Haematologica, Volume 107, Issue 6 by Haematologica

haematologica Journal of the Ferrata Storti Foundation

haematologica.org

VOL.

107 JUNE 2022

ISSN 0390 - 6078

haematologica Editor-in-Chief Jacob M. Rowe (Jerusalem)

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

Associate Editors Hélène Cavé (Paris), Monika Engelhardt (Freiburg), Steve Lane (Brisbane), Pier Mannuccio Mannucci (Milan), Pavan Reddy (Ann Arbor), David C. Rees (London), Francesco Rodeghiero (Vicenza), Gilles Salles (New York), Kerry Savage (Vancouver), Aaron Schimmer (Toronto), Richard F. Schlenk (Heidelberg), Sonali Smith (Chicago)

Statistical Consultant Catherine Klersy (Pavia)

Editorial Board Walter Ageno (Varese), Sarit Assouline (Montreal), Andrea Bacigalupo (Roma), Taman Bakchoul (Tübingen), Pablo Bartolucci (Créteil), Katherine Borden (Montreal), Marco Cattaneo (Milan), Corey Cutler (Boston), Kate Cwynarski (London), Mary Eapen (Milwaukee), Francesca Gay (Torino), Ajay Gopal (Seattle), Alex Herrera (Duarte), Shai Izraeli (Ramat Gan), Martin Kaiser (London), Marina Konopleva (Houston), Johanna A. Kremer Hovinga (Bern), Nicolaus Kröger (Hamburg), Austin Kulasekararaj (London), Shaji Kumar (Rochester), Ann LaCasce (Boston), Anthony R. Mato (New York), Matthew J. Maurer (Rochester), Neha Mehta-Shah (St. Louis), Alison Moskowitz (New York), Yishai Ofran (Haifa), Farhad Ravandi (Houston), John W. Semple (Lund), Liran Shlush (Toronto), Sara Tasian (Philadelphia), Pieter van Vlieberghe (Ghent), Ofir Wolach (Haifa), Loic Ysebaert (Toulouse)

Managing Director Antonio Majocchi (Pavia)

Editorial Office Lorella Ripari (Office & Peer Review Manager), Simona Giri (Production & Marketing Manager), Paola Cariati (Graphic Designer), Giulia Carlini (Graphic Designer), Igor Poletti (Graphic Designer), Marta Fossati (Peer Review), Diana Serena Ravera (Peer Review), Laura Sterza (Account Administrator)

Assistant Editors Britta Dost (English Editor), Rachel Stenner (English Editor), Bertie Vitry (English Editor), Massimo Senna (Information technology), Idoya Lahortiga (Graphic artist)

Haematologica | 107 - June 2022

Brief information on Haematologica 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, Original articles, Review articles, Perspective articles, Editorials, Guideline articles, Letters to the Editor, Case reports & Case series and Comments. 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 at 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. Subscription. Detailed information about subscriptions is available at www.haematologica.org. Haematologica is an open access journal and access to the online journal 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 printed edition for the year 2022 are as following: 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 | 107 - June 2022

Table of Contents Volume 107, Issue 6: June 2022

About the cover 1229

Images from the Haematologica Atlas of Hematologic Cytology: abnormalities of platelet shape Carlo L. Balduini and Alessandro Pecci https://doi.org/10.3324/haematol.2022.280787

Landmark paper in Hematology 1230

Post-transplant cyclophosphamide: overcoming the HLA barrier for hemopoietic stem cell transplants Andrea Bacigalupo https://doi.org/10.3324/haematol.2022.281256

1232

INTERFERing with the progression of T-cell acute lymphoblastic leukemia: a multifaceted therapy Daniel Herranz https://doi.org/10.3324/haematol.2021.279549

1235

Does the world need germline editing for b-thalassemia? Andreas E. Kulozik https://doi.org/10.3324/haematol.2021.279998

1237

When timing is more important than quantity in COVID-19 vaccination Raul Cordoba https://doi.org/10.3324/haematol.2021.280264

Editorials

Review Series on the Treatment of Thrombocytopenias 1239

Introduction to a review series on the treatment of thrombocytopenic disorders: something old, something new Francesco Rodeghiero https://doi.org/10.3324/haematol.2022.280920

1243

Treatment of chemotherapy-induced thrombocytopenia in patients with non-hematologic malignancies David J. Kuter https://doi.org/10.3324/haematol.2021.279512

1264

Treatment of drug-induced immune thrombocytopenias Irene Marini et al. https://doi.org/10.3324/haematol.2021.279484

1278

Treatment of inherited thrombocytopenias Carlo L. Balduini https://doi.org/10.3324/haematol.2022.280856 Haematologica | 107 - June 2022

III

Articles 1293

Acute Lymphoblastic Leukemia Either IL-7 activation of JAK-STAT or BEZ inhibition of PI3K-AKT-mTOR pathways dominates the single-cell phosphosignature of ex vivo treated pediatric T-cell acute lymphoblastic leukemia cells Daniela Kuzilková et al. https://doi.org/10.3324/haematol.2021.278796

1311

Acute Myeloid Leukemia AXL/MERTK inhibitor ONO-7475 potently synergizes with venetoclax and overcomes venetoclax resistance to kill FLT3-ITD acute myeloid leukemia Sean M. Post et al. https://doi.org/10.3324 haematol.2021.278369

1323

Bone Marrow Failure Transforming growth factor-b signaling modifies the hematopoietic acute inflammatory response to drive bone marrow failure Jose Javier et al. https://doi.org/10.3324 haematol.2020.273292

1335

Chronic Lymphocytic Leukemia Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients Jennifer R. Brown et al. https://doi.org/10.3324/haematol.2021.278901

1347

Hematopoiesis Profound systemic alteration of the immune phenotype and an immunoglobulin switch in Erdheim-Chester disease in 78 patients from a single center Fleur Cohen Aubart et al. https://doi.org/10.3324/haematol.2021.279118

1358

Hemostasis Critical role of peroxisome proliferator-activated receptor α in promoting platelet hyperreactivity and thrombosis under hyperlipidemia Li Li et al. https://doi.org/10.3324 haematol.2021.279770

1374

Hemostasis Proline-rich tyrosine kinase Pyk2 regulates deep vein thrombosis Stefania Momi et al. https://doi.org/10.3324/haematol.2021.279703

1384

Non-Hodgkin Lymphoma CCR6 activation links innate immune responses to mucosa-associated lymphoid tissue lymphoma development Boguslawa Korona et al. https://doi.org/10.3324/haematol.2021.280067

1397

Plasma Cell Disorders Isatuximab plus carfilzomib and dexamethasone versus carfilzomib and dexamethasone in relapsed multiple myeloma patients with renal impairment: IKEMA subgroup analysis Marcelo Capra et al. https://doi.org/10.3324/haematol.2021.279229

Haematologica | 107 - June 2022

1410

Plasma Cell Disorders Apoptosis reprogramming triggered by splicing inhibitors sensitizes multiple myeloma cells to Venetoclax treatment Debora Soncini et al. https://doi.org/10.3324/haematol.2021.279276

1427

Red Cell Biology & its Disorders Correction of RNA splicing defect in b654-thalassemia mice using CRISPR/Cas9 gene-editing technology Dan Lu et al. https://doi.org/10.3324/haematol.2020.278238

1438

Red Cell Biology & its Disorders Ionophore-mediated swelling of erythrocytes as a therapeutic mechanism in sickle cell disease Athena C. Geisness et al. https://doi.org/10.3324 haematol.2021.278666

1448

Direct and indirect anti-leukemic properties of activity-on-target interferons for the treatment of T-cell acute lymphoblastic leukemia Steven Goossens et al. https://doi.org/10.3324/haematol.2021.278913

1454

1460

The highly selective Bruton tyrosine kinase inhibitor acalabrutinib leaves macrophage phagocytosis intact Jonathan J. Pinney et al. https://doi.org/10.3324/haematol.2021.279560

1466

Diagnosis of acute promyelocytic leukemia based on routine biological parameters using machine learning Estelle Cheli et al. https://doi.org/10.3324/haematol.2021.280406

1470

1474

1479

Reduced immunogenicity of a third COVID-19 vaccination among recipients of allogeneic hematopoietic stem cell transplantation Sigrun Einarsdottir et al. https://doi.org/10.3324/haematol.2021.280494

Letters

Haematologica | 107 - June 2022

ABOUT THE COVER

C. Balduini and A. Pecci

Images from the Haematologica Atlas of Hematologic Cytology: abnormalities of platelet shape Carlo L. Balduini1 and Alessandro Pecci2 1 Ferrata Storti Foundation and 2University of Pavia, Pavia, Italy E-mail: carlo.balduini@unipv.it https://doi.org/10.3324/haematol.2022.280787

One of the mechanisms by which megakaryocytes release platelets is the formation of long, thin cytoplasmic extensions that are released into the circulation and eventually transformed into individual platelets. In vitro studies suggested that platelet biogenesis requires the formation of preplatelets (discoid or sausage-shaped giant elements) that have the capacity to convert into barbell-shaped proplatelets and undergo fission into platelets. Although no systematic study has been performed so far, personal experience indicates that preplatelets and proplatelets are nearly never identified in peripheral blood films from healthy subjects, while they are sometimes observed in those from individuals with some forms of inherited thrombocytopenia and in conditions with accelerated platelet turnover, such as immune thrombocytopenia and thrombotic microangiopathies. The images above are blood films from members of a family with ACTN1-related thrombocytopenia. The elongated elements in A-C recall preplatelets, while the barbell-shaped or elongated elements in D-I recall proplateles.1 Disclosures No conflicts of interest to disclose.

References 1. Balduini CL, Pecci A. Inherited thrombocytopenias. Haematologica. 2020;105(Suppl 1):237-247. Haematologica | 107 - June 2022

1229

LANDMARK PAPER IN HEMATOLOGY

A. Bacigalupo

Post-transplant cyclophosphamide: overcoming the HLA barrier to hematopoietic stem cell transplants Andrea Bacigalupo Sezione di Ematologia, Dipartimento di Scienze Radiologiche ed Ematologiche, Università Cattolica del Sacro Cuore, Roma, Italy. E-mail: apbacigalupo@yahoo.com https://doi.org/10.3324/haematol.2022.281256

TITLE

HLA-haploidentical bone marrow transplantation for hematologic malignancies using nonmyeloablative conditioning and high-dose, posttransplantation cyclophosphamide.

AUTHORS Luznik L, O'Donnell PV, Symons HJ, et al. JOURNAL Biology of Blood and Marrow Transplantation. 2008;14(6):641-650. PMID: 18489989 Matching for human leukocyte antigens (HLA) for transplants between donors and recipients has been a major area of research over the past half century: in the early seventies an HLA-matched sibling was required for optimal results. Recipients of transplants from HLA mismatched donors had dismal outcomes, with high mortality due to either rejection or acute graft-versus-host disease (GvHD).1 The restriction of donors to HLA-matched subjects, initially siblings, prompted the institution of the International Unrelated Donor Registry, now comprising over 35 million individuals. In donors, and more so, in unrelated donors, allele matching for at least HLA A,B,C,DRB1, but possibly DQB1 and DP, has been shown to produce the best results and the lowest transplant-related mortality.2 Ex vivo T-cell depletion was developed to reduce GvHD and allow the use of related HLA haploidentical donors: GvHD could be prevented, although graft rejection and immune reconstitution, with a high infectious-related

mortality, remained problems for decades. A new sophisticated technology, capable of selective T-cell depletion (ab), is now available and has improved the outcome of ex vivo T-cell-depleted haploidentical grafts.3 In a landmark paper published in 2008,4 Luznik et al. reported an innovative way of removing alloreactive donor T cells in patients receiving a haploidentical graft, thereby allowing for successful engraftment with little or no GvHD. The method (I call it the revolution) is posttransplant high-dose cyclophosphamide (PTCY), 50 mg/kg on day +3 and day +4, followed by a calcineurin inhibitor and mycophenolate mofetil. The study came from Baltimore, where George Santos in 1966 had shown that high-dose PTCY could prevent GvHD from mismatched grafts in an animal model. However, despite these encouraging pre-clinical results, for over 30 years nobody really had the nerve to give 100 mg/kg of cyclophosphamide on days +3 and +4 post-transplant, because of the

Figure 1. The original Baltimore protocol, designed for haploidentical related marrow grafts, as published in the Biology of Blood and Marrow Transplantation (2008). In this study post-transplant cyclophosphamide was given on days +3 and +4 in one group of patients (Baltimore) and on day +3 only in another group (Seattle), followed by tacrolimus and mycophenolate mofetil. The cumulative incidence of chronic graft-versus-host disease (GvHD) (shown on the right) suggested that two doses of cyclophosphamide are superior to one dose, in protecting patients from chronic GvHD. The conditioning regimen, although nonmyeloablative, is highly immunosuppressive and allows the engraftment of mismatched grafts. Cy: cyclophosphamide; TBI: total body irradiation; G-CSF: granulocyte colony-stimulating factor; MMF: mycophenolate mofetil. Haematologica | 107 - June 2022

1230

LANDMARK PAPER IN HEMATOLOGY

A. Bacigalupo

fear that the infused donor stem cells could be damaged. However, when the Baltimore group published their clinical trial with haploidentical transplants in 2008,4 it became immediately clear that high-dose PTCY was going to be a revolution in the field of allogeneic hematopoietic stem cell transplantation: in one move the authors had successfully crossed the HLA barrier, not with expensive sophisticated technology, but with a simple drug, accessible to everybody, given at a specific time, at a specific dose (that is the secret). PTCY kills alloreactive donor T cells, 72 hours after infusion, and leaves the remaining Tcell repertoire untouched: this is “selective in vivo T-cell depletion”. This is why the use of PTCY has spread rapidly and successfully throughout the world.5 PTCY is currently

used not only in haploidentical grafts, but increasingly so in HLA-identical grafts, both from unrelated and related donors, in malignant and non-malignant disorders. PTCY has made haploidentical transplants as successful as HLA-related transplants (I could not believe my eyes when we started using PTCY in haploidentical grafts in 2009!). There is one last curious fact about this landmark paper: it was rejected by the New England Journal of Medicine, the Journal of Clinical Oncology and Blood, to appear (..only) in Biology of Blood and Marrow Transplantation, and to become the most cited allo-transplant paper in the past 20 years. Disclosures No conflicts of interest to disclose.

References 1. Powles RL, Morgenstern GR, Kay HE, et al. Mismatched family donors for bone-marrow transplantation as treatment for acute leukaemia. Lancet. 1983;1(8325):612-615. 2. Arora M, Weisdorf DJ, Spellman SR, et al. HLA-identical sibling compared with 8/8 matched and mismatched unrelated donor bone marrow transplant for chronic phase chronic myeloid leukemia. J Clin Oncol. 2009;27(10):1644-1652. 3. Merli P, Pagliara D, Galaverna F, et al. TCRαβ/CD19 depleted HSCT from an HLA-haploidentical relative to treat children with

different nonmalignant disorders. Blood Adv. 2022;6(1):281-292. 4. Luznik L, O'Donnell PV, Symons HJ, et al. HLA-haploidentical bone marrow transplantation for hematologic malignancies using nonmyeloablative conditioning and high-dose, posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2008;14(6):641-650. 5. Mussetti A, Paviglianiti A, Parody R, Sureda A. Is post-transplant cyclophosphamide the new methotrexate? J Clin Med. 2021;10(16):3548.

Haematologica | 107 - June 2022

1231

EDITORIAL

D. Herranz

INTERFERing with the progression of T-cell acute lymphoblastic leukemia: a multifaceted therapy Daniel Herranz1,2 1 Rutgers Cancer Institute of New Jersey, Rutgers University, New Brunswick and 2Department of Pharmacology, Robert Wood Johnson Medical School, Rutgers University, Piscataway, NJ, USA. E-mail: dh710@cinj.rutgers.edu https://doi.org/10.3324/haematol.2021.279549

In this issue of Haematologica, a new manuscript by Goosens et al.1 elegantly dissects the direct and indirect therapeutic effects of type-I interferons (IFN-I) in the treatment of T-cell acute lymphoblastic leukemia (T-ALL). Even if advances in T-ALL treatment in the last decades have resulted in high cure rates, 20-50% of patients still relapse and ultimately die, underscoring the need to identify novel therapeutic strategies and to properly stratify patients who might respond to specific targeted agents.2 Interferons have been widely used in the treatment of both solid and hematologic tumors because of their multiple anticancer properties, which include direct cancer cell-intrinsic cytostatic/cytotoxic effects, as well as immune system-mediated cancer cell-extrinsic effects.3 However, IFN-I therapy in cancer has typically resulted in uneven and unreliable results given the poor anticancer properties of these compounds in some tumors together with complex side effects due to their pleiotropic activity.4 In order to assess the direct anticancer activity of IFN-I in T-ALL, Goossens and colleagues treated different human TALL cell lines as well as T-ALL patient-derived xenografts with human IFN-I, both in vitro and in vivo. Consistent with previous literature,5 the antileukemic effects of IFN-I stimulation were only observed in samples that showed JAK/STAT1 activation upon treatment with IFN-I, as measured by pSTAT1 intracellular staining. These results suggest that this fast and easy method to analyze pSTAT1 levels in patients’ cells in vitro could be used as a biomarker to stratify patients who might respond to IFN-I treatment. In order to assess the indirect anticancer effects of IFN-I in T-ALL, authors then used a model of PTEN-null and IFN-Isensitive mouse primary leukemia, upon transplantation into immunocompetent or immunodeficient recipients. These experiments showed that, even if murine IFN-I treatment resulted in antileukemic effects with extended survival in both settings, its therapeutic effects were much stronger in the presence of an intact immune system, demonstrating its significant immune-mediated cellextrinsic antileukemic effects. Next, authors used activityon-target interferons (AcTaferons; AFN)6 in order to specifically direct the activity of IFN-I to CD8+ murine cells (mCD8-AFN), given that roughly half of T-ALL are CD8+ and, moreover, CD8 is also expressed by mouse classical dendritic cells type I (cDC1), which are relevant for

triggering a CD8 cytotoxic response (CTL) upon IFN-I stimulation.7 In this context, as expected, mCD8-AFN treatment in immunodeficient mice resulted in antileukemic effects only when these mice harbored CD8+, not CD8–, mouse leukemias. However, rather unexpectedly, similar results were also obtained when these CD8+ or CD8–cells were transplanted into immunocompetent mice. By contrast, when authors used a different AFN directed at Clec9a (mClec9a-AFN), which has been shown to elicit a cDC1-mediated antitumor response in other tumors,7 significant antileukemic effects were observed in vivo in immunocompetent mice harboring both CD8+ or CD8– leukemias. Importantly, mClec9a treatment even resulted in 20-40% cure rates, while no leukemic mice were cured either by mCD8-AFN or mIFN treatment itself. Interestingly, and as expected (given that Clec9a is not expressed in normal or malignant T cells), this antileukemic effect was completely absent if these leukemias were transplanted into immunodeficient mice, highlighting that mClec9a-AFN antileukemic effects are driven exclusively by immune system-mediated antitumor responses. These results showing strong indirect effects for mClec9aAFN but reduced/absent effects for mCD8-AFN are intriguing. Previous studies showed that IFN signaling in dendritic cells, but not in T cells, is required for AFN antitumor activity, however, optimal antitumor effects of AFN are still dependent on the presence of CD8+ cytotoxic T cells (CTL),8 and priming and activation of CTL requires prior activation and maturation of dendritic cells. One possible explanation of these discordant results might be that binding of the mCD8-AFN on CTL could neutralize their cytotoxic properties; however, this is unlikely since mCD8-AFN was previously shown to have significant additive antitumor effect in combination with tumor necrosis factor-based targeted therapy.9 Moreover, in the study by Goosens et al.1 mCD8-AFN treatment seemed to translate into improved antileukemic effects in CD8+ leukemias transplanted into immunocompetent mice, as compared to immunocompromised mice. Another interesting but bizarre possibility to reconcile these results might be that, in order to elicit its indirect immunemediated effects, mCD8-AFN treatment might first require some direct cell-intrinsic effects to take place, which would thus explain indirect effects being observed only on CD8+ leukemias. Finally, it is also possible that mCD8-AFN does not activate cDC1 cells to the same extent as

Haematologica | 107 - June 2022

1232

EDITORIAL

D. Herranz

Figure 1. Direct and indirect antileukemic effects of different interferon treatments. (A) Effects of IFN-I treatment. (B) Effects of IFN-I treatment targeted to CD8+ cells using AcTaferons (mCD8-AFN). (C) Effects of IFN-I treatment targeted to Clec9a+ cells using AcTaferons (mClec9a-AFN). IFN-1: type-1 interferons; T-ALL: T-cell acute lymphoblastic leukemia: cDC: classical dendritic cells; CTL: cytotoxic lymphocytes; AFN: activity-on-target interferons.

mClec9a-AFN, or that a different Clec9a+ hematologic population might be more relevant in order to mediate the therapeutic effects observed. Related to this, it would be interesting to test the potential synergistic effects of mCD8-AFN and mClec9a-AFN when used concomitantly to treat CD8+ leukemias. Further research is therefore warranted to uncover the biological reasons for these differences. Regardless, the important findings of Goossens and colleagues serve to revitalize the field of interferons for the treatment of T-cell malignancies, as AFN show

significantly reduced side effects as compared to interferon itself, and both mCD8-AFN and mClec9a-AFN showed direct and/or indirect antileukemic properties which could be exploited for the treatment of interferonsensitive leukemias alone or in combination with classical chemotherapy regimens which, in turn, might help to reduce or prevent relapses in these patients. Disclosures No conflicts of interest to disclose.

Haematologica | 107 - June 2022

1233

EDITORIAL

D. Herranz

References 1. Goossens S, Cauwels, A, Pieters, T, et al. Direct and indirect anti-leukemic properties of activity-on-target interferons for the treatment of T-cell acute lymphoblastic leukemia. Haematologica. 2022;107(6):1448-1453. 2. Hunger SP, Mullighan CG. Acute lymphoblastic leukemia in children. N Engl J Med. 2015;373(16):1541-1552. 3. Parker BS, Rautela J, Hertzog PJ. Antitumour actions of interferons: implications for cancer therapy. Nat Rev Cancer. 2016;16(3):131-144. 4. Zhang X, Wang S, Zhu Y, et al. Double-edged effects of interferons on the regulation of cancer-immunity cycle. Oncoimmunology. 2021;10(1):1929005. 5. Lesinski GB, Anghelina M, Zimmerer J, et al. The antitumor effects of IFN-alpha are abrogated in a STAT1-deficient mouse. J

Clin Invest. 2003;112(2):170-180. 6. Garcin G, Paul F, Staufenbiel M, et al. High efficiency cell-specific targeting of cytokine activity. Nat Commun. 2014;5:3016. 7. Cauwels A, Van Lint S, Paul F, et al. Delivering type I interferon to dendritic cells empowers tumor eradication and immune combination treatments. Cancer Res. 2018;78(2):463-474. 8. Cauwels A, Van Lint S, Garcin G, et al. A safe and highly efficient tumor-targeted type I interferon immunotherapy depends on the tumor microenvironment. Oncoimmunology. 2018;7(3):e1398876. 9. Huyghe L, Van Parys A, Cauwels A, et al. Safe eradication of large established tumors using neovasculature-targeted tumor necrosis factor-based therapies. EMBO Mol Med. 2020;12(2):e11223.

Haematologica | 107 - June 2022

1234

EDITORIAL

A.E. Kulozik

Does the world need germline editing for b-thalassemia? Andreas E. Kulozik Chairman, Department of Pediatric Oncology, Hematology and Immunology, University of Heidelberg, Hopp-Children Cancer Center Heidelberg (KiTZ), Germany E-mail: andreas.kulozik@med.uni-heidelberg.de https://doi.org/10.3324/haematol.2021.279998

In this issue of Haematologica, Lu and co-workers report the correction, by CRIPSR-Cas9 gene editing, of the b-globin gene in the germ cells of a mouse model carrying the human b-globin gene with the IVS2-654 thalassemia (b654) mutation that is common in East Asian patients with bthalassemia. This mutation creates a new donor splice site in intron 2 of the b-globin gene which co-operates with a cryptic splice site further downstream to insert an abnormal exon containing a premature stop codon. This mutation thus results in the almost complete inactivation of the affected b-globin allele. Lu and co-workers designed two sgRNA that simultaneously target and delete both the novel 5’ donor splice site and the cryptic acceptor splice site. Microinjection of the two sgRNA into fertilized murine eggs together with the nuclease Cas9 produced 12/37 viable mice with editing of the target locus, half of which carried the desired deletion of the target region. Remarkably, the peripheral blood of seven of the 12 edited mice showed correct splicing of the b-globin gene, and six of these seven exclusively expressed correctly processed RNA and normal human globin. Finally, the authors demonstrated that mice with successfully edited b654 human b-globin genes have much-improved hematologic parameters and survival compared to nonedited mice.1 These results indicate that the faulty RNA processing induced by the common b654 thalassemia mutation can be corrected by a complex and innovative editing strategy. b-thalassemia is one of the most common genetic disorders worldwide and has been a target for the development of gene therapy for decades. In fact, the b-globin gene was the first human gene to be cloned more than 40 years ago and more than 300 mutations resulting in thalassemia have been described since then. Early attempts at gene therapy were not successful.2 However, with the advent of modern vector technology the first reports of successful somatic thalassemia gene therapy of hematopoietic cells began to emerge some 10 years ago.3 More recently, systematic clinical studies have employed self-inactivating lentiviral vectors containing a therapeutic b-globin gene. Such constructs are used to transduce human hematopoietic stem cells mobilized and isolated from affected patients thus adding a functional b-globin gene into the genome of these cells playing a key role in the pathogenesis in thalassemia.4 The game-changing efficacy and safety of this procedure convinced the European Medicine Agency (EMA) to license such a product for

the treatment of a defined group of patients with transfusion dependent b-thalassemia. Reversing the perinatal hemoglobin switch from fetal to adult globin synthesis by CRIPSR-Cas9-mediated inactivation of BCL11A, the central erythroid-specific negative regulator of g-globin gene expression, in hematopoietic stem cells has been reported to induce high-level HbF synthesis resulting in transfusion independence.5 However, both gene addition and gene editing strategies may potentially carry the risk of serious long-term complications by insertional mutagenesis.6 Concerns about the safety of gene editing have recently been raised by reports describing that the double-stranded DNA breaks induced by Cas9 can trigger a TP53-mediated DNA damage response and major structural changes of the DNA resulting in the formation of micronuclei and chromothripsis, one of the major mechanisms of carcinogenesis.7,8 While these safety concerns have so far not been an issue in either gene addition or gene editing studies in patients with bthalassemia, long-term follow-up will inform us whether such concerns will be relevant in the long run. When considering the therapeutic use of genetic engineering it is conceptually important to distinguish between somatic gene therapy targeting a disease-relevant cell type or tissue and germline engineering that introduces heritable genetic changes. While the former strategy has been used to develop novel treatments of several genetic and acquired diseases, the latter is commonly used in animal models aimed at the understanding of key pathological mechanisms. In fact, manipulation to introduce heritable changes into human germ lines has been viewed very critically by several European, NorthAmerican and Chinese scientific societies and is legally banned in the European Union.9,10 Despite this, Lu and co-workers consider that the results of their manipulation of murine germ cells “provide a groundwork for the exploration of b654-thalassemia therapy in the future”. Notably, these authors report that the manipulation of fertilized murine eggs induced several unexpected structural variants including inversions, unexpected single nucleotide substitutions and larger deletions than those the pair of sgRNA were designed to generate. These findings thus highlight the potential of gene editing to induce unexpected genetic variants that go beyond simple off-target effects induced by sequence similarities between the guide RNA and other loci of the genome. Consistent with the findings of Pellman’s group in edited

Haematologica | 107 - June 2022

1235

EDITORIAL

A.E. Kulozik

hematopoietic stem cells7 these findings indicate that gene editing is more imprecise than widely thought and that its clinical use should be developed with caution. While this is certainly true for the development of somatic gene therapy, these findings are clearly particularly relevant when heritable manipulation of human germ cells is considered for the exploration of b-thalassemia therapy. Even with a perfectly precise technology the ethical concerns of introducing heritable changes into human germ cells have been grave enough to ban human germline engineering for therapeutic purposes. And to those who do not share the concept of ethical reservations against germline manipulation,

the substantial technical uncertainties relating to the lack of specificity of the procedure should be reason enough to stand off. One might therefore wonder whether Lu and coworkers may want to adapt their innovative technical approach to the development of somatic gene therapy, which may also carry the potential of risk but does not cause heritable changes of germ cells thus limiting any potential unwanted outcome to the individual patient. Disclosures AEK has received honoraria from Novartis, Sanofi, Novo Nordisk, Celgene and Bluebird Bio.

References 1. Lu D, Gong X, Fang Y, et al. Correction of RNA splicing defect in ß654-thalassemia mice using CRISPR/Cas9 gene-editing technology. Haematologica. 2022;107(6)1427-1437. 2. Kolata GB, Wade N. Human gene treatment stirs new debate. Science. 1980;210(4468):407. 3. Cavazzana-Calvo M, Payen E, Negre O, et al. Transfusion independence and HMGA2 activation after gene therapy of human b-thalassaemia. Nature. 2010;467(7313):318-322. 4. Thompson AA, Walters MC, Kwiatkowski J, et al. Gene therapy in patients with transfusion-dependent β-thalassemia. N Engl J Med. 2018;378(16):1479-1493. 5. Frangoul H, Altshuler D, Cappellini MD, et al. CRISPR-Cas9 gene editing for sickle cell disease and β-thalassemia. N Engl J Med. 2021;384(3):252-260.

6. Kunz JB, Kulozik AE. Gene therapy of the hemoglobinopathies. Hemasphere. 2020;4(5):e479. 7. Leibowitz ML, Papathanasiou S, Doerfler PA, et al. Chromothripsis as an on-target consequence of CRISPR-Cas9 genome editing. Nat Genet. 2021;53(6):895-905. 8. Haapaniemi E, Botla S, Persson J, Schmierer B, Taipale J. CRISPR-Cas9 genome editing induces a p53-mediated DNA damage response. Nat Med. 2018;24(7):927-930. 9. Coller BS. Ethics of human genome editing. Annu Rev Med. 2019;70:289-305. 10. Büning H, Griesenbach U, Fehse B, et al. Consensus statement of European societies of gene and cell therapy on the reported birth of genome-edited babies in China. Hum Gene Ther. 2018;29(12):1337-1338.

Haematologica | 107 - June 2022

1236

EDITORIAL

Raul Cordoba

When timing is more important than quantity in COVID-19 vaccination Raul Cordoba Department of Haematology, Fundacion Jimenez Diaz University Hospital, Health Research Institute IIS-FJD, Madrid, Spain E-mail: raul.cordoba@fjd.es https://doi.org/10.3324/haematol.2021.280264

In this issue of Haematologica, Kohn et al. report the results of a prospective, single-center, non-randomized study,1 shedding more light on the efficacy of vaccination against coronavirus disease 2019 (COVID-19) in patients with non-Hodgkin lymphoma (NHL) or chronic lymphocytic leukemia (CLL) who have received prior treatment with anti-CD20 antibody-containing therapies and/or have low levels of serum immunoglobulins. Patients with hematologic malignancies may suffer from impairments of primary immunity due to biological features of the disease but also from secondary immunodeficiencies related to therapies. These patients are at higher risk of severe infections, which may result in a worse survival.2 One of the best examples is CLL, in which infections are a main contributor to morbidity and mortality driven by an impaired immune system. Moreover, treatment initiation is likely to induce a more profound immune dysfunction that further increases the risk of severe infections.3 With regard to infection by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), patients with NHL or CLL are at higher risk of developing severe and/or prolonged forms of COVID-19.4,5 In the EPICOVIDEHA study, one of the largest registries of COVID-19 in patients with hematologic diseases published so far, patients with NHL and CLL showed a mortality rate of 31.8% and 28.3%, respectively.6 It is known that the seroconversion rate after SARS-CoV2 infection is low in these groups of patients after recent treatment with anti-CD20 monoclonal antibodies. The same poor serological response has been shown after two doses of mRNA SARS-CoV-2 vaccination, with a rate of 5% in patients with CLL recently treated with antiCD20 antibodies7 and 3% in NHL patients vaccinated within 45 days after administration of the last dose of monoclonal antibody, although the rate reached up to 80% in patients vaccinated more than 1 year after this therapy.8 One burning question is whether a third dose of mRNA SARS-CoV-2 vaccine may enhance the serological response in this group of poor responders. The study by Kohn et al. analyses the outcome of 100 patients with NHL or CLL who received a third dose of mRNA SARS-CoV-2 vaccine at the discretion of each phy-

sician in a non-randomized study. Serology was performed at least 2 weeks after the last vaccination, with a median interval between serology and the last vaccine injection of 47 days. Half of the patients did not show a serological response to vaccination. Patients who did not have a serological response had significantly lower lymphocyte counts, B-cell counts and IgG levels than those patients with a demonstrated serological response. These factors may, therefore, be considered for further vaccination strategies in these groups of patients. Patients who had received any treatment within the year before their first vaccine injection were at higher risk than other patients of not developing a serological response, with anti-CD20 therapy being strongly associated with the risk of absence of a serological response. Among the patients in whom the last anti-CD20 administration was within 1 year prior to their first vaccine injection, 74% did not seroconvert despite 16 out of 25 patients having received a third dose. In this study, Kohn et al. did not find an association between the number of vaccine injections and seroconversion rate, with an absence of serological response in 58.3% and 46.9% of patients who received three or two doses, respectively. Regardless of the number of vaccine administrations, patients who did not receive anti-CD20 antibodies within 1 year prior to their first vaccine injection had higher levels of anti-spike IgG levels. These findings are supported by those of another study which included patients with CLL given a third dose of mRNA SARS-CoV-2 vaccine and showed a moderate increase in SARS-CoV-2 anti-spike IgG levels after the third dose in patients treated for CLL, although the increase in IgG levels had a limited impact on the prevalence of antispike IgG ≥30 BAU/mL in patients treated for CLL, which rose from 5% after two doses to 45% after receiving the third dose.9 Administration of a third dose of mRNA SARSCoV-2 vaccine seems not to overcome the poor serological response observed in patients who had anti-CD20 treatment within 1 year prior to their first vaccine injection or low IgG levels. Disclosures No conflicts of interest to disclose.

Haematologica | 107 - June 2022

1237

EDITORIAL

Raul Cordoba

References 1. Kohn M, Delord M, Chbat M, et al. A third anti-SARS-CoV-2 mRNA dose does not overcome the pejorative impact of antiCD20 therapy and/or low immunoglobulin levels in patients with lymphoma or chronic lymphocytic leukemia Haematologica. 2022;107(6):1454-1459. 2. Jolles S, Smith BD, Vinh DC, et al. Risk factors for severe infections in secondary immunodeficiency: a retrospective US administrative claims study in patients with hematological malignancies. Leuk Lymphoma. 2022;63(1):64-73. 3. Langerbeins P, Eichhorst B. Immune dysfunction in patients with chronic lymphocytic leukemia and challenges during COVID-19 pandemic. Acta Haematol. 2021;144(5):508-518. 4. Duléry R, Lamure S, Delord M, et al. Prolonged in-hospital stay and higher mortality after Covid-19 among patients with nonHodgkin lymphoma treated with B-cell depleting immunotherapy. Am J Hematol. 2021;96(8):934-944. 5. Mato AR, Roeker LE, Lamanna N, et al. Outcomes of COVID-19 in

patients with CLL: a multicenter international experience. Blood. 2020;136(10):1134-1143. 6. Pagano L, Salmanton-García J, Marchesi F, et al. COVID-19 infection in adult patients with hematological malignancies: a European Hematology Association Survey (EPICOVIDEHA). J Hematol Oncol. 2021;14(1):168. 7. Benjamini O, Rokach L, Itchaki G, et al. Safety and efficacy of the BNT162b mRNA COVID-19 vaccine in patients with chronic lymphocytic leukemia. Haematologica. 2022;107(3):625-634 8. Gurion R, Rozovski U, Itchaki G, et al. Humoral serologic response to the BNT162b2 vaccine is abrogated in lymphoma patients within the first 12 months following treatment with anti-CD20 antibodies. Haematologica. 2022;107(3):715-720. 9. Marlet J, Gatault P, Maakaroun Z, et al. Antibody responses after a third dose of COVID-19 vaccine in kidney transplant recipients and patients treated for chronic lymphocytic leukemia. Vaccines (Basel). 2021;9(10):1055.

