/kg), and reached a hemoglobin (Hb) of 11.3 g/dL at 23 days old. After the first transfusion, he showed a slow and progressive reduction in Hb (Figure 1B). The first suspected condition was the sequelae of fetal maternal hemolytic diseases, but the immune-hematological tests were negative. Due to the early onset of the condition, the patient was suspected of having a hereditary anemia. Accordingly, a first level of investigation was performed on the parents. Both parents showed normal red blood cell parameters without abnormal hemoglobins. Thus, a second level of investigation was performed, as genetic testing, using a targeted next-generation sequencing (t-NGS) 86-gene custom panel for hereditary anemias. This panel is an updated version of a similar previously published one.4 The genetic analysis of the proband revealed the presence of the rare nonsense variant c.832C>T, p.Arg278* (rs147431446, alternative allele frequency <0.0001 in GnomAD_exome database;

Table 1. Clinical characteristics of the proband affected by sideroblastic anemia type II.

Characteristics

Units of measure

Red blood cell count Hemoglobin Hematocrit Mean corpuscular volume Mean corpuscular hemoglobin Mean corpuscular hemoglobin concentration Red cell distribution width Reticulocyte absolute count White blood cell count Platelet count Mean platelet volume Total bilirubin Unconjugated bilirubin Lactate dehydrogenase Haptoglobin Alanine aminotransferase Serum iron Transferrin saturation Ferritin Erythropoietin

×106/mL g/dL % fL pg g/dL % ×103/mL ×103/mL ×103/mL fL mg/dL mg/dL U/L mg/L U/L mmol-/L % mg/L mU/mL

Patient clinical data (reference range)a according to age at analyses, months 1 6 to 9 3.5 (4.3–5.9) 6.8 (13–17) 21.9 (40–52) 63.4 (80–96) 19.8 (27–31) 31.2 (32–36) 23.5 (11–16) 17.0 (31–82) 9.5 (4–10.0) 792 9.4 (7.2–13) 0.5 (0.3–1.2) 0.25 (<0.6) 290 (<600) n.a. 26 (10–45) 29 (17.9–47.8) 9 747 (21–445) 114 (5–25)

2.8 (4.0–5.6) 8.1 (11.2–14.2) 22.2 (35.5–43.9) 78.0 (71–89) 28.6 (23–30) 36.7 (31–35) 18.2 (11.0–15.5) 8.4 (30–100) 9.7 (5.7–14.5) 304 (150–450) 17.3 (7.0–10.2) 0.3 (0.3–1.2) n.a. (0.2–0.8) 292 (120–300) 1,090 (50–480) 21 (10–45) 42 (7.2–17.9) 13 (15–39) 838 (7–75) n.a. (10–30)

reference range from Sapienza University, Rome, Italy; n.a.: not available.

haematologica | 2020; 105(12)

2883

Case Reports

NM_017875.4; predicted as pathogenic by the ACMG/AMP 2015 guideline) as homozygous (Figure 1A, C). The assumed mechanism of the mutation p.Arg278* is non-expression of the protein due to nonsense mediated decay. Analysis of the inheritance pattern was carried out for both parents. According to the recessive inhertiance pattern, the father was a heterozygous carrier of the variant. However, no mutations in the SLC25A38 gene were identified for the mother (Figure 1A, C). We also confirmed the t-NGS data by Sanger sequencing analysis on DNA from the proband and parents. Interestingly, the proband was homozygous for all of the

variants located on chromosome 3. In particular, beyond the SLC25A38 nonsense variant, two additional single nucleotide variants (SNV) showed a Mendelian violation of inheritance: rs1049296, a coding variant in the TF gene (NM_001063.3:c.1765C>T, p.Pro589Ser); and rs16861582, an intronic variant in the CP gene (NM_000096.3:c.2662-12T>C). These variants resulted in the homozygous state for the proband and the heterozygous state for the father, but none of these were identified in the mother (Figure 1C). Thus, a third level of investigation was performed, using single nucleotide polymorphism (SNP) array in the proband and parents, to

Figure 1. Description of a patient with sideroblastic anemia type II, and characterization of uniparental isodisomy. (A) Genetic pedigree of the family. Square: male; circle: female; solid symbols: affected person; arrow: proband. According to uniparental isodisomy, only the father carried the variant c.832C>T, p.Arg278* in the SLC25A38 gene. (B) Hemoglobin levels of the patient after blood transfusions and erythropoietin (EPO) supplementation. (C) Alignment track of next-generation sequencing analysis of the proband, mother, and father, showing presence of the c.832C>T variant as homozygous in the proband (mutated allele [T] frequency, 97% [184/190 reads]; wild-type [wt] allele (C) frequency, 3% [6/190 reads]), as heterozygous in the father, and absent in the mother. (D) Agilent CytoGenomic view of the single nucleotide polymorphism (SNP) data. Top panel: number of uncut alleles. Bottom panel: comparative genomic hybridization (CGH) data (log2 ratios). Data from GenetiSure Postnatal Research CGH+SNP 2x400 Array (Agilent), which show large regions of homozygosity on chromosome 3. (E) Diagram of the SLC25A38 protein structure and the pathogenic variants, as obtained from the Human Gene Mutation Database Professional (updated in June 2020). Circle colors define the mutation types (for different mutation types at a single position, the color defines the most frequent) (https://www.cbioportal.org/mutation_mapper).