Haematologica | 107 - June 2022

1238

INTRODUCTION TO REVIEW SERIES

Introduction to a review series on the treatment of thrombocytopenic disorders: something old, something new Francesco Rodeghiero

Correspondence:

Hematology Project Foundation, affiliated to the Department of Hematology, San Bortolo Hospital, Vicenza, Italy

Rodeghiero Francesco rodeghiero@hemato.ven.it Received: March 23, 2022. Accepted: March 28, 2022. https://doi.org/10.3324/haematol.2022.280920 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Thrombocytopenia may be a major component of several disorders, exposing patients to an increased risk of bleeding and/or heralding more complex, but often underestimated, clinical scenarios. In a series of reviews in Haematologica, leading authors comprehensively discuss the management of thrombocytopenia in three distinct clinical areas generally neglected and still lacking specific treatments.1-3 Inherited thrombocytopenias,1 drug-induced immune thrombocytopenia2 and chemotherapy-induced thrombocytopenia3 were chosen as illustrative epitomes. In the review on drug-induced immune thrombocytopenia, three subsections expand beyond classical cases of drugassociated thrombocytopenia to include, in addition to the well-known heparin-induced thrombocytopenia, two rare complications recently reported in association with vaccination against coronavirus disease 2019 (COVID-19) – vaccine-induced thrombotic thrombocytopenia4 and vaccine-associated immune thrombocytopenia.5 The readers will learn how to best suspect, diagnose and improve the outcome of these disorders. At the same time, they will gain some insight into the fascinating world of platelets, sometimes considered a Cinderella of hematology, and envisage the long but promising way toward new scientific discoveries for a better understanding and treatment of platelet disorders. Phylogenetic reconstruction, together with morphological and functional studies in avian and mammalian species, suggest that the small anucleated cells named platelets (Plättchen in German) by the Italian scientist Giulio Bizzozero6 back in 1881, represent the evolutionary vestige of the more primitive "nucleated thrombocyte”. Nucleated thrombocytes of our birds evolved from these ancestors and, reminiscent of their primitive function, remain mostly committed to innate defense mechanisms and wound healing. With the appearance of mammals, some 150 million of years ago, evolutionary forces, driven by the need to ensure more efficient protection from blood loss in-

duced by parturition and trauma, led to the development of anucleated platelets endowed with augmented hemostatic responses.7 As an undesirable consequence, this hemostatic advantage turned out to be a primary cause of the high rate of thrombotic disorders that plague our times. The importance of the more primitive hemostatic mechanism stands out in the natural experiments represented by those clinical disorders in which thrombocytopenia is a sufficient cause of bleeding, despite an intact coagulation system. Quite surprisingly, contrary to what is expected, there are clinical situations accompanied by a reduced number of platelets in which thrombosis is also a major factor of morbidity or mortality, posing dramatic management dilemmas (see, for example, heparin-induced thrombocytopenia and vaccine-induced thrombotic thrombocytopenia). Conversely, thrombocytosis may be accompanied by a hemorrhagic tendency as in essential thrombocythemia, whose first denomination was ‘hemorrhagic thrombocythemia’.8 Finally, as detailed in Balduini’s review,1 in a significant proportion of cases with inherited thrombocytopenias an associated impaired platelet function may further aggravate the bleeding tendency. Even worse, in rarer cases, the genetic lesions have impacts beyond thrombopoiesis and thrombocytopenia is just one component of more complex syndromic forms or may herald the future development of hematologic malignancies or bone marrow aplasia or fibrosis, raising ethical dilemmas on how to best inform patients on their condition. Figure 1 depicts the many inherited and acquired causes that can lead to thrombocytopenia and the main pathogenic mechanisms involved. It is quite evident that a precise diagnosis is essential not only for directing prognosis and treatment, but also for excluding any underlying or associated disorder requiring prompt identification. As Balduini mentions in his review,1 the prevalence of in-

Haematologica | 107 - June 2022

1239

REVIEW SERIES - A review series on thrombocytopenia treatment

F. Rodeghiero

Figure 1. Pathogenic mechanisms of congenital and acquired thrombocytopenia. The pathogenic mechanisms responsible for thrombocytopenia can be grouped into three categories: reduced production of platelets, increased consumption/destruction in the bloodstream, and increased destruction/entrapment in the spleen. The causes of thrombocytopenia are shown in the figure and many of them act through more than one pathogenic mechanism. Increased awareness of the many distinct forms of thrombocytopenia is a prerequisite for appropriate prevention, as in drug-associated thrombocytopenia, or for making a precise prognosis or planning comprehensive treatment exploiting as best possible the limited options available for the various different types of congenital and acquired thrombocytopenia.

herited thrombocytopenias is reported to be as high as more than two cases per 100.000 persons, an order of magnitude similar to that of immune thrombocytopenic purpura (ITP). Nevertheless, only a very tiny proportion of patients is diagnosed with inherited thrombocytopenias at major hematology centers, thus highlighting the underdiagnosis of this pathology. No prevalence data are available for drug-induced immune thrombocytopenia, but its diagnosis is very difficult without a high index of suspicion and underdiagnosis is anticipated also for this condition. Indeed, even in expert centers, several cases of isolated thrombocytopenia are initially misdiagnosed as ITP with potentially dangerous consequences.9 For most of the disorders reviewed in the series, reducing or aborting bleeding – the enemy platelets were evolved to fight against - is the primary goal that should dictate management. Whereas many agents can negatively interfere with the thrombotic potential of platelets, thus reducing platelet-dependent thrombotic risk at the cost of slightly increasing the hemorrhagic risk, there are no drugs that can augment platelet hemostatic activities

without increasing thrombotic risk. Some “hemostatic” agents act in an indirect way by reducing fibrinolysis or by potentiating the clotting mechanism. These include desmopressin, which releases high molecular weight multimers of von Willebrand factor and factor VIII into the circulation, and recombinant activated factor VII, which can directly activate the extrinsic pathway of coagulation. Unfortunately, there are no robust data to support the use of these agents in thrombocytopenic patients and recombinant activated factor VII may lead to an increased risk of thrombosis. Educating patients to avoid risky situations and drugs that may interfere with platelet function remains a principal duty of treating physicians, but ultimately the only practical way to reduce the hemorrhagic risk is to increase platelet concentration in the circulation by platelet transfusions or by stimulating platelet production by megakaryocytes with thrombopoietin-receptor agonists (TPO-RA) or with human recombinant thrombopoietin, currently available only in China. It is still debatable what is an optimal platelet count sufficient to avoid spontaneous bleeding or to prevent

Haematologica | 107 - June 2022

1240

REVIEW SERIES - A review series on thrombocytopenia treatment hemorrhages associated with menstruation or parturition or in preparation for surgery or dental extraction. We have to admit that despite intensive research we still do not understand why patients with the same platelet count manifest bleeding at different frequencies and with different severity. The seminal research by Harker and Slichter10 proved that the bleeding time (a surrogate measurement in vivo of the bleeding tendency) is differently correlated to the actual platelet count in different disorders. Indeed, at parity of platelet count, the bleeding time was most prolonged in Wiskott-Aldrich syndrome, followed by aplastic anemia and only moderately increased in ITP. A correlation was found with the platelet volume, in keeping with the increased volume of the more active and young, regenerating platelets. Experience with ITP also suggests that platelet count and platelet activity are not the only determinants of bleeding severity. To escape from vessels, blood must go through the endothelium and basal membrane before a platelet plug is formed. Below a critical number of platelets, the steadystate trophic effects on the endothelium are impaired and the multimolecular vascular endothelium cadherin complex breaks down, with subsequent loss of the intercellular barrier, permitting extravasation of red cells into the surrounding tissues.11 Furthermore, basal membrane breakage may be favored by a second hit such as inflammation. Remarkably, in both humans and rabbits, glucocorticoid treatment corrects endothelium abnormalities together with the bleeding manifestations associated with severe thrombocytopenia,12 in accordance with the pioneering clinical observation of the early 1950s that corticosteroids rapidly improved bleeding manifestations of ITP within the first 2 days of treatment and before an increase in platelet count could be observed. Unfortunately, this beneficial effect is not found in the clinical situations described in this review, once again reinforcing the concept that different pathogenic mechanisms and different clinical circumstances require distinct therapeutic approaches. Once again, ITP provides illustrative examples. In this autoimmune disorder, corticosteroids, intravenous immunoglobulins, immunosuppressive agents, splenectomy and TPO-RA are all variably effective. However, none of them is indicated in classical drug-induced immune thrombocytopenia in which the only treatment is to stop the incriminated drug. In the case of heparin-induced thrombocytopenia or vaccine-induced thrombotic thrombocytopenia, it is not sufficient to stop heparin or its congeners and anticoagulation with non-heparin agents remains the mainstay of treatment. Unless there is bleeding, no attempt to increase platelet count is recommended, but intravenous administration of immunoglobulins is felt useful by some investigators. For inherited thrombocytopenias, splenectomy and administration of TPO-RA (still an experimental therapy) may be

F. Rodeghiero

appropriate in very selected cases but, unfortunately, platelet transfusion is still of fundamental importance in cases of severe bleeding. The potential utility of TPO-RA in chemotherapy-induced thrombocytopenia is an ongoing subject of intensive investigation, and is a major focus of Kuter’s review on the treatment of chemotherapy-induced thrombocytopenia in non-hematologic malignancies.3 Chemotherapy regimens for cancer treatment are designed to deliver the maximum tolerable dose intensity, an objective often limited by organ toxicity or by myelotoxicity, forcing reductions in the dosage with the risk of decreasing response rate and survival. Blood transfusion and growth factors may help to reduce anemia and neutropenia, but until the advent of TPO-RA (romiplostim and eltrombopag and, more recently, avatrombopag) no treatment was available to increase platelet production. There is no doubt that the occurrence of thrombocytopenia during chemotherapy for solid tumors (but similar considerations also apply to chemotherapy for malignant hematologic diseases) has several negative effects such as forcing the delivery of less than optimal chemotherapy, in terms of dosage and timing, thus leading to a reduction of relative dose intensity, and causing an increased risk of bleeding, with a negative impact on quality of life. Moreover, platelet transfusion may become necessary even in the absence of hemorrhage when the platelet count falls below a minimal threshold considered sufficiently safe (usually set at 10x109/L or 20x109/L in febrile patients) with related risks of infection and refractoriness due to alloimmunization and the additional burden of hospitalization. Many studies have been and are being conducted with traditional TPO-RA (romiplostim and eltrombopag) or with more recently introduced ones, such as avatrombopag, lusutrombopag and hetrombopag, in efforts to fully exploit the potential of TPO-RA to eliminate thrombocytopenia as an additional limiting factor to giving the desired dose intensity of chemotherapy.13 However, although chemotherapy-induced thrombocytopenia may affect a sizable proportion of patients with solid tumors, grade 3-4 thrombocytopenia occurs in less than 10% and appears to be isolated in less than 1%.14 Thus, considering that other cytopenias or toxicities may lead to a reduction of dose intensity, it is difficult to envisage how much the additional reduction or elimination of chemotherapy-induced thrombocytopenia will contribute to improving relative dose intensity. One can only concur with Kuter’s conclusion that the beneficial effect of TPO-RA on relative dose intensity, tumor response, transfusion, bleeding and survival have not yet been adequately demonstrated and that we still do not have sufficient evidence for a widespread adoption of TPO-RA use.3 On the other hand, several studies have shown that platelet count is, on average, higher when TPORA are administered although these studies differ as to de-

Haematologica | 107 - June 2022

1241

REVIEW SERIES - A review series on thrombocytopenia treatment sign, patient selection, type of chemotherapy, timing of TPO-RA administration and with regard to single or composite endpoints. It is to be hoped that building on the partial success of these studies might give impetus to sounder investigational protocols. We are confident that increased awareness of the relevance of the disorders discussed in this review series may stimulate further basic and clinical research. Cutting-edge technologies and bioinformatics applied to genomics will contribute to the identification of an increasing number of inherited thrombocytopenias and to make their diagnosis easier and more precise. Perhaps, more importantly, we need a reinvigorated international collaboration to produce large cohorts of patients with long-term follow-up to better capture the most relevant outcomes of these disorders. As to drug-induced immune thrombocytopenia, the

F. Rodeghiero

recent unexpected emergence of thrombocytopenic disorders after COVID-19 vaccination or after an increasing list of innovative drugs, such as immune checkpoint inhibitors, calls for supranational pharmacovigilance systems and creation of easily accessible, updated repertories of drugs identified or suspected to cause immune or non-immune thrombocytopenias. Waiting for artificial platelets to become available in clinical settings and possibly make platelet transfusion an obsolete practice, the potential benefit of increasing natural thrombopoiesis in chemotherapy-induced thrombocytopenia and other thrombocytopenias needs to be further exploited by well-designed prospective studies investigating old and new TPO-RA. Disclosures No conflicts of interest to disclose.

References 1. Balduini C. Treatment of inherited thrombocytopenias. Haematologica. 2022;107(6):1278-1292. 2. Marini I, Uzun G, Jamal K, Bakchoul T. Treatment of druginduced immune thrombocytopenias. Haematologica. 2022;107(6):1264-1277. 3. Kuter DJ. Treatment of chemotherapy-induced thrombocytopenia in patients with nonhematological malignancies. Haematologica. 2022;107(6):1243-1263. 4. Cines DB, Bussel JB. SARS-CoV-2 Vaccine–Induced Immune Thrombotic Thrombocytopenia. N Engl J Med. 2021;384(23):2254-2256. 5. Lee EJ, Beltrami-Moreira M, Al-Samkari H, et al. SARS-CoV-2 Vaccination and immune thrombocytopenia in de novo and preexisting ITP patients. Blood. 2022;139(10):1564-1574. 6. Bizzozero G. Ueber einen neuen Forrnbestandteil des Blutes und dessen Rolle bei der Thrombose und Blutgerinnung. Virchows Archiv für Pathologische Anatomie und Physiologie und für Klinische Medizin. 1882;90:261-332. 7. Schmaier AA, Stalker TJ, Runge JJ, et al. Occlusive thrombi arise in mammals but not birds in response to arterial injury: evolutionary insight into human cardiovascular disease. Blood. 2011;118(13):3661-3669.

8. Epstein E, Goedel A. Hämorrhagische thrombocythämie bei vasculärer schrumpfmilz. Virchows Archiv für Pathologische Anatomie und Physiologie und für Klinische Medizin. 1934;292:233-248. 9. Arnold DM, Nazy I, Clare R, et al. Misdiagnosis of primary immune thrombocytopenia and frequency of bleeding: lessons from the McMaster ITP Registry. Blood Adv. 2017;1(25):2414-2420. 10. Harker LA, Slichter SJ. The bleeding time as a screening test for evaluation of platelet function. N Engl J Med. 1972;287(4):155-159. 11. Nachman RL, Shahin R. Platelets, petechiae, and preservation of the vascular wall. N Engl J Med. 2008;359(12):1261-1270. 12. Kitchens CS. Human thrombocytopenia is associated with structural abnormalities of the endothelium that are ameliorated by glucocorticosteroid administration. Blood. 1986;67(1):203-206. 13. Lozano ML, Rodeghiero F. Thrombopoietin receptor agonist in chemotherapy-induced thrombocytopenia. Lancet Haematol. 2022;9(3):e168-e169. 14. Ten Berg MJ, van den Bemt PM, Shantakumar S, et al. Thrombocytopenia in adult cancer patients receiving cytotoxic chemotherapy: results from a retrospective hospital-based cohort study. Drug Saf. 2011;34(12):1151-1160.

Haematologica | 107 - June 2022

1242

REVIEW SERIES

Treatment of chemotherapy-induced thrombocytopenia in patients with non-hematologic malignancies David J. Kuter

Correspondence: Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA

David J. Kuter dkuter@mgh.harvard.edu Received: December 29, 2021. Accepted: January 28, 2022. https://doi.org/10.3324/haematol.2021.279512 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract Chemotherapy-induced thrombocytopenia (CIT) is a common complication of the treatment of non-hematologic malignancies. Many patient-related variables (e.g., age, tumor type, number of prior chemotherapy cycles, amount of bone marrow tumor involvement) determine the extent of CIT. CIT is related to the type and dose of chemotherapy, with regimens containing gemcitabine, platinum, or temozolomide producing it most commonly. Bleeding and the need for platelet transfusions in CIT are rather uncommon except in patients with platelet counts below 25x109/L in whom bleeding rates increase significantly and platelet transfusions are the only treatment. Nonetheless, platelet counts below 70x109/L present a challenge. In patients with such counts, it is important to exclude other causes of thrombocytopenia (medications, infection, thrombotic microangiopathy, post-transfusion purpura, coagulopathy and immune thrombocytopenia). If these are not present, the common approach is to reduce chemotherapy dose intensity or switch to other agents. Unfortunately decreasing relative dose intensity is associated with reduced tumor response and remission rates. Thrombopoietic growth factors (recombinant human thrombopoietin, pegylated human megakaryocyte growth and development factor, romiplostim, eltrombopag, avatrombopag and hetrombopag) improve pretreatment and nadir platelet counts, reduce the need for platelet transfusions, and enable chemotherapy dose intensity to be maintained. National Comprehensive Cancer Network guidelines permit their use but their widespread adoption awaits adequate phase III randomized, placebo-controlled studies demonstrating maintenance of relative dose intensity, reduction of platelet transfusions and bleeding, and possibly improved survival. Their potential appropriate use also depends on consensus by the oncology community as to what constitutes an appropriate pretreatment platelet count as well as identification of patient-related and treatment variables that might predict bleeding.

Introduction Thrombocytopenia is a common problem in patients with cancer, whether due to the underlying disease, infection, other medications or cancer treatment with chemotherapy or radiation. Thrombocytopenia creates a number of problems in the care of the cancer patient. At platelet counts less than 10x109/L, spontaneous bleeding may be increased. At platelet counts less than 50x109/L, surgical procedures are often complicated by bleeding. At platelet counts under 100x109/L, chemotherapy and radiation therapy may be administered with caution thereby decreasing dose intensity and clinical outcome.1 Therapeutic and prophylactic platelet transfusions create the additional risk

of infusion-related complications and might be immunosuppressive.2 Finally, thrombocytopenia instills in the patient a sense of anxiety and fear of bleeding which exacerbates that due to the cancer diagnosis itself. The clinician’s response to thrombocytopenia in a cancer patient is varied. Reduction of relative dose intensity (RDI) of chemotherapy or radiation is common; less effective regimens may be chosen; treatment may even be precluded. For some, treatment of another cause of thrombocytopenia (e.g., stopping the offending antiviral agent) may be effective. Platelet transfusion is often the only immediately available treatment. There is increasing interest in using recombinant human thrombopoietin (rhTPO) or thrombopoietin receptor agonists (TPO-RA) such as romi-

Haematologica | 107 - June 2022

1243

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia plostim, eltrombopag, avatrombopag, lusutrombopag, and hetrombopag to enhance platelet production and platelet counts.3 Recognizing that chemotherapy creates other hematologic problems (e.g., neutropenia) that may also limit the ability to administer chemotherapy, the focus here will be primarily on situations in which thrombocytopenia is a major limiting variable. This review will discuss the general approach to the patient with a non-hematologic malignancy receiving non-myeloablative chemotherapy, the pathophysiology of chemotherapy-induced thrombocytopenia (CIT), and options for treating CIT including TPO-RA. The use of TPO-RA therapy in myeloablative settings (stem cell transplantation and acute myeloid leukemia induction) has been discussed separately.4 A discussion of thrombocytopenia secondary to therapeutic irradiation or treatment of hematologic malignancies is beyond the scope of this review. CIT is here defined as a platelet count less than 100x109/L and divided into grades as follows; grade 1: 75x109/L to less than 100x109/L; grade 2: 50x109/L to less than 75x109/L; grade 3: 25x109/L to less than 50x109/L; and grade 4: less than 25x109/L.5,6

The platelet count is an imprecise predictor of bleeding risk in cancer patients The main reason to check the platelet count in a cancer patient receiving chemotherapy is to attempt to predict the bleeding risk. A biological estimate of the lowest effective platelet count for effective hemostasis comes from the work of Slichter and colleagues,7-9 who used chromium-51 labeled red blood cells to quantify fecal blood loss in stable thrombocytopenic aplastic patients treated only with anabolic steroids. At platelet counts above 10x109/L, blood loss was normal at less than 5 mL/day. At platelet counts of 5x109/L to 10x109/L, blood loss rose slightly to 9±7 mL/day; however at platelet counts below 5x109/L, the loss was markedly elevated to 50±20 mL/day. Subsequent platelet kinetic studies10 found a fixed minimum requirement for 7.1x109 platelets/L/day to maintain vascular integrity; 18% of the normal daily turnover of 41.2x109 platelets/L/day. This is consistent with in vitro data showing that thrombin generation appears to be maximal as long as the platelet count is above 10x109/L; below that value thrombin generation declines in direct proportion to the platelet count.11,12 A recent trial assessed the relation of platelet count to bleeding (using a validated bleeding scale6) in thrombocytopenic patients undergoing myeloablative chemotherapy for leukemia or stem cell transplantation (Figure 1).13

D.J. Kuter

It clearly showed bleeding of grade 2 or higher on 25% of days with platelet counts of 5x109/L or less, on 17% of days with platelet counts from 6–80x109/L (P<0.001 for platelet counts ≤5x109/L vs. counts of 6–80x109/L), on 13% of days with platelet counts of 81–100x109/L (P=0.001 for platelet counts of 81–100x109/L vs. counts of 6–80x109/L), and on 8% of days with platelet counts above 100x109/L (P<0.001 for platelet counts >100x109/L vs. counts of 6–80x109/L). In cancer patients undergoing chemotherapy, thrombocytopenic bleeding and bleeding grade have been inadequately studied; in general both increase once the platelet count drops below 75x109/L (odds ratio=3.1; 95% confidence interval [95% CI]: 1.9–5.1).14 Roughly, when the platelet count falls below 50x109/L the probability of bleeding is 0–9.6%; rises to 10.1–17.7% when the count is below 20x109/L; and rises again to 18.4–40.1% when below 10x109/L.15 The incidence of CIT rises with each subsequent cycle of chemotherapy.16 Unfortunately, using just the platelet count to predict bleeding risk for the cancer patient is an over simplification. Platelet function may be altered by other medications, antipyretics, chemotherapy drugs themselves, and renal insufficiency. Platelets in patients with CIT lack the increased size and function of the "young" platelets in immune thrombocytopenia (ITP) which tends to mitigate the bleeding risk seen at comparably low platelet counts in ITP.17 Other patient-related variables markedly affect the hemostatic risk (Table 1). The bleeding risk for each patient needs to be personalized

The importance of maintaining chemotherapy dose intensity Most chemotherapy regimens have been developed to provide the greatest therapeutic benefit with acceptable toxicity. As demonstrated in Table 2, in patients with metastatic breast cancer, chemotherapy dose intensity is important; when cyclophosphamide/methotrexate/5-fluorouracil dose intensity was decreased by 50%, there was less thrombocytopenia but there was also a significant decrease in median survival.18 Reductions in RDI (due to either reductions in dose or dosing frequency) decrease response rate and survival. In a retrospective study of patients with metastatic colorectal disease being treated with FOLFIRI (folinic acid, fluorouracil, irinotecan) or modified FOLFOX6 (folinic acid, fluorauracil, oxaliplatin), patients receiving a high RDI of irinotecan had a median performance-free survival of 9.9 months compared with 5.6 months for those receiving a low RDI of irinotecan (hazard ratio [HR]=3.18; 95% CI: 1.47–6.88; P<0.01); median overall survival was also reduced from 26.7 months down to 12.9 months for

Haematologica | 107 - June 2022

1244

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia

D.J. Kuter

Figure 1. Relation between bleeding (measured using the World Health Organization bleeding scale) and the platelet count in patients with hypoproliferative thrombocytopenia. The percentage of days on which patients had bleeding of grade 2 or greater is shown, along with the associated 95% confidence intervals (dashed lines), according to the morning platelet count category. (From Slichter et al.13 and reproduced with the permission of the Massachusetts Medical Society.)

the same two groups (HR=2.72; 95% CI: 1.22–6.04; P=0.01].19 Similar effects of reduction in RDI leading to decreased overall survival have been demonstrated in the adjuvant treatment of non-small cell lung cancer,20,21 breast cancer,22 ovarian cancer,22 and non-Hodgkin lymphoma.23 In elderly patients with advanced non-small cell lung cancer, those receiving a RDI of 80% or more had a higher response rate (55.2% vs. 33.3%) and overall survival than those who received a RDI less than 80%.24 In advanced epithelial ovarian cancer25 and metastatic breast cancer26 improved survival was associated with a RDI of 85% or more. Many factors other than thrombocytopenia determine the feasibility of delivering an adequate RDI for any particular cancer patient: anemia, neutropenia, mucositis, nausea, intravenous access, need for concurrent radiation therapy, and performance status. However, for a significant number of chemotherapy recipients, thrombocytopenia is the major limiting variable, as discussed below. So, for the cancer patient whose chemotherapy regimen is primarily limited by thrombocytopenia, the potential bleeding risk with chemotherapy becomes a complex calculation in which the platelet count and other hemostatic variables need to be weighed in the context of potential benefit from maintaining chemotherapy RDI. Furthermore, there is no reliable predictor of the degree of nadir thrombocytopenia (and hence risk of bleeding) based on pretreatment platelet count or other patient-related or chemotherapy regimen variables. In general, chemotherapy is administered at pretreatment platelet counts over 100x109/L with many chemotherapy regimens and physicians challenged when platelet counts are below 70x109/L, in particular below 50x109/L. There is no evi-

dence as to what constitutes an "adequate pretreatment platelet count" for any specific chemotherapy for solid tumors; although 70x109/L or higher is widely accepted.

Not all cases of thrombocytopenia in chemotherapy recipients are due to chemotherapy: the clinical approach to thrombocytopenia in the cancer patient Although chemotherapy and radiation are by far the major causes of thrombocytopenia in the cancer patient, other etiologies should be considered in all patients. In general, the following “checklist” should be considered in cancer patients with platelet counts less than 100x109/L. • Is the underlying disease the cause of the thrombocytopenia? Tumor that metastasizes to bone marrow is common in patients with breast and lung cancer as well as in those with primary hematologic malignancies such as lymphoma. Most such patients also demonstrate pancytopenia and these cytopenias generally occur when over 80% of the bone marrow is infiltrated. As discussed below, patients with infiltrated bone marrow respond poorly to TPO-RA.27 • Is there an associated immune thrombocytopenia? Up to 1% of patients with Hodgkin disease,28,29 2 to 10% of patients with chronic lymphocytic leukemia,30-32 and 0.76% (range, 0–1.8%) of patients with other non-Hodgkin lymphomas29 develop a secondary ITP. These patients respond to steroids, rituximab, splenectomy, and TPO-RA just like pa-

Haematologica | 107 - June 2022

1245

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia tients with primary ITP,33 although treatment of the underlying lymphoma may be effective.29 • Has there been a recent infection? While infection may produce consumptive coagulopathies (e.g., disseminated intravascular coagulation), some bacteria release neuraminidase that actually reduces platelet survival by removing the sialic acids coating platelets and thereby increasing their clearance by the liver Küpffer cell type C lectin receptor (CLEC4F) or hepatic Ashwell-Morell receptor.3436 Viral infections (e.g., cytomegalovirus) in compromised patients may inhibit bone marrow production of platelets. Such thrombocytopenias improve with adequate treatment of the infection. • Has the patient received a new medication? Heparin-induced thrombocytopenia should be considered.37 Antibiotics (e.g., vancomycin,38 linezolid39) and antiviral agents (e.g., ganciclovir40,41) commonly induce thrombocytopenia either by direct bone marrow toxicity or by immune drug-dependent antibody clearance.38,42 • Has there been a recent transfusion? Post-transfusion purpura is a rare complication of transfusion of red blood cells and platelets with the platelet count usually dropping below 10x109/L.43,44 Post-transfusion purpura occurs in the 1% of patients who lack the common platelet antigen PLA-1 (HPA-1a) and is usually seen in women previously sensitized by pregnancy. Upon transfusion of HPA-1a-positive platelets into sensitized HPA-1a-negative patients, antibody destroys the transfused platelets and by an as yet unclear mechanism also destroys the patient’s own HPA-1a-negative platelets. This under-recognized complication of transfusion responds readily to intravenous immunoglobulin. • Does the patient have a coagulopathy? In addition to infections, some tumors (e.g., gastric and pancreatic adenocarcinomas) can cause chronic disseminated intravascular coagulation.45,46 Such thrombocytopenic patients usually have elevated D-dimer and prothrombin fragment 1.247 with a low fibrinogen, but often have minimally prolonged prothrombin time and partial thromboplastin time.48 Treatment of this is often difficult. Heparin may improve the coagulopathy, but most patients do not improve

D.J. Kuter

without effective treatment of the underlying tumor. • Is there a chemotherapy- or transplant-related thrombotic microangiopathy? Mitomycin-C and gemcitabine induce endothelial injury with a resultant thrombotic microangiopathy whose major manifestation is renal failure and thrombocytopenia, best referred to as a chemotherapy-related hemolytic uremic syndrome.49 Patients with such a microangiopathy usually have thrombocytopenia, microangiopathic hemolytic anemia, and organ dysfunction with a normal level of ADAMTS13;50 most improve with supportive care and discontinuation of the chemotherapy. Plasma exchange, rituximab, or steroids are not indicated.51 It is unclear whether complement inhibition with eculizumab or ravulizumab will help.52 • Is the thrombocytopenia temporally related to chemotherapy or radiation therapy? When was the last chemotherapy or radiation therapy administered? The platelet has a normal lifespan of 8 to 10 days. After many types of chemotherapy, the platelet count generally starts to drop by day 7 and reaches its nadir at day 14 with a gradual return back to baseline by day 28 to 35.53

Table 1. Variables potentially increasing bleeding risk in cancer patients.

Platelet count

Platelet function

Patient-specific variables

Liver disease

Antibiotics

Fever

Splenomegaly

Renal insufficiency.

Infection

Antipyretics

Procedures

Chemotherapy drugs

Age

Anticoagulants

Coagulation factor abnormalities Indwelling catheters Surgery Malignant lesions Mucositis/esophagitis

Table 2. The effect of dose intensity on platelet count and chemotherapy response.

Low dose

High dose

Mean (SEM) dose intensity for cyclophosphamide*

46 (1)

80 (2)

Mean (SD) nadir absolute neutrophil count, x109/L

2.850 (0.180)

0.760 (0.080)

<0.001

240 (10)

180 (8)

<0.001

12.8

15.6

0.009

Mean (SEM) nadir platelet count, x109/L Median survival (months)

P-value

Low dose intensity: cyclophosphamide 300 mg/m2, methotrexate 20 mg/m2, 5-fluorouracil 300 mg/m2; high dose intensity: cyclophosphamide 600 mg/m2, methotrexate 40 mg/m2, 5-fluorouracil 600 mg/m2. *Expressed as percentage of cyclophosphamide prescribed for the higher dose arm and averaged over the entire course of treatment. Dose intensity for the other drugs was similar to that of cyclophosphamide. Table based on data from Tannock et al.18 SEM: standard error of mean; SD: standard deviation.

Haematologica | 107 - June 2022

1246

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia Depending upon the dose and duration of radiation therapy, the onset of thrombocytopenia is generally at day 7 to day 10 with a longer duration of thrombocytopenia, sometimes lasting for 30 to 60 days. • What dose and type of chemotherapy was given? As reviewed next, the incidence, severity, and duration of thrombocytopenia vary with the chemotherapy regimen. Most non-myeloablative chemotherapy regimens were developed to minimize thrombocytopenia and the need for platelet transfusions. As such, most standard regimens have relatively low rates of dose-limiting thrombocytopenia; when thrombocytopenia occurs, it is often of short duration (4 to 6 days). Most cases respond well to platelet transfusion.

D.J. Kuter

in melanoma patients. The median time to the first decline in platelet count was about 1 to 2 weeks after starting chemotherapy, except for platinum-based regimens which was usually longer than 2 weeks. As expected, other cytopenias also occurred in most patients but 17.7% of CIT patients had only thrombocytopenia. In comparison, CIT was much more common in patients with hematologic malignancies (28% of 2,537 patients) with grade 3 and grade 4 thrombocytopenia in 16.3% and 12.4%, respectively. The incidence of thrombocytopenia-related bleeding and platelet transfusions has been very poorly studied in CIT. Bleeding events were not measured for the 62,071 chemotherapy regimens described above, but platelet transfusion data were available for 10,582; transfusions occurred in 2.5% of patients (1.0% for platinum-based The incidence of thrombocytopenia regimens, 0.6% for anthracycline-based regimens, 1.8% for and the use of platelet transfusions gemcitabine-based regimens, and 0.3% for taxane-based vary with the type of chemotherapy regimens). Table 3 provides an overview of the reported frequencies of thrombocytopenia and platelet transAlthough many factors may lead to reduction of chemo- fusions with various chemotherapy regimens. therapy RDI, it is difficult to identify how much is attributed to thrombocytopenia. In a review of different chemotherapy regimens in 614 patients with solid tumors, Pathophysiology of chemotherapyCIT (a platelet count <100x109/L) was seen in 21.8% of all subjects; thrombocytopenia was unaccompanied by other induced thrombocytopenia cytopenias in 6.2%.54 Grade 3 thrombocytopenia was seen in 3.6% and grade 4 thrombocytopenia in 3.3%. CIT oc- Not all chemotherapy drugs cause thrombocytopenia in curred in 82% of those receiving only carboplatin, and in the same way. In reviewing the mechanism of CIT, it is 58%, 64% and 59% of those receiving combination ther- helpful to understand how platelets are made (Figure 2). apies with carboplatin, gemcitabine or paclitaxel, respect- Stem cells differentiate into cells committed to megaively. karyocyte differentiation (megakaryocyte colony-forming In a retrospective analysis of 43,995 patients (including cells). At some stage, these cells stop their mitotic divithose with solid tumors or hematologic malignancies) who sions and enter a process called endomitosis, in which received 62,071 chemotherapy regimens in the USA be- DNA replication occurs but with no subsequent division of tween 2000 and 2007,55 CIT occurred in 13,304 (21.4%) the nucleus or the cell. This gives rise to polyploid preregimens. Grade 3 and grade 4 thrombocytopenia oc- cursor cells with 2, 4, 8, 16, or 32 times the normal diploid curred in 2,660 (4.3%) and 2,087 (3.4%) regimens, respect- DNA content. These polyploid megakaryocyte precursor ively: 7.8% and 3.4% of gemcitabine-based regimens; 6.5% cells then stop synthesizing DNA and mature into large, and 4.1% of platinum-based regimens; 3.0% and 2.2% of morphologically identifiable megakaryocytes. Mature anthracycline-based regimens; and 1.4% and 0.5% of ta- megakaryocytes then produce platelets by a mechanism xane-based regimens. that is still poorly defined. In its simplest iteration, porIn a more recent analysis of 15,521 patients with solid tu- tions of the megakaryocyte membrane bud off into the mors,56 12.8% (95% CI: 12.3–13.4%) had CIT: grade 2 in bone marrow sinusoid to produce platelets.57-59 Other 6.4%; grade 3 in 4.2% and grade 4 in 1.9%. CIT was more models suggest that mature megakaryocytes extrude long common in patients with solid tumors who received gem- cytoplasmic processes through endothelial cells and large citabine- and platinum-based regimens (14.8% and 13.5%, strands of platelet material (proplatelets) are released into respectively) than in patients treated with anthracycline- the circulation, eventually becoming mature platelets, or taxane-based regimens (9.3% and 6.5%, respectively). possibly through fragmentation in the lung.60 More recently With regard to tumor type, CIT occurred in 21.4% of mel- it has been suggested that some intact megakaryocytes anoma patients, 14.3% of lung cancer patients, 13.5% of migrate to the lung where they may account for up to 50% colorectal cancer patients, 12.9% of pancreatic cancer pa- of platelet production.61 If not consumed in hemostasis, tients, and 9.6% of breast cancer patients. Grade 3 (13.3%) the mature platelet undergoes programmed cell death and 4 (5.0%) thrombocytopenia occurred most commonly (apoptosis) determined by a “platelet clock”.62 This platelet Haematologica | 107 - June 2022

1247

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia clock depends upon the presence of the anti-apoptotic protein Bcl-x(L), a protein that restrains the pro-apoptotic proteins Bax and Bak.62-65 When Bcl-x(L) declines, the activity of Bax and Bak increases, which triggers platelet apoptosis. The apoptotic platelets are cleared by the reticuloendothelial cell system, probably in the liver; the spleen plays little role in normal platelet homeostasis.66 Different chemotherapy drugs affect the megakaryocyte and platelet production pathway at different steps (Figure 2). Alkylating agents such as busulfan and carboplatin affect pluripotential stem cells.67,68 Cyclophosphamide spares hematopoietic stem cells because of their abundant levels of aldehyde dehydrogenase, but affects later megakaryocyte progenitors.69 The antibody-drug conjugate T-DM1 (trastuzumab [T] coupled to the microtubule toxin emtansine [DM1]) causes grade 3 or higher thrombocytopenia within 1 week in about 13% of patients by inhibiting megakaryocyte growth and differentiation. T-DM1 is internalized into megakaryocytes via the FcgRIIa receptor or by pinocytosis where it releases DM1, which inhibits megakaryocyte polyploidization and growth.70 Bortezomib has no effect on stem cells71 or megakaryocyte maturation but does inhibit NF-kB, a critical regulator of platelet shedding.72 This probably explains the relatively short duration of thrombocytopenia following its administration.72 Not all chemotherapy drugs reduce platelet production; some can actually increase the rate of platelet destruction. Indeed, platelet survival itself may be altered by some chemotherapy agents. The experimental chemotherapy agent ABT-737 reduces the activity of the platelet clock Bcl-xL and rapidly induces platelets to undergo apoptosis.63,73 After a single dose of ABT-737, platelets dropped to 30% of baseline by 2 h, to 5% of baseline by 6 h, started to recover to 10% of baseline by 24 h, and returned to baseline after 72 h.73 This was not due to platelet activation. Rather, caspase-mediated apoptosis was induced with rapid appearance of phosphatidylserine on the platelet surface and clearance of these cells from the circulation by the reticuloendothelial system in the liver. Although not evaluated for most standard chemotherapy drugs, etoposide also increases platelet apoptosis by reducing Bcl-x(L) activity.73 The antibody-drug conjugates gemtuzumab ozogamicin and inotuzumab ozogamicin are both associated with acute thrombocytopenia (platelet counts dropping by 86% in 3–4 days in monkeys) and sinusoidal obliteration syndrome due to acute hepatic sequestration of platelets.74 Finally, chemotherapy may enhance platelet clearance by immune mechanisms. In the treatment of many lymphomas, the administration of single-agent fludarabine has been noted to produce an antiplatelet antibody-mediated ITP in up to 4.5% of patients.75 This ITP is commonly responsive to rituximab.76 Platelet destruction is also increased when chemotherapy drugs produce a

D.J. Kuter

drug-dependent antiplatelet antibody-mediated secondary ITP, but this effect is uncommon.

Current approaches to the treatment of chemotherapy-induced thrombocytopenia The response to significant CIT has not been codified in guidelines and there are few studies to describe the appropriate approach to CIT. Much depends upon the underlying goals of the treatment of the cancer patient; different levels of risk assessment need to be brought into play for those being treated for a cure compared to those being treated for palliation. Overall, it is reasonable when confronted with CIT first to assess the underlying need for chemotherapy and the goals of treatment for that particular patient. A clinical assessment of bleeding risk for patients is also important to undertake, especially if patients are receiving anticoagulant drugs or other therapies that might increase bleeding. What follows is the synthesis of the data and the author's personal experience over the past four decades: • If possible, treat any other underlying cause of thrombocytopenia: stop antibiotics, treat infections, and control coagulopathy. • Reduce chemotherapy dose, frequency or alter the chemotherapy regimen, especially if chemotherapy is not standard or not of curative intent. • Platelet transfusion support can be used if maintenance of dose intensity is vital for response or survival. Platelet transfusions are indicated if the patient is bleeding or to prevent major bleeding if platelet counts are less than 10x109/L (<20x109/L if febrile).13,77 However, in the outpatient setting, this transfusion trigger needs to be reconsidered; transfusing at higher platelet counts on the Friday before a long weekend has its advantages. • Antifibrinolytic agents such as e-aminocaproic acid (Amicar®) or tranexamic acid (Lysteda®) have been used in some thrombocytopenic cancer patients to improve hemostasis when platelet transfusions do not work, but are of unproven benefit.78-80 Total daily doses of 2–24 g (mean: 6 g) of e-aminocaproic acid given in three or four divided doses have been used.79 Tranexamic acid doses of 4–6 g/day given as three or four divided doses have also been studied.80 However, the use of antifibrinolytic agents in cancer patients might exacerbate the underlying increased risk of thrombosis. A recent randomized, prospective, blinded study reported that addition of tranexamic acid (1,000–1,300 mg every 8 h) had no benefit in reducing grade ≥2 bleeding, platelet transfusions, or days without grade ≥2 bleeding in thrombocytopenic (platelets <30x109/L) patients undergoing treatment for hematologic malignancies. Of interest, while there was no increase in non-catheter thromboses, an increased inci-

Haematologica | 107 - June 2022

1248

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia

D.J. Kuter

Table 2. Thrombocytopenia and platelet transfusions in commonly used chemotherapy regimens.