2884

haematologica | 2020; 105(12)

Case Reports

define either a heterozygous deletion or UPD. No genomic deletions or rearrangements at or around the SLC25A38 locus were identified in the proband or parents at 3p22.1. The SNP array identified large regions of homozygosity that involved the whole paternal chromosome 3, which included the SLC25A38 locus: 3p26. 3p25. 3(54290_9540289), 3p25. 3p12. 1 (11549397_ 84682450), 3q11. 2q12.3 (94565518_101899846), and 3q12. 3q29 (102540948_197959451) (Figure 1D). Currently, only one case of paternal UPD for whole chromosome 3 has been described with no apparent disease phenotype, suggesting the absence of paternal imprinted genes on this chromosome.5 Homozygous or compound heterozygous mutations in the SLC25A38 gene are causal for SA type II.1 A recent retrospective multicenter European study of a cohort of patients with childhood-onset congenital SA showed that in SA, SLC25A38 is the second most commonly mutated gene after ALAS2.6,7 Different recessive loss-offunction mutations (i.e., nonsense, frameshift, splicing, missense) have been reported for the SLC25A38 along the entire gene (Figure 1E).1,6-12 These variants impair the transport of glycine in the mitochondria. Indeed, SLC25A38 is a mitochondrial carrier that is expressed in erythroblasts.8 It is located on the mitochodrial inner membrane and imports glycine into the erythroid mitochondria. The glycine condenses with succinylCoA to form δ-aminolevulinic acid (ALA), a substrate for heme that it is then exported into the cytosol.8,13 Thus, mutations in SLC25A38 result in alterations to heme synthesis and the hematological phenotype characterized by hypochromic microcytic anemia, and iron depositions around the nucleus in the mitochondria of erythroblasts in the bone marrow.1 Patients with SLC25A38 mutations show neonatal microcytic anemia and high ferritin levels (higher than those of patients with mutated ALAS2), as described for the patient here.7 Of note, patients who carry biallelic pathogenic variants in SLC25A38 gene show iron overload not only in the liver, but also in the myocardium,6 and the follow-up of this patient will include such evaluation. The same amino acid, arginine 278, was mutated in glycine in a patient previously described (Figure 1E).9 Comparison of our case and others with nonsense mutations with the patients with biallelic missense mutations reveals no differences. Most of the patients with SLC25A38 mutations present a severe or moderate microcytic anemia, with early onset, transfusion dependency and severe or moderate hepatic iron overload. All patients with a mutated SLC25A38 gene require transfusion at a mean rate of 13.3 transfusions/year (range: 0.3–20/year), with chelation therapy and vitamin B6 supplementation.6 Accordingly, the patient considered here continued to receive erythrocyte transfusions from the age of 6 months, with a median interval of 30 days. His median pre-transfusion Hb was 7.3 g/dL, and this transfusion regimen maintained his Hb at ~8.5 g/dL. He also received vitamin supplementation at a standard dose, without high-dose pyridoxine. Pharmacological supplementation with glycine and folate might improve his heme synthesis. In yeast and zebrafish models, exogenous glycine in combination with folate ameliorated the heme levels.14 This combined supplementation was also administered to three patients with SLC25A38 mutations, although without any improvements.14 So, to date, the only therapeutic strategy here is chronic blood transfusions and iron chelation. The potentially curative therapy remains hematopoietic stemcell transplantation.7 Due to religious beliefs, families can sometimes refuse haematologica | 2020; 105(12)

permission for blood transfusions, therefore, prior to obtaining the results of the genetic testing, we started the patient on EPO therapy. Initially, this resulted in an apparent lengthening of the transfusion interval (Figure 1B), but later he showed a reduction in Hb. Thus, the treatment with EPO was stopped. To date, no cases of patients with SA being treated with EPO have been described in the literature. Indeed, EPO is not expected to be not effective in this disease, as down-regulation of SLC25A38 is described in these patients, which is similar to patients with myelodysplastic syndrome with ring sideroblasts.15 In our patient, we noted a similar picture, with an apparent initial response to EPO, with progressive lengthening of the transfusion interval. Unfortunately, EPO treatment did not improve his Hb. Currently, the patient is 9 months old, and his psychophysical development is normal. He does not require iron chelation at present. For his sexual phenotype, he has just undergone subcutaneous substitutive testosterone therapy. Testosterone stimulates erythropoiesis, and we believe that this patient will benefit from its administration, for both his anemia and his hypogonadism, although over a short period. According to the literature, the clinical features in this patient confirm that SA represent a severe transfusiondependent disease with no valid options for treatment. From a diagnostic perspective, this case highlights the importance of the evaluation of the possible occurrence of UPD for patients with SA due to SLC25A38 mutations, and of an assessment of the inheritance pattern of the identified variants. Indeed, the incidence of UPD is higher for rare autosomal recessive diseases compared to common autosomal recessive diseases with higher carrier frequencies.3 Nevertheless, in the presence of a normal karyotype, the recurrence risk of a rare autosomal recessive disease caused by UPD of a whole chromosome is negligible since it is a rare, generally sporadic event.3 Immacolata Andolfo,1,2 Stefania Martone,1,2 Michela Ribersani,3 Simona Bianchi,3 Francesco Manna,2 Rita Genesio,1 Antonella Gambale,2,4 Piero Pignataro,1 Anna Maria Testi,3 Achille Iolascon1,2 and Roberta Russo1,2 1 Dipartimento di Medicina Molecolare e Biotecnologie Mediche, Università degli Studi di Napoli ‘Federico II’, Naples; 2CEINGE Biotecnologie Avanzate, Naples; 3Department of Translational and Precision Medicine, Sapienza University, Rome and 4Dipartimento assistenziale integrato di Medicina di Laboratorio, UOC Genetica Medica, Azienda Ospedaliera ‘Federico II’, Naples, Italy Correspondence: IMMACOLATA ANDOLFO - andolfo@ceinge.unina.it doi:10.3324/haematol.2020.258533 Acknowledgments: the authors thank the CEINGE Service Facility platforms of the Sequencing Core and Oligo Synthesis. Funding: this work was supported by an EHA Junior Research Grant to Immacolata Andolfo (3978026), and by a Bando Star Linea 1 - Junior Principal Ivestigator Grants - COINOR, Università degli Studi di Napoli ‘Federico II’ to Roberta Russo. The authors also thank the parents of the patient for granting their permission for the case to be communicated to the scientific community.