Regimen

Ibritumomab tiuxetana Bortezomibb Carboplatinc Cisplatind Cisplatin/etoposidee Cisplatin/etoposidef Cisplatin/tegafurg Gemcitabineh Gemcitabini Gemcitabinej Docetaxelk Docetaxel/doubletl Temozolomidem Gemcitabine/cisplatinn Gemcitabine/cisplatino Gemcitabine/cisplatinp Gemcitabine/cisplatinq Gemcitabine/oxaliplatinr Gemcitabine/carboplatins Gemcitabine/carboplatint Pemetrexate/cisplatinu Pemetrexate/carboplatinv R-CHOP 21w R-CHOP 14x ICEy MAIDz MAIDaa MAID-Intensiveab 5FU/doxorubicin/cyclophosphamidacac FOLFOX FOLFOX4ae FOLFIRIaf Lapatinib/capecitabineag T-DM1ah T-DM1ai T-DM1aj T-DM1ak Palbociclib/fulvestrantal Topotecanam Topotecan/cisplatinan Cyclophosphamide/adriamycin/ vincristineao Oxaliplatin/tegafurap Capcitabine/oxaliplatinaq ad

Cancer

Thrombocytopenia

NHL139 Myeloma140 Various141 Unknown primary142 SCLC143 SCLC144 Gastric145 Pancreas146 Pancreas147 Pancreas148 Breast149 NSCLC150 Glioblastoma151 Unknown primary142 NSCLC152 Pancreas147 NSCLC153 Pancreas148 NSCLC154 SCLC143 NSCLC153 NSCLC154 NHL155 NHL155 NHL125 Sarcoma156 Sarcoma157 Sarcoma157

Grade 3

Grade 4

87% (est) 28%

13% 3%

Platelet transfusion

30%

23% 4% 4% 17% 5% 3.4%

6% 3% 1.7% 3.7%

12%

1% 1.9% 7.5% 11% 37% 21.2% 2.1% 4.5%

10% 32%

15%

4.5% 1% 24%

11%

1.8% 3%

22% 4.1% 13% 5% 9% 35% 52% 34% 79%

Breast158

23%

10%

Colon Colorectal160 Colorectal160 Breast161 Breas161 Breast162 Breast163 Breast164 Breast165 SCLC166 SCLC144

3.4%

1.7% 28.8% 29%

0.6% 9.8% 9%

5.8%

SCLC166

4.9%

1.4%

1.9%

Gastric145 Colorectal167

10% 9%

3% 0%

159

2.0% 0.5%

1.0% 0% 0.2% 12.9%

4.2% 7.3%

8.3% 1.8%

Table updated from that previously published.168 aIbritumomab tiuxetan (11 MBq/kg) x 1. bBortezomib 1.3 mg/m2 IV days 1, 4, 8 and 11; repeated every 21 days. cCarboplatin 500 mg/m2 IV every 4 weeks. dCisplatin 100 mg/m2 IV every 3 weeks. eCisplatin 60 mg/m2 IV day 1, etoposide 120 mg/m2 IV day 1 and 100 mg PO BID days 2 and 3; 21-day cycle. fCisplatin 80 mg/m2 IV day 1, etoposide 100 mg/m2 IV days 1-3; 21-day cycle. gCisplatin 60 mg/m2 IV day 1, tegafur potassium oxonate 40 mg/m2 PO BID days 1-14; 21-day cycle. hGemcitabine 1,000 mg/m2 IV on days 1, 8, and 15, repeated every 4 weeks. iGemcitabine 1,000 mg/m2 IV on days 1 and 8, repeated every 3 weeks. jGemcitabine 1,000 mg/m2 IV weekly. kDocetaxel 75 mg/m2 IV every 3 weeks. lDocetaxel 75 mg/m2 IV q3weeks with either IV cetuximab, vandetanib, apecitabine, gemcitabine, oxaliplatin, carboplatin, or iri-

Legend continued on following page. Haematologica | 107 - June 2022

1249

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia

D.J. Kuter

notecan. mTemozolomide 75 mg/m2/d PO during radiotherapy; thereafter 150-200 mg/m2 PO days 1-5 of each 28-day cycle. nGemcitabine 1,250 mg/m2 IV on days 1 and 8 and cisplatin 100 mg/m2 IV on day 1, repeated every 3 weeks. oGemcitabine 1,000 mg/m2 IV on days 1 and 8 and cisplatin 75 mg/m2 IV on day 1, repeated every 3 weeks. pGemcitabine 1,000 mg/m2 IV on days 1 and 8; cisplatin 60 mg/m2 IV on day 1, repeated every 3 weeks. qCisplatin 75 mg/m2 IV on day 1 and gemcitabine 1,250 mg/m2 IV on days 1 and 8, repeated every 3 weeks. rGemcitabine 1,000 mg/m2 IV on day 1 plus oxaliplatin 100 mg/m2 IV on day 2, repeated every 14 days. sGemcitabine 1,000 mg/m2 IV on days 1 and 8 plus carboplatin IV (AUC= 5) on day 1, repeated every 3 weeks. tGemcitabine 1200 mg/m2 IV on days 1 and 8; carboplatin (AUC 5) IV on day 1. uCisplatin 75 mg/m2 IV and pemetrexed 500 mg/m2 IV on day 1, repeated every 3 weeks. vPemetrexed 500 mg/m2 IV plus carboplatin (AUC= 5) IV on day 1, repeated every 3 weeks. wCyclophosphamide 750 mg/m2 IV, doxorubicin 50 mg/m2 IV, vincristine 1.4 mg/m2 IV, rituximab 375 mg/m2 IV on day 1; prednisone 40 mg/m2 PO days 1-5; repeated every 21 days. xCyclophosphamide 750 mg/m2 IV, doxorubicin 50 mg/m2 IV, vincristine 1.4 mg/m2 IV, rituximab 375 mg/m2 IV on day 1; prednisone 40 mg/m2 PO days 1-5; repeated every 14 days. yEtoposide 100 mg/m2/day IV on days 1 to 3; carboplatin (AUC=5) IV on day 2; ifosfamide 5 g/m2 via continuous infusion for 24 hours beginning on day 2; repeated every 2 weeks. zIfosfamide 2,500 mg/m2 IV, adriamycin 20 mg/m2 IV, and dacarbazine 300 mg/m2 IV on days 1-3; repeated every 4 weeks. aaIfosfamide 2,500 mg/m2 IV, adriamycin 20 mg/m2 IV, and dacarbazine 300 mg/m2 IV on days 1-3; repeated every 3 weeks. abIfosfamide 3,000 mg/m2 IV, adriamycin 25 mg/m2 IV, and dacarbazine 400 mg/m2 IV on days 1-3; repeated every 4 weeks. acFluorouracil 600 mg/m2 IV, doxorubicin 60 mg/m2 IV, and cyclophosphamide 750 mg/m2 IV on day 1; repeated every 21 days. adLeucovorin 400 mg/m2 IV, flourauracil 400 mg/m2 IV, oxaliplatin 85mg/m2 IV all on day 1 followed by 2,400 mg/m2 IV over 46 hours; administered every 2 weeks. aeFluorouracil 400 mg/m2 IV bolus with 600 mg/m2 continuous infusion for 2 days; leucovorin 200 mg/m2 IV on day 1; oxaliplatin 85 mg/m2 IV on day 1. afFluorouracil 400 mg/m2 IV bolus with 2400 mg/m2 continuous infusion for 2 days; leucovorin 400 mg/m2 IV on day 1; irinotecan 180 IV mg/m2 on day 1. agLapatninb 1250 mg PO qd and capcitabine 1000 mg/m2 PO BID days 1 to 14, every 21 days. ahTrastuzumab emtansine (T-DM1) 3.6 mg/kg IV every 21 days. aiT-DM1 0.3-4.8 mg/kg IV every 21 days. ajT-DM1 3.6 mg/kg IV every 21 days. akT-DM1 3.6 mg/kg IV every 21 days. al Palbociclib 125 mg/day PO days 1-21; fulvestrant 500 mg IM every 2 weeks; 28-day cycle. amTopotecan 1.5 mg/m2 IV days 1-5 every 21 days. anTopotecan 1.75 mg/m2 PO days 1-5; cisplatin 60 mg/m2 IV day 5. aoCyclophosphamide 1000 mg/m2 IV, doxorubicin 45 mg/m2 IV, vincristine 2 mg IV all on day 1; every 21 days. apOxliplatin 130 mg/m2 IV day 1; tegafur potassium oxonate 40 mg/m2 PO BID days 1-14; 21-day cycle. aqCapecitabine 900 mg/m2 PO BID days 1-14; oxaliplatin 130 mg/m2 IV day 1; 21-day cycle. NHL: non-Hodgkin lymphoma; SCLC: small-cell lung cancer; NSCLC: nonsmall cell lung cancer; est: estimated; PO: per os; IV: intravenous; AUC: area under the curve; BID: bis in die.

Figure 2. The production of platelets from bone marrow stem cells. Stem cells differentiate into cells committed to megakaryocyte differentiation (megakaryocyte colony-forming cells, MK-CFC) which are mitotically active. MK-CFC then stop mitosis and start endomitosis producing immature megakaryocytes (MK) with polyploid nuclei. The immature MK then stop their endomitosis and mature into large, morphologically identifiable MK that then migrate to the bone marrow sinusoids and produce platelets.

dence of central line occlusion requiring clearing with tissue plasminogen activator was observed.81 • Recombinant interleukin 11 (oprelvekin, Neumega®) was shown to reduce the need for platelet transfusions from 96% to 70% of patients who had been transfused with platelets in a prior cycle and who then received additional chemotherapy; but it has significant adverse effects.82 This drug was approved by the Food and Drug Administration for the prevention of thrombocytopenia with chemotherapy but is no longer manufactured for use in North America; it is still available in other parts of the world. • Despite the lack of Food and Drug Administration approval for these agents, rhTPO and TPO-RA might be considered in patients for whom thrombocytopenia prevents maintenance of dose intensity crucial for remission or survival, and who cannot be supported by platelet transfusions. Recently the National Comprehensive Cancer Network endorsed the use of TPO-RA, notably romiplostim, for treatment of CIT. They

recommended initiation of TPO-RA for platelet counts 30– 50x109/L with discontinuation when the platelet count recovers to 50–100x109/L based on an unsupported concern over thrombosis.83

The use of thrombopoietin and thrombopoietin receptor agonists in chemotherapy-induced thrombocytopenia The pathophysiology of thrombopoietin in chemotherapy-induced thrombocytopenia Thrombopoietin (TPO) is the major regulator of platelet production. In animals or humans deficient in TPO or its receptor, the platelet count is 10–15% of normal.84,85 Megakaryocyte, erythroid, and myeloid precursor cells are

Haematologica | 107 - June 2022

1250

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia all reduced in such knock-out animals but their white blood cell and red blood cell counts are normal.86 TPO is produced by the liver usually at a constant rate and its production is decreased by liver disease87 and increased by interleukin-6 in rare conditions such as that associated with ovarian cancer.88,89 TPO has no storage form and is secreted into the circulation and cleared by TPO receptors on platelets. In disorders such as in CIT, TPO levels are inversely related to the rate of platelet production and rise in a log-linear fashion (Figure 3).90-93 TPO binds to its receptor on many hematopoietic cells (Figure 4) and exerts its effects on most stages of megakaryocyte growth (Figure 2). It is necessary for the viability of hematopoietic stem cells; when the TPO receptor is absent, humans are born with thrombocytopenia and develop pancytopenia over subsequent years.94-96 TPO stimulates mitosis of megakaryocyte colony-forming cells. Its major effect (at exceedingly low concentrations) is to increase megakaryocyte endomitosis and increase megakaryocyte ploidy, greatly expanding the megakaryocyte mass. It then stimulates megakaryocyte maturation. It is unclear whether TPO plays any role in platelet shedding.97 An under-appreciated effect of TPO is that it prevents apoptosis of early and late megakaryocytes,98 an effect that may play a major protective role in patients receiving radiation and chemotherapy. Before considering the use of TPO agents in patients with cancer, it is important to note that solid tumors appear not to possess functional TPO receptors.99,100 In one study using reverse transcriptase polymerase chain reaction on 39 human cell lines and 20 primary normal and malignant human tissues, TPO receptor (c-mpl) transcripts were found in all megakaryocytic cell lines tested (DAMI, CMK, CMK-2B, CMK-2D, SO), the CD34+ leukemia cell line KMT-2, and the hepatocellular carcinoma cell line Hep3B.100 While

D.J. Kuter

fetal liver and brain cells had detectable levels of c-mpl mRNA, none was found in primary tumors. In a more extensive study, microarray detected TPO receptor mRNA in 0/118 breast tumors and at very low levels in 14/29 lung tumors.99 Low but detectable levels of TPO receptor mRNA were found by quantitative polymerase chain reaction in some normal (14–43%) and malignant (3–17%) breast, lung, and ovarian tissues but in none of these tissues was TPO receptor protein detectable by immunohistochemistry. Culture of breast, lung, and ovarian carcinoma cell lines with TPO-RA showed no stimulation of growth. Finally, in none of the human clinical studies described next was there any stimulation of tumor growth by the administration of rhTPO or TPO-RA. Development of recombinant thrombopoietin and thrombopoietin receptor agonists The development of clinically relevant TPO molecules has occurred in two phases: the early recombinant TPO molecules and the recent TPO-RA.3,101 With the discovery of TPO in 1994, two recombinant TPO molecules were developed (Table 4). rhTPO is a fully glycosylated TPO protein made in Chinese hamster ovary cells. The other, pegylated recombinant human megakaryocyte growth and development factor (PEG-rhMGDF), is a non-glycosylated protein comprising the first 153 amino acids of TPO coupled to polyethylene glycol. Both molecules are potent stimulators of platelet production with half-lives of about 40 h. In healthy volunteers both demonstrated the same time course of platelet response after a single dose: by day 3 megakaryocyte ploidy increased, by day 5 platelet counts started to rise, by days 10–14 a peak platelet count was obtained, and by day 28 platelet counts returned to their baseline.102

Figure 3. Log-linear correlation of thrombopoietin level with platelet counts. Patients undergoing double umbilical cord blood transplantation had serial platelet counts and thrombopoietin levels determined over their hospital course. As platelet counts declined, thrombopoietin levels rose. Figure created from data in the study by deFilipp et al.91 Haematologica | 107 - June 2022

1251

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia Between 1995 and 2000, both recombinant TPO underwent extensive clinical development in CIT.103 As discussed more below, both produced an earlier and higher nadir platelet count, shortened the duration of thrombocytopenia, re-

D.J. Kuter

duced platelet transfusions, and enabled non-myeloablative chemotherapy to be given on schedule. Development of both was stopped in the West around 2000 because of concerns over neutralizing antibody formation

Figure 4. Recombinant human thrombopoietin and the thrombopoietin receptor agonists bind to and activate the thrombopoietin receptor in different ways. The thrombopoietin (TPO) receptor has been proposed to exist as an inactive preformed dimer (left side) with a proximal and distal hematopoietic receptor domain (HRD1 and HRD2, respectively). Upon binding of romiplostim or recombinant human TPO (not pictured) to the distal HRD2, the conformation of the receptor changes (right side) and a number of signal transduction pathways are activated which increase platelet production. The other TPO receptor agonists bind to the transmembrane region of the receptor and activate many of the same signal transduction pathways. TPO: thrombopoietin; HRD: hematopoietic receptor domain protein; STAT: signal transducer and activator of transcription; JAK: Janus kinase; GRP2: growth receptor bound protein 2; SOS: son of sevenless (a guanine nucleotide exchange factor); RAS: rat sarcoma virus (a small GTP-ase); RAF: rapidly accelerated fibrosarcoma (a serine/threonine kinase); MAPK: mitogen-activated protein kinase.

Table 4. Thrombopoietic therapies under development for treating chemotherapy-induced thrombocytopenia.

TPO-R binding Chelates iron Off-target effects Potency in healthy subjects Route Dietary effect Interacts with cations Decrease dose if East Asian Decrease dose if liver dysfunction Increases liver function tests Use in renal failure Use in pregnancy Regulatory approval for CIT

rhTPO

Romiplostim

Eltrombopag

Avatrombopag

Hetrombopag

Distal No ?+ ++++++++ SC No No No No No OK OK Yes, China and 7 other countries

Distal No ?+ ++++++++ SC No No No No No OK No

Transmembrane + +++ + Oral Yes ++ Yes Yes + Probably OK No

Transmembrane ? ? ++++ Oral No No No No No Probably OK No

Transmembrane No ? +++ Oral Yes ? Yes Yes ? No data No

rhTPO: recombinant human thrombopoietin; TPO-R: thrombopoietin receptor; SC: subcutaneous; CIT: chemotherapy-induced thrombocytopenia. Haematologica | 107 - June 2022

1252

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia against PEG-rhMGDF.104 In 525 healthy volunteers given up to three monthly doses of PEG-rhMGDF, 13 (2.5%) developed thrombocytopenia due to the formation of antibodies to PEG-rhMGDF that cross-reacted with endogenous TPO, creating TPO deficiency and thrombocytopenia. All subjects recovered, but some required immunosuppressive treatment.105,106 Development of rhTPO (TPIAO®) continued in China where it is now licensed for treatment of CIT (https://www.mims.com/thailand/drug/info/tpiao) and ITP.107 Despite the failure of one of the recombinant TPO molecules, interest turned to developing newer TPO molecules (now called TPO-RA) with novel properties and less risk of antibody formation.1,3 Romiplostim is a “peptibody” created by inserting a 14 amino acid peptide that activates the TPO receptor into an IgG4 heavy chain (Table 4).108 Romiplostim is approved in many countries for the treatment of ITP and in Japan for the treatment of aplastic anemia. A separate approach identified a number of small molecules (eltrombopag, avatrombopag, lusutrombopag, hetrombopag) that bind and activate the TPO receptor (Table 4).109 All of these small molecule TPO-RA bind the TPO receptor in the transmembrane region, an area different from where TPO or romiplostim binds, and activate the TPO receptor in a different fashion (Figure 4).110-112 All of these small molecules have undergone extensive clinical development and most (eltrombopag, avatrombopag, hetrombopag) are licensed for treating patients with ITP, of whom they increase the platelet count in over 85%.113-119 Eltrombopag is also approved for the treatment of thrombocytopenia in patients with hepatitis C infection requiring anti-viral treatment120 and in patients with aplastic anemia in whom immunosuppressive therapy has failed.121,122 In the latter disease, treatment was also associated with an increase in white blood cells and red blood cells. Avatrombopag and lusutrombopag are both approved for treating thrombocytopenic patients with chronic liver disease about to undergo procedures.109 PEG-rhMGDF, rhTPO, romiplostim, eltrombopag, avatrombo-

D.J. Kuter

pag and hetrombopag have all been studied in patients with CIT. Although a 2017 meta-analysis stated that there was insufficient evidence to support their use in CIT,15 considerable new evidence has emerged since then which challenges that conclusion. Clearly all TPO increase the platelet count in CIT patients but the question is whether that translates into clinical benefit in maintaining RDI, reducing bleeding, and improving response or survival. What follows is a summary of current data. Treatment of chemotherapy-induced thrombocytopenia with pegylated recombinant human growth and development factor When PEG-rhMGDF was administered for up to 16 days after treatment of lung cancer patients with carboplatin and paclitaxel, the median platelet count nadir was 188x109/L (range, 68–373x109/L) versus the count in the placebo-treated group of 111x109/L (range, 21–307x109/L; P=0.013) (Figure 5). The nadir platelet count occurred earlier in the patients treated with PEG-rhMGDF; the median time to nadir was 7 versus 15 days (P<0.001). The platelet count recovered to baseline in 14 days in the patients given PEGrhMGDF as compared with more than 21 days in those receiving placebo (P<0.001).123 There was no effect on platelet transfusions or bleeding; only one patient in the placebo group required a platelet transfusion. Thromboses were not increased and the patients’ survival was not affected. In a second major study,124 patients with advanced malignancy were treated with intravenous carboplatin 600 mg/m2 and cyclophosphamide 1,200 mg/m2 in their first cycle. In subsequent cycles they received, in addition, PEG-rhMGDF for 1, 3 or 7 days after chemotherapy. Those receiving the same chemotherapy dose on a subsequent cycle had a significantly higher platelet nadir than in cycle 1, (48x109/L vs. 36x109/L; P=0.003) and the duration of grade 3 or 4 thrombocytopenia was significantly shorter (0 vs. 3 days; P=0.004). However, there was no difference

Figure 5. Pegylated recombinant human megakaryocyte growth and development factor increases the platelet count in patients undergoing chemotherapy. Lung cancer patients being treated with carboplatin and paclitaxel were also given either placebo (purple circles) or pegylated recombinant human megakaryocyte growth and development factor (yellow circles) in a double-blind, randomized study.123 Platelet counts were measured daily. The inset shows the probability of recovery of the platelet count back to baseline in the two treatment groups. Figure adapted from published data.123 PEG-rhMGDF: pegylated recombinant human megakaryocyte growth and development factor.

Haematologica | 107 - June 2022

1253

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia in the time to platelet recovery. Administration of PEGrhMGDF prior to chemotherapy did not show any benefit. A third study has suggested a possible survival benefit from treatment with PEG-rhMGDF.125 In the treatment of patients with relapsed non-Hodgkin lymphoma with ifosfamide, carboplatin, and etoposide (ICE) chemotherapy, maintenance of RDI correlates with improved survival. In a study of 38 non-Hodgkin lymphoma patients randomized to placebo (n=16) or PEG-rhMGDF (n=22), ICE was given on schedule to 42% of those on placebo and 75% of those on PEG-rhMGDF (P=0.008) with overall survival of 21% and 31%, respectively (P=0.06), after a median followup of 8.5 years. Patients on placebo were 4.4 times more likely to have a dose delay, which was due to thrombocytopenia in 83%. Grade 4 thrombocytopenia was seen in 35% of placebo recipients versus 15% of patients given PEG-rhMGDF (P=0.02) with platelet nadirs of 20x109/L and 49x109/L (P=0.008), respectively. Platelet transfusions were administered in 23% of placebo cycles and 8% of PEG-rhMGDF cycles (P=0.04).

D.J. Kuter

As described above, PEG-rhMGDF development stopped in 2000 due to the appearance of antibodies against PEGrhMGDF which cross-reacted with endogenous TPO and caused thrombocytopenia.105 Treatment of chemotherapy-induced thrombocytopenia with recombinant human thrombopoietin In an early study, rhTPO was administered on days 2, 4, 6, and 8 after a second cycle of carboplatin chemotherapy for patients with a gynecological malignancy (Figure 6). The mean platelet count nadir in the second cycle was higher than that in the first cycle, during which no rhTPO was administered (44x109/L vs. 20x109/L; P=0.002); the number of days with platelet count less than 20x109/L was lower (1 vs. 4 days, P=0.002); the number of days with a platelet count less than 50x109/L was lower (4 vs. 7 days; P=0.006). The need for platelet transfusion in the group receiving rhTPO was reduced from 75% of patients in cycle 1 to 25% of patients in cycle 2 (P=0.013). Recovery to a platelet count 100x109/L or greater was faster (20 days for

Figure 6. Recombinant human thrombopoietin reduces the need for platelet transfusions in patients undergoing carboplatin chemotherapy for gynecological malignancy. (A) Platelet count time course for patients in cycle 2 (treated with recombinant human thrombopoietin [rhTPO]) compared to that in patients in cycle 1 (treated without rhTPO). Figure provided by Pharmacia, Inc. (Peapack, NJ, USA). (B) Platelet counts and platelet transfusions in cycle 2 (treated with rhTPO) compared with cycle 1 (treated without rhTPO). Figure created from data in the study by Vadhan-Raj et al.126 d: days.

Haematologica | 107 - June 2022

1254

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia

D.J. Kuter

Table 5. Outcomes of 62 patients with chemotherapy-induced thrombocytopenia in a cross-over study who received placebo in one chemotherapy cycle and recombinant human thrombopoietin in an adjacent cycle. Study data on file at 3sBIO Inc, Shenyang, China. Nadir platelet count (x10 /L), mean (SD) Highest recovery platelet count (x109/L), mean (SD) Platelet count difference (x109/L), mean (SD) Duration platelets <50x109/L post-chemotherapy, mean (SD) days Time for platelets to rise >75x109/L, mean (SD) days Time for platelets to rise >100x109/L, mean (SD) days Platelet transfusion events, mean (SD) Platelet units transfused, mean (SD) 9

Placebo

rhTPO

P value

49 (35) 153 (82) 104 (74) 4.7 (6.9) 17.2 (10.9) 21.2 (10.6) 0.6 (1.3) 3.0 (6.6)

61 (51) 261 (164) 199 (158) 3.1 (4.4) 12.0 (9.7) 15.8 (9.7) 0.3 (0.7) 1.5 (4.4)

0.015 <0.001 <0.001 <0.05 <0.001 <0.001 <0.05 <0.05

rhTPO: recombinant human thrombopoietin; SD: standard deviation.

rhTPO in cycle 2; 23 days without rhTPO in cycle 1; P<0.001).126 Numerous CIT studies have been conducted in China with rhTPO but with few results published in non-Asian medical journals.127,128 In one study (Study 005) made available to this author in English, 62 cancer patients (42% with lymphoma and the rest with solid tumors) with a platelet count less than 75x109/L on a prior cycle and requiring two more chemotherapy cycles at the same dose were randomly treated in a crossover study of CIT prophylaxis. Twenty-eight patients in group A (n=28) received rhTPO treatment (1 mg/kg rhTPO subcutaneously daily for 14 days starting 6–24 h after chemotherapy) during cycle 1 and received no rhTPO during cycle 2. Group B (n=34) received no rhTPO in the first cycle but received rhTPO treatment during cycle 2. As shown in Table 5 there was a modest improvement in nadir platelet count, with a shorter duration of platelet count below 50x109/L, a much higher recovery platelet count by day 24, and fewer transfusion events. No data were provided regarding maintenance of RDI, bleeding, or whether platelet transfusions were standardized. Similar improvements in platelet counts were obtained in an additional crossover study (Study 006) done in 213 cancer patients. In a study of 58 lymphoma patients undergoing high-dose cytarabine therapy, two different forms of prophylactic rhTPO were prospectively evaluated. A group that was given rhTPO daily for 10 days after chemotherapy was compared to a group that was given two doses prior to chemotherapy and eight doses after. Those receiving rhTPO before the chemotherapy were less likely to have grade 4 thrombocytopenia (26.9% vs. 48.1%), experienced a shorter duration of grade 4 thrombocytopenia (0.58 days vs. 1.23 days), and required fewer transfusions (13.5% vs. 25%).129 In a small study of rhTPO in 30 patients undergoing adjuvant chemotherapy for gastric or colorectal cancer with regimens containing gemcitabine and capecitabine, patients were randomized to receive either prophylactic treatment with rhTPO 15,000 U/day beginning 4 days before chemotherapy or to receive treatment only when platelet

counts after chemotherapy dropped below 75x109/L. Those receiving prophylactic rhTPO experienced better outcomes in terms of a higher mean [standard deviation, SD] nadir platelet count (76x109/L [27x109/L] vs. 53x109/L [17x109/L]; P<0.001], fewer days of dose interruption (1.72 [2.78] vs. 3.72 [3.38]; P=0.002), and a shorter recovery to platelet counts over 100x109/L (4.6 [4.7] vs. 8.9 [2.3] days; P<0.001).130 In a recent meta-analysis131 of 12 CIT studies, mostly in China, when compared with patients treated with placebo or interleukin-11, those treated with rhTPO had modest decreases in the duration of days with platelet counts under 50x109/L and 75x109/L but no difference in days with platelet counts under 100x109/L. Other clinical endpoints were not reported. Although rhTPO is approved for CIT in China, the studies above document a modest increase in platelet count with treatment but it is hard to assess from published data whether any clinical endpoints such as RDI, remission rate, bleeding or survival were affected. Treatment of chemotherapy-induced thrombocytopenia with romiplostim In one retrospective study, cancer patients were selected who had a platelet count less than 100x109/L and who had a more than 4-week delay in their chemotherapy or had dose reductions/modifications in two or more prior cycles of chemotherapy.132 These patients were treated with 2 mg/kg of romiplostim weekly. Platelet counts improved in all and 19/20 had platelet counts of 100x109/L or more. Fifteen patients resumed chemotherapy and all but one of these continued for two or more further cycles without dose modifications. Three of 20 patients developed deep vein thrombosis. In a phase II, randomized prospective trial (NCT02052882: An Open Label Phase II Study of Romiplostim for Chemotherapy Induced Thrombocytopenia) in solid tumor patients with CIT, patients with platelet counts below 100x109/L for more than 4 weeks despite dose reduction or delay were randomized to receive either romiplostim weekly or observation.133 The primary endpoint was attain-

Haematologica | 107 - June 2022

1255

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia ing a platelet count of 100x109/L or more by week 3. After enrolling 23 patients, it was found that 14/15 (93%) patients who received romiplostim attained the primary endpoint, whereas only 1/8 (12.5%) of those on observation did so (P<0.001). After 2–3 weeks of treatment, those receiving romiplostim had a mean platelet count of 141x109/L versus 57x109/L for those on observation. The randomized portion of the study was discontinued and 37 subsequent patients all received romiplostim. Of the 52 patients who received romiplostim, 44 (85%) met the primary endpoint. All of these 44 patients resumed chemotherapy supported with romiplostim and only three (6.8%) developed subsequent CIT. Twenty-eight patients continued on romiplostim for more than 6 months at a mean dose of 3.3 mg/kg. Six of the 59 patients (10.2%) developed a venous thromboembolism during the first year of romiplostim therapy; none discontinued romiplostim. At four Boston cancer centers, supportive care with romiplostim has been utilized for almost 10 years for CIT patients.27,134 Patients eligible for this program had to have a platelet count below 100x109/L for at least 3 weeks after their last chemotherapy treatment or a dose delay of longer than 1 week. Overall, 173 CIT patients (153 with solid tumors, 20 with lymphoma or myeloma) were treated, with 170 (90%) undergoing a median of four (range, 1–36) chemotherapy cycles on romiplostim. The primary outcome was a platelet response defined as a median on-romiplostim platelet count of 75x109/L or more and at least 30x109/L higher than pretreatment

D.J. Kuter

baseline. Among all the solid tumor patients 71% had a platelet response, 79% avoided chemotherapy dose reductions/delays and 89% avoided platelet transfusions. The median baseline platelet count of 60x109/L rose to 116x109/L (P=0.001). The median weekly romiplostim dose was 3 mg/kg (interquartile range, 3–5 mg/kg). Solid tumor patients who failed to respond were characterized by extensive bone marrow involvement by tumor, prior pelvic radiotherapy or treatment with temozolomide and their low response rate (<10%) was comparable to that of patients with lymphoma and myeloma (Figure 7). Bleeding rates (7.1/100 patient-years on romiplostim) were less than those of historical controls and there was no apparent increase in thrombosis (11/100 patient-years on romiplostim). Two different dosing algorithms were explored: weekly romiplostim dosing including on days of chemotherapy administration versus a less intense regimen with weekly dosing except on days of chemotherapy administration. Patients on weekly dosing had a significantly higher median platelet count (143x109/L vs. 106x109/L; P<0.001) and a higher rate of achieving a platelet response (81% vs. 63%; P=0.006). Other clinical outcomes including the extent of chemotherapy RDI reduction and bleeding were better in patients receiving weekly treatment. There are two ongoing trials of romiplostim in CIT with the primary endpoint being the incidence of either a chemotherapy dose delay or reduction (defined as “no thrombocytopenia-induced modification [dose delay, reduction,

Figure 7. Platelet counts of patients with chemotherapy-induced thrombocytopenia treated with romiplostim. Median weekly platelet counts for solid tumor patients (n=122, blue) with no predictors of non-response (no PNR); solid tumor patients (n=31, gray) with predictors of nonresponse (PNR: bone marrow invasion by tumor, prior pelvic irradiation, or prior temozolomide treatment); aggressive lymphoma patients (n=13, red); and myeloma patients (n=7, purple). Data reproduced with permission from Al-Samkari H et al.27 Haematologica | 107 - June 2022

1256

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia omission or chemotherapy treatment discontinuation due to platelet counts <100 x 109/L] of any myelosuppressive agent in the second and third cycles of the planned onstudy chemotherapy regimen”): NCT03937154: A Phase 3 Randomized Placebo-controlled Double-blind Study of Romiplostim for the Treatment of Chemotherapy-induced Thrombocytopenia in Patients Receiving Chemotherapy for Treatment of Non-small Cell Lung Cancer (NSCLC), Ovarian Cancer, or Breast Cancer; and NCT03362177: A Phase 3 Randomized Placebo-controlled Double-blind Study of Romiplostim for the Treatment of Chemotherapy-Induced Thrombocytopenia in Patients Receiving Oxaliplatin-based Chemotherapy for Treatment of Gastrointestinal, Pancreatic, or Colorectal Cancer. Treatment of chemotherapy-induced thrombocytopenia with eltrombopag In a CIT prophylaxis study (NCT00102726), 183 patients received either placebo or eltrombopag 50 mg, 75 mg or 100 mg on days 2 through 11 for at least two 21-day chemotherapy cycles. Eltrombopag was well tolerated. The primary endpoint (the difference in platelet count from day 1 in cycle 2 to the platelet nadir in cycle 2) was not met but postnadir platelet counts were higher for cycles 1 and 2 than in patients in the placebo group.135 In an early but informative blinded, placebo-controlled phase I study to prevent CIT, patients with solid tumors and a platelet count 300x109/L or below receiving up to six cycles of either gemcitabine alone (9 patients) or gemcitabine plus either cisplatin or carboplatin (10 patients) were randomized (3:1) to receive eltrombopag 100 mg or placebo on days -5 to -1 and days 2-6 starting from cycle 2; no study drug was administered for cycle 1.136 For patients receiving gemcitabine alone, the mean (SD) nadir platelet count for cycles 2-6 was 143x109/L (82x109/L) for eltrombopag versus 103x109/L (64x109/L) for placebo; for those receiving gemcitabine plus cisplatin or carboplatin the nadir was 115x109/L (83x109/L) versus 53x109/L (7x109/L) for placebo; 14% of all eltrombopag patients and 50% of placebo patients required dose reductions or delays in cycles 3-6. Three thromboembolic events were reported and felt to be related to other disease characteristics. Platelet counts greater than 400x109/L were reported more frequently in eltrombopag-treated versus control patients (92/19 [4.8 events/patient] vs. 18/7 [2.6 events/patient]) and the decision was made not to increase the dose of eltrombopag above 100 mg/day. A larger phase II study investigated eltrombopag prophylaxis for the prevention of CIT in patients receiving either gemcitabine alone (42 patients) or gemcitabine with carboplatin or cisplatin (32 patients) over six cycles of chemotherapy. Patients were randomized (1:2) to receive either placebo or eltrombopag 100 mg/day for 5 days before and again daily for 5 days after the chemotherapy.137 The primary endpoint

D.J. Kuter

was the mean pretreatment platelet count over six cycles of chemotherapy. The treatment was well tolerated with no increased risk of thrombosis (5/52 [9.6%] on eltrombopag and 2/23 [8.7%] on placebo) but was complicated by a 65% withdrawal rate. The geometric mean platelet count of the 48 eltrombopag–treated patients was 246x109/L compared with 193x109/L for the 23 placebo patients, but the difference did not attain statistical significance (P=0.103). Patients receiving eltrombopag had a slightly lower rate of grade 3/4 thrombocytopenia (27/50 [54%] vs. 16/23 [70%]) and slightly higher nadir platelet counts than patients receiving placebo. There were fewer dose reductions, dose delays and missed doses due to thrombocytopenia in cycles 2-6 for eltrombopag (15/38 [39%]) than for placebo (10/19 [53%]). A real-world retrospective observational study assessed the response of lymphoma patients whose platelet counts dropped below 30x109/L and who were then treated with eltrombopag (n=51), rhTPO (n=50) or no platelet growth factor support (n=52).138 The baseline platelet counts for all three groups was 24x109/L. After 10 days there was a significantly higher median [SD] platelet count in those on eltrombopag and rhTPO than the untreated patients (131x109/L [71x109/L], 147x109/L [68x109/L], and 76x109/L [40x109/L], respectively; P<0.001); the median (SD) duration of platelet counts <50x109/L was 6.25 (2.61), 5.48 (2.62), and 8.33 (3.98) days, respectively (P=0.036); the mean (SD) days required for recovery to greater than 50x109/L was 6.33 (2.31), 5.44 (2.57), and 8.32 (2.53) days, respectively (P=0.001). Patients receiving eltrombopag or rhTPO were less likely to have grade 2/3 bleeding (5.9% and 4.0%) compared with untreated patients (11.5%); with fewer platelet transfusions (55% and 50%) compared with untreated patients (75%). One eltrombopag CIT study is currently being conducted: NCT04600960: Eltrombopag for Chemotherapy-induced Thrombocytopenia: A Prospective Single-center One-arm Study. Treatment of chemotherapy-induced thrombocytopenia with avatrombopag Avatrombopag has been studied in patients with CIT: NCT03471078: Randomized, Double-blind, Placebo-controlled Study With Open-label Extension to Evaluate the Efficacy and Safety of Avatrombopag for the Treatment of Chemotherapy-induced Thrombocytopenia in Subjects With Active Non-Hematological Cancers. This was a double-blind, placebo-controlled phase III prospective study assessing the safety and efficacy of avatrombopag in patients with CIT (grade ≥2 in a prior cycle) who were receiving chemotherapy for ovarian cancer, small cell lung cancer, non-small cell lung cancer, or bladder cancer. One hundred twenty-two patients were enrolled who had developed grade 3 or 4 thrombocytopenia following treatment with chemotherapy in a prior cycle. Preliminary results of the study showed that avatrombopag failed to

Haematologica | 107 - June 2022

1257

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia

D.J. Kuter

meet the composite primary endpoint (avoiding platelet transfusions, chemotherapy dose reductions by ≥15%, and chemotherapy dose delays by ≥4 days). In the intent-totreat population, 69.5% of patients who received avatrombopag and 72.5% of those who received placebo were responders for the primary endpoint (P=0.72). In the perprotocol population, these rates were 85.0% and 84.4% for avatrombopag and placebo, respectively, (P=0.96). Other avatrombopag CIT studies are being planned with different entry requirements.

CIT but their beneficial effect on RDI, tumor response, transfusion, bleeding or survival have not yet been adequately demonstrated. It is too early to assess the costeffectiveness of this form of supportive care. There remains a need for consensus as to what is an adequate pretreatment platelet count in patients being given chemotherapy; such a count will certainly vary with the chemotherapy regimen and patient’s pretreatment variables. This reviewer feels that in most situations, a pretreatment platelet count over 50x109/L is usually adequate.

Treatment of chemotherapy-induced thrombocytopenia with hetrombopag CIT studies are ongoing in China and will soon be started in the West. The current ongoing study is: NCT03976882: A Randomized, Double-blind, Placebo-controlled Multicentre Study With an Open-label Extension to Evaluate the Efficacy and Safety of Hetrombopag for the Treatment of Chemotherapy-induced Thrombocytopenia in Subjects With Malignancy.

Disclosures The author has received research funding from Actelion (Syntimmune), Agios, Alnylam, Amgen, Argenx, Bristol Myers Squibb (BMS), Immunovant, Kezar, Principia, Protalex, Rigel, and Takeda (Bioverativ) UCB; and has served in a consulting role for Actelion (Syntimmune), Agios, Alnylam, Amgen, Argenx, Bristol Myers Squibb (BMS), Caremark, CRICO, Daiichi Sankyo, Dova, Genzyme, Immunovant, Incyte, Kyowa-Kirin, Merck Sharp Dohme, Momenta, Novartis, Pfizer, Platelet Disorder Support Association, Principia, Protalex, Protalix, Rigel, Sanofi, Genzyme, Shionogi, Shire, Takeda (Bioverativ), UCB, Up-To-Date, and Zafgen.

Conclusion CIT is a common complication of non-myeloablative chemotherapy in patients with solid tumors and its incidence varies with the chemotherapy regimen used. Bleeding is generally associated with the degree of thrombocytopenia. Studies that include other variables in addition to the platelet count are needed to predict bleeding in chemotherapy patients; artificial intelligence predictive algorithms may help here. Other non-chemotherapy-related causes for thrombocytopenia should be assessed and treated in all patients. Platelet transfusions are the main therapy for bleeding in CIT patients. Chemotherapy dose or frequency reductions are the mainstay of CIT treatment but may compromise RDI and therapeutic effect. rhTPO and TPO-RA increase the pretreatment and nadir platelet counts in most patients with

Contributions The author is the sole individual responsible for the design, writing and opinions in this manuscript. No writing assistance was provided. Acknowledgments The author is pleased to acknowledge the many helpful discussions and comments on this topic with Dr. Gerald Soff, Dr. Hanny Al-Samkari, Dr. Irene Kuter and the members of the Massachusetts General Hospital Cancer Center. Funding The author has received no financial or other support for the preparation of this article.

References 1. Kuter D. General aspects of thrombocytopenia, platelet transfusions, and thrombopoietic growth factors. In: Kitchens CS, Kessler CM, Konkle BA, et al. editors. Consultative Hemostasis and Thrombosis, 4th edition. Philadelphia: Elsevier; 2019:108-126. 2. Schmied L, Hoglund P, Meinke S. Platelet-mediated protection of cancer cells from immune surveillance - possible implications for cancer immunotherapy. Front Immunol. 2021;12:640578. 3. Kuter D. Clinical applications of thrombopoietic growth factors. In: Plow T, editor. UpToDate. http://www.uptodate.com (accessed October 26, 2021). 4. Kuter DJ. What is the role of novel thrombopoietic agents in the

management of acute leukemia? Best Pract Res Clin Haematol. 2016;29(4):372-378. 5. National Institutes of Health NCI. Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. In: U.S: Department of Health and Human Services, 2017. 6. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer. 1981;47(1):207-214. 7. Slichter SJ. Optimizing platelet transfusions in chronically thrombocytopenic patients. Semin Hematol. 1998;35(3):269-278. 8. Slichter SJ, LeBlanc R, Jones MK. Quantitative analysis of bleeding risk in cancer patients (PTS) prophylactically transfused at platelet (PLT) counts (CTS) of, 5,000, 10,000, or 20,000. Blood. 1999;94(3):376a.