References 1. Ducamp S, Fleming MD. The molecular genetics of sideroblastic anemia. Blood. 2019;133(1):59-69. 2. Liu G, Guo S, Anderson GJ, Camaschella C, Han B, Nie G. Heterozygous missense mutations in the GLRX5 gene cause sideroblastic anemia in a Chinese patient. Blood. 2014;124(17):27502751. 3. Niida Y, Ozaki M, Shimizu M, Ueno K, Tanaka T. Classification of

2885

Case Reports

4. 5.

6. 7. 8.

2886

uniparental isodisomy patterns that cause autosomal recessive disorders: proposed mechanisms of different proportions and parental origin in each pattern. Cytogenet Genome Res. 2018;154(3):137146. Russo R, Andolfo I, Manna F, et al. Multi-gene panel testing improves diagnosis and management of patients with hereditary anemias. Am J Hematol. 2018;93(5):672-682. Xiao P, Liu P, Weber JL, Papasian CJ, Recker RR, Deng HW. Paternal uniparental isodisomy of the entire chromosome 3 revealed in a person with no apparent phenotypic disorders. Hum Mutat. 2006; 27(2):133-137. Fouquet C, Le Rouzic MA, Leblanc T, et al. Genotype/ phenotype correlations of childhood-onset congenital sideroblastic anaemia in a European cohort. Br J Haematol. 2019;187(4):530-542. Cazzola M, Malcovati L. Diagnosis and treatment of sideroblastic anemias: from defective heme synthesis to abnormal RNA splicing. Hematology Am Soc Hematol Educ Program. 2015;2015:19-25. Guernsey DL, Jiang H, Campagna DR, et al. Mutations in mitochondrial carrier family gene SLC25A38 cause nonsyndromic autosomal recessive congenital sideroblastic anemia. Nat Genet. 2009; 41(6):651-653. Kannengiesser C, Sanchez M, Sweeney M, et al. Missense

10. 11.

12. 13. 14.

15.

SLC25A38 variations play an important role in autosomal recessive inherited sideroblastic anemia. Haematologica. 2011;96(6):808-813. Wong WS, Wong HF, Cheng CK, et al. Congenital sideroblastic anaemia with a novel frameshift mutation in SLC25A38. J Clin Pathol. 2015;68(3):249-251. Le Rouzic MA, Fouquet C, Leblanc T, et al. Non-syndromic childhood onset congenital sideroblastic anemia: a report of 13 patients identified with an ALAS2 or SLC25A38 mutation. Blood Cells Mol Dis. 2017;66:11-18. 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. Furuyama K, Kaneko K. Iron metabolism in erythroid cells and patients with congenital sideroblastic anemia. Int J Hematol. 2018; 107(1):44-54. LeBlanc MA, Bettle A, Berman JN, et al. Study of glycine and folic acid supplementation to ameliorate transfusion dependence in congenital SLC25A38 mutated sideroblastic anemia. Pediatr Blood Cancer. 2016;63(7):1307-1309. Del Rey M, Benito R, Fontanillo C, et al. Deregulation of genes related to iron and mitochondrial metabolism in refractory anemia with ring sideroblasts. PLoS One. 2015;10(5):e0126555.

haematologica | 2020; 105(12)

Case Reports

Eculizumab for complement mediated thrombotic microangiopathy in sickle cell disease Despite being the first genetic disease described, sickle cell disease (SCD) continues to afflict patients with immense pain, significant comorbidities and premature death. SCD has only recently benefited from new interventions with L-glutamine (2017), voxelotor (2019) and crizanlizumab (2019) representing the first Food and Drug Administration approved medications for SCD since hydroxyurea in 1997. These interventions have demonstrated some ability to reduce vaso-occlusive pain crisis episodes, improve hemoglobin (HGB), or reduce markers of hemolysis and have largely been used as preventative care measures. While these and additional approaches, such as hematopoietic stem cell transplant and gene therapy, can improve SCD care, many patients with SCD continue to suffer from severe acute SCD complications that can result in organ damage and early death.1,2 Unfortunately, in these situations, supportive care remains the primary approach to alleviate complications. The lack of more targeted approaches in part reflects an incomplete understanding of the pathophysiology and accompanying pharmacological targets that could specifically mitigate acute disease complications. We present a summary of three cases of children with SCD who developed significant acute complications that demonstrate underlying complement-mediated thrombotic microangiopathy (CM-TMA). These cases include a

delayed hemolytic transfusion reaction (DHTR), vasoocclusive crisis (VOC) and drug-induced immune hemolytic anemia (DIIHA). Patient #1 is a 14-year-old male with a history of two episodes of DHTR. At the age of 12 years, he received a transfusion pre-operatively for hip core-decompression. He presented 7 days later with severe diffuse body pain, hemoglobinuria, fever, and total HGB of 9.4 g/dL with hemoglobin A (HbA) at 17% (Figure 1A). Further testing revealed a previously detected anti-U alloantibody, negative direct antiglobulin test (DAT) and evidence for intravascular hemolysis. On the night of admission, he became hypertensive with headache and sluggish mentation. A brain magnetic resonance imaging scan was normal. HGB dropped to 5.6 g/dL within 30 hours of admission, and the patient rapidly deteriorated to multiorgan failure (MOF) with thrombocytopenia (Figure 1A). Due to strong suspicion for DHTR with hyperhemolysis and CM-TMA, he was treated with eculizumab 600 mg intravenously (IV) and erythropoietin 150 IU/kg to augment erythropoiesis. Over the next 24 hours, he developed a new consolidation in the left lung consistent with acute chest syndrome (ACS). He received one unit of crossmatch compatible, U-negative red blood cells (RBC) after a dose of 1 g/kg intravenous immunoglobulin (IVIG) on day 9. Over the next few days, he made a gradual improvement in clinical and laboratory status. Eculizumab 600 mg was continued weekly for a total of four doses. Subsequent analysis revealed significant alternative complement pathway (ACP) activation at the peak

Table 1. Evaluation of complement pathway during acute sickle cell crises.

DHTR Patient #1 Day 8 Day 22 C3 (71-150 mg/dL) C4 (15.7-47 mg/dL) CH50 (101-300 units) Bb (0.49-1.42 mcg/mL)* C3a (25-88.2 ng/mL) C5a (2.74-16.33 ng/mL) SC5b-9 (≤ 244 ng/mL) Day of eculizumab administration

VOC Episodes Patient #2 2 3 4

Day 5

DIIHA Patient #3 Day 22

8 weeks

110

173

133

143

167

14.2

21.6

362

260

9.75

1.41

1.9

1.2

1.3

1.9

23.6*

1.63

2.185

375.5

127.2

124

76.4

115

164

210.4

102.5

182.9

>37.3

31.26

13.5

17.2

>38.0

35.53

27.6

1208

307

145

112

153

>1800

554

394

None

same day

none

Days 8, 15, 22 and 29

Days 5, 12, 18, 26, then every 2 weeks*

DHTR: delayed hemolytic transfusion reaction; VOC: vaso-occlusive crisis, DIIHA: drug induced immune hemolytic anemia; C3: complement component 3; C4: complement component 4; CH50: screening test for total complement activity; Bb: complement component fragment Bb; C3a: complement component fragment 3a; C5a: complement component fragment 5a; sC5b-9: soluble membrane attack complex; ND: not done. *8-week laboratory analysis were performed few hours prior to eculizumab administration. The whole blood sample for all testing was collected in EDTA anticoagulant tubes and plasma was obtained within 2 hours of collection, and stored in -80°C until they were ready for analysis with single thawing. This method of plasma collection in EDTA results in chelation of calcium and magnesium, thus preventing any in vitro complement activation. All testing was obtained in a Clinical Laboratory Improvement Amendments (CLIA) certified hospital-based clinical laboratory. All normal values are in parentheses under each value except in patient #3, day 5* Bb normal ranges were 1.32–4.18 mcg/mL due to variability seen with different enzyme-linked immunosorbent assay kits. Patient #1, day 8 and day 22 signifies the complement evaluation a few hours prior to respective dosing of eculizumab. Patient#2 presented with five episodes of VOC (each column represents episodes 1 through 5) with complement function testing coinciding with a drop in hemoglobin of greater than 2 g/dL from baseline. Eculizumab were dosed a few hours after samples were collected during episodes #3 and #5. Patient #3, days 5, 22 and 8 weeks reflect the complement evaluation. Complement proteins C3 and C4 signify the quantitative serum levels. Fragment Bb is a serine protease that in combination with hydrolyzed C3 (C3H2O) generates C3 convertase (C3bBb), which augments the cleavage of C3 to generate C3a and C3b. Anaphylatoxins, C3a and C5a are involved in local inflammation and tissue damage, while C3b results in red blood cell opsonization and deposition on endothelium. Terminal complex, C5b-9 contributes to intravascular hemolysis.

haematologica | 2020; 105(12)

2887

Case Reports

Figure 1. A graphic representation of a delayed hemolytic transfusion reaction, vaso-occlusive crisis and drug induced immune hemolytic anemia episode. (1A) DHTR (delayed hemolytic transfusion reaction): the patient received a transfusion (black arrows) on day one and presented with DHTR on day 7 with a hemoglobin (HGB) of 9.4 g/dL, which dropped to a nadir of 5.6 g/dL by day 8 along with absolute reticulocytopenia, as expected in DHTR. Thrombocytopenia coincided with this severe anemia and elevations in lactate dehydrogenase (LDH), total bilirubin (T.Bili), and aspartate transaminase (AST). Other evidence for intravascular hemolysis included reduced haptoglobin (<14; reference 30-120 mg/dL) and elevated plasma free HGB (120; reference < 30 mg/dL) not shown. Eculizumab (E) 600 mg was initiated on day 8 and given weekly for a total of four doses. The patient developed acute chest syndrome (ACS) and received one unit of crossmatch compatible, U-negative red blood cells (RBC) after a dose of intravenous immunoglobulin (IVIG) on day 9. The resolution of hemolysis was evidenced by improvement in LDH and HGB following the first dose of eculizumab and maintained throughout the hospital stay. Thrombocytopenia recovered within 1 week of complement inhibition. The patient required an additional transfusion on day 19 from rebound anemia likely secondary to frequent blood draws, since markers of hemolysis and complement activation on day 22 did not worsen. Immunohematology work-up during this DHTR episode: direct antiglobulin test (DAT) negative, historical anti-U detected. Note: the initial three LDH values were greater than 4,000 U/L (upper limit of our clinical laboratory detection). Eculizumab was dosed based on the weight of the patient, as per the guidelines for loading dose for children with atypical hemolytic uremic syndrome (aHUS). (1B) Vaso-occlusive crisis (VOC): grey colored dotted bars depict the time points when the patient presented with VOC and also corresponds to complement testing (numbered #1 through #5). Eculizumab was administered for episodes #3 and #4 as described. Lower graph represents episode #4 of VOC and shows an initial drop in HGB and platelets by day 4, when erythropoietin (P) 210 IU/kg was commenced along with a dose of IVIG (I) 1 g/kg on days 4 and 5, respectively. Given the continued deterioration in HGB to a nadir of 3.6 g/dL along with severe headache and worsening hypoxia, eculizumab (E) 900 mg was administered on day 7 after the blood for complement function analyses was collected. Rapid improvement in hemolysis and clinical status was observed within 48 hours of a single eculizumab dose. Eculizumab was dosed based on the weight of the patient, as per the guidelines for loading dose for children with aHUS. (1C) Drug induced immune hemolytic anemia (DIIHA): the patient presented on day 0 to the hospital with fever and received ceftriaxone; HGB dropped within 3 hours to 4 g/dL. Black arrows on the graph denote blood transfusions. Additional laboratory workup showed evidence of intravascular hemolysis, confirmed by elevated LDH >4,000 U/L, presence of schistocytes on blood smear and elevated plasma free HGB (not shown). Multiorgan failure was evident with a peak of creatinine at 2.36 mg/dL (baseline 0.3 mg/dL), and AST/alanine transaminase at 1,002/70 U/L along with rise in T.Bili. Black colored bars at the bottom of the graph depict the time points when various supportive care measures and eculizumab were administered. Shortly after the initiation of eculizumab, hemolysis decreased, as shown by the rapid drop in LDH, and the patient required less transfusion support. Thrombocytopenia improved. He had initial improvement in creatinine on continuous renal replacement therapy (CRRT), with a brief increase when CRRT was weaned. This rise was not sustained, and creatinine levels decreased promptly without any additional intervention except continued eculizumab therapy. By day 25, blood counts and chemistry were within normal limits. The rebound thrombocytosis persisted for few weeks before trending back to the patientâ&#x20AC;&#x2122;s baseline. Note: the initial three LDH values were greater than 4,000 U/L (upper limit of our clinical laboratory detection). Eculizumab was dosed based on the weight of the patient, as per the guidelines for loading dose for children with aHUS. Blood smears from days 4 and 5 showing the presence of schistocytes and helmet cells (black arrows), along with paucity of platelets. ARC: absolute reticulocyte count.