Haematologica | 107 - June 2022

1258

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia 9. Slichter SJ, Harker LA. Thrombocytopenia: mechanisms and management of defects in platelet production. Clin Haematol. 1978;7(3):523-539. 10. Hanson SR, Slichter SJ. Platelet kinetics in patients with bone marrow hypoplasia: evidence for a fixed platelet requirement. Blood. 1985;66(5):1105-1109. 11. Chantarangkul V, Clerici M, Bressi C, Giesen PL, Tripodi A. Thrombin generation assessed as endogenous thrombin potential in patients with hyper- or hypo-coagulability. Haematologica. 2003;88(5):547-554. 12. Chantarangkul V, Clerici M, Bressi C, Tripodi A. Standardization of the endogenous thrombin potential measurement: how to minimize the effect of residual platelets in stored plasma. Br J Haematol. 2004;124(3):355-357. 13. Slichter SJ, Kaufman RM, Assmann SF, et al. Dose of prophylactic platelet transfusions and prevention of hemorrhage. N Engl J Med. 2010;362(7):600-613. 14. Elting LS, Rubenstein EB, Martin CG, et al. Incidence, cost, and outcomes of bleeding and chemotherapy dose modification among solid tumor patients with chemotherapy-induced thrombocytopenia. J Clin Oncol. 2001;19(4):1137-1146. 15. Zhang X, Chuai Y, Nie W, Wang A, Dai G. Thrombopoietin receptor agonists for prevention and treatment of chemotherapyinduced thrombocytopenia in patients with solid tumours. Cochrane Database Syst Rev. 2017;11(11):CD012035. 16. Hassan BA, Yusoff ZB, Hassali MA, Bin Othman S. Treatment patterns and outcomes in management of solid cancer patients suffering from thrombocytopenia in Penang hospital. Asian Pac J Cancer Prev. 2011;12(11):2841-2845. 17. Harker LA, Slichter SJ. The bleeding time as a screening test for evaluation of platelet function. N Engl J Med. 1972;287(4):1551559. 18. Tannock IF, Boyd NF, DeBoer G, et al. A randomized trial of two dose levels of cyclophosphamide, methotrexate, and fluorouracil chemotherapy for patients with metastatic breast cancer. J Clin Oncol. 1988;6(9):1377-1387. 19. Nakayama G, Tanaka C, Uehara K, et al. The impact of dose/time modification in irinotecan- and oxaliplatin-based chemotherapies on outcomes in metastatic colorectal cancer. Cancer Chemother Pharmacol. 2014;73(4):847-855. 20. Ramsden K, Laskin J, Ho C. Adjuvant chemotherapy in resected stage II non-small cell lung cancer: evaluating the impact of dose intensity and time to treatment. Clin Oncol (R Coll Radiol). 2015;27(7):394-400. 21. Crawford J, Denduluri N, Patt D, et al. Relative dose intensity of first-line chemotherapy and overall survival in patients with advanced non-small-cell lung cancer. Support Care Cancer. 2020;28(2):925-932. 22. Denduluri N, Lyman GH, Wang Y, et al. Chemotherapy dose intensity and overall survival among patients with advanced breast or ovarian cancer. Clin Breast Cancer. 2018;18(5):380-386. 23. Bosly A, Bron D, Van Hoof A, et al. Achievement of optimal average relative dose intensity and correlation with survival in diffuse large B-cell lymphoma patients treated with CHOP. Ann Hematol. 2008;87(4):277-283. 24. Luciani A, Bertuzzi C, Ascione G, et al. Dose intensity correlate with survival in elderly patients treated with chemotherapy for advanced non-small cell lung cancer. Lung Cancer. 2009;66(1):94-96. 25. Hanna RK, Poniewierski MS, Laskey RA, et al. Predictors of reduced relative dose intensity and its relationship to mortality in women receiving multi-agent chemotherapy for epithelial ovarian cancer. Gynecol Oncol. 2013;129(1):74-80.

D.J. Kuter

26. Loibl S, Skacel T, Nekljudova V, et al. Evaluating the impact of relative total dose intensity (RTDI) on patients' short and longterm outcome in taxane- and anthracycline-based chemotherapy of metastatic breast cancer- a pooled analysis. BMC Cancer. 2011;11:131. 27. Al-Samkari H, Parnes AD, Goodarzi K, Weitzman JI, Connors JM, Kuter DJ. A multicenter study of romiplostim for chemotherapyinduced thrombocytopenia in solid tumors and hematologic malignancies. Haematologica. 2021;106(4):1148-1157. 28. Dimou M, Angelopoulou MK, Pangalis GA, et al. Autoimmune hemolytic anemia and autoimmune thrombocytopenia at diagnosis and during follow-up of Hodgkin lymphoma. Leuk Lymphoma. 2012;53(8):1481-1487. 29. Hauswirth AW, Skrabs C, Schutzinger C, et al. Autoimmune thrombocytopenia in non-Hodgkin's lymphomas. Haematologica. 2008;93(3):447-450. 30. Zent CS, Ding W, Reinalda MS, et al. Autoimmune cytopenia in chronic lymphocytic leukemia/small lymphocytic lymphoma: changes in clinical presentation and prognosis. Leuk Lymphoma. 2009;50(8):1261-1268. 31. Kyasa MJ, Parrish RS, Schichman SA, Zent CS. Autoimmune cytopenia does not predict poor prognosis in chronic lymphocytic leukemia/small lymphocytic lymphoma. Am J Hematol. 2003;74(8):1-8. 32. Hodgson K, Ferrer G, Pereira A, Moreno C, Montserrat E. Autoimmune cytopenia in chronic lymphocytic leukaemia: diagnosis and treatment. Br J Haematol. 2011;154(1):14-22. 33. Visco C, Rodeghiero F, Romano A, et al. Eltrombopag for immune thrombocytopenia secondary to chronic lymphoproliferative disorders: a phase 2 multicenter study. Blood. 2019;134(20):1708-1711. 34. Grewal PK, Aziz PV, Uchiyama S, et al. Inducing host protection in pneumococcal sepsis by preactivation of the Ashwell-Morell receptor. Proc Natl Acad Sci U S A. 2013;110(50):20218-20223. 35. Grewal PK, Uchiyama S, Ditto D, et al. The Ashwell receptor mitigates the lethal coagulopathy of sepsis. Nat Med. 2008;14(6):648-655. 36. Kuter DJ. Novel therapies for immune thrombocytopenia. Br J Haematol. 2022;196(6):1311-1328. 37. Kuter DJ, Konkle BA, Hamza TH, et al. Clinical outcomes in a cohort of patients with heparin-induced thrombocytopenia. Am J Hematol. 2017;92(8):730-738. 38. Von Drygalski A, Curtis BR, Bougie DW, et al. Vancomycininduced immune thrombocytopenia. N Engl J Med. 2007;356(9):904-910. 39. Kuter DJ, Tillotson GS. Hematologic effects of antimicrobials: focus on the oxazolidinone, linezolid. Pharmacotherapy. 2001;21(8):1010-1013. 40. Wang Y, Smith KP. Safety of alternative antiviral agents for neonatal herpes simplex virus encephalitis and disseminated infection. J Pediatr Pharmacol Ther. 2014;19(2):72-82. 41. Danziger-Isakov L, Mark Baillie G. Hematologic complications of anti-CMV therapy in solid organ transplant recipients. Clin Transplant. 2009;23(3):295-304. 42. Reese JA, Li X, Hauben M, et al. Identifying drugs that cause acute thrombocytopenia: an analysis using 3 distinct methods. Blood. 2010;116(12):2127-2133. 43. Gonzalez CE, Pengetze YM. Post-transfusion purpura. Curr Hematol Rep. 2005;4(2):154-159. 44. Hawkins J, Aster RH, Curtis BR. Post-transfusion purpura: current perspectives. J Blood Med. 2019;10:405-415. 45. Levi M. Cancer and DIC. Haemostasis. 2001;31(Suppl 1):47-8. 46. Saba HI, Morelli GA, Saba RI. Disseminated intravascular

Haematologica | 107 - June 2022

1259

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia coagulation (DIC) in cancer. Cancer Treat Res. 2009;148:137-156. 47. Rosen EA, Vallurupalli M, Choy E, Lennerz JK, Kuter DJ. Management of disseminated intravascular coagulation in a patient with hepatic angiosarcoma: a case report. Medicine (Baltimore). 2018;97(47):e13321. 48. Kuter DJ, Rosenberg RD. Disorders of hemostasis. In: Beck WS, editor. Hematology. Cambridge, Mass: MIT Press. 1991:577-598. 49. Humphreys BD, Sharman JP, Henderson JM, et al. Gemcitabineassociated thrombotic microangiopathy. Cancer. 2004;100(12):2664-2670. 50. Weitz IC. Thrombotic microangiopathy in cancer. Semin Thromb Hemost. 2019;45(4):348-353. 51. Schwartz J, Winters JL, Padmanabhan A, et al. Guidelines on the use of therapeutic apheresis in clinical practice-evidencebased approach from the Writing Committee of the American Society for Apheresis: the sixth special issue. J Clin Apher. 2013;28(3):145-284. 52. Jodele S, Fukuda T, Vinks A, et al. Eculizumab therapy in children with severe hematopoietic stem cell transplantationassociated thrombotic microangiopathy. Biol Blood Marrow Transplant. 2014;20(4):518-525. 53. Shimazaki C, Inaba T, Uchiyama H, et al. Serum thrombopoietin levels in patients undergoing autologous peripheral blood stem cell transplantation. Bone Marrow Transplant. 1997;19(8):771-775. 54. Ten Berg MJ, van den Bemt PM, Shantakumar S, et al. Thrombocytopenia in adult cancer patients receiving cytotoxic chemotherapy: results from a retrospective hospital-based cohort study. Drug Saf. 2011;34(12):1151-1160. 55. Wu Y, Aravind S, Ranganathan G, Martin A, Nalysnyk L. Anemia and thrombocytopenia in patients undergoing chemotherapy for solid tumors: a descriptive study of a large outpatient oncology practice database, 2000-2007. Clin Ther. 2009;31(Pt 2):24162432. 56. Shaw JL, Nielson CM, Park JK, Marongiu A, Soff GA. The incidence of thrombocytopenia in adult patients receiving chemotherapy for solid tumors or hematologic malignancies. Eur J Haematol. 2021;106(54):662-672. 57. Potts KS, Farley A, Dawson CA, et al. Membrane budding is a major mechanism of in vivo platelet biogenesis. J Exp Med. 2020;217(9):e20191206. 58. Wright JH. A rapid method for the differential staining of blood films and malarial parasites. J Medical Res. 1902;7(1):138-144. 59. Wright JH. The origin and nature of blood plates. Boston Med Surg J. 1906;154:643-645. 60. Machlus KR, Thon JN, Italiano JE Jr. Interpreting the developmental dance of the megakaryocyte: a review of the cellular and molecular processes mediating platelet formation. Br J Haematol. 2014;165(2):227-236. 61. Lefrancais E, Ortiz-Munoz G, Caudrillier A, et al. The lung is a site of platelet biogenesis and a reservoir for haematopoietic progenitors. Nature. 2017;544(7648):10510-10519. 62. Dowling MR, Josefsson EC, Henley KJ, Hodgkin PD, Kile BT. Platelet senescence is regulated by an internal timer, not damage inflicted by hits. Blood. 2010;116(10):1776-1778. 63. Josefsson EC, James C, Henley KJ, et al. Megakaryocytes possess a functional intrinsic apoptosis pathway that must be restrained to survive and produce platelets. J Exp Med. 2011;208(10):2017-2031. 64. Josefsson EC, White MJ, Dowling MR, Kile BT. Platelet life span and apoptosis. Methods Mol Biol. 2012;788:59-71. 65. White MJ, Schoenwaelder SM, Josefsson EC, et al. Caspase-9 mediates the apoptotic death of megakaryocytes and platelets, but is dispensable for their generation and function. Blood.

D.J. Kuter

2012;119(18):4283-4290. 66. Berger G, Hartwell DW, Wagner DD. P-selectin and platelet clearance. Blood. 1998;92(11):4446-4452. 67. Fitchen JH, Deregnaucourt J, Cline MJ. An in vitro model of hematopoietic injury in chronic hypoplastic anemia. Cell Tissue Kinet. 1981;14(1):8590. 68. McManus PM, Weiss L. Busulfan-induced chronic bone marrow failure: changes in cortical bone, marrow stromal cells, and adherent cell colonies. Blood. 1984;64(5):1036-1041. 69. DeZern AE, Petri M, Drachman DB, et al. High-dose cyclophosphamide without stem cell rescue in 207 patients with aplastic anemia and other autoimmune diseases. Medicine (Baltimore). 2011;90(2):89-98. 70. Uppal H, Doudement E, Mahapatra K, et al. Potential mechanisms for thrombocytopenia development with trastuzumab emtansine (T-DM1). Clin Cancer Res. 2015;21(1):123133. 71. Fitzgerald M, Fraser C, Webb I, Schenkein D, Esseltine D, Weich N. Normal hematopoietic stem cell function in mice following treatment with bortezomib. Biol Blood Marrow Transplant. 2003;3:193. 72. Lonial S, Waller EK, Richardson PG, et al. Risk factors and kinetics of thrombocytopenia associated with bortezomib for relapsed, refractory multiple myeloma. Blood. 2005;106(12):3777-3784. 73. Zhang H, Nimmer PM, Tahir SK, et al. Bcl-2 family proteins are essential for platelet survival. Cell Death Differ. 2007;14(5):943951. 74. Guffroy M, Falahatpisheh H, Biddle K, et al. Liver microvascular injury and thrombocytopenia of antibody-calicheamicin conjugates in Cynomolgus monkeys-mechanism and monitoring. Clin Cancer Res. 2017;23(7):1760-1770. 75. Leach M, Parsons RM, Reilly JT, Winfield DA. Autoimmune thrombocytopenia: a complication of fludarabine therapy in lymphoproliferative disorders. Clin Lab Haematol. 2000;22:175178. 76. Hegde UP, Wilson WH, White T, Cheson BD. Rituximab treatment of refractory fludarabine-associated immune thrombocytopenia in chronic lymphocytic leukemia. Blood. 2002;100(3):2260-2262. 77. Rebulla P, Finazzi G, Marangoni F, et al. The threshold for prophylactic platelet transfusions in adults with acute myeloid leukemia. N Engl J Med. 1997;337(3):1870-1875. 78. Antun AG, Gleason S, Arellano M, et al. Epsilon aminocaproic acid prevents bleeding in severely thrombocytopenic patients with hematological malignancies. Cancer. 2013;119(21):37843787. 79. Kalmadi S, Tiu R, Lowe C, Jin T, Kalaycio M. Epsilon aminocaproic acid reduces transfusion requirements in patients with thrombocytopenic hemorrhage. Cancer. 2006;107(1):136140. 80. Wardrop D, Estcourt LJ, Brunskill SJ, et al. Antifibrinolytics (lysine analogues) for the prevention of bleeding in patients with haematological disorders. Cochrane Database Syst Rev. 2013;(7):CD009733. 81. Gernsheimer TB, Brown SP, Triulzi DJ, et al. Effects of tranexamic acid prophylaxis on bleeding outcomes in hematologic malignancy: the a-TREAT trial. Blood. 2020;136(Suppl 1):1-2. 82. Tepler I, Elias L, Smith JW 2nd, et al. A randomized placebocontrolled trial of recombinant human interleukin-11 in cancer patients with severe thrombocytopenia due to chemotherapy. Blood. 1996;87(9):3607-3614. 83. Griffiths EA, Alwan LM, Bachiashvili K, et al. Considerations for

Haematologica | 107 - June 2022

1260

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia use of hematopoietic growth factors in patients with cancer related to the COVID-19 pandemic. J Natl Compr Canc Netw. 2020 Sep 1. [Epub ahead of print] 84. de Sauvage FJ, Carver-Moore K, Luoh SM, et al. Physiological regulation of early and late stages of megakaryocytopoiesis by thrombopoietin. J Exp Med. 1996;183(2):651-656. 85. de Sauvage FJ, Villeval JL, Shivdasani RA. Regulation of megakaryocytopoiesis and platelet production: lessons from animal models. J Lab Clin Med. 1998;131(6):496-501. 86. Carver-Moore K, Broxmeyer HE, Luoh SM, et al. Low levels of erythroid and myeloid progenitors in thrombopoietin- and cmpl-deficient mice. Blood. 1996;88(3):803-808. 87. Peck-Radosavljevic M, Wichlas M, Zacherl J, et al. Thrombopoietin induces rapid resolution of thrombocytopenia after orthotopic liver transplantation through increased platelet production. Blood. 2000;95(3):795-801. 88. Kaser A, Brandacher G, Steurer W, et al. Interleukin-6 stimulates thrombopoiesis through thrombopoietin: role in inflammatory thrombocytosis. Blood. 2001;98(9):2720-2275. 89. Stone RL, Nick AM, McNeish IA, et al. Paraneoplastic thrombocytosis in ovarian cancer. N Engl J Med. 2012;366(7):610-618. 90. Yang C, Li J, Kuter DJ. The physiological response of thrombopoietin (c-Mpl ligand) to thrombocytopenia in the rat. Br J Haematol. 1999;105(2):478-485. 91. DeFilipp Z, Makar RS, Brown J, et al. Endogenous thrombopoietin levels are elevated following double cord blood unit transplantation. Bone Marrow Transplant. 2020;55(6):11781180. 92. Kuter DJ. The physiology of platelet production. Stem Cells. 1996;14 (Suppl 1):88-101. 93. Emmons RV, Reid DM, Cohen RL, et al. Human thrombopoietin levels are high when thrombocytopenia is due to megakaryocyte deficiency and low when due to increased platelet destruction. Blood. 1996;87(10):4068-4071. 94. Muraoka K, Ishii E, Tsuji K, et al. Defective response to thrombopoietin and impaired expression of c-mpl mRNA of bone marrow cells in congenital amegakaryocytic thrombocytopenia. Br J Haematol. 1997;96(2):287-292. 95. Ihara K, Ishii E, Eguchi M, et al. Identification of mutations in the c-mpl gene in congenital amegakaryocytic thrombocytopenia. Proc Natl Acad Sci U S A. 1999;96(6):3132-3136. 96. Ballmaier M, Germeshausen M, Schulze H, et al. c-mpl mutations are the cause of congenital amegakaryocytic thrombocytopenia. Blood. 2001;97(1):139-146. 97. Choi ES, Hokom MM, Chen JL, et al. The role of megakaryocyte growth and development factor in terminal stages of thrombopoiesis. Br J Haematol. 1996;95(2):227-233. 98. Zauli G, Vitale M, Falcieri E, et al. In vitro senescence and apoptotic cell death of human megakaryocytes. Blood. 1997;90(6):2234-2243. 99. Erickson-Miller CL, Pillarisetti K, Kirchner J, et al. Low or undetectable TPO receptor expression in malignant tissue and cell lines derived from breast, lung, and ovarian tumors. BMC Cancer. 2012;12:405. 100. Columbyova L, Loda M, Scadden DT. Thrombopoietin receptor expression in human cancer cell lines and primary tissues. Cancer Res. 1995;55(16):3509-3512. 101. Kuter DJ. Milestones in understanding platelet production: a historical overview. Br J Haematol. 2014;165(2):248-258. 102. Harker LA, Marzec UM, Hunt P, et al. Dose-response effects of pegylated human megakaryocyte growth and development factor on platelet production and function in nonhuman

D.J. Kuter

primates. Blood. 1996;88(2):511-521. 103. Kuter DJ, Begley CG. Recombinant human thrombopoietin: basic biology and evaluation of clinical studies. Blood. 2002;100(10):3457-3469. 104. Li J, Xia Y, Bertino A, Coburn J, Kuter DJ. Characterization of an anti-thrombopoietin antibody that developed in a cancer patient following the injection of PEG-rHuMGDF. Blood. 1999;94:51a. 105. Li J, Yang C, Xia Y, et al. Thrombocytopenia caused by the development of antibodies to thrombopoietin. Blood. 2001;98(12):3241-3248. 106. Basser RL, O'Flaherty E, Green M, et al. Development of pancytopenia with neutralizing antibodies to thrombopoietin after multicycle chemotherapy supported by megakaryocyte growth and development factor. Blood. 2002;99(7):2599-2602. 107. Kong Z, Qin P, Xiao S, et al. A novel recombinant human thrombopoietin therapy for the management of immune thrombocytopenia in pregnancy. Blood. 2017;130(9):1097-1103. 108. Molineux G. The development of romiplostim for patients with immune thrombocytopenia. Ann N Y Acad Sci. 2011;1222:55-63. 109. Virk ZM, Kuter DJ, Al-Samkari H. An evaluation of avatrombopag for the treatment of thrombocytopenia. Expert Opin Pharmacother. 2021;22(3):273-280. 110. Erickson-Miller C, Delorme E, Iskander M, et al. Species specificity and receptor domain interaction of a small molecule TPO receptor agonist. Blood. 2004;104(11):2909. 111. Erickson-Miller CL, Delorme E, Tian SS, et al. Preclinical activity of eltrombopag (SB-497115), an oral, nonpeptide thrombopoietin receptor agonist. Stem Cells. 2009;27(2):424-430. 112. Erickson-Miller CL, DeLorme E, Tian SS, et al. Discovery and characterization of a selective, nonpeptidyl thrombopoietin receptor agonist. Exp Hematol. 2005;33(1):85-93. 113. Kuter DJ. Biology and chemistry of thrombopoietic agents. Semin Hematol. 2010;47(3):243-248. 114. Kuter DJ. The biology of thrombopoietin and thrombopoietin receptor agonists. Int J Hematol. 2013;98(1):10-23. 115. Kuter DJ. General aspects of thrombocytopenia, platelet transfusions, and thrombopoietic growth factors. In: Kitchens C, Kessler C, Konkle B, editors. Consultative Hemostasis and Thrombosis. 3rd ed. Philadelphia: Elsevier Saunders. 2013:103116. 116. Kuter DJ, Bussel JB, Newland A, et al. Long-term treatment with romiplostim in patients with chronic immune thrombocytopenia: safety and efficacy. Br J Haematol. 2013;161(3):411-423. 117. Kuter DJ, Rummel M, Boccia R, et al. Romiplostim or standard of care in patients with immune thrombocytopenia. N Engl J Med. 2010;363(20):1889-1899. 118. Bussel JB, Buchanan GR, Nugent DJ, et al. A randomized, double-blind study of romiplostim to determine its safety and efficacy in children with immune thrombocytopenia. Blood. 2011;118(1):28-36. 119. Syed YY. Hetrombopag: first approval. Drugs. 2021;81(13):15811585. 120. McHutchison JG, Dusheiko G, Shiffman ML, et al. Eltrombopag for thrombocytopenia in patients with cirrhosis associated with hepatitis C. N Engl J Med. 2007;357(22):2227-2236. 121. 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. 122. Olnes MJ, Scheinberg P, Calvo KR, et al. Eltrombopag and improved hematopoiesis in refractory aplastic anemia. N Engl J

Haematologica | 107 - June 2022

1261

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia Med. 2012;367(1):11-19. 123. Fanucchi M, Glaspy J, Crawford J, et al. Effects of polyethylene glycol-conjugated recombinant human megakaryocyte growth and development factor on platelet counts after chemotherapy for lung cancer. N Engl J Med. 1997;336(6):404-409. 124. Basser RL, Underhill C, Davis I, et al. Enhancement of platelet recovery after myelosuppressive chemotherapy by recombinant human megakaryocyte growth and development factor in patients with advanced cancer. J Clin Oncol. 2000;18(15):28522861. 125. Moskowitz CH, Hamlin PA, Gabrilove J, et al. Maintaining the dose intensity of ICE chemotherapy with a thrombopoietic agent, PEG-rHuMGDF, may confer a survival advantage in relapsed and refractory aggressive non-Hodgkin lymphoma. Ann Oncol. 2007;18(11):1842-1850. 126. Vadhan-Raj S, Verschraegen CF, Bueso-Ramos C, et al. Recombinant human thrombopoietin attenuates carboplatininduced severe thrombocytopenia and the need for platelet transfusions in patients with gynecologic cancer. Ann Intern Med. 2000;132(5):364-368. 127. Bai CM, Xu GX, Zhao YQ, Han SM, Shan YD. [A multi-center clinical trial of recombinant human thrombopoietin in the treatment of chemotherapy-induced thrombocytopenia in patients with solid tumor]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao. 2004;26(4):437-441. 128. Bai CM, Zou XY, Zhao YQ, Han SM, Shan YD. Thrombopoietin Clinical Trial Cooperation Group. [The clinical study of recombinant human thrombopoietin in the treatment of chemotherapy-induced severe thrombocytopenia]. Zhonghua Yi Xue Za Zhi. 2004;84(5):397-400. 129. Wang Z, Fang X, Huang H, et al. Recombinant human thrombopoietin (rh-TPO) for the prevention of severe thrombocytopenia induced by high-dose cytarabine: a prospective, randomized, self-controlled study. Leuk Lymphoma. 2018;59(12):2821-2828. 130. Li Q, Jin G, Jiang C, et al. Prophylactic administration of recombinant human thrombopoietin attenuates XELOX or SOX regimen-induced thrombocytopaenia. Arch Med Sci. 2021;17(5):1440-1446. 131. Abudurousuli R, Luo T, Liu R, et al. Research on the recombinant human thrombopoietin in the treatment of thrombocytopenia caused by tumor chemotherapy: a metaanalysis. In J Intell Syst. 2121;3:184-192. 132. Parameswaran R, Lunning M, Mantha S, et al. Romiplostim for management of chemotherapy-induced thrombocytopenia. Support Care Cancer. 2014;22(5):1217-1222. 133. Soff GA, Miao Y, Bendheim G, et al. Romiplostim treatment of chemotherapy-induced thrombocytopenia. J Clin Oncol. 2019;37(31):2892-2898. 134. Al-Samkari H, Marshall AL, Goodarzi K, Kuter DJ. The use of romiplostim in treating chemotherapy-induced thrombocytopenia in patients with solid tumors. Haematologica. 2018;103(4):e169-e172. 135. Kellum A, Jagiello-Gruszfeld A, Bondarenko IN, Patwardhan R, Messam C, Mostafa Kamel Y. A randomized, double-blind, placebo-controlled, dose ranging study to assess the efficacy and safety of eltrombopag in patients receiving carboplatin/paclitaxel for advanced solid tumors. Curr Med Res Opin. 2010;26(10):2339-2346. 136. Winer ES, Safran H, Karaszewska B, et al. Eltrombopag with gemcitabine-based chemotherapy in patients with advanced solid tumors: a randomized phase I study. Cancer Med. 2015;4(1):16-26.

D.J. Kuter

137. Winer ES, Safran H, Karaszewska B, et al. Eltrombopag for thrombocytopenia in patients with advanced solid tumors receiving gemcitabine-based chemotherapy: a randomized, placebo-controlled phase 2 study. Int J Hematol. 2017;106(6):765-776. 138. Zhu Q, Yang S, Zeng W, et al. A real-world observation of eltrombopag and recombinant human thrombopoietin (rhTPO) in lymphoma patients with chemotherapy induced thrombocytopenia. Front Oncol. 2021;11:701539. 139. Wiseman GA, Gordon LI, Multani PS, et al. Ibritumomab tiuxetan radioimmunotherapy for patients with relapsed or refractory non-Hodgkin lymphoma and mild thrombocytopenia: a phase II multicenter trial. Blood. 2002;99(12):4336-4342. 140. Richardson PG, Barlogie B, Berenson J, et al. A phase 2 study of bortezomib in relapsed, refractory myeloma. N Engl J Med. 2003;348(26):2609-2617. 141. Budd GT, Ganapathi R, Adelstein DJ, et al. Randomized trial of carboplatin plus amifostine versus carboplatin alone in patients with advanced solid tumors. Cancer. 1997;80(6):1134-1140. 142. Gross-Goupil M, Fourcade A, Blot E, et al. Cisplatin alone or combined with gemcitabine in carcinomas of unknown primary: results of the randomised GEFCAPI 02 trial. Eur J Cancer. 2012;48(5):721-727. 143. Lee SM, James LE, Qian W, et al. Comparison of gemcitabine and carboplatin versus cisplatin and etoposide for patients with poor-prognosis small cell lung cancer. Thorax. 2009;64(1):75-80. 144. Eckardt JR, von Pawel J, Papai Z, et al. Open-label, multicenter, randomized, phase III study comparing oral topotecan/cisplatin versus etoposide/cisplatin as treatment for chemotherapy-naive patients with extensive-disease small-cell lung cancer. J Clin Oncol. 2006;24(13):2044-2051. 145. Lee KW, Chung IJ, Ryu MH, et al. Multicenter phase III trial of S-1 and cisplatin versus S-1 and oxaliplatin combination chemotherapy for first-line treatment of advanced gastric cancer (SOPP trial). Gastric Cancer. 2021;24(1):156-167. 146. Ozaka M, Matsumura Y, Ishii H, et al. Randomized phase II study of gemcitabine and S-1 combination versus gemcitabine alone in the treatment of unresectable advanced pancreatic cancer (Japan Clinical Cancer Research Organization PC-01 study). Cancer Chemother Pharmacol. 2012;69(5):1197-204. 147. Choi JH, Oh SY, Kwon HC, et al. Gemcitabine versus gemcitabine combined with cisplatin treatment locally advanced or metastatic pancreatic cancer: a retrospective analysis. Cancer Res Treat. 2008;40(1):22-26. 148. Poplin E, Feng Y, Berlin J, et al. Phase III, randomized study of gemcitabine and oxaliplatin versus gemcitabine (fixed-dose rate infusion) compared with gemcitabine (30-minute infusion) in patients with pancreatic carcinoma E6201: a trial of the Eastern Cooperative Oncology Group. J Clin Oncol. 2009;27(23):37783785. 149. Stemmler HJ, Harbeck N, Groll de Rivera I, et al. Prospective multicenter randomized phase III study of weekly versus standard docetaxel (D2) for first-line treatment of metastatic breast cancer. Oncology. 2010;79(3-4):197-203. 150. Jin Y, Sun Y, Shi X, et al. Meta-analysis to assess the efficacy and toxicity of docetaxel-based doublet compared with docetaxel alone for patients with advanced NSCLC who failed first-line treatment. Clin Ther. 2014;36(12):1980-1990. 151. Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;352(10):987-996. 152. Inal A, Kaplan MA, Kucukoner M, Urakci Z, Karakus A, Isikdogan A. Cisplatin-based therapy for the treatment of elderly patients

Haematologica | 107 - June 2022

1262

REVIEW SERIES - Treatment of chemotherapy-induced thrombocytopenia with non-small-cell lung cancer: a retrospective analysis of a single institution. Asian Pac J Cancer Prev. 2012;13(5):1837-1840. 153. Scagliotti GV, Parikh P, von Pawel J, et al. Phase III study comparing cisplatin plus gemcitabine with cisplatin plus pemetrexed in chemotherapy-naive patients with advancedstage non-small-cell lung cancer. J Clin Oncol. 2008;26(21):3543-3551. 154. Gronberg BH, Bremnes RM, Flotten O, et al. Phase III study by the Norwegian lung cancer study group: pemetrexed plus carboplatin compared with gemcitabine plus carboplatin as first-line chemotherapy in advanced non-small-cell lung cancer. J Clin Oncol. 2009;27(19):3217-3224. 155. Cunningham D, Hawkes EA, Jack A, et al. Rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisolone in patients with newly diagnosed diffuse large B-cell non-Hodgkin lymphoma: a phase 3 comparison of dose intensification with 14-day versus 21-day cycles. Lancet. 2013;381(9880):1817-1826. 156. Ogura K, Goto T, Imanishi J, et al. Neoadjuvant and adjuvant chemotherapy with modified mesna, adriamycin, ifosfamide, and dacarbazine (MAID) regimen for adult high-grade non-small round cell soft tissue sarcomas. Int J Clin Oncol. 2013;18(1):170176. 157. Fayette J, Penel N, Chevreau C, et al. Phase III trial of standard versus dose-intensified doxorubicin, ifosfamide and dacarbazine (MAID) in the first-line treatment of metastatic and locally advanced soft tissue sarcoma. Invest New Drugs. 2009;27(5):482-489. 158. Jones SE, Schottstaedt MW, Duncan LA, et al. Randomized double-blind prospective trial to evaluate the effects of sargramostim versus placebo in a moderate-dose fluorouracil, doxorubicin, and cyclophosphamide adjuvant chemotherapy program for stage II and III breast cancer. J Clin Oncol. 1996;14(11):2976-2983. 159. Allegra CJ, Yothers G, O'Connell MJ, et al. Initial safety report of NSABP C-08: a randomized phase III study of modified FOLFOX6 with or without bevacizumab for the adjuvant treatment of patients with stage II or III colon cancer. J Clin Oncol.

D.J. Kuter

2009;27(19):3385-3390. 160. Kilpatrick K, Shaw JL, Jaramillo R, et al. Occurrence and management of thrombocytopenia in metastatic colorectal cancer patients receiving chemotherapy: secondary analysis of data from prospective clinical trials. Clin Colorectal Cancer. 2021;20(2):170-176. 161. Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine for HER2-positive advanced breast cancer. N Engl J Med. 2012;367(19):1783-1791. 162. Krop IE, Beeram M, Modi S, et al. Phase I study of trastuzumabDM1, an HER2 antibody-drug conjugate, given every 3 weeks to patients with HER2-positive metastatic breast cancer. J Clin Oncol. 2010;28(16):2698-2704. 163. Krop IE, LoRusso P, Miller KD, et al. A phase II study of trastuzumab emtansine in patients with human epidermal growth factor receptor 2-positive metastatic breast cancer who were previously treated with trastuzumab, lapatinib, an anthracycline, a taxane, and capecitabine. J Clin Oncol. 2012;30(26):3234-3241. 164. Krop IE, Kim SB, Gonzalez-Martin A, et al. Trastuzumab emtansine versus treatment of physician's choice for pretreated HER2-positive advanced breast cancer (TH3RESA): a randomised, open-label, phase 3 trial. Lancet Oncol. 2014;15(7):689-699. 165. Turner NC, Ro J, Andre F, et al. Palbociclib in hormone-receptorpositive advanced breast cancer. N Engl J Med. 2015;373(3):209-219. 166. von Pawel J, Schiller JH, Shepherd FA, et al. Topotecan versus cyclophosphamide, doxorubicin, and vincristine for the treatment of recurrent small-cell lung cancer. J Clin Oncol. 1999;17(2):658-667. 167. Chiu J, Tang V, Leung R, et al. Efficacy and tolerability of adjuvant oral capecitabine plus intravenous oxaliplatin (XELOX) in Asian patients with colorectal cancer: 4-year analysis. Asian Pac J Cancer Prev. 2014;14(11):6585-6590. 168. Kuter DJ. Managing thrombocytopenia associated with cancer chemotherapy. Oncology (Williston Park). 2015;29(4):282-294.

Haematologica | 107 - June 2022

1263

REVIEW SERIES

Treatment of drug-induced immune thrombocytopenias Irene Marini, Gunalp Uzun, Kinan Jamal and Tamam Bakchoul

Correspondence:

Centre for Clinical Transfusion Medicine, Medical Faculty of Tübingen, University of Tübingen, Germany

Tamam Bakchoul tamam.bakchoul@med.uni-tuebingen.de Received: December 30, 2021. Accepted: February 15, 2022. https://doi.org/10.3324/haematol.2021.279484 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract Several therapeutic agents can cause thrombocytopenia by either immune-mediated or non-immune-mediated mechanisms. Non-immune-mediated thrombocytopenia is due to direct toxicity of drug molecules to platelets or megakaryocytes. Immune-mediated thrombocytopenia, on the other hand, involves the formation of antibodies that react to platelet-specific glycoprotein complexes, as in classic drug-induced immune thrombocytopenia (DITP), or to platelet factor 4, as in heparin-induced thrombocytopenia (HIT) and vaccine-induced immune thrombotic thrombocytopenia (VITT). Clinical signs include a rapid drop in platelet count, bleeding or thrombosis. Since the patient's condition can deteriorate rapidly, prompt diagnosis and management are critical. However, the necessary diagnostic tests are only available in specialized laboratories. Therefore, the most demanding step in treatment is to identify the agent responsible for thrombocytopenia, which often proves difficult because many patients are taking multiple medications and have comorbidities that can themselves also cause thrombocytopenia. While DITP is commonly associated with an increased risk of bleeding, HIT and VITT have a high mortality rate due to the high incidence of thromboembolic complications. A structured approach to drug-associated thrombocytopenia/thrombosis can lead to successful treatment and a lower mortality rate. In addition to describing the treatment of DITP, HIT, VITT, and vaccine-associated immune thrombocytopenia, this review also provides the pathophysiological and clinical information necessary for correct patient management.

Introduction Thrombocytopenia may develop in patients receiving several therapeutic agents. According to the pathogenesis of the thrombocytopenia, two categories can be distinguished: immune-mediated and non-immune-mediated thrombocytopenia.1 In the latter, direct cytotoxic effects of the drug molecules on the megakaryocytes may impair megakaryopoiesis or, more rarely, cause an accelerated platelet elimination by directly stimulating platelet apoptosis.2 In contrast, in drug-induced immune thrombocytopenia (DITP) a humoral immune response against platelet antigens causes increased platelet destruction/ consumption and/or impaired platelet production.3 Platelet specific glycoprotein (GP) complexes such as GPIIb/IIIa and GPIb/IX are most common target antigens in cases of DITP. As in primary immune thrombocytopenia (ITP), these antibodies may be associated to severe thrombocytopenia and bleeding while life-threatening thromboembolic complications are common in heparininduced thrombocytopenia4 (HIT) and vaccine induced-

thrombotic thrombocytopenia (VITT).5 The identification of the compound responsible for the reduced platelet count is essential to define the correct therapeutic approach in these patients. However, this is often challenging in many patients who are taking multiple medications and have comorbidities that can by themselves lead to thrombocytopenia. In this review, we will provide an update on treatment of DITP, HIT, VITT and vaccine-associated immune thrombocytopenia. Since the therapeutic approach must take into account both the pathophysiologic mechanisms and the clinical characteristics of the individual patient, these aspects will also be briefly dealt with.

Drug-induced immune thrombocytopenia Thrombocytopenia with a definite or probable causal relation to drug administration has been reported for many drugs. A freely available data bank called “Platelets on the

Haematologica | 107 - June 2022

1264

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias Web” (https://www.ouhsc.edu/platelets/ditp.html) contains more than 300 drugs with at least one confirmed/suspected case of DITP. Pathophysiology of drug-induced immune thrombocytopenia In DITP, IgG (less commonly IgM/A) antibodies bind to platelets leading to their destruction via either complement activation6,7 and/or phagocytosis8,9 (Figure 1). According to their binding mechanisms, at least six types of antibodies have been identified3 (Table 1): 1) quinine-type drug-dependent antibodies (DDAbs); 2) hapten-dependent DDAbs; 3) fiban-type DDAbs; 4) drug-specific DDAbs (against chimeric antibodies); 5) autoantibody; and 6) immune complexes (see below; sections on HIT/VITT).

I. Marini et al.

nia induced by chimeric GPIIb/IIIa antagonists. In this case, DITP may present within hours of exposure (early onset) due to already preformed naturally occurring antibodies against murine compounds of the therapeutic monoclonal antibodies. Bleeding is a frequent complication of DITP. Incidence rates of 9% and 0.8% have been reported for major and fatal bleeding in patients with DITP, respectively.10

Diagnosis of drug-induced immune thrombocytopenia Five clinical criteria have been proposed to help physicians to determine the diagnosis of DITP1: 1) exposure to new drugs 5-10 days before onset of thrombocytopenia; 2) recovery from thrombocytopenia after discontinuing the candidate drug; 3) other drugs were continued or reintroduced after discontinuation of the candidate drug with a Clinical features of drug-induced immune sustained normal platelet count; 4) other causes for thrombocytopenia thrombocytopenia were excluded; and 5) re-exposure to Thrombocytopenia usually occurs approximately 5-10 days the candidate drug (a practice that may cause unwarafter initial drug administration and the median nadir pla- ranted risk) resulted in recurrent thrombocytopenia.11 telet count is <20x109/L. An exception is thrombocytope- Nevertheless, considering that acquired thrombocytopenia

Figure 1. Schematic representation of the pathophysiology of drug-induced thrombocytopenia (DITP), immune checkpoint inhibitor-induced thrombocytopenia (ICI-induced ITP), heparin-induced thrombocytopenia (HIT), and vaccine-induced immune thrombotic thrombocytopenia (VITT). Ab: antibody; PLT: platelet; GP: glycoprotein; Hep: heparin; PF4: platelet factor 4; CI: checkpoint inhibitor; PD-L1: like programmed death-ligand 1. Haematologica | 107 - June 2022

1265

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

I. Marini et al.

Table 1. Binding mechanisms of antibodies in drug-induced immune thrombocytopenia (DITP).

Mechanism Quinine-type

Drug binds DDAb and subsequently to platelet integrin

Example of drugs Quinine, sulfonamide antibiotics, non-steroidal anti-inflammatory drugs

Drug links covalently to membrane protein and induces drug-specific binding by DDAb

Penicillin, some cephalosporin antibiotics

Drug reacts with GPIIb/IIIa and induces neoepitope(s) for the DDAb

Tirofiban, Eptifibatide

Drug-specific

DDAb recognize murine component of chimeric Fab fragment specific for GPIIIa

Abciximab

Autoantibody

Drug induces antibody that reacts with autologous platelets in absence of drug

Gold salts, procainamide

Antibodies form immune complexes with their target antigens

Heparin, an as yet unidentified component of adenoviral vector-based vaccine against COVID-19

Hapten-dependent Fiban-type

Immune complexes

DDAbs: drug dependent antibodies; GP: glycoprotein.

is common in patients who are taking multiple medications and have comorbidities that can lead to thrombocytopenia, identifying the drug responsible for thrombocytopenia based uniquely on clinical information may prove very difficult. Therefore, in vitro confirmation of DDAbs is recommended, even if the available tests do not always allow a laboratory to confirm the diagnostic suspicion. Laboratory assays have been developed to detect antibodies that bind to platelets in the presence of drugs or a drug metabolite. These assays have moderate sensitivity due to the insolubility of some drugs, and are not always feasible since some antibodies are specific for drug metabolites that are not available for testing.10,12

treatment has been stopped. Notably, as long as the drug or its metabolite(s) are present in plasma, transfusion of platelets is unsuccessful.13 Administration of high doses of IVIG (1 g/kg body weight) can be used to accelerate platelet recovery under two circumstances: in patients with severe thrombocytopenia and bleeding or in those at high risk of bleeding.14,15 A suggested approach to manage cases of suspected DITP is reported in Figure 2. Of note, well-documented reports indicate that folk medicines, herbal preparations, and even foods occasionally trigger acute thrombocytopenia.16 However, it is not clear whether immune mechanisms are responsible for the decrease in platelet count in all these cases.

Treatment of patients suspected of having drug-induced immune thrombocytopenia Immediate discontinuation of the drug responsible is the initial step for the treatment of DITP. In patients receiving multiple medications, if feasible, all compounds started within the last 5-10 days should be stopped and replaced. Discontinuation of the drugs could also be done sequentially, possibly depending on the a priori probability that a specific drug is incriminated. It is important to consider testing for drug-dependent antibodies before any therapeutic initiative is taken; therefore, a serum sample should be collected before DITP-treatment to subsequently confirm the clinical suspicion. In this context, it should be noted that intravenous immunoglobulin (IVIG) interferes with immunoassays that are commonly used to DDAb. In our experience, blood samples should be collected either before or at least 48 hours after IVIG administration in case the patient received this treatment (see below). Generally, after 4-5 half-lives of the offending drug or drug metabolite, the platelet count begins to recover once

Immune thrombocytopenia induced by immune checkpoint inhibitors Immune checkpoint inhibitors (ICI) were developed as a novel treatment approach for cancer and are currently used as immune therapy. ICI are monoclonal antibodies against proteins (PD-L1 and CTLA417) used by tumor cells to escape the immunosurveillance system. Complications during immune therapy with ICI are primarily inflammatory conditions and autoimmune disorders. However, ICI-induced ITP has been observed in some cases.18 This condition is rare, with an incidence between 0.2% to 2.8%, and is almost never severe or fatal.19 Pathophysiology of ICI-induced immune thrombocytopenia Immune-induced adverse events are thought to be triggered by a revival of exhausted T cells that can be induced on ICI administration (Figure 1). Whereas the exact pathomechanism of ICI-induced ITP is still unclear, it is likely

Haematologica | 107 - June 2022

1266

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

I. Marini et al.

Figure 2. Suggested algorithm to manage cases with suspected drug-induced thrombocytopenia (DITP). (Adapted from Arnold et al.80).

that ICI trigger the expression of antiplatelet antibodies Diagnosis of ICI-induced immune thrombocytopenia similarly to primary ITP.20 Obtaining a clear diagnosis of ICI-induced ITP is complex for several reasons. First, it is characterized by symptoms simiClinical manifestations of ICI-induced immune lar to all immune thrombocytopenias. Moreover, the abthrombocytopenia sence of specific diagnostic tests and/or biomarkers Patients who develop ICI-induced ITP show symptoms represents a critical limitation. Consequently, the diagnosis common to other types of thrombocytopenia, such as can only be made by a process of exclusion. In addition, it petechiae, easy bruising, spontaneous bleeding, and hema- must be remembered that the onset of the majority of the turia.20 However, more severe manifestations, such as cer- symptoms appears within the first 12 weeks of treatment, ebral hemorrhage with subsequent neurological deficits, which is an identical time-frame to other more common have been also reported in some rare fatal cases.21 types of acquired thrombocytopenias.22 Haematologica | 107 - June 2022

1267

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias Risk management of patients with ICI-induced immune thrombocytopenia Due to the uncertainties of a laboratory or clinical diagnosis, there is no reliable strategy to lower the risk of this condition. However, the search for autoantibodies in patients with a personal or family history of autoimmune diseases is recommended.20 Treatment of ICI-induced immune thrombocytopenia There is neither a general or a specific treatment for ICI-induced ITP. However, the American Society of Clinical Oncology published a guideline for the management of ICI-induced ITP depending on platelet count.23 In this, Grade 1 thrombocytopenia (platelets<100x109/L) should be managed by continuation of the immune therapy with a tightly scheduled follow up and laboratory evaluation. Restarting therapy is possible after recovery of a normal platelet count. In cases of Grade 2 thrombocytopenia (<75x109/L), maintain treatment but monitor for improvement; if not resolved, interrupt ICI. When the platelet count decreases to below 50x109/L (Grades 3 and 4), the advice of a consultant hematologist should be sought. Usually, in cases of Grades 3 and 4, the ICI treatment should be withheld and the use of highdose corticosteroids and/or IVIG over at least four weeks is suggested. If symptoms and platelet count do not improve within 48-72 hours of high-dose corticosteroid therapy, administration of infliximab (a chimeric monoclonal antibody neutralizing TNF-a) is recommended. In the absence of any recovery to Grade 1, the immune therapy must be definitively discontinued. There are other treatments, (e.g., rituximab or thrombopoietin receptor agonists) which can be considered in cases of patients refractory to initial therapy.

Heparin-induced thrombocytopenia Unfractionated heparin (UFH) or low molecular weight heparin (LMWH) binds to platelet factor 4 (PF4) and forms immunogenic complexes that induce the production of antibodies specific for PF4/heparin complexes in a subgroup of treated patients. A minority of these immunized patients could develop HIT type II (hereafter referred to as HIT), which is characterized by a drop in platelet count starting between days 5-10 of heparin therapy, with or without thromboembolic events.4 The use of LMWH instead of UFH is associated with a lower risk of HIT. In contrast to HIT, nonimmune thrombocytopenia (HIT type I) can also occur in 1030% of patients treated with heparin due to direct binding to platelets resulting in mild platelet activation. In fact, it has been shown that binding of UFH, LMWH or fondaparinux to the GPIIb/IIIa complex potentiates outside-in platelet signaling resulting in HIT type I,1 which typically occurs within the first few days (earlier than day 5) of heparin therapy, and is usually mild and without major clinical consequences.

I. Marini et al.

Platelet count often remains above 80-100x109/L and returns to basal levels spontaneously within a few days despite continuous heparin treatment. Pathophysiology of heparin-induced thrombocytopenia A subset of anti-PF4/heparin antibodies binds with their Fcdomains to Fc gamma Receptor IIA (FcgRIIA) on platelets,24 leading to platelet activation (Figure 1). Subsequently, the activated platelets release their granular contents, generate microparticles, promote thrombin formation, and finally aggregate.25,26 The activation of endothelium and monocytes, together with the expression of tissue factor, contributes to the pathophysiology of HIT.27 In addition, there is increasing evidence to suggest that neutrophils are also involved in thrombus formation. In fact, they are activated by P-selectin28 on platelets and via FcgRIIA by anti-PF4/heparin antibodies29 leading to neutrophil extracellular trap (NET) formation and release of their prothrombotic molecules (NETosis). Clinical manifestations of heparin-induced thrombocytopenia The key clinical feature of HIT is a drop in platelet count of >50% (from the highest value upon the first heparin administration), usually starting 5-10 days after beginning heparin treatment. However, HIT can be observed earlier in patients who are re-exposed to heparin after receiving it in the previous 100 days (rapid onset HIT).30,31 In a subgroup of HIT patients, characterized by disseminated intravascular coagulation (DIC), a bigger drop in platelet count (<20x109/L) might be observed.25,30 Thrombosis represents the most severe complication of HIT and contributes to disease-related morbidity and mortality. A new thrombotic complication is observed in almost 50% of patients with acute HIT;32 deep-vein thrombosis (DVT) is the most common. Notably, patients with HIT may develop asymptomatic thrombosis in veins in the lower and upper extremities, highlighting the importance of examining all four extremities in HIT.33 Rare complications of HIT include thrombosis in cerebral and splanchnic veins, as well as catheter-associated thrombosis. Although less frequent, arterial thrombosis has also been reported and usually involves lower-limb, cerebral, coronary, mesenteric and brachial arteries.26 In a few cases, severe HITassociated DIC triggers microthrombosis and critical limb ischemia. Finally, skin necrosis at the heparin injection sites and adrenal hemorrhagic necrosis are rare clinical manifestations of HIT.4 Diagnosis of heparin-induced thrombocytopenia The diagnosis of HIT can be challenging. The most commonly used scoring system is the 4Ts system,34 which includes four typical clinical peculiarities of HIT: (i) thrombocytopenia; (ii) timing of onset of thrombocytopenia;

Haematologica | 107 - June 2022

1268

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

I. Marini et al.

Table 2. The 4Ts scoring system to evaluate the pretest risk for heparin-induced thrombocytopenia (HIT).

4Ts Thrombocytopenia Timing of fall in platelet count

2 points

1 point

0 point

Platelet count falls >50% and platelet nadir ≥20x109/L

Platelet count falls 30-50% or platelet nadir 10-19x109/L

Platelet count falls <30% or platelet nadir <10x109/L

Clear onset between days 5-10 or falls ≤1 day*

Onset after day 10 or falls ≤1 day**

Platelet count falls <4 days or >14 days after exposure

Suspected thrombosis (not proven)

None

Possible

Definite

Thrombosis or other sequelae New thrombosis*** (confirmed) Other causes of thrombocytopenia

None apparent

The resulting clinical probability score is divided into high (6–8 points), intermediate (4–5 points), and low (≤3 points). *In case of prior heparin exposure during the last 30 days. **In case of prior heparin exposure within 30-100 days previously. ***Also skin necrosis, acute systemic reaction post-intravenous unfractionated heparin (UFH) bolus, progressive or recurrent thrombosis, non-necrotizing (erythematous) skin lesions.

(iii) thrombosis or other clinical sequelae; and (iv) the likelihood of other causes of thrombocytopenia (Table 2). The 4Ts system has modest positive predictive values and high negative predictive values.35 The laboratory diagnosis of HIT is based on the demonstration of the presence of anti-PF4/heparin antibodies and their ability to activate platelets in the presence of heparin. Several enzyme-linked immunosorbent assays (ELISA) and particle-based immunoassays are commercially available to detect antibody binding. Recent data showed high negative predictive values of these assays, and therefore they are effective for excluding HIT, while their low specificity makes a positive test result of little use in confirming a clinical suspicion.36 The presence of platelet-activating antibodies can be determined only by using functional assays that combine both high sensitivity and specificity for clinically-relevant HIT antibodies. The Heparin-Induced Platelet Activation37 assay and Serotonin Release Assay38 are currently considered the “gold standard”. Management of patients with suspected heparin-induced thrombocytopenia The first step in the management of patients with suspicion of HIT is the immediate discontinuation of all heparins. However, discontinuation of heparin alone is not sufficient to prevent the development or progression of the thrombosis because PF4 is an autoantigen and may form complexes with endogenous polyanions. Thus, antibody-mediated platelet activation will still take place even in the absence of heparin. Consequently, use of non-heparin anticoagulants for patients with high clinical suspicion of HIT is recommended while awaiting laboratory results. Different anticoagulants are available to treat patients with HIT (Table 3).39 However, the lack of clinical experience with some nonheparin anticoagulants that are rarely used outside HIT enhances the risk of complications related to under-dosing (thrombosis) or over-dosing (bleeding). Therefore, it is ex-

tremely important that physicians are able to discriminate between patients who actually have HIT and those who, although they may have PF4/heparin-specific antibodies and some degree of thrombocytopenia, actually do not. Diagnostic algorithms are available for diagnosis of HIT40,41 (see Figure 3 for an example). Alternative anticoagulants for heparin-induced thrombocytopenia Parenteral anticoagulant: activated factor X (Xa) inhibitors (danaparoid, fondaparinux) - Danaparoid is a low molecular weight heparinoid consisting of a mixture of heparan sulphate, dermatan sulphate and chondroitin sulphate. In a prospective randomized trial that analyzed patients with confirmed HIT, it was shown that danaparoid is efficient in preventing new, progressive, or recurrent thromboembolic complications (including thrombotic death) or limb amputation.42 Danaparoid has a low cross reactivity with HIT antibodies and can also mitigate HIT antibody-induced platelet activation through disruption of PF4/heparin complexes and replacing them from the platelet surface. Fondaparinux is a synthetic heparin pentasaccharide and has a potent indirect anti-Xa inhibitor effect. The off-label use of fondaparinux in HIT has recently increased43 and it has been shown that it is safe for patients with acute thrombosis and heparin-dependent platelet-activating antibodies.44 The efficacy and safety of fondaparinux are similar to those of argatroban and danaparoid in patients with suspected HIT.45 Despite the lack of randomized clinical trials, there is increasing evidence to indicate that fondaparinux is an efficient and safe option for the treatment of HIT.46 Direct thrombin inhibitors - Argatroban is a synthetic direct thrombin inhibitor that binds to the thrombin active site in a reversible manner. It is capable of inhibiting both free and clot-associated thrombin. Clinical trials reported that treatment with argatroban reduces death, amputation and thrombosis compared to historical controls.47,48 Bivalirudin is another synthetic peptide composed of two

Haematologica | 107 - June 2022

1269

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

I. Marini et al.

Table 3. Non-heparin alternative anticoagulants that may be used to treat heparin-induced thrombocytopenia (HIT) patients.

Anticoagulant

Mechanism of action

Application

Clearance

Half-life

Monitoring

Parenteral Argatroban

Direct thrombin inhibition

Intravenously

Hepatic

40-50 minutes

aPTT

Bivalirudin

Direct thrombin inhibition

Intravenously

Renal

25 minutes

aPTT

Danaparoid

Indirect inhibition of FXa

Intravenously, subcutaneously

Renal

24 hours

Danapariodanti-Xa

Fondaparinux

Indirect inhibition of FXa

Subcutaneously

Renal

17-24 hours

Fondaparinuxanti-Xa

Oral Direct thrombin inhibition

Peroral

Renal (ca. 85%)

12-14 hours

Not required

Rivaroxaban

Direct inhibition of FXa

Peroral

Renal (ca. 33%)

5-9 hours

Not required

Apixaban

Direct inhibition of FXa

Peroral

Renal (ca. 25%)

8-15 hours

Not required

Endoxaban

Direct inhibition of FXa

Peroral

Renal (ca. 50%)

10-14 hours

Not required

Dabigatran

APTT: activated partial thromboplastin time; Xa: activated Factor X; Fxa: Factor Xa.

short hirudin peptide fragments. It is widely used for nonHIT patients with coronary disease because some studies showed that it reduced bleeding complications in the setting of invasive cardiology.49 Direct thrombin inhibitors - Argatroban is a synthetic direct thrombin inhibitor that binds to the thrombin active site in a reversible manner. It is capable of inhibiting both free and clot-associated thrombin. Clinical trials reported that treatment with argatroban reduces death, amputation and thrombosis compared to historical controls.47,48 Bivalirudin is another synthetic peptide composed of two short hirudin peptide fragments. It is widely used for nonHIT patients with coronary disease because some studies showed that it reduced bleeding complications in the setting of invasive cardiology.49 Direct oral anticoagulants - Rivaroxaban, apixaban and endoxaban directly block Xa, while dabigatran is a direct thrombin inhibitor. There is increasing evidence to suggest the safety and efficacy of several direct oral anticoagulants (DOAC) in HIT.46 In one small multicenter prospective study good safety and efficacy of rivaroxaban, without new thrombotic events, were observed.50 In another case series, the use of rivaroxaban, apixaban or dabigatran showed no recurrent arterial, venous thrombosis or bleeding complications.51,52 Although these observations seem encouraging, robust clinical studies investigating the efficacy and safety of these drugs in patients with acute HIT are limited, and this precludes any evidence-based recommendations. Im-

portantly, the observed low trough levels of the drugs may not offer adequate protection for some HIT patients. Intravenous immunoglobulin treatment The use of high doses of IVIG was shown to inhibit HIT antibody-mediated platelet activation. Evidence emerging from case reports indicates that patients with prolonged thrombocytopenia who are refractory to standard treatment may benefit from IVIG therapy.53 However, more data are needed to clarify the indications and the appropriate IVIG dosing. Choice and duration of the anticoagulation Several factors should be considered when selecting the type and duration of non-heparin anticoagulant: the patient’s clinical stability, hepatic and renal functions; the physician’s expertise also plays a role. For example, fondaparinux should be avoided in patients with impaired renal function, and a cautious approach should be adopted in patients with hepatic insufficiency when argatroban is used. All the advantages and disadvantages of non-heparin anticoagulants should be weighed up when selecting an agent for the treatment of HIT patients (Table 3). Anticoagulation should be given at therapeutic concentration to HIT patients who have suffered from thromboembolic complication for at least three months. The treatment can be initiated using parenteral agents and a later switch can be made to oral anticoagulants, DOAC, or a vitamin K antagonist. While the timing for the switch has

Haematologica | 107 - June 2022

1270

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

I. Marini et al.

Figure 3. Management of patients with suspected heparin-induced thrombocytopenia (HIT) based on clinical assessment supported by complementary laboratory investigations. Screening platelet factor 4 (PF4)-dependent immunoassays are indicated for patients with at least intermediate probability of HIT. If the ELISA assay is positive, heparin should be stopped, an alternative anticoagulant should be started, and thromboembolic complications should be excluded. Next, functional assays should also be performed to confirm or refute a diagnosis of HIT. HIPA: heparin-induced platelet activation assay; SRA: serotonin release assay; FC: flow cytometric assay; DOAC: direct oral anticoagulants; Vit-K: vitamin K; DIC: disseminated intravascular coagulation. (Figure modified, with permission, from Bakchoul and Marini1).

not standardized, vitamin K antagonists should always be given together with parenteral anticoagulants for at least five days and until an international normalized ratio (INR) of over 2.0 is achieved. Since vitamin K antagonist and argatroban prolong prothrombin time/INR beyond that of vitamin K antagonist alone, combined therapy should be given until INR is >4. There is, however, no evidence of the use of therapeutic anticoagulation in patients with isolated thrombocytope-

nia who have no other indication for anticoagulation. In such cases, we prefer to treat the patient (serologically confirmed HIT type II with no suggestion of thrombosis) with non-heparin anticoagulants at a therapeutic concentration until the platelet count has reached pre-heparin values, or at least 150x109/L. Before ending the treatment, it is also recommended to re-evaluate the thrombotic risk and exclude the presence of asymptomatic thrombosis.

Haematologica | 107 - June 2022

1271

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

COVID-19 vaccines and thrombocytopenia Vaccine-induced thrombotic thrombocytopenia Coronavirus disease 2019 (COVID-19) is an infectious disease caused by coronavirus type 2 (SARS-CoV-2). In a very short time, different types of vaccines (mRNA vaccine, adenoviral vector-based vaccine, and protein-based vaccine) were developed. Reports of the development of thrombocytopenia and thrombosis at unusual sites with fatal outcome started to emerge shortly after the rollout of adenoviral vector-based vaccines; this phenomenon is now termed a vaccine-induced immune thrombotic thrombocytopenia (VITT). Definition of vaccine-induced immune thrombotic thrombocytopenia - Vaccine-induced immune thrombotic thrombocytopenia is characterized by thrombosis, thrombocytopenia, and the presence of PF4 reactive antibodies developed 4-30 days from the administration of an adenoviral vector-based vaccine.5 An incidence of from 2.5 to 38 per million doses has been reported after the administration of the first dose of ChadOx1 nCov1954 (https://www.gov.uk/government/publications/coronaviruscovid-19-vaccine-adverse-reactions). See and co-workers recently reported an incidence of 3.5 per million doses after vaccination with Ad26.Cov2.S in the USA.55 Although even more rarely, VITT can also develop after a second dose of ChadOx1 nCov19 with an incidence of 1.9 per million doses.54 Early case reports indicated young women (<40 years of age) represented the high risk group, but recent case series reported an even distribution of VITT across gender groups.5 Pathophysiology vaccine-induced immune thrombotic thrombocytopenia - IgG antibodies reactive to PF4 are the drivers of the pathophysiology of VITT56 (Figure 1). However, the link between vaccination and the development of antiPF4 antibodies is not clear. One of the proposed mechanisms is the generation of a neoantigen with the binding of PF4 to vaccine components such as human and nonstructural viral proteins or free DNA.57 Another theory suggests cross-reactivity between spike protein of the virus and PF4.58 However, anti-PF4 antibodies do not cross-react with spike protein,58 and no correlation between antiSARS-CoV-2 antibodies and anti-PF4 antibodies has been found.59 Anti-PF4 antibodies bind to PF4 on the platelet surface and these complexes activate platelets by engaging their FcγRIIA, thus causing the generation of procoagulant platelets and promoting thrombus formation.56,60 Clinical manifestations of vaccine-induced immune thrombotic thrombocytopenia - In most cases, the first physical signs occur within two weeks of vaccination, but a delay in presentation is also possible. Petechiae, bruising or even

I. Marini et al.

hematoma can be seen in patients with severe thrombocytopenia. Cerebral venous sinuses (CVST) are the most common sites of thrombosis and the first symptom is a severe headache.5 Intracranial hemorrhage is present in almost one-third of patients with CVST, which is associated with a high mortality rate.61 Patients with cerebral ischemia and hemorrhage present with altered mental status and/or focal neurological deficits. Abdominal pain is a sign of splanchnic vein thrombosis, while dyspnea and chest pain suggest pulmonary artery embolism. Patients with lower extremity DVT have leg pain or swelling. A mortality rate as high as 60% has been reported in an initial case series62 and recent case series report this to still be over 20%.61 Diagnosis of vaccine-induced immune thrombotic thrombocytopenia - The suggested clinical and laboratory criteria include: 1) symptom onset within 5-30 days of vaccination with an adenoviral vector-based COVID-19 vaccine (ChadOx1 nCov19 and Ad26.COV2.S); 2) documented venous or arterial thrombosis; 3) thrombocytopenia (<150x109/L); 4) D-dimer >4,000 ng/mL fibrinogen equivalent units; and 5) positivity of anti-PF4 IgG ELISA test and modified functional platelet activation assay.5,63 If all five criteria are met, the diagnosis of VITT is considered definite; if one criterion is missing, the diagnosis is considered probable; in these cases, anticoagulation and IVIG may be considered according to the clinical and laboratory findings.5 A suggested approach for the diagnosis and initial management of patients with suspected VITT is reported in Figure 4. Of note, rapid immunoassays are not suitable for the detection of anti-PF4 antibodies in VITT64 and a sensitive anti-PF4 ELISA is recommended.63,64 The Scientific and Standardization Committee on Platelet Immunology of the International Society on Thrombosis and Haemostasis also recommends additional laboratory confirmation of the diagnosis by functional platelet activation assays (heparininduced platelet activation assay, serotonin release assay or P-selectin expression assays).63 It is important to know that some patients do not have thrombosis at presentation but may later develop it.65 Moreover, a subgroup of patients may have a normal platelet count at presentation but they develop thrombocytopenia after a couple of days.5,66 Treatment of vaccine-induced immune thrombotic thrombocytopenia - Several societies have published recommendations regarding the management of patients with VITT67 that are mainly based on experience with the treatment of HIT patients. Because of the rapid deterioration in their clinical condition and the high mortality rate, patients with suspected or confirmed CVST should be transferred to a center with capability for neurosurgical intervention (see below). To avoid further thrombotic complications, anticoagulation in therapeutic doses is required.68 It is recommended to avoid heparin and LMWH in patients with VITT,69 although

Haematologica | 107 - June 2022

1272

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias successful use of heparin has been reported in some cases.70 It has not yet been demonstrated that heparinbased treatments worsen the clinical condition of patients with VITT, but one of the following non-heparin anticoagulants may be preferred: DOACs such as apixaban or rivaroxaban; direct thrombin inhibitors like argatroban,

I. Marini et al.

dabigatran, bivalirudin, and fondaparinux. Bleeding risk, renal or hepatic impairment, and the need for oral or parenteral application should be considered when choosing the anticoagulant. In patients with CSVT, parenteral agents should be preferred over DOAC due to the increased bleeding risk in these patients.68 The anticoagulant can be

Figure 4. Suggested approach to diagnosis and initial management of patients with suspected vaccine-induced immune thrombotic thrombocytopenia (VITT). FEU: fibrin equivalent units; PT: prothrombin time; aPTT: activated partial thromboplastin time; HIPA: heparin-induced platelet activation assay; SRA: serotonin release assay; PEA: P-selectin expression assay. Haematologica | 107 - June 2022

1273

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias switched to an oral alternative after the acute phase, or at discharge in patients receiving a parenteral anticoagulant in hospital,68 and anticoagulation should continue for at least three months after the platelet count has normalized. Patients without a documented thrombosis but with severe headache suggestive of CSVT should also receive anticoagulation.71 Intravenous immunoglobulin is the only available therapy that can interfere with the pathological anti-PF4 antibodies and limit the progression of VITT.72,73 High-dose IVIG treatment (1 g/kg/day) should be started promptly in suspected cases without waiting for the results of confirmatory tests. Additional IVIG administrations should be performed on the second day in case of CSVT and splanchnic thrombosis, or later in patients not responding to the first dose.71 It has been shown that IVIG inhibits VITT sera-induced procoagulant platelet formation and activation in functional assays.56,72,73 In accordance with these in vitro data, several case reports showed a rapid increase in platelet count after IVIG administration.72,73 In a recent case series investigating VITT patients with CSVT, a significantly lower mortality rate in the subgroup receiving IVIG than in those who did not (40% vs. 73%; P=0.022) has been reported.70 Furthermore, steroids (e.g., prednisone 1-2 mg/kg/day or dexamethasone 40 mg/day for 4 days) could also be considered to mitigate the immune response when IVIG is not available,71 although the benefit of steroids in VITT is uncertain.70 Based on the successful use of plasma exchange in a few refractory patients,74 this treatment might be applied to patients not responding to IVIG. The rationale for this approach is that plasma exchange not only removes the IgG antibodies causing VITT from the circulation, but also replaces factors consumed during the process of thrombosis.75 In addition, endovascular treatment and neurosurgical interventions might be required in selected cases with CSVT. While platelet transfusion should generally be avoided, it must be considered in cases where life-threatening bleeding occurs or when immediate major surgery is required.71 Finally, it is important to consider that patients may develop a new thrombosis during the course of the disease. Therefore, patients should be constantly monitored for clinical signs of thrombosis and platelet counts should be checked. Treatment of female VITT patients who are pregnant or breastfeeding requires special considerations. IVIG can be used safely in these patients. If IVIG is not available, shortterm corticosteroid treatment may be considered in consultation with obstetricians.71 The preferred anticoagulation is heparin/LMWH, but danaparoid or fondaparinux may also be used in these patient populations.71 In contrast, direct oral factor Xa inhibitors should be avoided.

I. Marini et al.

ported after COVID-19 vaccines. It is not yet clear whether ITP after receiving COVID-19 vaccines is coincidental or consequential. De novo ITP development has been reported after both mRNA and adenoviral vector vaccines. The mechanism(s) of vaccine-associated ITP are yet to be clarified. Molecular mimicry and underlying predisposition to autoimmunity have been proposed as potential etiological factors.76 Furthermore, exacerbation of a previously undetected ITP after vaccination could be another explanation. Most of the patients present skin purpura or oral mucosal bleeding, but severe bleeding seems to be rare. In the largest case series reported so far, Lee and coworkers found a response rate of over 90% to first-line treatments (IVIG, corticosteroid, platelet transfusion).76 In another case series, 5 out of 9 patients responded to firstline treatment with IVIG and corticosteroid, but most of the patients remained on corticosteroids for at least 30 days.77 Further studies are needed to evaluate the longterm outcome of patients developing ITP after receiving COVID-19 vaccines. Vaccination in patients with immune thrombocytopenia Case series reported ITP exacerbation after COVID-19 vaccination. Visser and co-workers reported exacerbation of ITP in 13.8% of ITP patients. Factors associated with exacerbation after vaccination were platelet count <50x109/L, young age, and ITP treatment at the time of vaccination. Five patients (2.2%) suffered from a bleeding event.78 Similarly, in another study, exacerbation was observed in 19 out of 109 (17%) ITP patients after the first vaccination and in 14 out of 70 (20%) after the second dose.76 Splenectomy and refractory cases (>5 lines of therapy) were risk factors for ITP exacerbation. If ITP patients tolerate the first vaccine well, there is less likely to be an exacerbation of the ITP after the second dose.76 Furthermore, based on the currently available data, patients with ITP exacerbation after vaccination responded favorably to treatment.76-79 In conclusion, ITP patients should receive vaccination for COVID-19. A platelet count 3-7 days before and another after the vaccination will help to identify any immediate drop in platelet values.79 A second dose should be avoided in ITP patients who experienced a major exacerbation of ITP after the first dose of the vaccine.79

Conclusion and future aspects

Thrombocytopenia after drug administration can be associated with bleeding or thrombosis, depending on the pathophysiology of platelet destruction. Significant progress has been made over the last two decades in understanding the Vaccine-associated immune thrombocytopenia pathogenetic mechanism of DITP. However, there are still Recently, the development of de novo ITP has been re- numerous diagnostic and treatment challenges, especially Haematologica | 107 - June 2022

1274

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias

I. Marini et al.

in critically ill patients, including the difficulty in distinguish- the current work. Other authors declare no competing fiing drug-associated thrombocytopenia from secondary nancial interests. thrombocytopenia caused by underlying disorders. Contributions Disclosures IM conducted the literature search, created the figures, and TB has received research funding from CoaChrom Diagnos- wrote the sections on the pathophysiology and diagnosis tica GmbH, DFG, Robert Bosch GmbH, Stiftung Transfusion- of DITP and HIT. GU conducted the literature search and smedizin und Immunhämatologie e.V.: Ergomed, Surrey, wrote the sections on clinical manifestations and treatDRK Blutspendedienst, Deutsche Herzstiftung, Ministerium ment of VITT. KJ conducted the literature search and wrote fuer Wissenschaft, Forschung und Kunst Baden-Wuerttem- the section about ICI-induced ITP. TB designed the original bergm, has received lecture honoraria from Aspen Germany layout and edited the manuscript. All authors approved the GmbH, Bayer Vital GmbH, Bristol-Myers Squibb GmbH & final version of the manuscript. Co., Doctrina Med AG, Meet The Experts Academy UG, Schoechl medical education GmbH, Mattsee, Stago GmbH, Acknowledgment Mitsubishi Tanabe Pharma GmbH, Novo Nordisk Pharma This work was supported by grants from the German ReGmbH, has provided consulting services to: Terumo, has search Foundation and from the Herzstiftung to TB provided expert witness testimony relating to heparin in- (BA5158/4 and TSG-Study), by special funds from the state duced thrombocytopenia (HIT) and non‐HIT thrombocyto- of Baden-Wuerttemberg for coagulation research and the penic and coagulopathic disorders. All of these are outside blood donation service of the German Red Cross to TB.

References 1. Bakchoul T, Marini I. Drug-associated thrombocytopenia. Hematology Am Soc Hematol Educ Program. 2018;2018(1):576583. 2. Towhid ST, Tolios A, Munzer P, et al. Stimulation of platelet apoptosis by balhimycin. Biochem Biophys Res Commun. 2013;435(2):323-326. 3. Curtis BR. Drug-induced immune thrombocytopenia: incidence, clinical features, laboratory testing, and pathogenic mechanisms. Immunohematology. 2014;30(2):55-65. 4. Gruel Y, De Maistre E, Pouplard C, et al. Diagnosis and management of heparin-induced thrombocytopenia. Anaesth Crit Care Pain Med. 2020;39(2):291-310. 5. Pavord S, Scully M, Hunt BJ, et al. Clinical features of vaccineinduced immune thrombocytopenia and thrombosis. N Engl J Med. 2021;385(18):1680-1689. 6. Najaoui A, Bakchoul T, Stoy J, et al. Autoantibody-mediated complement activation on platelets is a common finding in patients with immune thrombocytopenic purpura (ITP). Eur J Haematol. 2012;88(2):167-174. 7. Peerschke EI, Andemariam B, Yin W, Bussel JB. Complement activation on platelets correlates with a decrease in circulating immature platelets in patients with immune thrombocytopenic purpura. Br J Haematol. 2010;148(4):638-645. 8. Shulman NR. Immunoreactions involving platelets. IV. Studies on the pathogenesis of thrombocytopenia in drug purpura using test doses of quinidine in sensitized individuals; their implications in idiopathic thrombocytopenic purpura. J Exp Med. 1958;107(5):711-729. 9. McKenzie SE, Taylor SM, Malladi P, et al. The role of the human Fc receptor Fc gamma RIIA in the immune clearance of platelets: a transgenic mouse model. J Immunol. 1999;162(7):4311-4318. 10. Vayne C, Guery EA, Rollin J, et al. Pathophysiology and diagnosis of drug-induced immune thrombocytopenia. J Clin Med. 2020;9(7):2212. 11. Selleng K, Schutt A, Selleng S, et al. Studies of the anti-platelet factor 4/heparin immune response: adapting the enzyme-linked

immunosorbent spot assay for detection of memory B cells against complex antigens. Transfusion. 2010;50(1):32-39. 12. Warkentin TE. Laboratory diagnosis of heparin-induced thrombocytopenia. Int J Lab Hematol. 2019;41 Suppl 1:15-25. 13. George JN, Aster RH. Drug-induced thrombocytopenia: pathogenesis, evaluation, and management. Hematology Am Soc Hematol Educ Program. 2009;153-158. 14. Tvito A, Bakchoul T, Rowe JM, et al. Severe and persistent heparin-induced thrombocytopenia despite fondaparinux treatment. Am J Hematol. 2015;90(7):675-678. 15. Ray JB, Brereton WF, Nullet FR. Intravenous immune globulin for the treatment of presumed quinidine-induced thrombocytopenia. DICP. 1990;24(7-8):693-695. 16. Lavy R. Thrombocytopenic purpura due to Lupinus termis bean. J Allergy. 1964;35:386-389. 17. Bagchi S, Yuan R, Engleman EG. Immune checkpoint inhibitors for the treatment of cancer: clinical impact and mechanisms of response and resistance. Annu Rev Pathol. 2021;16:223-249. 18. Michot JM, Lazarovici J, Tieu A, et al. Haematological immunerelated adverse events with immune checkpoint inhibitors, how to manage? Eur J Cancer. 2019;122:72-90. 19. Haddad TC, Zhao S, Li M, et al. Immune checkpoint inhibitorrelated thrombocytopenia: incidence, risk factors and effect on survival. Cancer Immunol Immunother. 2022;71(5):1157-1165. 20. Calvo R. Hematological side effects of immune checkpoint inhibitors: the example of immune-related thrombocytopenia. Front Pharmacol. 2019;10:454. 21. Hadfield MJ, Mui G. A fatal case of immune thrombocytopenia secondary to the immune checkpoint inhibitor ipilimumab in a patient with BRAF wild type metastatic melanoma. J Oncol Pharm Pract. 2020;26(6):1530-1532. 22. Shiuan E, Beckermann KE, Ozgun A, et al. Thrombocytopenia in patients with melanoma receiving immune checkpoint inhibitor therapy. J Immunother Cancer. 2017;5:8. 23. Brahmer JR, Lacchetti C, Schneider BJ, et al. Management of immune-related adverse events in patients treated with immune checkpoint inhibitor therapy: American Society of

Haematologica | 107 - June 2022

1275

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias Clinical Oncology Clinical Practice Guideline. J Clin Oncol. 2018;36(17):1714-1768. 24. Rollin J, Pouplard C, Gruel Y. Risk factors for heparin-induced thrombocytopenia: focus on Fcgamma receptors. Thromb Haemost. 2016;116(5):799-805. 25. Greinacher A, Selleng K, Warkentin TE. Autoimmune heparininduced thrombocytopenia. J Thromb Haemost. 2017;15(11):2099-2114. 26. Greinacher A. Clinical practice. Heparin-induced thrombocytopenia. N Engl J Med. 2015;373(3):252-261. 27. Kasthuri RS, Glover SL, Jonas W, et al. PF4/heparin-antibody complex induces monocyte tissue factor expression and release of tissue factor positive microparticles by activation of FcgammaRI. Blood. 2012;119(22):5285-5293. 28. Perdomo J, Leung HHL, Ahmadi Z, et al. Neutrophil activation and NETosis are the major drivers of thrombosis in heparininduced thrombocytopenia. Nat Commun. 2019;10(1):1322. 29. Xiao Z, Visentin GP, Dayananda KM, Neelamegham S. Immune complexes formed following the binding of anti-platelet factor 4 (CXCL4) antibodies to CXCL4 stimulate human neutrophil activation and cell adhesion. Blood. 2008;112(4):1091-1100. 30. Warkentin TE. High-dose intravenous immunoglobulin for the treatment and prevention of heparin-induced thrombocytopenia: a review. Expert Rev Hematol. 2019;12(8):685-698. 31. Warkentin TE. HITlights: a career perspective on heparininduced thrombocytopenia. Am J Hematol. 2012;87 Suppl 1:S92-99. 32. Warkentin TE, Roberts RS, Hirsh J, Kelton JG. An improved definition of immune heparin-induced thrombocytopenia in postoperative orthopedic patients. Arch Intern Med. 2003;163(20):2518-2524. 33. Harris EI, Zurbriggen LD, Brunner MJ, Williams EC. Doppler ultrasound screening in patients with newly diagnosed heparininduced thrombocytopenia. Blood Adv. 2021;5(22):4575-4577. 34. Lo GK, Juhl D, Warkentin TE, et al. Evaluation of pretest clinical score (4 Ts) for the diagnosis of heparin-induced thrombocytopenia in two clinical settings. J Thromb Haemost. 2006;4(4):759-765. 35. Nagler M, Fabbro T, Wuillemin WA. Prospective evaluation of the interobserver reliability of the 4Ts score in patients with suspected heparin-induced thrombocytopenia. J Thromb Haemost. 2012;10(1):151-152. 36. Linkins LA, Bates SM, Lee AY, et al. Combination of 4Ts score and PF4/H-PaGIA for diagnosis and management of heparininduced thrombocytopenia: prospective cohort study. Blood. 2015;126(5):597-603. 37. Greinacher A, Michels I, Kiefel V, Mueller-Eckhardt C. A rapid and sensitive test for diagnosing heparin-associated thrombocytopenia. Thromb Haemost. 1991;66(6):734-736. 38. Sheridan D, Carter C, Kelton JG. A diagnostic test for heparininduced thrombocytopenia. Blood. 1986;67(1):27-30. 39. Cuker A, Arepally GM, Chong BH, et al. American Society of Hematology 2018 guidelines for management of venous thromboembolism: heparin-induced thrombocytopenia. Blood Adv. 2018;2(22):3360-3392. 40. Pishko AM, Cuker A. Diagnosing heparin-induced thrombocytopenia: the need for accuracy and speed. Int J Lab Hematol. 2021;43 Suppl 1:96-102. 41. Hogan M, Berger JS. Heparin-induced thrombocytopenia (HIT): review of incidence, diagnosis, and management. Vasc Med. 2020;25(2):160-173. 42. Lubenow N, Warkentin TE, Greinacher A, et al. Results of a systematic evaluation of treatment outcomes for heparin-