2888

haematologica | 2020; 105(12)

Case Reports

Figure 2. A model for complement- mediated thrombotic microangiopathy in sickle cell disease. Increased understanding of complement-mediated conditions such as atypical hemolytic uremic syndrome (aHUS) and paroxysmal nocturnal hemoglobinuria has renewed interest in understanding the specific role of complement in hyperhemolysis and innate immunity. Sickle cell disease (SCD) is a prototypical disease of chronic hemolysis, where increased levels of free plasma hemoglobin and heme with an insufficient, and thus ineffective scavenging mechanism by haptoglobin and hemopexin, leads to a state that is primed for complement activation. In addition, sickle erythrocytes themselves appear to be uniquely susceptible to complement-induced hemolysis, thereby further amplifying complement activation and additional hemolysis. Like in aHUS, the already inflamed endothelium in patients with SCD can be further modulated from increased hemolysis, complement activation, coagulation dysfunction, and other plasma proteins. Further genetic studies focusing on complotype are needed to help understand its role in SCD, as they can modulate the homeostatic balance of complement activity. Triggers such as pain crisis, acute chest syndrome, infection, etc. can drive this vulnerable state very quickly into a complement activated state, which, if unregulated, can result in common and terminal complement pathway activation that can lead to catastrophic damage in end organs and even death. Irrespective of the instigating cause, once complement-mediated hyperhemolysis is set off, it could result in a positive feedback loop causing further complement activation and a precipitous drop in hemoglobin and risk of sudden death. The table on the right parallels the resting and enabling state seen in SCD to patients with aHUS. CM-TMA: complement-mediated thrombotic microangiopathy; MAHA: microangiopathic hemolytic anemia; RBC: red blood cell.

of hemolysis, as evidenced by increased complement component fragment Bb (Bb) levels, anaphylatoxins (C3a and C5a) and terminal complement complex (C5b-9) (Table 1). Testing for complement regulatory genes (CFH, CFI, MCP (CD46), CFB, CFHR5, C3, THBD, DGKE, PLG, ADAMTS13 and MMACHC) revealed a homozygous deletion of complement factor H-related protein (CFHR) 3 and CFHR1, but criteria for DEAP-HUS (deficiency of CFHR plasma proteins and autoantibody positive form of HUS) were not met due to the absence of factor H autoantibodies. Patient #2 is a 15-year-old female with a history of VOC episodes and DHTR, who was transferred to our care at the age of 9 years. Since transfer, she suffered from multiple episodes of VOC, each accompanied with a drop in HGB. Five of these episodes are shown in the dotted lines (Figure 1B, top). During episode #3, with no recent RBC transfusion history, the patient presented with an HGB of 6.7 g/dL which decreased to 3.8 g/dL over 5 days, accompanied by lethargy, hypoxia and respiratory distress. Given her past history of DHTR and rapid decompensation, one dose of eculizumab 900 mg IV was administered along with erythropoietin 210 IU/kg daily and IVIG at 1 g/kg. The patientâ&#x20AC;&#x2122;s symptoms resolved with a rise in HGB and she was discharged home within 4 days. Six months later, she presented again (#4) with VOC, worsening hypoxia, and HGB 7.9 g/dL (Figure 1B, bottom). On admission, she received erythropoietin and IVIG at the doses outlined above. The HGB dropped to a nadir of 3.6 g/dL over 3 days, associated with severe headaches. Due to inadequate response to the above measures, eculizumab 900 mg was administered with rapid improvement in 48 hours. Complement analyses (Table 1) indicated clear evidence of ACP activation during episodes #1 and #5, as shown by elevation of Bb lehaematologica | 2020; 105(12)

vels. During episodes #3 and #4, eculizumab was administered in anticipation of worsening organ function, which could explain why the complement activation markers were not significantly elevated (unlike in other episodes) and why her rapid HGB response and prompt reversal of organ dysfunction were observed. Complement gene evaluation revealed a heterozygous deletion of CFHR3/CFHR1. Patient #3 is a 3-year-old male with a history of splenectomy at the age of 2 years for recurrent acute splenic sequestration. He presented with fever, tachypnea, HGB of 7 g/dL, with a negative chest x-ray (CXR). He developed increased work of breathing within 3 hours of receiving ceftriaxone. Repeat CXR revealed bilateral infiltrates requiring emergent intubation. HGB dropped to 4 g/dL, and he received extended phenotypematched and crossmatch-compatible RBC transfusions. Additional laboratory work-up revealed intravascular hemolysis and MOF (Figure 1C). The DAT was positive for complement component 3 (C3) only. The patient required escalation of respiratory support to high frequency oscillatory ventilation and nitric oxide. Oliguric acute kidney injury and hypertension required the use of continuous renal replacement therapy and multiple antihypertensives. Multiple common and rare causes for his rapid multi-organ failure were entertained including sepsis (negative blood/respiratory cultures), cold agglutinin syndrome (negative testing for mycoplasma serology and Donath-Landsteiner antibody), hemophagocytic lymphohistiocytosis (normal soluble interleukin-2 receptor, CD107a, perforin/granzyme B, Epstein Barr and cytomegaloviral load) and ceftriaxone-induced hemolysis. As his clinical course was consistent with CM-TMA, complement inhibition with eculizumab 600 mg IV was initiated on day 5. Following eculizumab, he demonstrat2889

Case Reports

ed a rapid response with weaning of his ventilator support and dialysis, along with the reduced need for blood products. Follow-up testing was notable for strongly positive ceftriaxone-dependent antibodies consistent with DIIHA. Complement analyses confirmed significant activation of ACP (factor Bb elevation, see Table 1). Proximal and terminal complement pathway activation were likewise observed as indicated by increased C3 and C5 activation and C5b-9, respectively. Additionally, hypocomplementemia with reduced C3, C4, and CH50, seen in this patient, suggests worse disease. These markers improved following complement inhibition, which correlated with improved clinical status within 11 days of initial presentation. Eight months after this episode, this patient’s renal function is gradually improving. He remains on eculizumab 300 mg every 2 weeks pending full renal recovery. Complement genetic analysis was negative. All patients and/or their guardians of cases described in this report provided written consent for the off-label use of eculizumab. These patients received meningococcal and pneumococcal vaccinations as part of routine SCD standard-of-care or given right before eculizumab and continued on a prophylactic antibiotic regimen while on treatment. CM-TMA from an underlying mutation involving the complement regulatory genes is traditionally called atypical hemolytic uremic syndrome or ‘atypical HUS’ (aHUS). CM-TMA occurring secondary to complement amplifying disorders such as hematopoietic stem cell transplantation, malignancy, infections, or autoimmune diseases are termed ‘secondary HUS’.3,4 The above clinical vignettes suggest that some complications of SCD may also reflect complement activation-induced secondary HUS (Figure 2). As SCD is a chronic hemolytic condition and plasma free HGB and heme can activate complement,5,6 additional complement activation during episodes of disease exacerbation may lead to increased hemolysis and a sustained positive feedback loop that leads to life-threatening anemia. As sickle erythrocytes are uniquely susceptible to complement-mediated damage,7 this may also result in further exacerbation of the disease. Heme-dependent endothelial dysfunction seen in SCD can also be modulated by complement activation.8 In addition, heme can regulate the coagulation system, which likely reflects another important feature of SCD pathophysiology that contributes to inflammation and thrombosis.9 Increased C5a production has also been shown to cause acute lung injury and vasoocclusion in animal models.10,11 Prior reports suggested the role of ACP in the pathophysiology of SCD, and a protective effect of eculizumab in some settings.12-15 However, the present data suggest that some complications in SCD may not only reflect exuberant ACP activation, but actually represent an underlying CM-TMA. In these situations, chronic hemolysis and endothelial dysfunction may saturate scavenging and detoxifying mechanisms, reducing the capacity of patients with SCD to manage elevated plasma HGB and heme during periods of crises. In two of our patients, we detected variants in the complement genes. The CFH gene encodes soluble plasma factor H, which is a principal inhibitor of ACP. Further genetic studies focusing on complotype are needed to help understand its role in SCD. Figure 2 compares salient clinical features seen in aHUS and CM-TMA in SCD. In this way, the underlying pathophysiology of SCD may prime individuals for secondary HUS through ACP activation, with a subset of these patients reaching an inflection point during periods of crises (second-hit) 2890

that lead to additional hyperhemolysis and MOF. Case #3 is the first report of successful use of eculizumab to treat life-threatening DIIHA. Ceftriaxone-induced hemolysis often occurs secondary to immunglobulin M (IgM) anticeftriaxone antibodies, which typically engage the classical complement pathway and induce hemolysis. However, the impact of eculizumab did not appear to reflect inhibition of IgM-induced hemolysis as marked activation of APC was evidenced by increased levels of factor Bb, which likewise responded to eculizumab, consistent with a positive feedback loop that drives additional APC activation in these patients. While rare, this event in particular holds public health and preventative importance, as ceftriaxone is a widely used medication in SCD, and there is a high prevalence of ceftriaxone-induced RBC antibodies in these patients.16 This collection of cases therefore emphasizes a previously underappreciated role of CM-TMA and complement across a broad range of SCD presentations, and may reflect novel insights into the pathophysiology of acute exacerbations of SCD that may be sensitive to complement inhibition to avoid severe hemolytic complications in SCD. Satheesh Chonat,1,2 Sara Graciaa,2 H. Stella Shin,1,3 Joanna G. Newton,1,2 Maa-Ohui Quarmyne,1,2 Jeanne Boudreaux,1,2 Amy Tang,1,2 Patricia E. Zerra,2,4 Margo R. Rollins,2,4 Cassandra D. Josephson,2,4 Clark Brown,1,2 Clinton H. Joiner,1,2 Ross M. Fasano2,4 and Sean R. Stowell4 1 Department of Pediatrics, Emory University School of Medicine; 2 Aflac Cancer and Blood Disorders Center; 3Division of Pediatric Nephrology, Children’s Healthcare of Atlanta and 4Center for Transfusion and Cellular Therapy, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA, USA. Correspondence: SATHEESH CHONAT - satheesh.chonat@emory.edu . doi:10.3324/haematol.2020.262006 Acknowledgments: the authors would like to thank the hematology, critical care medicine, nephrology and transfusion medicine teams at Children’s Healthcare of Atlanta for their assistance with patient care. No funding was used to conduct this study.