I. Marini et al.

induced thrombocytopenia in patients receiving danaparoid, ancrod, and/or coumarin explain the rapid shift in clinical practice during the 1990s. Thromb Res. 2006;117(5):507-515. 43. Schindewolf M, Steindl J, Beyer-Westendorf J, et al. Frequent off-label use of fondaparinux in patients with suspected acute heparin-induced thrombocytopenia (HIT)--findings from the GerHIT multi-centre registry study. Thromb Res. 2014;134(1):2935. 44. Warkentin TE, Davidson BL, Buller HR, et al. Prevalence and risk of preexisting heparin-induced thrombocytopenia antibodies in patients with acute VTE. Chest. 2011;140(2):366-373. 45. Kang M, Alahmadi M, Sawh S, et al. Fondaparinux for the treatment of suspected heparin-induced thrombocytopenia: a propensity score-matched study. Blood. 2015;125(6):924-929. 46. Linkins LA, Hu G, Warkentin TE. Systematic review of fondaparinux for heparin-induced thrombocytopenia: when there are no randomized controlled trials. Res Pract Thromb Haemost. 2018;2(4):678-683. 47. Lewis BE, Wallis DE, Leya F, et al.; Argatroban-915 Investigators. Argatroban anticoagulation in patients with heparin-induced thrombocytopenia. Arch Intern Med. 2003;163(15):1849-1856. 48. Tardy-Poncet B, Nguyen P, Thiranos JC, et al. Argatroban in the management of heparin-induced thrombocytopenia: a multicenter clinical trial. Crit Care. 2015;19:396. 49. Warkentin TE, Greinacher A, Koster A. Bivalirudin. Thromb Haemost. 2008;99(5):830-839. 50. Linkins LA, Warkentin TE, Pai M, et al. Rivaroxaban for treatment of suspected or confirmed heparin-induced thrombocytopenia study. J Thromb Haemost. 2016;14(6):1206-1210. 51. Ng HJ, Than H, Teo EC. First experiences with the use of rivaroxaban in the treatment of heparin-induced thrombocytopenia. Thromb Res. 2015;135(1):205-207. 52. Mirdamadi A. Dabigatran, a direct thrombin inhibitor, can be a life-saving treatment in heparin-induced thrombocytopenia. ARYA Atheroscler. 2013;9(1):112-114. 53. Dougherty JA, Yarsley RL. Intravenous immune globulin (IVIG) for treatment of autoimmune heparin-induced thrombocytopenia: a systematic review. Ann Pharmacother. 2021;55(2):198-215. 54. Schultz NH, Sorvoll IH, Michelsen AE, et al. Thrombosis and thrombocytopenia after ChAdOx1 nCoV-19 vaccination. N Engl J Med. 2021;384(22):2124-2130. 55. See I, Su JR, Lale A, et al. US case reports of cerebral venous sinus thrombosis with thrombocytopenia after Ad26.COV2.S vaccination, March 2 to April 21, 2021. JAMA. 2021;325(24):24482456. 56. Althaus K, Moller P, Uzun G, et al. Antibody-mediated procoagulant platelets in SARS-CoV-2-vaccination associated immune thrombotic thrombocytopenia. Haematologica. 2021;106(8):2170-2179. 57. Greinacher A, Selleng K, Palankar R, et al. Insights in ChAdOx1 nCoV-19 vaccine-induced immune thrombotic thrombocytopenia. Blood. 2021;138(22):2256-2268. 58. Greinacher A, Selleng K, Mayerle J, et al. Anti-platelet factor 4 antibodies causing VITT do not cross-react with SARS-CoV-2 spike protein. Blood. 2021;138(14):1269-1277. 59. Uzun G, Althaus K, Bakchoul T. No correlation between anti-PF4 and anti-SARS-CoV-2 antibodies after ChAdOx1 nCoV-19 vaccination. N Engl J Med. 2021;385(14):1334-1336. 60. Azzarone B, Veneziani I, Moretta L, Maggi E. Pathogenic mechanisms of vaccine-induced immune thrombotic thrombocytopenia in people receiving anti-COVID-19 adenoviralbased vaccines: a proposal. Front Immunol. 2021;12:728513. 61. van de Munckhof A, Krzywicka K, Aguiar de Sousa D, et al. Declining mortality of cerebral venous sinus thrombosis with

Haematologica | 107 - June 2022

1276

REVIEW SERIES - Treatment of drug-induced immune thrombocytopenias thrombocytopenia after SARS-CoV-2 vaccination. Eur J Neurol. 2022;29(1):339-344. 62. Greinacher A, Thiele T, Warkentin TE, et al. Thrombotic thrombocytopenia after ChAdOx1 nCov-19 vaccination. N Engl J Med. 2021;384(22):2092-2101. 63. Nazy I, Sachs UJ, Arnold DM, et al. Recommendations for the clinical and laboratory diagnosis of VITT against COVID-19: communication from the ISTH SSC Subcommittee on Platelet Immunology. J Thromb Haemost. 2021;19(6):1585-1588. 64. Sachs UJ, Cooper N, Czwalinna A, et al. PF4-dependent immunoassays in patients with vaccine-induced immune thrombotic thrombocytopenia: results of an interlaboratory comparison. Thromb Haemost. 2021;121(12):1622-1627. 65. Salih F, Schonborn L, Kohler S, et al. Vaccine-induced thrombocytopenia with severe headache. N Engl J Med. 2021;385(22):2103-2105. 66. Gabarin N, Patterson S, Pai M, et al. Venous thromboembolism and mild thrombocytopenia after ChAdOx1 nCoV-19 vaccination. Thromb Haemost. 2021;121(12):1677-1680. 67. Zazzeron L, Rosovsky RP, Bittner EA, Chang MG. Comparison of published guidelines for the diagnosis and the management of vaccine-induced immune thrombotic thrombocytopenia. Crit Care Explor. 2021;3(9):e0519. 68. Klok FA, Pai M, Huisman MV, Makris M. Vaccine-induced immune thrombotic thrombocytopenia. Lancet Haematol. 2021;9(1):e73-e80. 69. Oldenburg J, Klamroth R, Langer F, et al. Diagnosis and management of vaccine-related thrombosis following AstraZeneca COVID-19 vaccination: Guidance Statement from the GTH. Hamostaseologie. 2021;41(3):184-189. 70. Perry RJ, Tamborska A, Singh B, et al. Cerebral venous

I. Marini et al.

thrombosis after vaccination against COVID-19 in the UK: a multicentre cohort study. Lancet. 2021;398(10306):1147-1156. 71. Greinacher A, Langer F, Makris M, et al. Vaccine-induced immune thrombotic thrombocytopenia (VITT): update on diagnosis and management considering different resources. J Thromb Haemost. 2022;20(1):149-156. 72. Uzun G, Althaus K, Singh A, et al. The use of IV immunoglobulin in the treatment of vaccine-induced immune thrombotic thrombocytopenia. Blood. 2021;138(11):992-996. 73. Bourguignon A, Arnold DM, Warkentin TE, et al. Adjunct immune globulin for vaccine-induced immune thrombotic thrombocytopenia. N Engl J Med. 2021;385(8):720-728. 74. Patriquin CJ, Laroche V, Selby R, et al. Therapeutic plasma exchange in vaccine-induced immune thrombotic thrombocytopenia. N Engl J Med. 2021;385(9):857-859. 75. Rock G, Weber V, Stegmayr B. Therapeutic plasma exchange (TPE) as a plausible rescue therapy in severe vaccine-induced immune thrombotic thrombocytopenia. Transfus Apher Sci. 2021;60(4):103174. 76. Lee EJ, Beltrami Moreira M, Al-Samkari H, et al. SARS-CoV-2 vaccination and immune hrombocytopenia in de novo and preexisting ITP patients. Blood. 2022;139(10):1564-1574. 77. Choi PY, Hsu D, Tran HA, et al. Immune thrombocytopenia following vaccination during the COVID-19 pandemic. Haematologica. 2021 Aug 26. [Epub ahead of print] 78. Visser C, Swinkels M, van Werkhoven E, et al. COVID-19 vaccination in patients with immune thrombocytopenia. Blood Adv. 2022;6(6):1637-1644. 79. Kuter DJ. Exacerbation of immune thrombocytopenia following COVID-19 vaccination. Br J Haematol. 2021;195(3):365-370.

Haematologica | 107 - June 2022

1277

REVIEW SERIES

Treatment of inherited thrombocytopenias Carlo L. Balduini

Correspondence:

Ferrata-Storti Foundation and University of Pavia, Pavia, Italy

Carlo L. Balduini carlo.balduini@unipv.it Received: February 14, 2022. Accepted: March 7, 2022. https://doi.org/10.3324/haematol.2022.280856 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract The new techniques of genetic analysis have made it possible to identify many new forms of inherited thrombocytopenias (IT) and study large series of patients. In recent years, this has changed the view of IT, highlighting the fact that, in contrast to previous belief, most patients have a modest bleeding diathesis. On the other hand, it has become evident that some of the mutations responsible for platelet deficiency predispose the patient to serious, potentially lifethreatening diseases. Today's vision of IT is, therefore, very different from that of the past and the therapeutic approach must take these changes into account while also making use of the new therapies that have become available in the meantime. This review, the first devoted entirely to IT therapy, discusses how to prevent bleeding in those patients who are exposed to this risk, how to treat it if it occurs, and how to manage the serious illnesses to which patients with IT may be predisposed.

Introduction Until a few years ago, a review entirely dedicated to the therapy of inherited thrombocytopenias (IT) would have been unthinkable because the therapeutic armamentarium was extremely limited and no clinical studies were available to evaluate the efficacy of different treatments. On the other hand, very few forms of IT were known and, consequently, the series of patients with these diseases were also very small. In the last 20 years, the number of well-defined forms has rapidly increased, passing from a handful of diseases to almost 50 different forms. In addition, case series with reports on hundreds of patients have appeared in the literature. It was, therefore, possible to learn more about these diseases and identify new treatments. These advances now allow me to write a review entirely dedicated to IT. Being a niche topic, I think it is useful to devote a few lines to what is the current vision of IT. In addition, I believe it useful to include a table reporting the essential characteristics of the diseases mentioned in the paper (Table 1). For a detailed description of IT, the reader is invited to refer to the exhaustive reviews that have recently appeared in the literature.1-3

The changing view of inherited thrombocytopenias Until twenty years ago, IT were considered extremely rare, and were almost always characterized by a severe hemorrhagic diathesis. In 2012, their prevalence was estimated to be around 2.7 in 100,000, similar to that of severe hemophilia A and myelofibrosis with myeloid metaplasia. In the last 10 years, many new forms of IT have been discovered, and therefore the rate calculated 10 years ago very probably underestimates the real prevalence of these diseases. IT are, therefore, although rare, not so rare as originally thought. Today, we also know that most of the affected subjects have a moderate, mild or even absent spontaneous bleeding diathesis, although bleeding may occur during hemostatic challenges, such as surgery, childbirth, or taking medications that further hinder hemostasis. From this point of view, the clinical phenotype is, therefore, less serious than previously thought. However, an unpleasant surprise has emerged from the analysis of large case series: many IT predispose to the development of additional diseases, including hematological malignancies, bone marrow aplasia, and severe non-hematological illnesses, which strongly influences the prognosis of af-

Haematologica | 107 - June 2022

1278

REVIEW SERIES - Treatment of inherited thrombocytopenias

C.L. Balduini

Table 1. Essential features of the inherited thrombocytopenias mentioned in this review.

Disease (abbreviation, OMIM entry)

Freq.1

Gene

Peculiar features

FORMS WITH ONLY THROMBOCYTOPENIA GP1BA GP1BB GP9

Severely defective platelet function.

NBEAL2

Defective platelet function. Development of progressive bone marrow fibrosis and splenomegaly. Platelet count decreases over time.

ITGA2B/ITGB3-related thrombocytopenia (ITGA2B/ITGB3-RT, 187800)

ITGA2B ITGB3

Defective platelet function.

SLFN14-related thrombocytopenia (SLFN14-RT, 616913)

SLFN14

Defective platelet function.

FLI1-related thrombocytopenia

FLI1

Defective platelet function.

Bernard-Soulier syndrome, biallelic form (bBSS, 231200)

++++

Bernard-Soulier syndrome, monoallelic form (mBSS, 153670)

+++

Gray platelet syndrome (GPS, 139090)

SYNDROMIC FORMS Jacobsen syndrome (JBS, 147791), Paris-Trousseau thrombocytopenia (TCPT, 188025) Thrombocytopenia with absent radii (TAR, 274000)

++++ +++

Wiskott-Aldrich syndrome (WAS, 301000)

++++

Deletions in 11q23

Physical growth delay, intellectual disability and various malformations. Defective platelet function.

RBM8A

Bilateral radial aplasia +/- other upper and lower limb bone abnormalities. Platelet count spontaneously increases over time.

WAS

Severe immunodeficiency leading to early death. Eczema. Increased risk of malignancies and autoimmunity. Mild immunodeficiency. Mild and transient eczema. Increased risk of malignancies and autoimmunity.

X-linked thrombocytopenia (XLT or THC1, 313900)

FORMS PREDISPOSING TO ADDITIONAL DISEASES MYH9-related disease (MYH9-RD, 155100)

++++

MYH9

Most patients develop extra-hematological manifestations, i.e., sensorineural deafness, nephropathy evolving into kidney failure, and/or cataracts.

ANKRD26-related thrombocytopenia (ANKRD26-RT or THC2, 188000)

+++

ANKRD26

Propensity to acquire myeloid malignancies (about 10% of reported patients).

RUNX1

Propensity to acquire myeloid malignancies (over 40% of reported patients). Increased risk of T-cell acute lymphoblastic leukemia. Defective platelet function.

Familial platelet disorder with propensity to acute myelogenous leukemia (FPD-AML, 601399)

+++

ETV6-related thrombocytopenia (ETV6-RT or THC5, 616216)

ETV6

Propensity to acquire hematological malignancies (about 30% of reported patients), especially childhood B-cell acute lymphoblastic leukemia.

Congenital amegakaryocytic thrombocytopenia (CAMT, 604498)

+++

MPL

Evolution to bone marrow aplasia during infancy or childhood.

Congenital amegakaryocytic thrombocytopenia variant due to biallelic THPO mutation (na, na)

THPO

Evolution to bone marrow aplasia during infancy or childhood.

MECOM

Evolution to bone marrow aplasia during infancy or childhood. Bilateral radioulnar synostosis is frequent. Possible other skeletal anomalies, cardiac and/or renal malformations, and/or sensorineural deafness.

MECOM-associated syndrome, including radioulnar synostosis with amegakaryocytic thrombocytopenia 2 (RUSAT2, 616738).

Continued on following page. Haematologica | 107 - June 2022

1279

REVIEW SERIES - Treatment of inherited thrombocytopenias

C.L. Balduini

Radioulnar synostosis with amegakaryocytic thrombocytopenia 1 (RUSAT1, 605432)

HOXA11

Possible evolution to marrow aplasia during infancy or childhood. Bilateral radioulnar synostosis and other skeletal abnormalities and/or sensorineural deafness

GNE-related disorder (GNE-RD, na)

GNE

Biallelic mutations can cause thrombocytopenia and late-onset muscle disorder while monoallelic mutations cause sialuria

DIAPH1

Patients develop progressive sensorineural deafness during infancy or childhood. Recurrent mild neutropenia in some cases.

DIAPH1-related disorder (DIAPH1-RD, 124900)

The presence of clinically relevant defect of platelet function is in red. In these cases the bleeding risk is higher than predicted by the platelet count and patient management must take into account this evidence. 1Freq.: frequency; +: less than 10 reported families; ++: more than 10 reported families; +++: more than 50 reported families; ++++: more than 200 reported families.

fected subjects (Table 1). Altogether, the prevalence of IT predisposing to other diseases is near 50% of the known forms (Figure 1). The most feared risk for subjects with IT is, therefore, no longer that of bleeding, but of developing potentially fatal diseases (Figure 2). This has become even truer in recent years due to the identification of treatments capable of increasing the number of platelets and/or reducing the bleeding risk in many IT (Table 2).

Treatments to stop hemorrhages Bleeding from accessible sites Localized measures are often successful for mucocutaneous hemorrhages.4,5 Nasal packing and/or endoscopic cauterization may be effective for stopping epistaxis. Suturing often stops hemorrhages from accidental or surgical wounds (e.g., bleeding after tooth extraction). Compression and application of gelatin sponges or gauzes soaked in tranexamic acid can help stop bleeding from superficial wounds. Mouthwash with tranexamic acid may be useful for gum bleeding.

Platelet transfusion Platelet transfusions are the most effective therapeutic intervention to stop bleeding in people with IT. However, the risk of acute reactions, transmissions of infectious agents, transfusion-associated graft-versus-host disease (TA-GvHD) and alloimmunization with consequent refractoriness to subsequent platelet transfusions is intrinsic to this treatment.4,6 The formation of antibodies against donor HLA antigens is a highly detrimental event in IT patients because most of them are destined to remain thrombocytopenic for life, and alloimmunization could compromise the efficacy of future platelet support and put their lives at risk. The use of platelet transfusions should, therefore, be limited to bleeding that cannot be controlled with localized measures, to life-threatening hemorrhages, or to bleeding at critical sites. When platelet transfusion is essential, the risk of alloimmunization can be significantly reduced by two approaches: 1) leukoreduced platelet concentrates; and 2) HLA-matched donors.7 Although clinical studies in patients with IT are not available, experience in other clinical conditions provides useful indications.

Figure 1. Inherited thrombocytopenias (IT) predisposing to other disorders. Based on personal experience (303 consecutive families with IT), 45% of patients with known IT have one of the forms that predispose to additional disorders, including glomerulonephritis, bone marrow aplasia, leukemia and myelodysplastic syndromes. Recognizing patients with the IT highlighted in the figure is important because they need a specific follow-up and can benefit from effective treatments. The prevalence of the forms of IT mentioned in this manuscript are given in Table 1. Haematologica | 107 - June 2022

1280

REVIEW SERIES - Treatment of inherited thrombocytopenias A clinical study performed many years ago showed that leukoreduction decreased alloimmunization from 45% to 17%, and refractoriness from 13% to 3%.8 A more recent study demonstrated that even better results were obtained by the combined use of ABO-identical transfusion (platelets expressing ABO antigens) and leukoreduction.9 The advantage of using HLA-match donors to prevent alloimmunization in this system is obvious; but this is not practically feasible in the case of unforeseen serious bleeding. In the case of patients who have already developed alloantibodies, crossmatch-compatible platelet units should be used.10 Although rare, TA-GvHD is an important complication of platelet transfusion because it is almost always fatal. TA-GvHD develops: 1) when there are differences in histocompatibility between recipient and donor; 2) in the presence of immunocompetent T cells in the blood component; and 3) when the recipient is unable to reject the immunocompetent cells.11 In practice, newborns and children in the first year of life, patients with immunodeficiency and after bone marrow transplantation, as well as those receiving HLA-matched platelets or products from first- or second-degree relatives are at risk of this complication. Since g or X irradiation results in the inactivation of T lymphocytes, irradiated platelet concentrates should be administered to these categories of patients. More recent studies suggested that some techniques used for the reduction of pathogens in platelet

C.L. Balduini

Figure 2. Risk of death in hereditary thrombocytopenias (IT). The collection of large series of patients with IT revealed that the majority of patients have modest to no bleeding diathesis. In addition, available therapies are often able to reduce the risk of bleeding in those with severe forms of IT. The greatest risk is, therefore, that deriving from the possible development of further diseases to which some IT predispose; BM: bone marrow.

concentrates achieve similar, or even better, results.12 Patients with the classical, biallelic form of Bernard Soulier syndrome (bBSS) who completely lack the platelet GPIb/IX/V complex often develop isoimmunization against this complex. In this case, immunosuppression and/or plasmapheresis may restore the efficacy of platelet transfusions.13

Table 2. Summary of the most relevant treatments for inherited thrombocytopenias.

Indications

Comments

Platelet transfusions

Leukoreduction of platelet concentrates All inherited thrombocytopenias. To stop and HLA-matched donors reduce the risk bleedings when local measures failed. of alloimmunization and refractoriness to To prepare patients for surgery platelet transfusion

Splenectomy

-Wiskott–Aldrich syndrome -X-linked thrombocytopenia

Increases platelet count but also the already high risk of infections

Preparation for hemostatic challenges of patients with: - MYH9-related disease - Wiskott–Aldrich syndrome/X-linked thrombocytopenia - monoallelic Bernard-Soulier syndrome - ANKRD26-related thrombocytopenia

Efficacy in other conditions to be tested The efficacy and safety of long-term treatments (life-long?) remain to be demonstrated

TPO-receptor agonists

Variant of congenital amegakaryocytic Restore entire hemopoiesis thrombocytopenia (THPO mutation)

Hematopoietic stem cell transplantation

-Wiskott–Aldrich syndrome -Congenital amegakaryocytic thrombocytopenia (MPL mutation) -Severe Bernard-Soulier syndrome -MECOM-associated syndrome

Gene therapy

-Wiskott–Aldrich syndrome Haematologica | 107 - June 2022

1281

Can cure patients and is the first-line treatment for patients with poor prognosis

Can cure patients. Efficacy in other conditions not yet tested

REVIEW SERIES - Treatment of inherited thrombocytopenias Other treatments for stopping hemorrhages As for the preparation for hemostatic challenges (see below), the efficacy of systemic administration of tranexamic acid in helping to stop bleeding in the general population has been demonstrated in numerous studies.14 Even without clear evidence of its effectiveness in IT, many authors use this drug for stopping bleeding also in these conditions.4,5,15 Particular caution should be used in the case of hematuria because clot formation in the urinary tract has been reported in some patients.16 Also recombinant activated factor VII (rFVIIa) has been used successfully to stop bleeding in a few patients with bBSS.17,18 Due to the potential severe side effect of thromboembolic events,19 this drug should be considered in IT only if platelet transfusions have not proved effective.

Treatments to prevent bleeding As already discussed, most patients with IT have mild, moderate, or even absent spontaneous bleeding, and intervention is required only on the occasion of hemostatic challenges. Simple and cheap general measures can effectively reduce the risk of bleeding in everyday life Preventing hemostatic challenges is an important general recommendation to be given to patients with IT. They need to know that many drugs inhibit platelet function and facilitate bleeding. Aspirin and non-steroidal anti-inflammatory agents should certainly be avoided, but many other frequently used drugs, e.g., some antibiotics and antidepressants, can also affect platelet function.20 Patients with IT should, therefore, consult their physician before taking any medication and discuss the risk-to-benefit ratio of each treatment. Oral contraceptives are usually effective in preventing or controlling menorrhagia and their use should be considered in women with heavy menstruation.21 In case they develop iron deficiency anemia, this must be corrected with iron administration, not only to improve their quality of life, but also because it has been suggested that anemia facilitates bleeding. In fact, in vitro studies showed that red blood cells push the platelets towards the vessel wall and facilitate their hemostatic effect.22 Dental interventions are among the most frequent hemostatic challenges, and maintaining good oral hygiene with regular dental visits are simple measures to prevent the need for these invasive procedures. Moreover, adherence to screening programs for early cancer detection is particularly recommended, as thrombocytopenia can be a major obstacle to the treatment of advanced malignancies. Finally, activities such as contact sports should be discouraged in patients with severe thrombocytopenia

C.L. Balduini

(<20x109 platelets/L) or with bleeding episodes following even minor trauma. Fortunately, this is only rarely required. Prophylactic treatments are effective in reducing the risk of bleeding on the occasion of hemostatic challenges When a hemostatic challenge can be programmed, the need for therapeutic measures to improve hemostasis must be carefully evaluated. Two large retrospective clinical studies provided useful information on the bleeding risk of patients with IT on the occasion of two common hemostatic challenges: childbirth and surgery. The first study analyzed the course of pregnancy and the outcome of childbirth in 181 women with 13 different forms of IT who had 339 pregnancies.23 There was no difference in gestation to that of healthy subjects in terms of miscarriages, fetal bleeding, and pre-term births, while the frequency of delivery-related maternal bleeding was increased. Of note, 46 women received spinal or epidural anesthesia without bleeding complications. No significant differences were found between vaginal and cesarean delivery in terms of maternal bleeding. Regarding the risk of bleeding in the newborn, fatal cerebral hemorrhages were observed in only two infants, both born by vaginal delivery to two mothers with MYH9-related disease (MYH9-RD): one infant had MYH9-RD and was severely thrombocytopenic, the other was not tested. The risk of hemorrhage for newborns therefore appears to be small, but the low number of observed events does not allow us to recommend a preferred method for delivery. However, it is worth mentioning here an old study that enrolled 162 pregnant women with immune thrombocytopenia (ITP).24 No intracranial hemorrhage was observed in 71 newborns (at risk of maternal antibody-induced thrombocytopenia) delivered by cesarean section, while intracranial hemorrhages were reported in 2 of 17 newborns after vaginal delivery. Both the IT and ITP studies therefore seem to suggest that cesarean delivery is safer than vaginal delivery when the newborn is at risk of being thrombocytopenic. While awaiting further data to provide statistically significant evidence, my personal suggestion is to take advantage of the observation that the number of platelets of newborns with IT is similar to that of the mother who transmitted the disease.25 Therefore, if the mother has severe thrombocytopenia, I suggest a cesarean section; in other cases, the type of delivery will be suggested by obstetric considerations. The search for parameters predicting delivery-related bleeding in the mother suggested that hemorrhages requiring blood transfusion were more frequent in women with a history of severe bleedings before pregnancy and with platelet count <50x109/L. A surprising finding was the similar incidence of bleeding at delivery in those mothers receiving and those not receiving prophylactic platelet transfusions. However, platelet count was lower in women

Haematologica | 107 - June 2022

1282

REVIEW SERIES - Treatment of inherited thrombocytopenias given transfusions, and this suggests that prophylactic platelet infusions were effective in reducing the frequency of hemorrhages. The second study analyzed 829 surgical procedures carried out in 423 patients with well-defined forms of inherited platelet disorders, including 185 subjects with IT.15 Also in this context, bleeding was more frequent in IT than in healthy subjects, and a platelet count <68x109/L, as well as a history of previous hemorrhages, predicted these events. The presence of functional platelet defects associated with thrombocytopenia (Table 1) also had a negative prognostic value. As in the study on pregnancy, platelet transfusions were given more frequently to patients with severe thrombocytopenia, and this is probably the reason why bleeding was similar in those subjects receiving and those not receiving this prophylactic treatment. Of note, this study also suggested that prophylactic treatments other than platelet transfusions were effective in reducing bleeding at surgery (see below). Recommended use of platelet transfusions

The recommendations for the safe level of platelets on the occasion of hemostatic challenges derive from retrospective and observational studies, expert opinion or clinical practice reviews, and are, therefore, based on low-quality evidence. Furthermore, the only studies conducted in patients with IT are the two described above on pregnancy23 and surgery15; however, even these are retrospective and therefore do not provide any concrete evidence. Having said that, prophylactic platelet transfusion is recommended in preparation for surgery and childbirth in IT patients with <70 and <50x109 platelets/L, respectively. For surgeries where bleeding can have more serious consequences (e.g., neurosurgery or posterior eye surgery), it is reasonable to achieve higher platelet counts (at least 100x109/L).25,26 Higher platelet counts are also required for spinal epidural anesthesia (<70x109/L) because of the theoretical risk of hematoma formation and neurological damage.27 On the occasion of scheduled hemostatic challenges, a platelet count much higher than 50x109/L may be indicated also in subjects with platelet functional defects associated with thrombocytopenia, such as bBSS, gray platelet syndrome (GPS), ITGA2B/ITGB3-related thrombocytopenia (ITGA2B/ITGB3-RT), SLFN14-related thrombocytopenia, Jacobsen syndrome/Paris-Trousseau thrombocytopenia, FLI1-related thrombocytopenia, and familial platelet disorder with propensity to acute myelogenous leukemia (FPD/AML) (Table 1). The recommendations already provided for platelet concentrates to stop bleeding also apply to preparing patients for hemostatic challenges that can be programmed. In this case, however, there is time to search for an HLA donor who offers the greatest compatibility possible to reduce the risk of alloimmunization.

C.L. Balduini

Prophylactic administration of platelets is not required for the many IT patients who have platelet counts above those levels required for the invasive procedure, although it is prudent to ensure the immediate availability of platelet concentrates should unexpected excessive bleeding occur. Platelet count is not the only parameter to consider when deciding whether to administer platelet transfusions; the patient's medical history also plays an important role. In fact, subjects who had severe bleeding in the past are more at risk of hemorrhagic complications at the programmed hemorrhagic challenge than those without.15,23 If a patient has a safe platelet level but had a major bleeding event in the past, they should be evaluated further to identify any other condition that may explain this discrepancy. Short-term treatment with thrombopoietin receptor agonists is an attractive alternative to platelet transfusion in preparation to surgery

Two thrombopoietin receptor agonists (TPO-RA) have been used in IT: eltrombopag and romiplostim. Eltrombopag is a small, non-peptide molecule given orally (on an empty stomach), while romiplstim is a recombinant protein requiring subcutaneous injection. Both bind to the TPO receptor of megakaryocytes (although the binding site is different for the two drugs) and stimulate platelet production. Eltrombopag also acts as a calcium chelator, but the clinical results of this action are still little known.28 Eltrombopag is currently approved for ITP, HCV-related thrombocytopenia, and aplastic anemia,29 but recent studies showed that it was able to increase platelet count in all the forms of IT in which it was tested. In two, small, non-randomized clinical trials,30,31 36 patients with MYH9RD, ANKRD26-related thrombocytopenia (ANKRD26-RT), X-linked thrombocytopenia/Wiskott-Aldrich syndrome (XLT/WAS), and monoalleic BSS (mBSS) received a from 3- to 6-week course of eltrombopag. Treatment was well tolerated and the vast majority of patients achieved a platelet count higher than the values required for safe surgery (Figure 3). Based on these results, a monocentric, prospective study evaluated the ability of eltrombopag to replace platelet transfusion in preparation for 11 surgeries of 5 consecutive patients with severe MYH9-RD and high bleeding risk.32 A safe platelet count was obtained in 10 of 11 cases, and patients underwent invasive procedures without the need for platelet support and without any bleeding complications. Of note, eltrombopag caused a durable increase in platelet count throughout the perioperative period and allowed patients to receive standard antithrombotic prophylaxis. Finally, eltrombopag was effective in preparing for hip arthroplasty a patient with DIAPH1-related disorder (DIAPH1-RD) and severely reduced platelet count who had developed multiple anti-

Haematologica | 107 - June 2022

1283

REVIEW SERIES - Treatment of inherited thrombocytopenias HLA immunoglobulin G alloantibodies and platelet refractoriness because of previous platelet transfusions.33 Experience in IT with romiplostim (a drug currently approved for ITP) is more limited and relates to its efficacy in 67 patients with WAS >15 years of age.34 Patients received the drug at a dose of 9 mg/kg weekly for at least four weeks and 60% of them had an increase in platelet count. In particular, 33% of patients had a complete response, and platelet count increased from a median of 30x109/L pretreatment to a peak median value of 247x109 /L by the third week. Twenty-seven percent achieved partial responses, with the average platelet count rising from 17x109 /L to 73x109 /L by the second week. Even if none of these patients underwent surgery, the platelet values achieved would have been sufficient for the most common surgical interventions. Due to the high cost of clinical trial registration and the small number of IT patients who could use this drug, eltrombopag and romiplostim have not yet been approved for any IT, and future approval remains unlikely. Nevertheless, it seems reasonable to propose that short-term administration of TPO-RA is regarded as the first-line option to cover elective surgery in those IT where it has been shown to be effective. Moreover, the use of TPO-RA should be considered in other forms of IT without a clinically relevant defect in platelet function (Table 1) and with no contraindication to this drug. TPO-RA are not expected to work in congenital amegakaryocytic thrombocytopenia

Figure 3. First clinical study with eltrombopag in inherited thrombocytopenias. This figure shows the effect of eltrombopag in 12 patients with MYH9-related disease. Most of them had a clinically relevant increase in platelet count after a few weeks of treatment. (From Pecci et al.30 with permission.)

C.L. Balduini

(CAMT) because the TPO receptor (MPL) is defective in this condition. As for eltrombopag, the drug for which there is more experience in IT, I suggest that patients start treatment with 50 mg/day three weeks before the scheduled surgery (Figure 4) and, depending on platelet count, continue this treatment for 3-7 days after the intervention. If the platelet count does not rise sufficiently after two weeks, eltrombopag can be increased to 75 mg/day. If the desired platelet count has still not been achieved, the patient should receive alternative prophylactic treatment. Does short-term treatment with thrombopoietin receptor agonists or thrombopoietin represent an alternative to platelet transfusion in preparation for childbirth?

Since TPO-RA pass the placental barrier, the possible harmful effects on the fetus have so far discouraged their use in pregnant women with thrombocytopenia. However, a recent, small, retrospective study assessed the safety and efficacy of eltrombopag and romiplostim in 15 pregnant women (17 pregnancies) with chronic ITP who had not responded to 27 previous treatment lines.35 In 58% of pregnancies, TPO-RA were given in the third trimester in preparation for delivery. Most patients responded to treatment and neither the mothers nor the neonates had serious complications, except for transient thrombocytosis in one neonate. Case reports and small series have also not reported malformation in newborns of ITP mothers who had taken TPO-RA. Instead, low birth-weight infants were sometimes observed.36 Concerning IT, the only evidence from literature is that eltrombopag was successfully used to prepare a woman with MYH9-RD for cesarean delivery, without consequences for the newborn.37 Available data are, therefore, not yet sufficient to recommend the use of TPO-RA before delivery for women with acquired or inherited thrombocytopenia, even if the results reported so far seem promising in terms of safety for both the mother and the fetus. While eltrombopag and romiplostim pass through the placenta, the full-length recombinant human thrombopoietin (rhTPO), approved by the China State Food and Drug Administration for the treatment of refractory chronic ITP, is not expected to do so. This should eliminate doubts about its possible toxicity to the fetus. A prospective study in which 31 pregnant women with ITP received rhTPO in the second or third trimester of gestation demonstrated the absence of any congenital disease or developmental delays of newborns during a median follow-up of 53 weeks.38 Furthermore, rhTPO was well tolerated by the mothers and increased their platelet count. Results were, therefore, good, but the study was too small in size to be able to state that rhTPO can be safely used in pregnant women. Furthermore, the drug is only available in China. In conclusion, TPO-RA should not be routinely used as a

Haematologica | 107 - June 2022

1284

REVIEW SERIES - Treatment of inherited thrombocytopenias

C.L. Balduini

Figure 4. Eltrombopag for preparation for surgery in IT patients. Thrombopoietin receptor agonists is an advantageous alternative to platelet transfusions in the preparation for surgery of patients with forms of IT who respond to this treatment. The figure shows the trend of the number of platelets in one representative patient reported in the study by Zaninetti et al.31 who used eltrombopag to prepare for 11 surgical interventions. This was an adult subject with severe MYH9-related disease who received eltrombopag in preparation for orthopedic surgery and a rapid normalization of platelet count was achieved. This avoided excessive bleeding during surgery and safe administration of antithrombotic prophylaxis. (From Zaninetti et al.31 with permission.) LMWH: low molecular weight heparin.

substitute for platelet transfusions in preparation for de- surgery-related bleeding also in some IT.44,45 Furthermore, livery outside of clinical trials. In case of refractoriness to a retrospective study (see above) disclosed that prophyplatelet transfusions, however, their utilization should be lactic administration of DDAVP was associated with a considered. lower incidence of surgical bleeding in subjects with inherited functional platelet defect isolated or associated Other hemostatic treatments in preparation for hemostatic with thrombocytopenia.15 When considering the use of challenges DDAVP, it should be borne in mind that it rarely results in Only one study has investigated the efficacy of prophylactic hypotension, tachycardia, fluid retention, or hyponatremia, treatments other than platelet transfusions in preparation and is contraindicated in ischemic heart disease or conof IT patients for hemostatic challenges, and concluded gestive heart failure, infancy, and pregnancy. that antifibrinolytic agents were associated with a lower rFVIIa promotes hemostasis by activating the extrinsic post-surgical bleeding frequency.15 However, this is low- pathway of the coagulation cascade, and is approved in quality evidence because the study was retrospective and hemophilia A or B with inhibitors, congenital factor VII few patients received this type of treatment. In this regard deficiency, acquired hemophilia, and Glanzmann’s thromit is worth mentioning that the efficacy of antifibrinolytics basthenia with refractoriness to platelet transfusions. A in IT is supported by a few case reports describing the suc- few thromboembolic events have been reported. rFVIIa cess of tranexamic acid in major surgery.39,40 Moreover, the has been successfully used to cover major or minor suruse of tranexamic acid in IT is made plausible by its docu- gery in a few patients with bBSS17 and in one subject with mented efficacy in the general population in preventing thrombocytopenia and absent radii (TAR)46 with an adminbleeding in a variety of major surgical procedures, especially istration schedule similar to that recommended for Glanzcardiac surgery and major orthopedic operations.14 Lastly, mann thrombasthenia.47 Given the unproven efficacy in IT the Network for the Advancement of Patient Blood Man- and possible side effects, rFVIIa should only be considered agement, Hemostasis and Thrombosis (NATA) recom- when other prophylactic treatments are not possible. mended intravenous administration of tranexamic acid for women at increased risk of postpartum hemorrhage.41 Since tranexamic acid is not burdened with significant side ef- Treatments to stably increase platelet fects (in particular, it does not appear to increase thrombotic risk),42 its prophylactic use in IT can be considered count in subjects with severe bleeding diathesis when proven treatments are not available. Another drug to be considered is desmopressin (1-deamino-8-D-arginine vasopressin [DDAVP]), which pro- Although rare, some patients have a severe bleeding motes the release of von Willebrand factor from diathesis that interferes with normal daily activities and endothelial cells and enhances the procoagulant activity significantly reduces their quality of life. In addition, the of platelets.43 Desmopressin is approved for mild hemo- repeated platelet transfusions required by their frequent philia A and type 1 von Willebrand disease, and a few case bleeding episodes can result in alloimmunization and rereports have suggested that it was effective in preventing fractoriness to platelet transfusion. It would, therefore, be Haematologica | 107 - June 2022

1285

REVIEW SERIES - Treatment of inherited thrombocytopenias desirable in these cases to avoid platelet transfusions by steadily increasing the platelet count. In very rare cases, splenectomy achieves this, while TPO-RA could, theoretically, be effective in a higher number of patients. Splenectomy Several patients with IT underwent splenectomy because they were misdiagnosed with ITP (for instance, personal experience indicates that 15% of subjects with MYH9-RD, the most frequent IT, received this treatment); but the removal of the spleen had no favorable effect. Notable exceptions are WAS and X-linked thrombocytopenia (XLT), a mild form of WAS. In these IT, splenectomy often increases platelet count up to normal levels and maintains levels stable, but unfortunately the procedure aggravates the immunodeficiency typical of these diseases. Splenectomy, therefore, reduces the high risk of dying from bleeding, but increases the already high risk of dying from infections. Concerning WAS, an old retrospective study of 42 patients concluded that the advantages outweigh the disadvantages, as splenectomy increased survival.48 A more recent retrospective study of 173 patients (41 splenectomized) concluded that the advantage of the increase in the number of platelets was counterbalanced by the increase in infections, which were especially frequent in patients not receiving antibiotic prophylaxis.49 The net result was that splenectomy did not modify patients’ survival. We now know that hematopoietic stem cell transplantation (HSCT) is the treatment of choice for WAS and one of the therapeutic options for XLT (see below). Moreover, gene therapy has been given successfully in a few cases (see below). Thus, splenectomy should be considered only when more effective treatments are not possible. Once splenectomy has been performed, lifelong antibiotic therapy is essential to reduce the risk of infections. Splenectomy should be avoided in patients who are candidates for HSCT, or who have already been treated with HSCT, as it increases the incidence of serious post-transplant infections.50 Life-long thrombopoietin receptor agonists: a credible option? Thrombopoietin receptor agonists are proving effective in preparing patients with some forms of IT for hemostatic challenges. Why not administer them as chronic treatment to patients with severe bleeding diathesis? Results of a few studies support the long-term effectiveness of TPO-RA in IT. In a prospective trial, 4 patients (2 with MYH9-RD, one with WAS, and one with ITGA2B/ITGB3RT) with mucosal hemorrhages WHO grade 2 or 3 received eltrombopag for 19-22 weeks and obtained a stable increase in platelet count and remission of bleeding symptoms.31 Two patients achieved these results with the dosage of 25 mg/day and the other 2 with 50 mg/day. The

C.L. Balduini

WAS patient experienced worsening of a pre-existing cutaneous eczema and the drug was prudently discontinued. No other adverse events were recorded. A retrospective study analyzed 8 patients with XLT/WAS and severe thrombocytopenia (6 children and 2 adults) who received eltrombopag for 22-209 weeks. Five patients achieved a platelet response and experienced significant improvement in bleeding symptoms. No major adverse events were observed.51 Moreover, a case report described a 2-year-old male with WAS and life-threatening bleeding episodes with secondary anemia who received eltrombopag for 32 weeks as a ‘bridge’ therapy to HSCT.52 With 5075 mg/day of eltrombopag, the number of platelets went from baseline values between 10 and 20x109/L to values around 30x109/L. This small increase was enough to greatly reduce bleeding and abolished the need for platelet transfusions. Treatment was well tolerated. Romiplostim has also been administered successfully for prolonged treatments in some cases of IT. A study already mentioned above retrospectively evaluated 67 young WAS patients receiving this drug for 1-12 months while they were waiting for HSCT.34 Most short-term responders (38/40) had a sustained response to romiplostim over several months and no clinically relevant side effects were observed. Romiplostim was also used to treat 5 children affected with the congenital amegakaryocytic thrombocytopenia (CAMT) variant caused by THPO gene mutations (a form of IT with TPO deficiency and propension to develop bone marrow aplasia).53,54 In most of these children, treatment was started when symptoms of bone marrow aplasia were already present (transfusion-dependent hyporegenerative anemia and/or neutropenia with frequent infections). In all cases, romiplostim greatly increased platelet count. Notably, it induced a trilineage hematological response with improvement not only of platelet count, but also of hemoglobin concentration and neutrophil count. The response was maintained throughout the follow-up period (from 13 months to 6.5 years) (Figure 5). Lastly, successful long-term use of romiplostim has been observed in a case of MYH9-RD who had previously been diagnosed as ITP.55 In conclusion, the few available data seem to indicate that TPO-RA maintain efficacy even for very prolonged treatments without causing relevant side effects. However, this does not exclude the possibility that the side effects that have been hypothesized in other diseases may occur in IT patients who have been receiving TPO-RA for many years. The disease for which there is more experience of prolonged administration of TPO-RA is ITP, and some studies have shown an increase in thrombotic events in treated patients while others did not. The most recent and largest meta-analysis, including both eltrombopag- and romiplostim-treated patients, looked at 11 trials with a total of 740 patients enrolled in the intervention group and 352 in the

Haematologica | 107 - June 2022

1286

REVIEW SERIES - Treatment of inherited thrombocytopenias control group.56 More thromboembolic events were noted in the TPO-RA group (n=25) than in the control group (n=4), but this difference did not reach statistical significance. Another hypothesized risk of TPO-RA is that of facilitating the onset of acute leukemia in subjects with other pathologies that in themselves predispose to hematological malignancies. This hypothesis is particularly alarming for subjects with IT, because three frequent forms predispose to acute leukemia. A large and recent meta-analysis investigated the results of 8 randomized studies including 707 patients with different forms of myelodysplastic syndrome (MDS), who are at high risk of leukemic transformation.57 The authors concluded that neither romiplostim nor eltrombopag promoted progression to leukemia (Risk Ratio: 1.08 and 1.12, respectively). The few studies in patients with IT, and the broader evidence obtained in ITP and MDS, seem to indicate that, overall, prolonged treatment with TPO-RA maintains efficacy for a very long time, do not expose the patient to the risk of leukemia, and appear to result in an insignificant increase in the risk of thrombosis. Obviously, these reassurances refer to the prolonged use of these drugs and not to their lifelong administration. In conclusion, lifelong TPO-RA administration is a credible and attractive option for most severe forms of IT, but the efficacy and safety of this approach remain unproven. Furthermore, TPO-RA are not approved for any form of IT, and they are expensive, and these are serious obstacles to their use in chronic treatments.