References 1. Vidler JB, Gardner K, Amenyah K, Mijovic A, Thein SL. Delayed haemolytic transfusion reaction in adults with sickle cell disease: a 5-year experience. Br J Haematol. 2015;169(5):746-753. 2. Habibi A, Mekontso-Dessap A, Guillaud C, et al. Delayed hemolytic transfusion reaction in adult sickle-cell disease: presentations, outcomes, and treatments of 99 referral center episodes. Am J Hematol. 2016;91(10):989-994. 3. Riedl M, Fakhouri F, Le Quintrec M, et al. Spectrum of complementmediated thrombotic microangiopathies: pathogenetic insights identifying novel treatment approaches. Semin Thromb Hemost. 2014;40(4):444-464. 4. Le Clech A, Simon-Tillaux N, Provôt F, et al. Atypical and secondary hemolytic uremic syndromes have a distinct presentation and no common genetic risk factors. Kidney Int. 2019;95(6):1443-1452. 5. Frimat M, Tabarin F, Dimitrov JD, et al. Complement activation by heme as a secondary hit for atypical hemolytic uremic syndrome. Blood. 2013;122(2):282-292. 6. Merle NS, Grunenwald A, Rajaratnam H, et al. Intravascular hemolysis activates complement via cell-free heme and heme-loaded microvesicles. JCI Insight. 2018;3(12):1-18. 7. Wang RH, Phillips G, Medof ME, Mold C. Activation of the alternative complement pathway by exposure of phosphatidylethanolamine and phosphatidylserine on erythrocytes from sickle cell disease patients. J Clin Invest. 1993;92(3):1326-1335. 8. Merle NS, Paule R, Leon J, et al. P-selectin drives complement attack on endothelium during intravascular hemolysis in TLR-4/heme-

haematologica | 2020; 105(12)

Case Reports

9. 10. 11. 12.

dependent manner. Proc Natl Acad Sci U S A. 2019;116(13):62806285. Markiewski MM, Nilsson B, Nilsson Ekdahl K, Mollnes TE, Lambris JD. Complement and coagulation: strangers or partners in crime? Trends Immunol. 2007;28(4):184-192. Mulligan MS, Schmid E, Beck-Schimmer B, et al. Requirement and role of C5a in acute lung inflammatory injury in rats. J Clin Invest. 1996;98(2):503-512. Vercellotti GM, Dalmasso AP, Schaid Jr TR, et al. Critical role of C5a in sickle cell disease. Am J Hematol. 2019;94(3):327-337. Chonat S, Chandrakasan S, Kalinyak KA, Ingala D, Gruppo R, Kalfa TA. Atypical haemolytic uraemic syndrome in a patient with sickle cell disease, successfully treated with eculizumab. Br J Haematol. 2016;175(4):744-747.

haematologica | 2020; 105(12)

13. Dumas G, Habibi A, Onimus T, et al. Eculizumab salvage therapy for delayed hemolysis transfusion reaction in sickle cell disease patients. Blood. 2016;127(8):1062-1064. 14. Chonat S, Quarmyne M-O, Bennett CM, et al. Contribution of alternative complement pathway to delayed hemolytic transfusion reaction in sickle cell disease. Haematologica. 2018;103(10):e483-e485. 15. Thein SL, Pirenne F, Fasano RM, et al. Hemolytic transfusion reactions in sickle cell disease: underappreciated and potentially fatal. Haematologica. 2020;105(3):539-544. 16. Quillen K, Lane C, Hu E, Pelton S, Bateman S. Prevalence of ceftriaxone-induced red blood cell antibodies in pediatric patients with sickle cell disease and human immunodeficiency virus infection. Pediatr Infect Dis J. 2008;27(4):357-358.