C.L. Balduini

Treatments to cure inherited thrombocytopenias All IT are potentially curable with HSCT or by replacing hematopoietic progenitors carrying the mutation with gene-corrected autologous cells (gene therapy). The single exception is the form of CAMT caused by loss-of-function mutations in the gene for TPO, which is produced primarily in the liver, and does not benefit from replacing the patient's hemopoietic progenitors. In just a few IT, however, the risk:benefit ratio is in favor of these treatments. Hematopoietic stem cell transplantation Hematopoietic stem cell transplantation is the treatment of choice in CAMT, WAS and MECOM-associated syndrome because they almost always lead to death if left untreated. A recent study analyzed data of 86 patients with CAMT receiving HSCT collected by the Center for International Blood and Marrow Transplant Research from 2000 to 2018.58 In most cases, transplant was from HLA matched or mismatched unrelated donors; the remaining were from HLA-matched sibling and HLA-mismatched relative. The predominant graft types were bone marrow and cord blood. The 5-year overall survival was 86%, with mortality and graft failure higher with HLA mismatched donor. The 5-year incidence of chronic graft-versus-host disease was 33%, and in the vast majority of cases it occurred after transplant from an unrelated donor. Better results in term of survival were obtained in patients transplanted before the age of three years and within 12 months of diagnosis.

Figure 5. Romiplostim in patients with a form of inherited thrombocytopenia due to mutation in the gene for thrombopoietin (congenital amegakaryocytic thrombocytopenia [CAMT] variant due to THPO mutation). The figure is from an article reporting an Egyptian family with 4 siblings suffering from this recently identified form of inherited thrombocytopenia. The patient described in the figure is the one with the most severe clinical manifestations. He was born with thrombocytopenia but later developed pancytopenia from bone marrow aplasia. The administration of small doses of romiplostim rapidly increased the values of platelets, neutrophils and hemoglobin. The infections and bleeding stopped, and the patient no longer needed red blood cell transfusions. Of note, the treatment was still effective and well tolerated after 7 years. (From Pecci et al.53 with permission.) ANC: absolute neutrophil count; HgB: hemoglobin; PLT: platelets. Haematologica | 107 - June 2022

1287

REVIEW SERIES - Treatment of inherited thrombocytopenias Thus, early transplant from an HLA-matched donor offers the best chance of survival; but even in the absence of these characteristics, HSCT is indicated in CAMT patients who are otherwise not destined to reach adulthood. Improved prophylactic antimicrobials and immunoglobulin supplementation increased the survival of patients with the severe syndrome induced by WAS mutations, but nevertheless they are doomed to die from infection or bleeding in early adulthood.59 In one old study, the 20-year probability of overall survival was 0% for patients who fail to express WAS protein, while it was 92% for those expressing normal-sized protein.60 Thus, HSTC (or gene therapy; see below) is mandatory in subjects with WAS mutations and the composite phenotype of severe thrombocytopenia and immunodeficiency with autoimmunity. A scoring system can help identify patients who need to undergo this procedure.59 It uses five parameters: thrombocytopenia, the severity of eczema, infections, development of autoimmunity, and malignancy. HSCT (or gene therapy; see below) is recommended with a score ≥3, which indicates a severe phenotype. HSCT outcome in WAS has improved over time, and what we can expect today from this procedure is exemplified by a recent study that has taken into consideration 197 patients transplanted at European Bone Marrow Transplantation centers between 2006 and 2017.61 After a median follow-up of 44.9 months, 176 patients were alive, with a 3-year overall survival of 88.7%, and chronic GvHD-free survival of 81.7%. Conditioning regimen and donor type were unrelated to overall survival, whereas age <5 years at HSCT was associated with a more favorable outcome. Similar results have been obtained in a previous study evaluating the outcome of HSCT in 129 patients transplanted at 29 Primary Immune Deficiency Treatment Consortium centers from 2005 through 2015 (Figure 6).62 Hematopoietic stem cell transplantation is indicated also in MECOM-associated syndrome (RUSAT2), a rare form with amegakaryocytic thrombocytopenia due to MECOM mutation variably associated with other somatic defects and evolving to bone marrow failure during infancy or childhood.63 Four different reports described the outcome of HSCT in a total of 20 patients (median age: 9.5 months); 76.5% had a good outcome while the remaining 23.5% died.64-67 Although the results are less positive than in CAMT and WAS, HSCT must be considered in this rare disease. Patients with radioulnar synostosis with amegakaryocytic thrombocytopenia 1 (RUSAT1) caused by HOXA11 mutation have a phenotype similar to RUSAT2, but evolution to bone marrow aplasia is less frequent and HSCT may not be required.68 In rare cases, HSCT can be considered in IT other than CAMT, WAS and MECOM-associated syndrome when the patient has a severe bleeding diathesis and does not respond to any treatments. This was the case of 3 patients

C.L. Balduini

with bBSS69,70 and one with thrombocytopenia with TAR71 who successfully received HSCT from HLA-identical siblings because they had life-threatening hemorrhages and had become refractory to platelet transfusions. In one subject with GPS and severe pancytopenia due to the development of myelofibrosis, HSCT corrected both the myelofibrosis and the platelet defect.72 Finally, HSCT in a young girl with GNE-related disorder (GNE-RD) and severe bleeding tendency resulted in a successful outcome.73 Gene therapy Autologous hematopoietic stem cell (HSC) gene therapy is an attractive alternative to HSCT, addressing as it does the needs of patients who lack appropriate donors. In the field of IT, it has only been used up to now in WAS, in which it promises to correct not only thrombocytopenia but also severe immunodeficiency. Moreover, it has the advantages over HSCT of bearing a low risk of rejection or GvHD and not requiring immunosuppression or fully myeloablative conditioning, which further increase the already high risk of infection. However, the first clinical trial yielded very alarming results. It used a g-retroviral vector in which the WAS gene was under the control of a strong viral promoter. Treatment corrected blood cell defects in most cases, but 7 of 9 evaluable patients developed leukemia as a consequence of insertions of the retroviral vector in proto-oncogenes and overexpression of these genes.74 Thereafter, a self-inactivating lentiviral vector in which the WAS complementary DNA is under the control of a human WAS promoter was used to address the insertional mutagenesis risk.75 A recent paper reported the outcome of this approach in 8 WAS children after a median follow-up of 7.6

Figure 6. Hematopoietic stem cell transplantation (HSCT) in Wiskott-Aldrich syndrome (WAS). Patients with WAS on average die in early adulthood from infections or bleeding. HSCT cures this disease in the vast majority of cases. The best results are obtained when the procedure is performed early. (From Burroughs et al.62 with permission.) OS: overall survival.

Haematologica | 107 - June 2022

1288

REVIEW SERIES - Treatment of inherited thrombocytopenias years.76 The gene-corrected cells were engrafted and were stable, and no serious treatment-related side effects occurred. However, one patient not analyzed in this study died 7 months after gene therapy from pre-existing complications of infection.77 Expression of WAS protein was increased and great benefit in terms of recurrent infections, autoimmunity and eczema were observed. In addition, bleeding episodes were less frequent and no patient required platelet transfusions. However, platelet count normalized in only 3 subjects, in 2 of them after splenectomy. Interestingly, an accurate investigation of patients' platelet counts concluded that gene therapy only partially corrected the platelet compartment because structure and function remained defective. Similar clinical benefit was described after a 3.5-year follow-up in another series of 8 children receiving the same therapeutic approach.78 The good results obtained by these two studies in children open up the possibility of treating adult patients for whom allogeneic HSCT would be associated with high risk; the successful lentiviral gene therapy in a 30-year-old patient with severe WAS supports this hypothesis.79

Management of the manifestations that add to thrombocytopenia

C.L. Balduini

of initial kidney damage appear, patients have to receive angiotensin converting enzyme inhibitors and/or angiotensin receptor blockers, because these drugs have been reported to have a beneficial effect.86,87 A retrospective analysis of 10 MYH9-RD patients revealed that cochlear implantation was highly effective in restoring hearing function in patients who developed severe deafness. Better results were obtained when this procedure was performed early.40,88 Standard surgery for cataracts should be carried out when indicated in MYH9-RD subjects who have developed this complication. In patients with IT associated with congenital limb deformities (TAR, RUSAT), reconstructive orthopedic surgery may allow children to resume most daily activities.46,89

Future directions Despite the great advances, there is still plenty of room for improving IT treatment. Genotype/phenotype correlation - the ability to accurately define the prognosis in each single patient would be especially beneficial in IT with a predisposition to other diseases. For example, knowing the risk of an individual patient with RUNX1, ANKRD26 or ETV6 mutations of developing hematological malignancies could allow HSCT to be performed before onset. This is expected to reduce the risks associated with the procedure. Platelet transfusion - the scarcity of donors and the risk of developing alloimmunization limit the use of this important therapeutic support. In vitro production of poorly immunogenic human platelets and development of ‘artificial’ platelets could solve the problem. Platelets obtained in vitro from induced pluripotent stem cell (iPSC) have already been successfully used in one patient with aplastic anemia90 and this gives rise to the hope that genetic engineering can be used to create universal HLA class-I depleted platelets with very low immunogenicity. On the other hand, 'synthetic platelet' nanoparticles have already been manufactured and have been proved capable of supporting hemostasis in animal models.91 TPO-RA - these drugs increased platelet count in some forms of IT, but they are probably effective in many others as well; large, collaborative clinical studies are required to test this hypothesis. Clinical trials to test the safety of these drugs in pregnancy are also highly desirable. Gene therapy - genetic engineering has been shown to be effective in WAS, but other severe forms of IT could benefit from this approach when HSCT is not possible.

When the diagnosis of an IT with predisposition to hematological malignancies (FPD-AML, ETV6-RT, or ANKRD26-RT) is made, blood cell count, examination of blood smear, and physical examination are recommended. Some authors advise a baseline bone marrow study to rule out malignancy and for future comparative use, after which, follow-up evaluations, including basic blood investigation, and physical examination, should be carried out every 6-12 months.80,81 In cases in which patients develop a hematological malignancy, they are treated as patients with de novo forms. When HSCT is considered, it is imperative to avoid choosing as the donor a relative carrying the familial germline mutation, as this results in a high risk of donorderived malignancies and/or poor engraftment. Thus, the potential related donors should be investigated by molecular testing because a normal platelet count does not exclude the possibility that they have the same mutation as the patient.82-84 As already discussed, HSCT is required for patients with bone marrow aplasia diagnosed with CAMT and MECOMassociated syndrome. It is important to remember that subjects with the rare variant of CAMT caused by THPO mutation do not respond to this treatment.53,54 Patients with MYH9-RD and mutations at high risk of developing glomerulonephritis85 should be monitored every Disclosures 6-12 months for the occurrence of proteinuria. Should signs No conflicts of interest to disclose

Haematologica | 107 - June 2022

1289

REVIEW SERIES - Treatment of inherited thrombocytopenias

C.L. Balduini

References 1. Nurden AT, Nurden P. Inherited thrombocytopenias: history, advances and perspectives. Haematologica. 2020;105(8):20042019. 2. Palma-Barqueros V, Revilla N, Sánchez A, et al. Inherited platelet disorders: an updated overview. Int J Mol Sci. 2021;22(9):4521. 3. Pecci A, Balduini CL. Inherited thrombocytopenias: an updated guide for clinicians. Blood Rev. 2021;48:100784. 4. Bolton-Maggs PH, Chalmers EA, Collins PW, et al. A review of inherited platelet disorders with guidelines for their management on behalf of the UKHCDO. Br J Haematol. 2006;135(5):603-633. 5. Kumar R, Kahr WH. Congenital thrombocytopenia: clinical manifestations, laboratory abnormalities, and molecular defects of a heterogeneous group of conditions. Hematol Oncol Clin North Am. 2013;27(3):465-494. 6. Dupuis A, Gachet C. Inherited platelet disorders: management of the bleeding risk. Transfus Clin Biol. 2018;25(3):228‐235. 7. Garraud O, Cognasse F, Tissot JD, et al. Improving platelet transfusion safety: biomedical and technical considerations. Blood Transfus. 2016;14(2):109-122. 8. Trial to Reduce Alloimmunization to Platelets Study Group. Leukocyte reduction and ultraviolet B irradiation of platelets to prevent alloimmunization and refractoriness to platelet transfusions. N Engl J Med. 1997;337(26):1861-1869. 9. Cardillo A, Hea JMl, Henrichs K, et al. Reducing the need for HLA-matched platelet transfusion. N Engl J Med. 2021;384:24512452. 10. Cohn CS. Platelet transfusion refractoriness: how do I diagnose and manage? Hematology Am Soc Hematol Educ Program. 2020;2020(1):527-532. 11. Luban NL. Prevention of transfusion-associated graft-versushost disease by inactivation of T cells in platelet components. Semin Hematol. 2001;38:34-45. 12. Cid J. Prevention of transfusion-associated graft-versus-host disease with pathogen-reduced platelets with amotosalen and ultraviolet A light: a review. Vox Sang. 2017;112(7):607-613. 13. Peaceman AM, Katz AR, Laville M. Bernard-Soulier syndrome complicating pregnancy: a case report. Obstet Gynecol. 1989;73(3 Pt 2):457‐459. 14. Franchini M, Mannucci PM. The never ending success story of tranexamic acid in acquired bleeding. Haematologica. 2020;105(5):1201-1205. 15. Orsini S, Noris P, Bury L et al. Bleeding risk of surgery and its prevention in patients with inherited platelet disorders. Haematologica. 2017 Jul;102(7):1192-1203. 16. Schultz M, Van Der Lelie H. Microscopic haematuria as a relative contraindication for tranexamic acid. Br J Haematol. 1995;89(3):663-664. 17. Tefre KL, Ingerslev J, Sørensen B. Clinical benefit of recombinant factor VIIa in management of bleeds and surgery in two brothers suffering from the Bernard-Soulier syndrome. Haemophilia. 2009;15(1):281-284. 18. Hacihanefioglu A, Tarkun P, Gonullu E. Use of recombinant factor VIIa in the management and prophylaxis of bleeding episodes in two patients with Bernard-Soulier syndrome. Thromb Res 2007;120(3):455-457. 19. Poon MC, d'Oiron R, Zotz RB, et al. The international, prospective Glanzmann Thrombasthenia Registry: treatment and outcomes in surgical intervention. Haematologica. 2015;100(8):1038-1044.

20. Scharf RE. Drugs that affect platelet function. Semin Thromb Hemost. 2012;38(08):865-883. 21. Ahuja SP, Hertweck SP. Overview of bleeding disorders in adolescent females with menorrhagia. J Pediatr Adolesc Gynecol. 2010;23(6 Suppl):S15-21. 22. Barshtein G, Ben-Ami R, Yedgar S. Role of red blood cell flow behavior in hemodynamics and hemostasis. Expert Rev Cardiovasc Ther. 2007;5(4):743-752. 23. Noris P, Schlegel N, Klersy C, et al. Analysis of 339 pregnancies in 181 women with 13 different forms of inherited thrombocytopenia. Haematologica. 2014;99(8):1387-1394. 24. Samuels P, Bussel JB, Braitman LE, et al. Estimation of the risk of thrombocytopenia in the offspring of pregnant women with presumed immune thrombocytopenic purpura. N Engl J Med. 1990;323(4):229-235. 25. Estcourt LJ, Birchall J, Allard S, et al. Guidelines for the use of platelet transfusions. Br J Haematol. 2017;176(3):365-394. 26. Kaufman RM, Djulbegovic B, Gernsheimer T, et al. Platelet transfusion: a clinical practice guideline from the AABB. Ann Intern Med. 2015;162(3):205‐213. 27. Bauer ME, Arendt K, Beilin Y, et al. The Society for Obstetric Anesthesia and Perinatology Interdisciplinary Consensus Statement on Neuraxial Procedures in Obstetric Patients with Thrombocytopenia. Anesth Analg. 2021;132(6):1531-1544. 28. Rodeghiero F, Pecci A, Balduini CL. Thrombopoietin receptor agonists in hereditary thrombocytopenias. J Thromb Haemost. 2018;16(9):1700-1710. 29. Ghanima W, Cooper N, Rodeghiero F, Godeau B, Bussel JB. Thrombopoietin receptor agonists: ten years later. Haematologica. 2019;104(6):1112‐1123. 30. Pecci A, Gresele P, Klersy C, et al. Eltrombopag for the treatment of the inherited thrombocytopenia deriving from MYH9 mutations. Blood. 2010;116(26):5832-5837. 31. Zaninetti C, Gresele P, Bertomoro A, et al. Eltrombopag for the treatment of inherited thrombocytopenias: a phase II clinical trial. Haematologica 2020;105(3):820-828. 32. Zaninetti C, Barozzi S, Bozzi V, Gresele P, Balduini CL, Pecci A. Eltrombopag in preparation for surgery in patients with severe MYH9-related thrombocytopenia. Am J Hematol. 2019;94(8):E199‐E201. 33. Westbury SK, Downes K, Burney C, et al. Phenotype description and response to thrombopoietin receptor agonist in DIAPH1related disorder. Blood Adv. 2018;2(18):2341-2346. 34. Khoreva A, Abramova I, Deripapa E, et al. Efficacy of romiplostim in treatment of thrombocytopenia in children with WiskottAldrich syndrome. Br J Haematol. 2021;192(2):366-374. 35. Michel M, Ruggeri M, Gonzalez-Lopez TJ, et al. Use of thrombopoietin receptor agonists for immune thrombocytopenia in pregnancy: results from a multicenter study. Blood. 2020;136(26):3056-3061. 36. Howaidi J, AlRajhi AM, Howaidi A, AlNajjar FH, Tailor IK. Use of thrombopoietin receptor agonists in pregnancy: a review of the literature. Hematol Oncol Stem Cell Ther. 2021 Jun16. [Epub ahead of print] 37. Favier R, De Carne C, Elefant E, et al. Eltrombopag to treat thrombocytopenia during last month of pregnancy in a woman with MYH9-related disease: a case report. A A Pract. 2018;10(1):10-12. 38. Kong Z, Qin P, Xiao S, et al. A novel recombinant human thrombopoietin therapy for the management of immune thrombocytopenia in pregnancy. Blood. 2017;130(9):1097-1103.

Haematologica | 107 - June 2022

1290

REVIEW SERIES - Treatment of inherited thrombocytopenias 39. Nurden P, Nurden A, Favier R, Gleyze M. Management of pregnancy for a patient with the new syndromic macrothrombocytopenia, DIAPH1-related disease. Platelets. 2018;29(7):737-738. 40. Pecci A, Verver EJ, Schlegel N, et al. Cochlear implantation is safe and effective in patients with MYH9 related disease. Orphanet J Rare Dis. 2014;9:100. 41. Muñoz M, Stensballe J, Ducloy-Bouthors AS. Patient blood management in obstetrics: prevention and treatment of postpartum haemorrhage - NATA consensus statement. Blood Transfus. 2019;17(2):112-136. 42. Franchini M, Mengoli C, Marietta M. Safety of intravenous tranexamic acid in patients undergoing major orthopaedic surgery: a meta-analysis of randomised controlled trials. Blood Transfus. 2018;16(1):36-43. 43. Colucci G, Stutz M, Rochat S, et al. The effect of desmopressin on platelet function: a selective enhancement of procoagulant COAT platelets in patients with primary platelet function defects. Blood. 2014;123(12):1905‐1916. 44. Sehbai AS, Abraham J, Brown VK. Perioperative management of a patient with May-Hegglin anomaly requiring craniotomy. Am J Hematol. 2005;79(4):303-308. 45. Matzdorff AC, White JG, Malzahn K, Greinacher A. Perioperative management of a patient with Fetchner syndrome. Ann Hematol. 2001;80(7):436-439. 46. Coppola A, Simone CD, Palmieri NM, et al. Recombinant activated factor VII for hemostatic cover of orthopaedic interventions in a girl with thrombocytopenia with absent radii syndrome. Blood Coagul Fibrinolysis. 2007;18(2):199-201. 47. Di Minno G, Zotz RB, d'Oiron R, et al. The international, prospective Glanzmann Thrombasthenia Registry: treatment modalities and outcomes of non-surgical bleeding episodes in patients with Glanzmann thrombasthenia. Haematologica. 2015;100(8):1031‐1037. 48. Mullen CA, Anderson KD, Blaese RM. Splenectomy and/or bone marrow transplantation in the management of the WiskottAldrich syndrome: long-term follow-up of 62 cases. Blood. 1993;82(10):2961-2966. 49. Albert MH, Bittner TC, Nonoyama S, et al. X-linked thrombocytopenia (XLT) due to WAS mutations: clinical characteristics, long-term outcome, and treatment options. Blood. 2010;115(16):3231-3238. 50. Ozsahin H, Cavazzana-Calvo M, Notarangelo LD, et al. Longterm outcome following hematopoietic stem-cell transplantation in Wiskott-Aldrich syndrome: collaborative study of the European Society for Immunodeficiencies and European Group for Blood and Marrow Transplantation. Blood. 2008;111(1):439-445. 51. Gerrits AJ, Leven EA, Frelinger AL 3rd, et al. Effects of eltrombopag on platelet count and platelet activation in Wiskott-Aldrich syndrome/X-linked thrombocytopenia. Blood. 2015;126(11):1367-1378. 52. Gabelli M, Marzollo A, Notarangelo LD, Basso G, Putti MC. Eltrombopag use in a patient with Wiskott-Aldrich syndrome. Pediatr Blood Cancer. 2017;64(12). 53. Pecci A, Ragab I, Bozzi V, et al. Thrombopoietin mutation in congenital amegakaryocytic thrombocytopenia treatable with romiplostim. EMBO Mol Med. 2018;10(1):63-75. 54. Seo A, Ben-Harosh M, Sirin M, et al. Bone marrow failure unresponsive to bone marrow transplant is caused by mutations in thrombopoietin. Blood 2017;130(7):875-880. 55. Rabbolini DJ, Chun Y, Latimer M, et al. Diagnosis and treatment of MYH9-RD in an Australasian cohort with thrombocytopenia. Platelets. 2018;29(8):793-800.

C.L. Balduini

56. Tjepkema M, Amini S, Schipperus M. Risk of thrombosis with thrombopoietin receptor agonists for ITP patients: a systematic review and meta-analysis. Crit Rev Oncol Hematol. 2022;171:103581. 57. Meng F, Chen X, Yu S, et al. Safety and efficacy of eltrombopag and romiplostim in myelodysplastic syndromes: a systematic review and meta-analysis. Front Oncol. 2020;10:582686. 58. Cancio M, Hebert K, Kim S, et al. Outcomes in hematopoietic stem cell transplantation for congenital amegakaryocytic thrombocytopenia. Transplant Cell Ther. 2022;28(2):101.e1-101.e6. 59. Imai K, Morio T, Zhu Y, et al. Clinical course of patients with WASP gene mutations. Blood. 2004;103(2):456-464. 60. Albert MH, Notarangelo LD, Ochs HD. Clinical spectrum, pathophysiology and treatment of the Wiskott-Aldrich syndrome. Curr Opin Hematol. 2011;18(1):42-48. 61. Albert MH, Slatter MA, Gennery AR, et al. Hematopoietic stem cell transplantation for Wiskott-Aldrich syndrome: an EBMT inborn errors working party analysis. Blood. 2022;139(13):20662079 62. Burroughs LM, Petrovic A, Brazauskas R, et al. Excellent outcomes following hematopoietic cell transplantation for Wiskott-Aldrich syndrome: a PIDTC report. Blood. 2020;135(23):2094-2105. 63. Germeshausen M, Ancliff P, Estrada J, et al. MECOM-associated syndrome: a heterogeneous inherited bone marrow failure syndrome with amegakaryocytic thrombocytopenia. Blood Adv. 2018;2(6):586-596. 64. Niihori T, Ouchi-Uchiyama M, Sasahara Y, et al. Mutations in MECOM, encoding oncoprotein EVI1, cause radioulnar synostosis with amegakaryocytic thrombocytopenia. Am J Hum Genet. 2015;97(6):848-854. 65. Lord SV, Jimenez JE, Kroeger ZA, et al. A MECOM variant in an African American child with radioulnar synostosis and thrombocytopenia. Clin Dysmorphol. 2018;27(1):9-11. 66. Bluteau O, Sebert M, Leblanc T, et al. A landscape of germ line mutations in a cohort of inherited bone marrow failure patients. Blood. 2018;131(7):717-732. 67. Ripperger T, Hofmann W, Koch JC, et al. MDS1 and EVI1 complex locus (MECOM): a novel candidate gene for hereditary hematological malignancies. Haematologica. 2018;103(2):e55e58. 68. Thompson AA, Nguyen LT. Amegakaryocytic thrombocytopenia and radio-ulnar synostosis are associated with HOXA11 mutation. Nat Genet. 2000;26(4):397-398. 69. Locatelli F, Rossi G, Balduini C. Hematopoietic stem-cell transplantation for the Bernard-Soulier syndrome. Ann Intern Med. 2003;138(1):79. 70. Rieger C, Rank A, Fiegl M, et al. Allogeneic stem cell transplantation as a new treatment option for patients with severe Bernard-Soulier Syndrome. Thromb Haemost 2006;95(1):190-191. 71. Brochstein JA, Shank B, Kernan NA, Terwilliger JW, O'Reilly RJ. Marrow transplantation for thrombocytopenia-absent radii syndrome. J Pediatr. 1992;121(4):587-589. 72. Favier R, Roussel X, Audia S, et al. Correction of severe myelofibrosis, impaired platelet functions and abnormalities in a patient with gray platelet syndrome successfully treated by stem cell transplantation. Platelets. 2020;31(4):536‐540. 73. Zieger B, Boeckelmann D, Anani W, et al. Novel GNE gene variants associated with severe congenital thrombocytopenia and platelet sialylation defect. Thromb Haemost. 2022 Jan 20. [Epub ahead of print] 74. Braun CJ, Boztug K, Paruzynski A, et al. Gene therapy for Wiskott-Aldrich syndrome-long-term efficacy and genotoxicity.

Haematologica | 107 - June 2022

1291

REVIEW SERIES - Treatment of inherited thrombocytopenias Sci Transl Med. 2014;6(227):227ra33. 75. Aiuti A, Biasco L, Scaramuzza S, et al. Lentiviral hematopoietic stem cell gene therapy in patients with Wiskott-Aldrich syndrome. Science. 2013;341(6148):1233151. 76. Magnani A, Semeraro M, Adam F, et al. Long-term safety and efficacy of lentiviral hematopoietic stem/progenitor cell gene therapy for Wiskott-Aldrich syndrome. Nat Med. 2022;28(1):7180. 77. Hacein-Bey Abina S, Gaspar HB, Blondeau J, et al. Outcomes following gene therapy in patients with severe Wiskott-Aldrich syndrome. JAMA. 2015;313(15):1550-1563. 78. Ferrua F, Cicalese MP, Galimberti S. Lentiviral haemopoietic stem/progenitor cell gene therapy for treatment of WiskottAldrich syndrome: interim results of a non-randomised, open-label, phase 1/2 clinical study. Lancet Haematol. 2019;6(5):e239-e253. 79. Morris EC, Fox T, Chakraverty R, et al. Gene therapy for WiskottAldrich syndrome in a severely affected adult. Blood. 2017;130(11):1327-1335. 80. Godley LA, Shimamura A. Genetic predisposition to hematologic malignancies: management and surveillance. Blood. 2017;130(4):424-432. 81. Churpek JE, Artz A, Bishop M, Liu H, Godley LA. Correspondence regarding the Consensus Statement from the Worldwide Network for Blood and Marrow Transplantation Standing Committee on Donor Issues. Biol Blood Marrow Transplant. 2016;22(1):183-184. 82. Buijs A, Poddighe P, van Wijk R, et al. A novel CBFA2 singlenucleotide mutation in familial platelet disorder with propensity to develop myeloid malignancies. Blood. 2001;98(9):2856-2858. 83. Churpek JE, Lorenz R, Nedumgottil S, et al. Proposal for the clinical detection and management of patients and their family members with familial myelodysplastic syndrome/acute

C.L. Balduini

leukemia predisposition syndromes. Leuk Lymphoma. 2013;54(1):28-35. 84. Marconi C, Canobbio I, Bozzi V, et al. 5'UTR point substitutions and N-terminal truncating mutations of ANKRD26 in acute myeloid leukemia. J Hematol Oncol. 2017;10(1):18. 85. Pecci A, Klersy C, Gresele P, et al. MYH9-related disease: a novel prognostic model to predict the clinical evolution of the disease based on genotype-phenotype correlations. Hum Mutat. 2014;35(2):236-247. 86. Pecci A, Granata A, Fiore CE, Balduini CL. Renin-angiotensin system blockade is effective in reducing proteinuria of patients with progressive nephropathy caused by MYH9 mutations (Fechtner-Epstein syndrome). Nephrol Dial Transplant. 2008;23(8):2690-2692. 87. Tanaka M, Miki S, Saita H, al. Renin-angiotensin system blockade therapy for early renal involvement in MYH9-related disease with an E1841K mutation. Intern Med. 2019;58(20):2983‐2988. 88. Canzi P, Pecci A, Manfrin M, et al. Severe to profound deafness may be associated with MYH9-related disease: report of 4 patients. Acta Otorhinolaryngol Ital. 2016;36(5):415-420. 89. Al Kaissi A, Girsch W, Kenis V, et al. Reconstruction of limb deformities in patients with thrombocytopenia-absent radius syndrome. Orthop Surg. 2015;7(1):50-56. 90. Sugimoto N, Kanda J, Nakamura S, et al. The first-in-human clinical trial of iPSC-derived platelets (iPLAT1): autologous transfusion to an aplastic anemia patient with alloimmune platelet transfusion refractoriness. Blood 2021;138 (Supplement 1):351. 91. Sekhon UDS, Swingle K, Girish A, et al. Platelet-mimicking procoagulant nanoparticles augment hemostasis in animal models of bleeding. Sci Transl Med. 2022;14(629):eabb8975.

Haematologica | 107 - June 2022

1292

ARTICLE - Acute Lymphoblastic Leukemia

Either IL-7 activation of JAK-STAT or BEZ inhibition of PI3K-AKT-mTOR pathways dominates the single-cell phosphosignature of ex vivo treated pediatric T-cell acute lymphoblastic leukemia cells Daniela Kuzilková,1,2* Cristina Bugarin,3* Katerina Rejlova,1,2 Axel R. Schulz,4 Henrik E. Mei,4 Maddalena Paganin,5 Alessandra Biffi,5 Andrea Biondi,3 Tomas Kalina1,2# and Giuseppe Gaipa3# Childhood Leukemia Investigation Prague, Prague, Czech Republic; Department of Pediatric Hematology and Oncology, 2nd Faculty of Medicine, Charles University Prague, Prague, Czech Republic; 3Fondazione Tettamanti, Clinica Pediatrica Università degli Studi Milano Bicocca, Monza (MB), Italy; 4German Rheumatism Research Center Berlin (DRFZ), a Leibniz Institute, Berlin, Germany and 5Pediatric Hematology, Oncology and Stem Cell Transplant Division, Women and Child Health Department, University of Padova, Padova, Italy. 1

Correspondence: Andrea Biondi andrea.biondi@unimib.it Received: March 17, 2021 Accepted: October 12, 2021 Prepublished: October 21, 2021 https://doi.org/10.3324/haematol.2021.278796 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

*DK and CB contributed equally as co-first authors. # TK and GG contributed equally as co-senior authors.

Abstract T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive cancer arising from lymphoblasts of T-cell origin. While TALL accounts for only 15% of childhood and 25% of adult ALL, 30% of patients relapse with a poor outcome. Targeted therapy of resistant and high-risk pediatric T-ALL is therefore urgently needed, together with precision medicine tools allowing the testing of efficacy in patient samples. Furthermore, leukemic cell heterogeneity requires drug response assessment at the single-cell level. Here we used single-cell mass cytometry to study signal transduction pathways such as JAK-STAT, PI3K-AKT-mTOR and MEK-ERK in 16 diagnostic and five relapsed T-ALL primary samples, and investigated the in vitro response of cells to Interleukin-7 (IL-7) and the inhibitor BEZ-235. T-ALL cells showed upregulated activity of the PI3K-AKT-mTOR and MEK-ERK pathways and increased expression of proliferation and translation markers. We found that perturbation induced by the ex vivo administration of either IL-7 or BEZ-235 reveals a high degree of exclusivity with respect to the phospho-protein responsiveness to these agents. Notably, these response signatures were maintained from diagnosis to relapse in individual patients. In conclusion, we demonstrated the power of mass cytometry single-cell profiling of signal transduction pathways in T-ALL. Taking advantage of this advanced approach, we were able to identify distinct clusters with different responsiveness to IL-7 and BEZ-235 that can persist at relapse. Collectively our observations can contribute to a better understanding of the complex signaling network governing T-ALL behavior and its correlation with influence on the response to therapy.