2891

COMMENT SARS-CoV-2 severity in African Americans – a role for Duffy null? The scourge of the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2; COVID-19) infection in the United State has disproportionately affected African Americans, Hispanics, and Native Americans;1 a study in the United Kingdom found Blacks and Asians to have augmented risk.2 We here focus upon African Americans (AA), a subpopulation in which COVID-19 disease is more likely to occur and to result in disproportionately higher hospitalization and mortality rates. The latter is typically from pneumonitis progressing to a severe acute lung injury (ALI) syndrome. Identifying specific reasons for this disparity is challenging because severity is influenced by comorbidities as well as societal and environmental factors. Not surpri-singly, for the pandemic in general, emerging data are beginning to suggest contributory biological variabilities that reside within the genetic background.3 Here we suggest that a predisposition to severe COVID-19 pneumonitis amongst infected AA is established by an erythroid Duffy blood group polymorphism. Duffy blood group antigens, membrane proteins FYa and FYb, are allelic products of DARC (Duffy antigen chemokine receptor), now known as ACKR1 (atypical chemokine receptor 1).4 FY proteins are expressed on red blood cells (RBC) and endothelial cells of capillaries and venules (as well as on cerebellar neurons and epithelial cells of the kidney and lung). Normally, erythroid FY binds multiple inflammatory chemokines, probably functioning as a chemokine reservoir that helps regulate plasma levels.5 RBC FY can dampen leukocyte activation, and endothelial FY guides localization and presentation of chemokines. Specifically, endothelial FY located at sites of endothelial cell/cell contact regulates chemokine transcytosis and leukocyte diapedesis.4,6 Relevance to COVID-19 amongst AA derives from a DARC polymorphism (rs2814778, a T→C substitution in the promoter’s GATA box) that prevents erythroid expression of both FY proteins, while FY expression on endothelial cells is unaffected.4 Homozygosity results in the “Duffy null” phenotype. Because Duffy null protects from RBC invasion by Plasmodium vivax, its positive selection led to its current >95% prevalence in Western and South-Western sub-Saharan Africa.7 Its prevalence is somewhat lower, although still very high, in the remainder of sub-Saharan Africa. Consequently, perhaps 67% of African Americans are erythroid Duffy null. Under normal circumstances Duffy null accounts for benign ethnic neutropenia amongst AA, but it also exerts a pro-inflammatory effect that can promote leukocyte migration into the lung.4 This is the basic reason to suspect that erythroid Duffy null status would promote COVID-19 pneumonitis and accentuate its severity, resulting in ALI. In support, a 2012 analysis of three prior human studies revealed that, if inflammatory ALI deve-lops for some reason in Duffy null AA, it is significantly more severe than in either Duffy positive AA or European Americans. Specifically, erythroid Duffy null patients had a 17% higher risk of mortality, as well as fewer ventilator-free and organ failure-free days.8 We emphasize that there are likely multiple variables contributing to COVID-19 disease. Susceptibility (infectivity) and severity may have separate determinants, and various human subpopulations are likely to have their own unique risk factors. The role of genetics in influencing a great variety of infectious diseases is well known. For example, HIV 2892

(human immunodeficiency virus) attaches to Duffy protein on RBC which then present HIV to target cells.9 It is reported that in Duffy null individuals, HIV infectivity is substantially higher. We find no analogous studies addressing the infectivity issue for SARS-CoV-2. We do note, however, that the Duffy proteins carry sialic acid, and multiple other coronaviruses have been described as binding to sialic acidexposing proteins.10 Interestingly, interaction with Duffy sialic acids is the mechanism by which Plasmodium vivax begins to infect Duffy positive RBC. It will be interesting, when data become available, to see if COVID-19 is particularly severe in AA with sickle cell anemia (SCA). On the face of it, one might well expect so. However, SCA individuals already have a (subclinical) cytokine storm and activated leukocytes at baseline; they are, therefore, very likely at baseline to have abnormal activation of the vast number of leukocytes sequestered within the pulmonary microvasculature. Since they thus seem to be on the precipice of catastrophic pulmonary inflammation even at baseline, it seems entirely possible that in SCA COVID-19 is so very severe that any influence of Duffy status is simply overwhelmed by the SCA biology itself. A single study has looked at COVID-19 in SCA, but it registered only the presence/absence of specific disease features i.e., not their severity. In summary, we suggest that the greater severity of COVID-19 in African Americans reflects, at least in part, the biological impact of an underlying Duffy null state. Duffy status can easily be documented and, we speculate, this might aide in risk stratification at COVID-19 disease onset. Robert P. Hebbel and Gregory M. Vercellotti Division of Hematology-Oncology-Transplantation, Department of Medicine, University of Minnesota Medical School, Minneapolis, MN, USA Correspondence: ROBERT P. HEBBEL - hebbe001@umn.edu doi:10.3324/haematol.2020.269415

References 1. Tai DBG, Shah A, Doubeni CA, Sia IG, Wieland ML. The disproportionate impact of COVID-19 on racial and ethnic minorities in the United States. Clin Infect Dis. 2020 Jun 20. [Epub ahead of print]. 2. Raisi-Estabragh Z, McCracken C, Bethell MS, et al. Greater risk of severe COVID-19 in Black, Asian and Minority Ethnic populations is not explained by cardiometabolic, socioeconomic or behavioural factors, or by 25(OH)-vitamin D status: Study of 1326 cases from the UK Biobank. J Public Health. 2020;42(3):451-460. 3. Giudicessi JR, Roden DM, Wilde AAM, Ackerman MJ. Genetic susceptibility for COVID-19-assocated sudden cardiac death in African Americans. Heart Rhythm. 2020;17(9):1487-1492. 4. Rappoport N, Simon AJ, Amariglio N, Rechavi G. The Duffy antigen receptor for chemokines, ACKR1,- 'Jeanne DARC' of benign neutropenia. Br J Haematol. 2019;184(4):497-507. 5. Novitzky-Basso I, Rot A. Duffy antigen receptor for chemokines and its involvement in patterning and control of inflammatory chemokines. Front Immunol. 2012;3:266. 6. Girbl T, Lenn T, Perez L, et al. Distinct compartmentalization of the chemokines CXCL1 and CXCL2 and the atypical receptor ACKR1 determine discrete stages of neutrophil diapedesis. Immunity. 2018; 49(6):1062-1076. 7. Howes RE, Patil AP, Piel FB, et al. The global distribution of the Duffy blood group. Nature Comms. 2011;2:266. 8. Kangelaris KN, Sapru A, Calfee CS, et al. The association between a Darc gene polymorphism and clinical outcomes in African American patients with acute lung injury. Chest. 2012;141(5):1160-1169. 9. He W, Neil S, Kulkarni H, et al. Duffy antigen receptor for chemokines mediates trans-infection of HIV-1 from red blood cells to target cells and affects HIV-AIDS susceptibility. Cell Host Microbe. 2008;4(1):52-62. 10. Li W, Hulswit RJG, Widjaja I, et al. Identification of sialic acid-binding function for the Middle East respiratory syndrome coronavirus spike glycoprotein Proc Natl Acad Sci USA. 2017;114(40):E8508-E8517.

haematologica | 2020; 105(12)

haematologica — Vol. 105 n. 12 — December 2020 — 2699-2892