Introduction The incidence of acute lymphoblastic leukemia (ALL) in children peaks between 2 and 5 years of age, and approximately 15% of cases are of T-cell origin and lead to T-cell ALL (T-ALL). The survival rates of pediatric T-ALL patients have improved in recent years, approaching a 5-year event-free survival of 70-80%.1 However, 25% of children with T-ALL relapse with a generally more aggressive disease with a very poor outcome. Thus, efforts in the investigation of disease biology that might contribute to the

development of more effective and specific treatments are urgently needed.2,3 Glucocorticoids (GC) represent central components of T-ALL therapy, and the early response to GC-based therapy is an important predictor of longterm outcome.4 Accordingly, in Berlin-Frankfurt-Münsteroriented trials for childhood ALL, patients are stratified based on prednisone (PDN) response following 7 days of PDN monotherapy, and PDN poor responders (PPR) have a significantly worse outcome than PDN good responders (PGR),5 indicating that intrinsic differences in GC sensitivity at diagnosis impact the outcome. However, despite the

Haematologica | 107 - June 2022

1293

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL consolidated use of GC in ALL clinical management, a comprehensive understanding of the mechanisms underlying GC sensitivity is still lacking.6–8 The IL-7R-JAK-STAT5 signaling pathway plays a critical role in T-ALL, contributing to leukemogenesis in vivo9–11 or driving disease progression by regulating cell viability and proliferation.12–15 Somatic gain-of-function mutations in the specific IL-7Ra (CD127) subunit can occur in roughly 10-12% of pediatric and adult T-ALL cases, leading to constitutive activation of the receptor.16,17 Mutations may also occur in downstream effectors of IL-7/IL-7R-mediated signaling, such as JAK-STAT, PI3K-AKT-mTOR and RAS-MEK-ERK pathways which are found in a large proportion (30–50%) of T-ALL cases.18–21 IL-7R pathway mutations are enriched in subtypes of patients overexpressing HOXA and TLX17,18,22 or with early T-cell precursor ALL (ETP-ALL) phenotype.23,24 However mutations in IL-7R, JAK1 and JAK3 were also shown to associate with genetic lesions in WT1, PRC2 or PHF6 epigenetic regulators.25 Furthermore, the IL-7R-JAKSTAT5 pathway has been investigated for its role in GC resistance in T-ALL. Li et al.19 performed whole genome and targeted exome sequencing in patients with T-ALL and found that mutations in the IL-7R signaling components JAK1 and KRAS correlated with steroid resistance and poor outcome in a subset of T-ALL whereas Meyer et al.26 showed that IL-7 mediates an intrinsic and physiologic mechanism of GC resistance in normal thymocyte development that is retained during leukemogenesis in a subset of T-ALL. A large cohort of adult T-ALL cases studied by Kim and collaborators contained a subgroup of patients with IL-7R-pathway mutations with slow-response to remission induction regimen measured at day 8 and week 6.21 PI3K-AKT-mTOR is a major signaling pathway implicated in T-ALL malignant transformation promoting several functions including cell survival and proliferation.27 The major negative regulator of this pathway is the tumor suppressor lipid phosphatase and tensin homolog (PTEN), which is frequently inactivated in human cancer.28,29 PTEN non-sense or frame-shift mutations cluster in exon 7, resulting in a C-terminal truncated protein leading to a decreased or absent PTEN expression and activity30 resulting in PI3K-AKT aberrant activation. Gene expression profiling of T-ALL cases revealed subgroups of T-ALL, each characterized by a specific transcriptional profile and the ectopic expression of one particular transcription factor which could impact the differentiation process of the cells. The most represented subgroup shows ectopic TAL1 expression, whereas other major subgroups show mutual exclusive expression of TLX1, TLX3, HOXA9/10, LMO2, or NKX2-1.31–33 Furthermore, the ETP-ALL is characterized by aberrant expression of LYL12; hematopoietic transcription factors such as RUNX1

D. Kuzilková et al.

and ETV6 are frequently mutated in this genetically heterogeneous subgroup.24 In addition to the ectopic expression of transcription factors, which can represent an initiating step in T-ALL development, further mutations are required for the full leukemic transformation. These include mutations leading to dysregulation of the activity of PTEN, WT1, RAS, IL-7R and PHF6 affecting several signal transduction pathways implicated in the development of T-ALL.20 Relapsed disease represents one of the most challenging settings in the clinical management of T-ALL. Globally, recent studies have indicated that relapsed blasts can evade therapy by utilizing a variety of biological pathways.34 Despite the need for targeting molecules to treat resistant and highrisk pediatric T-ALL, in vitro drug testing may not be predictive of in vivo efficacy due to the complexity of the signaling network and due to the cell heterogeneity within the individual sample.35 Single-cell mass cytometry analysis represents a promising tool in the comprehension of signaling networks and their role in the response to drugs.36 Currently, mass cytometry allows the measurement of more than 40 parameters per single cell, rendering this approach an ideal method to assay drug candidate mechanisms of action and selectivity to cancer cells.37 Although many efforts have been made to characterize the genomic landscape of relapsed TALL, many aspects of leukemia resistance are still to be learned. Thus, the integration of genomic and epigenetic studies with new functional proteomic approaches at the single-cell level is essential to continue identifying potential drug targets or biomarkers. In this work, we aimed to evaluate the functional profiles of T-ALL cells collected from children with T-ALL at diagnosis or at relapse and correlate them with their biological and clinical features, including early response to therapy, by mass cytometry.

Methods Primary samples and diagnostic procedures This study was approved by the Institutional Review Board of San Gerardo Hospital (Monza, Italy), informed consent was obtained from all patients and their guardians in accordance with the Declaration of Helsinki. Bone marrow (BM) or peripheral blood (PB) mononuclear cells (MNC) were collected from 16 children with T-ALL enrolled in the AIEOP-ALL-2009 or R2006 protocols at the Pediatric Clinic of University Milano Bicocca at San Gerardo Hospital, stored and thawed as described previously.38 Immunophenotyping and evaluation of genetic aberrations were performed as described previously,38 and the T-ALL diagnosis was based on consensus criteria.39 We also collected PBMNC from three healthy subjects. The immunological classification as well as the clinical and genetic features are reported in Table 1.

Haematologica | 107 - June 2022

1294

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL In vitro prednisolone response assay In order to explore the functional involvement of individual or combined treatments on cell survival, we cultured 10 primary T-ALL samples in technical duplicate for 48 hours in complete medium in the presence of IL-7 (50 ng/mL, Peprotech), Methylprednisolon (50 ug/mL, Urbason, Sanofi), BEZ-235 (a dual PI3K/mTOR inhibitor, 800 nM) or without any treatment used as basal control. Cells were then stained with Annexin V/7-Aminoactinomycin D and absolute viable (and non-apoptotic) cell counts were performed by adding CountBright Absolute Counting Beads (Invitrogen) and calculated upon bead-based correction factor according to manufacturer´s instructions. Mass cytometry Sample preparation

An average of 0.5-1x106 cells per condition were starved in X-VIVO medium (37°C, 1 hour [h]), stimulated with IL-7 (50 ng/mL, 15 minutes [min]) or inhibited with BEZ-235 (800 nM, 30 min) at 37°C. Samples were fixed (MaxPar FIXI) and stored (10% glycerol in fetal bovine serum, -20°C, adapted from Watson et al.40). Sample staining

Monoclonal antibody (moAb) characteristics are listed in the Online Supplementary Table S1. In-house antibody conjugations with palladium, indium and platinum isotopes were performed as described previously.41,42 Two sets of antibody cocktails (surface moAb, intracellular moAb) were prepared and handled as described previously.43 The samples were thawed, washed with MaxPar Cell Staining Buffer (CSB, 800 g, 5 min, room temperature [RT]), and barcoding moAb were added (15 min, RT). The cells were washed, pooled together and washed in CSB (800 g, 5 min, RT). In order to minimize technical variability between stimulated and unstimulated samples, differently treated samples from particular donors were barcoded, mixed and further processed in one tube. In order to control for the batch effect, each patient sample was acquired together with barcoded healthy donor PBMNC as an internal spike-in control. The samples were stained according to the MaxPar Phosphoprotein Staining with Fresh Fix (Fluidigm) protocol as recommended by the manufacturer.

D. Kuzilková et al.

(v10.5, FlowJo LLC) or Infinicyt v2.0 (Cytognos) software (Online Supplementary Figure S1). Heatmaps were created in MeV software45 by displaying arcsinh-transformed absolute values of median intensities. The fold change was calculated by subtracting the arcsinh-transformed median of unstimulated samples from the arcsinh-transformed median of stimulated samples. Vaevictis dimensionality reduction For visual projection of our data, we used the deep learning-based tool Vaevictis (https://github.com/ stuchly/vaevictis/), which combines ideas adapted from Szubert et al.46 and Ding et al.47 with biased sampling. Using this approach, Vaevictis can produce reusable parametric mapping into the lower dimensional space to visualize both local and global relationships in very large datasets. Vaevictis dimensional reduction was calculated from a set of 13 parameters (p-4E-BP1, p-STAT5, p-AKT, p-p38, p-S6, p-LCK, p-CREB-1, p-ERK1/2, p-RB, Ki-67, PTEN, MYC, and BCL-2). Molecular characterization of T-cell acute lymphoblastic leukemia samples The somatic mutation analysis was performed on primary genomic T-ALL DNA BM samples. PTEN mutation analysis was performed by polymerase chain reaction (PCR) sequencing of PTEN exon 7 and PCR products were directly sequenced in both directions using Applied Biosystems ABI PRISM-3130 Genetic Analyzer instrument (Life Technologies). Alignment was performed using the Basic Local Alignment Search Tool database (BLAST, www.blast.ncbi.nlm.nih.gov). IL-7R mutations/deletions were studied by Sanger sequencing of the PCR products of IL7R exon 6 (FW: TGCATGGCTACTGAATGCTC, RV: CCCACACAATCACCCTCTTT). NOTCH1 mutation analysis was performed by PCR amplification and direct DNA sequencing of exons 26, 27, 28 and part of exon 34 on TAD and PEST domains.48 The sequences were BLAST aligned and manually checked for mutations identification. Statistical analysis Mann-Whitney U tests and Pearson correlation were performed in GraphPad Prism 5. A P-value below 0.05 was considered statistically significant.

Sample acquisition

The instrument (Helios, Fluidigm, CyTOF 6.7.1014 software) was prepared for acquisition according to the manufacturer’s recommendation. Analysis

Signal spillover between channels was corrected using the CATALYST R package as described previously.44 Compensated fcs files were further processed using FlowJo

Results T-cell acute lymphoblastic leukemia cells show constitutive activation of several signaling pathways relevant for proliferation, survival and translation activity We first evaluated the basal signaling profiles of proliferation and intracellular regulators in primary cells obtained

Haematologica | 107 - June 2022

1295

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

D. Kuzilková et al.

Table 1. Clinical and biological characteristics of studied T-cell acute lymphoblastic leukemia patients NOTCH1 (exons EGIL ETP immunoPDN MRD NUP214 IL7R PTEN 26-27-28-34Tscore phenotype response at day 15* /ABL1 exon 6 P)**

SEX

AGE (years)

WBC/mm3 at diagnosis

TALL-1

88,900

T-IV

PGR

TALL-2

169,570

TI/II

PPR

30.0%

mut

TALL-3

136,720

T-III

PGR

mut

TALL-4

60,050

T-III

PGR

mut

TALL-5

400,000

T-III

PGR

2.0%

mut

TALL-6

1026,000

T-IV

PPR

70.0%

mut

TALL-8

362,600

T-III

PGR

1.15%

mut

TALL-9

14,420

T-III

PGR

0.01%

mut

TALL-10

168,090

T-IV

PPR

24.0%

mut

TALL-11

15,800

T-III

PGR

0.14%

mut

TALL-12

159,580

TI/II

PPR

23.0%

mut

TALL-13

14,630

T-III

PGR

0.14%

mut

TALL-14

369,810

T-III

PPR

61.0%

mut

TALL-15

9,390

TI/II

PGR

1.30%

mut

TALL-16

1,700

TI/II

yes

PPR

90.0%

TALL-17

176,000

T-III

PPR

0.99%

Code

TALL: T-cell acute lymphoblastic leukemia; PDN: prednisone; PGR: prednisone good responder; PPR: prednisone poor responder; IR: intermediate risk; HR: high risk; SR: standard risk; wt: wild-type; mut: mutated; ETP: early T-cell phenotype according to Coustan-Smith et al.61 *According to the criteria by Basso et al.62 the % of flow cytometry minimal residual disease (MRD) in the bone marrow of day 15 identifies 3 patient’s risk groups: standard risk (SR, 0.1% blasts), intermediate risk (IR, 0.1 to 10%), and high risk (HR, ≥10); nd: not determined. **Details on the specific exons associated to mutation are reported in the Online Supplementary Table S3.

from the BM or PB of 16 pediatric T-ALL diagnostic specimens, 14 of which were reported previously.38 The clinical and laboratory characteristics of the patients are described in Table 1. Using cytometry by time of flight (CyTOF), we analyzed the samples with a panel of 30 moAb, including eight CD45 barcodes, 15 surface markers and 15 intracellular regulators of various pathways and cellular processes, such as the PI3K-AKT-mTOR, JAK-STAT, MEK-ERK, MAPK pathways, and the activation of T-cell receptors, the cell cycle, transcription and apoptosis (Table 2; Online Supplementary Table S1). Viable T-ALL cells were gated separately from residual non-malignant T cells (Online Supplementary Figure S1). T cells obtained from healthy donor PB (n=3) were used as controls. Compared to their healthy T-cell counterparts, T-ALL cells showed constitutive activation of several functional markers: phosphorylated forms of 4E-BP1, AKT and S6 (PI3K-AKT-mTOR pathway members); Ki-67, a cell cycle progression marker; p-RB, a regulator of the G1 to S transition; p-ERK1/2 and p-p38, members of the mitogen-activated kinase (MAPK) family; and CREB, a transcription

factor regulating proliferation, differentiation and survival (Figure 1). In contrast, the expression of BCL-2 (regulating anti-apoptotic activity) was significantly lower in T-ALL cells. Overall the phospho-signature and levels of intracellular regulators was similar in healthy donors’ peripheral T cells and residual nonmalignant T cells (internal control) and distinct from T-ALL. Single-cell analysis of functional parameters reveals intra- and interindividual heterogeneity of T-cell acute lymphoblastic leukemia cells In order to visualize the internal heterogeneity of the signaling network in T-ALL cells at the phospho-proteomic level, we applied deep learning-based Vaevictis dimensionality reduction (Figure 2A). Signaling, proliferation and internal regulators (13 parameters) were considered for the projection calculation. In the presented cohort of pediatric T-ALL samples, we observed heterogeneity in the signaling regulators, however there was a dominant subset (>50%) within the T-ALL cell compartment in the

Haematologica | 107 - June 2022

1296

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

D. Kuzilková et al.

Table 2. Cytometry by time of flight (CyTOF) panel Metal

Clone

Function

Detection prior or after metOH permeabilization

Vendor

Catalogue number

CD45

89Y

HI30

barcoding

prior

Fluidigm

3089003B

CD45

104Pd

Hi30

barcoding

prior

H.Mei

CD45

106Pd

Hi30

barcoding

prior

H.Mei

CD45

108Pd

Hi30

barcoding

prior

H.Mei

CD45

110Pd

Hi30

barcoding

prior

H.Mei

HLA-I

113In

W6/32

immunophenotype

prior

Exbio

11-422-C100

CD45

115In

MEM-28

barcoding

prior

Exbio

11-222-M001

CD19

142Nd

HIB19

immunophenotype

prior

Fluidigm

3142001B

CD5

143Nd

UCHT2

immunophenotype

prior

Fluidigm

3143007B

CD4

145Nd

MEM-241

immunophenotype

prior

Exbio

11-359-C100

CD8a

146Nd

RPA-T8

immunophenotype

prior

Fluidigm

3146001B

CD7

147Sm

CD7-6B7

immunophenotype

prior

Fluidigm

3147006B

CD34

148Nd

581

immunophenotype

prior

Fluidigm

3148001B

p-4E-BP1 [T37/T46]

149Sm

236B4

phospho-protein

after

Fluidigm

3149005A

p-Stat5 [Y694]

150Nd

phospho-protein

after

Fluidigm

3150005A

CD2

151Eu

TS1/8

immunophenotype

prior

Fluidigm

3151003B

p-Akt [S473]

152Sm

D9E

phospho-protein

after

Fluidigm

3152005A

BCL-2

153Eu

Bcl-2/100

apoptosis

after

Exbio

11-668-C100

CD56

155Gd

B159

immunophenotype

prior

Fluidigm

3155008B

p-p38 [T180/Y182]

156Gd

D3F9

phospho-protein

after

Fluidigm

3156002A

TSLP-R

158Gd

1B4

immunophenotype

prior

Fluidigm

3158026B

p-S6 [S240/244]

159Tb

D68F8

phospho-protein

after

CST

5364

CD38

160Gd

immunophenotype

prior

Exbio

11-366-C100

CD3

161Dy

UCHT1

immunophenotype

prior

Exbio

11-514-C100

p-Lck [T505]

162Dy

4/LCK-Y505

phospho-protein

after

Fluidigm

3162004A

Tdt

164Dy

E17-1519

immunophenotype

after

Fluidigm

3164015B

p-CREB [S133]

165Ho

87G3

phospho-protein

after

Fluidigm

3165009A

p-Rb [S807/811]

166Er

J112-906

proliferation

after

Fluidigm

3166011A

Ki-67

168Er

B56

proliferation

after

Fluidigm

3168007B

CD33

169Tm

WM53

immunophenotype

prior

Fluidigm

3169010B

Cl.Caspase3 (D175)

170Er

poly

apoptosis

after

CST

9661

p-Erk1/2 [T202/Y204]

171Yb

D13.14.4E

phospho-protein

after

Fluidigm

3171010A

PTEN

172Yb

A2B1

tumor supressor

after

559600

HLA-DR

174Yb

L243

immunophenotype

prior

Fluidigm

3174001B

c-Myc

176Yb

9E10

transcription factor

after

Fluidigm

3176012B

DNA intercalator

after

Fluidigm

201192B/201192A

Target

DNA

191Ir/193Ir

CD45

194Pt

MEM-28

barcoding

prior

Exbio

11-222-M001

HLA-I

198Pt

W6/32

barcoding

prior

Bxcell

BE0079

CST: cell signaling technology; BD: BD Biosciences; metOH: methanol.

Haematologica | 107 - June 2022

1297

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

D. Kuzilková et al.

Figure 1. Basal signaling profile of T-cell acute lymphoblastic leukemia and non-malignant T cells. T-cell acute lymphoblastic leukemia (T-ALL) cells show constitutive activation of various signaling pathways as well as proliferation markers as compared to either residual non-malignant T cells or T cells isolated from healthy donor’s (HD) peripheral blood (PB). Median intensity evaluated by mass cytometry of p-4E-BP, p-AKT, p-S6, p-ERK1/2 (top panels); p-RB, Ki-67, p-p38, p-CREB (middle panels); PTEN, p-STAT5, p-LCK, MYC and BCL-2 (bottom panels) in gated T-ALL cells obtained from bone marrow (BM) or PB of T-ALL patients (n=16), residual T cells from T-ALL patients (n=13) and T cells obtained from HD PB (n=3) are shown. T-ALL cells, residual T cells and HD T cells are depicted as circles, squares and triangles, respectively. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

majority of T-ALL patient specimens (Figure 2A; Online Supplementary Figure S2). In addition to the dominant subset, minor subsets of T-ALL cells were identified with a continuous transition between individual subsets. Notably, in some patients (e.g., TALL-10), a subset of T-ALL cells is clearly separated from the major subset, suggesting an intrinsic heterogeneity of T-ALL cells. We then compared interindividual differences in T-ALL patients’ signaling, proliferation and intracellular regulators. We observed a similar pattern in six PTEN exon 7 mutated (PTENmut) T-ALL patients (Figure 2A). Interestingly, on overlaid Vaevictis projections (Figure 2B), PTENmut patients are located in the same compartment of the map, forming a cluster. T-ALL samples with NUP214/ABL1 gene fusion are located within the same compartment but do not separate from PTEN wild-type (PTENwt) cells. As expected, PTENmut cells showed the absence of PTEN protein and a significantly higher MYC expression than PTENwt cells (Figure 2C). PTENwt cells showed only partial expression of MYC, documenting the maintenance of PTEN-MYC axis regulation. We also observed decreased levels of Ki-67 and p-p38 MAP kinase in PTENmut T-ALL

patients. Finally, T-ALL patients with the NUP214/ABL1 gene fusion showed constitutive activation of the JAKSTAT pathway, documented as p-STAT5, compared to TALL patients without this gene fusion (Figure 2C). The impact of both PTEN and MYC expression on the pattern of signaling, proliferation and intracellular regulators is shown on the Vaevictis map (Online Supplementary Figures S3 and S4). Further, we analyzed the correlation of MYC expression with NOTCH1 status in 15 of the 16 studied patients. The distribution of MYC expression among NOTCH1wt cases was variable with two clear clusters, one low (3 cases) and one high (5 cases) as illustrated in Figure 2D. Of the seven NOTCH1mut cases, six were MYC low and one was very high. We then analyzed the correlation between NOTCH1 and PTEN status. Six cases were PTENwt/NOTCH1mut, one case was PTENmut/NOTCH1mut, five cases PTENmut/NOTCH1wt and three cases PTENwt/NOTCH1wt (Table 1). As shown in Figure 2D, the level of MYC was low in all PTENwt cases regardless of NOTCH1 status (3 cases NOTCH1wt and 6 cases NOTCH1mut), whereas all PTENmut cases displayed high

Haematologica | 107 - June 2022

1298

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL MYC (5 cases NOTCH1wt and 1 case NOTCH1mut). In order to further investigate the intrinsic heterogeneity observed in the Vaevictis projections, we analyzed the TALL-10 sample with the most pronounced separated subsets, and we identified mosaic expression of both PTEN and MYC proteins. On the contrary homogeneous expression of these two proteins was observed in representative PTENwt (TALL-11) or PTENmut (TALL-14) patients (Figure 2E and F), indicating that PTEN-MYC regulation can be maintained even at a subclonal level in T-ALL. On the other hand, no differences in the phosphorylation of p38

D. Kuzilková et al.

or the expression of Ki-67 were observed in the two subclones of TALL-10 patient. Finally, in order to confirm the reliability of the approach, we evaluated interindividual variability in the signaling profiles of both proliferation and intracellular regulators in residual T cells (Online Supplementary Figure S4) and compared it to that of T-ALL cells. Residual T cells showed very low interpatient variability, while T-ALL cells showed a high degree of intersample heterogeneity, indicating that the activation signatures in T-ALL blasts were disease-specific and not caused by sample handling or processing.

Figure 2. Continued on following page. Haematologica | 107 - June 2022

1299

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

D. Kuzilková et al.

Figure 2. Single-cell and bulk analysis of functional parameters reveals intra- and interindividual heterogeneity of T-cell acute lymphoblastic leukemia cells.T-cell acute lymphoblastic leukemia (T-ALL) cells from pediatric patients with different genetic features are depicted: PTEN exon 7 mutated (PTENmut) T-ALL (n=6, orange label), T-ALL harboring NUP214/ABL1 gene fusion (n=2, red label) and T-ALL wild-type (wt) for both PTEN mutation and NUP214/ABL1 gene fusion (n=8, green label). (A) Deep learningbased Vaevictis dimensionality reduction (dim red) was applied on T-ALL cells obtained from 16 pediatric T-ALL patients at diagnosis. Signaling, proliferation and internal regulators (PTEN, MYC, BCL-2, p-4E-BP1, p-STAT5, p-AKT, p-p38, p-S6, p-LCK, p-CREB, p-RB, Ki-67, p-ERK1/2) were considered for the projection calculation. (B) Overlaid Vaevictis projection based on signaling, proliferation and internal regulators of T-ALL cells obtained from T-ALL patients (n=16) at diagnosis. Circles represent medians of particular T-ALL specimens. (C) Median intensity of deregulated proteins in PTENmut T-ALL samples (n=6, squares) as compared to PTENwt samples (n=10, circles) are shown. T-ALL samples carrying NUP214/ABL1 gene fusion (n=2) are depicted as red circles. (D) Median intensity of MYC protein in NOTCH1wt (left) and NOTCH1 mutated (NOTCH1mut) (right) patients. (E) The expression of the most deregulated proteins in PTENmut samples are shown. In the Vaevictis projection of T-ALL phospho-signature a compact sub-clone in TALL-10 specimen was identified (green gate). For comparison, 1 representative PTENmut (TALL-14, yellow label) and one representative PTENwt (TALL-11, green label) T-ALL sample are shown. Heat of each map corresponds with expression of indicated marker. Note the scale max differ for each marker and each sample. Mosaic expression of both, PTEN and MYC proteins were observed, while no differences in phosphorylation of p38 and expression of Ki-67 are present in the two sub-clones of TALL-10 patient. (F) Contour plots showing expression of PTEN, MYC, p-p38 and Ki-67 in TALL-10 in comparison with TALL-11 and TALL-14 samples are shown. Haematologica | 107 - June 2022

1300

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL In vitro response to IL-7 correlates with higher proliferation and predicts good response to prednisone induction The IL-7/IL-7R signaling axis activates three main pathways in T cells, the JAK-STAT, PI3K-AKT-mTOR and MEKERK pathways, ultimately promoting leukemia cell viability, cell cycle progression and growth. The contribution of each of these pathways to a particular functional outcome is still not completely known and appears to differ between normal and malignant states.20 For these reasons, we investigated the impact of IL-7/IL-7R signaling on the modulation of the before mentioned functional nodes upon IL-7 stimulation. For this purpose, we analyzed IL-7-induced ex vivo perturbation of nine phospho-epitopes in T-ALL cells, residual T cells and T cells obtained from healthy donor´s PB (Figure 3A). We resolved two response types: (i) IL-7-responsive T-ALL samples (n=6) and (ii) IL-7-non-responsive T-ALL samples (n=10), distinguished by their difference in p-STAT5 activation. P-STAT5 increased by 2.5-fold in IL-7 responders, while it remained unchanged in IL-7 non-responders. The fold increase in p-STAT5 was caused solely by IL-7 induction (Figure 3B), since the basal levels were comparable in both groups (Online Supplementary Figure S5). In contrast, in normal T cells, we observed a homogeneous pattern of the p-STAT5 response in 14 of 15

D. Kuzilková et al.

samples (Figure 3A; Online Supplementary Figure S6). However, apart from p-STAT5, no other tested nodes were differentially expressed between the two subgroups (Online Supplementary Figure S7). We next investigated the correlation between IL-7-induced STAT5 phosphorylation and the expression of the IL-7R in T-ALL cells (n=15) and residual T cells (n=9). The fold change in p-STAT5 upon IL7 stimulation quantitatively correlated with the percentage of CD127 (Figure 3C) but not with CD132 (common g chain, a subunit of the IL-7 receptor, Online Supplementary Figure S8) expression. However, the IL-7 non-responsiveness of T-ALL blasts cannot be explained solely by the lower expression of CD127 or CD132. For example, cases TALL-15 and TALL-9 expressed CD127 at 60% and >90% respectively but responded poorly to IL-7. TALL-15 was harboring NUP214/ABL1 gene fusion with basal p-STAT5 already at the maximal level, which can explain its non-responsiveness. By contrast TALL-9 (PTENwt, IL-7Rwt and NUP214/ABL1wt) displayed low responsiveness despite its low basal p-STAT5. Yet, STAT-5 was phosphorylated after pervanadate treatment showing that the method worked and also that STAT-5 could potentially be phosphorylated in these cells. Thus the lack of responsiveness to IL-7 in this specific case remains unclear. In order to further investigate the characteristics of IL-7 non-responsive T-ALL samples, we compared the profiles

Figure 3. Continued on following page. Haematologica | 107 - June 2022

1301

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

D. Kuzilková et al.

Figure 3. Response to IL-7 of T-cell acute lymphoblastic leukemia cells and non-malignant T cells. (A) Heatmap represents fold change in protein phosphorylation in ex vivo IL-7 stimulated cells as compared to their unstimulated state in T-cell acute lymphoblastic leukemia (T-ALL) cells (left) and residual T cells (right) obtained from T-ALL patients and healthy donors (in columns) at diagnosis. Heat in the squares represents the fold change in phosphorylation of proteins listed in rows. Hierarchical clustering with Euclidean distance metrics and an average linkage dissects two clusters of T-ALL and residual T cells: IL-7 responder and IL-7 non-responder. Values of 0 were replaced with 0.0001. (B) Median intensity of p-STAT5 in untreated and ex vivo IL-7 stimulated specimens from IL-7 responder T-ALL (circles, n=6) and IL-7 non-responder T-ALL (squares, 10) is shown. Median intensity of pSTAT5 increased after ex vivo IL-7 stimulation in IL-7 responder T-ALL but not in IL-7 non-responder.(C) Correlation between percentage of surface CD127 (IL-7RA) expression evaluated by flow cytometry and fold change of p-STAT5 after ex vivo IL-7 stimulation in T-ALL samples obtained from diagnostic T-ALL specimens. No difference in percentage of CD127 or CD132 (common g chain, subunit of IL-7R) surface expression was found between IL-7 responder and IL-7 non-responder T-ALL. (D) Median intensity of RB phosphorylation (left) and Ki-67 expression (right) in unstimulated specimens from IL-7 responder T-ALL (circles, n=6) and IL-7 non-responder T-ALL (squares, n=10) are shown. (E) White blood cell count (WBC) at diagnosis of T-ALL (d0, left), count of T-ALL cells in mm3 at day 8 (d8) and percentage of T-ALL cells in bone marrow at day 15 (d15) of T-ALL patients expressed as minimal residual disease (MRD) in IL-7 responder (circles, n=6) and IL-7 non-responder (squares, n=10) T-ALL samples are shown. Horizontal dashed line dissects prednisone (PDN) good responders (under the cut-off) and PDN poor responders (above the cutoff).

of all intracellular regulators in unstimulated T-ALL samples and we identified higher levels of p-RB and Ki67, both markers of cell cycle progression, in IL-7 responders compared to IL-7 non-responders (Figure 3D; Online Supplementary Figure S9). Interestingly, while both subgroups had comparable white blood cell counts at diagnosis, IL-7 responder T-ALL patients had significantly lower absolute blast cell counts at day 8 of treatment (Figure 3E), the predictor of prognosis in pediatric T-ALL,5 indicating that ex vivo IL-7 responsiveness might reflect the sensitivity to prednisone therapy in vivo. In order to experimentally test the sensitivity to prednisolone, we treated either IL-7 responder (TALL-1, -3, -4, -11, -13) or IL7 non-responder (TALL-10, -12, -14, -16, -17) samples ex vivo with IL-7 and prednisolone. Cells from IL-7 responder cases significantly increased their counts upon IL-7 ad-

ministration, while the cells from IL-7 non-responders did not (Online Supplementary Figure S10A). Next, we tested the ex vivo response to prednisolone in the absence or presence of IL-7. Prednisolone alone induced a significant reduction of viable cells in the culture of IL-7 responder cases but only an insignificant decrease in IL-7 non-responder cultures (Online Supplementary Figure S10B and C). However, in the presence of IL-7 the effect of prednisolone in responders was mitigated to the level observed at basal condition (Online Supplementary Figure S10B). The PI3K pathway can be inhibited in vitro by BEZ-235 in 11 of 16 T-cell acute lymphoblastic leukemia samples As reported previously,49 the level of basal phosphorylation of the PI3K pathway members p-4E-BP1, p-AKT and p-S6 was high in T-ALL (Figure 1). Therefore, we tested the

Haematologica | 107 - June 2022

1302

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL inhibitory potential of the dual PI3K/mTOR inhibitor BEZ235. Since no basal activation of the PI3K pathway was observed in the residual T cells or T cells obtained from PB of healthy donors, we did not include any of those in the analysis. Regarding the extent of BEZ-235-induced decrease in levels of p-4E-BP1, p-AKT and p-S6 (Figure 4A), we dissected two clusters of T-ALL patients: (i) ex vivo BEZ-235 responder T-ALL samples (n=11), which showed

D. Kuzilková et al.

a decreased level of p-4E-BP1 and (ii) ex vivo BEZ-235 non-responder T-ALL samples (n=5), which showed no to mild change in the level of PI3K pathway members´ phosphorylation. When we compared the basal phosphorylation of p-4E-BP1, p-AKT and p-S6 in the two clusters of T-ALL, we did not find any difference (Figure 4B). However, we observed significant inactivation of 4E-BP1 and AKT only in the BEZ-235 responder T-ALL samples. Then, to

Figure 4. Response of T-cell acute lymphoblastic leukemia cells to BEZ-235 inhibition. (A) Heatmap represents fold change in protein phosphorylation of PI3K pathway members in ex vivo BEZ-235 stimulated cells compared to their unstimulated state in T-cell acute lymphoblastic leukemia (T-ALL) patients samples (in columns) obtained at diagnosis. Only gated T-ALL cells were considered in the analysis. Squares represent the fold change in phosphorylation of proteins listed in rows. Hierarchical clustering with Euclidean distance metrics and an average linkage dissects two clusters of T-ALL: BEZ-235 responder and BEZ-235 nonresponder. (B) Median intensity of p-4E-BP1, p-AKT and p-S6 in untreated and ex vivo BEZ-235 treated specimens from BEZ-235 responder T-ALL (circles, n=11) and BEZ-235 non-responder T-ALL (squares, n=5) samples are shown. Median intensity of p-4EBP1 and p-AKT decreased after ex vivo BEZ-235 treatment in BEZ-235 responder T-ALL but not in BEZ-235 non-responder. (C) Median intensity of phosphorylated LCK in BEZ-235 responder T-ALL (circles, n=11) and BEZ-235 non-responder T-ALL (squares, n=5) is shown. Haematologica | 107 - June 2022

1303

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL further investigate the features associated with non-responsiveness to BEZ-235, we compared the profiles of all the analyzed intracellular regulators (Online Supplementary Figure S11). Only the inhibitory carboxy-terminal Tyr of Lck was hyperphosphorylated in BEZ-235 non-responder TALL patients under the basal conditions (Figure 4C). Next, we wanted to investigate whether the BEZ-235 response correlates with GC responsiveness and if the addition of BEZ-235 can increase cells’ sensitivity to prednisolone. We treated cells with BEZ-235 in the presence or absence of prednisolone. The decrease in viability induced by BEZ235 alone in four samples from the group considered BEZ235 responders (TALL-10, -14, -16, -17) was not significant. Furthermore, prednisolone + BEZ-235 combined treatment did not increase cells’ sensitivity to GC (Online Supplementary Figure S12A). Similar results were obtained when the four samples from the group of BEZ-235 nonresponders (TALL-1, 3, 4, 11) were tested, as shown in the Online Supplementary Figure S12B. JAK-STAT pathway activation and PI3K-AKT pathway inhibition are mutually exclusive features of T-cell acute lymphoblastic leukemia Interestingly, we observed a high degree of mutual exclu-

D. Kuzilková et al.

sivity in IL-7 and BEZ-235 responsiveness among the TALL samples: four of 16 samples (TALL-1, -3, -4, -11) were IL-7 responders but not BEZ-235 responders; in contrast, 9 of 16 samples (TALL-6, -8, -9, -10, -12, -14, -15, -16, -17) were BEZ-235 responders but IL-7 nonresponders. Two samples (TALL-5 and TALL-13) responded to both ex vivo treatments, and only one sample (TALL-2) was not responsive to any ex vivo treatment (Figure 5A; Online Supplementary Table S2), Furthermore, when we analyzed the composition of the IL-7- or BEZ-235 based clusters with regards to the in vivo response to GC (Figure 5B), we noticed that all the IL-7 responder samples were from PGR patients (Online Supplementary Table S2), while samples of BEZ-235 responders were enriched in PPR samples (6 of 9, 66.6%), suggesting a correlation between JAK-STAT or PI3K-AKT activation and the early response to treatment in vivo. Phospho-signatures and the dominance of the JAKSTAT or PI3K-AKT pathway are maintained at relapse Since phosphokinases are dynamic regulators of signaling networks, we investigated whether the phospho-signature present at diagnosis was maintained at relapse. Using the CyTOF panel, we analyzed five pairs of diagnostic and re-

Figure 5. Dominance of JAK-STAT or PI3K-AKT pathway in pediatric T-cell acute lymphoblastic leukemia. T-cell acute lymphoblastic leukemia (T-ALL) patient samples were defined as IL-7 and/or BEZ-235 responders based on phosphorylation/dephosphorylation of particular targets as described in Figure 3 and Figure 4. TALL-2 did not respond to any of those 2 ex vivo treatment, TALL-5 and TALL-13 were defined as dual responders to IL-7 and BEZ-235 ex vivo treatment. TALL-1, TALL-3, TALL-4 and TALL-11 are IL-7 responders (and BEZ-235 non-responders). TALL-6, TALL-8, TALL9, TALL-10, TALL-12, TALL-14, TALL-15, TALL-16 and TALL-17 are BEZ-235 responders (and IL-7 non-responders). Thirteen of 16 of T-ALL patients’ samples show mutual exclusivity of response. (B) Count of TALL cells/mm3 at day 8 (d8) in IL-7 responder (red circles, n=4), dual responder to IL-7 and BEZ-235 (purple hexagons, n=2), BEZ-235 responder (blue squares, n=9) and non-responder (grey cross, n=1) TALL samples are shown. Horizontal dashed line dissects prednisone (PDN) good responders (under the cut-off) and PDN poor responders (above the cut-off).

Haematologica | 107 - June 2022

1304

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL lapsed T-ALL patient samples (TALL-1, -2, -15, -16, -17). In order to visualize phospho-signatures, Vaevictis dimensionality reduction was applied (Figure 6A). Signaling, proliferation and internal regulators (13 parameters) were considered for the projection calculation. First, we applied a gating strategy to identify the dominant and secondary clones of T-ALL cells in the diagnostic samples. Next, we applied an identical gating strategy on the relapse samples and compared the expression patterns of our markers in all paired samples (Online Supplementary Figure S13). In TALL-1, TALL-15 and TALL-16, the major diagnostic clone (accounting for 72.1%, 58.9% and 79.3% of cells, respectively) was the subset characterized by highly phosphorylated RB and 4E-BP1, indicating active PI3K-AKT pathway and cell cycle progression. At relapse, the major diagnostic clone was still present in TALL-1 and TALL-15, although in a reduced proportion (24.2% and 8.93%, respectively); the major clones represented at relapse (62.3% and 67%, respectively) retained intermediate levels of p-RB and p-4E-BP1. In TALL-16, the dominant diagnostic clone (79.3%) was virtually absent at relapse (0.13%), showing a concomitant dramatic increase (84.2%) in the clone with intermediate levels of p-RB and p-4EBP1. In TALL-2 and TALL-17, the dominant clusters at diagnosis (70.6% and 60.8%, respectively), with intermediate levels of p-RB and p-4E-BP1, remained prevalent at relapse (53.5% and 82%). In summary, we observed that in four of five paired T-ALL samples (TALL-1, -2, -15, -17), the pattern of phospho-signatures was maintained at relapse, with a redistribution of cells from clones with highly phosphorylated RB and 4E-BP1 towards clones with lower p-RB. In one sample (TALL-16), the dominant cluster (with the highest activation level) disappeared at relapse. We next sought to assess whether these findings corresponded with the dominance of the JAK-STAT or PI3KAKT pathway. We performed the same ex vivo treatments applied to diagnostic samples to their relapsed counterparts. We observed that with regards to the fold change in phosphoproteins (n=9) after ex vivo IL-7 stimulation, all relapsed samples were allocated to the same cluster as their diagnostic counterparts, specifically (i) ex vivo IL-7 responder T-ALL samples containing paired samples from TALL-1 and (ii) ex vivo IL-7 non-responders T-ALL samples containing paired samples from patients with TALL-2, TALL-15, TALL-16 and TALL-17 (Figure 6B). Similarly, we analyzed the BEZ-235 induced PI3K-AKT response (p-4EBP1, p-AKT and p-S6) in the 5 paired T-ALL samples (Figure 6C); four of 5 relapsed samples were allocated to the same cluster as the diagnostic paired specimens, specifically: (i) ex vivo BEZ-235 responders TALL-15 and TALL-17; (ii) ex vivo BEZ-235 non-responders TALL-1 and TALL-2. An exception was TALL-16 in which the relapse sample shifted to the BEZ-235 non-responder cluster, while its diagnostic counterpart was originally located in the BEZ-

D. Kuzilková et al.

235 responder cluster. Notably, TALL-16 was the only sample that lost the dominant diagnostic clone (with a highly active PI3K-AKT pathway) at relapse (Figure 6A).

Discussion In the present study we provide original evidence showing that T-ALL blasts exhibit intrapatient heterogeneity of phospho-signaling regulators that are activated at different levels within interconnected cell subsets and that this regulatory activity is generally maintained at relapse. Our study accidentally identified one patient with a subclonal PTEN mutation that demonstrated unique heterogeneity of intracellular regulators documenting that single-cell analysis by mass cytometry can tackle this type of complex patterns. Furthermore, we observed that perturbation induced by the ex vivo administration of either IL-7 or BEZ235 reveals a high degree of mutual exclusivity with respect to the responsiveness to these agents at phospho-protein level. One limitation of the present study is the low number of patients. Also, we have selected the primary samples based on the availability of frozen vials in the cell bank, the number of cells/vial and the cell viability at the time of thawing. This resulted in a final selection which does not fully reflect the classical T-ALL case distribution. In particular, our series does not contain IL-7R mutated cases and is slightly enriched in PTEN mutated patients (37.5%) compared to the reported incidence of this subgroup in other reports such as those by Palomero et al.50 and by Zuurbier et al.30 who documented the incidence ranging from 16% to 17% of T-ALL. In addition to the bias in genetic mutations distribution, our series is slightly enriched in PPR patients, being 43.7% compared to 34% in historical prospective series of childhood T-ALL reported by Schrappe et al.51 Mass cytometry allowed us to apply a broad panel of moAb, demonstrating for the first time the feasibility and power of this technology in the detailed characterization of signaling networks in T-ALL samples. Mass cytometry phospho-profiling was pioneered by Bodenmiller,36 who also used cellular barcoding to study the signaling dynamics and cell-to-cell communication in human blood cells. Here, we confirm and extend our previous observations by phospho-flow cytometry of hyperactivated status of T-ALL.38 We demonstrate the potential value of this technology as a tool for the identification of the dominant features of the response to IL-7 activation or BEZ-235 inhibition (here, p-STAT5 and p-4E-BP1). Importantly, barcoding of samples and gating of residual T cells allowed for multiple levels of internal controls, assuring technical accuracy and proper context for interpretation. The running cost of mass cytometry technology with barcoding is comparable to conventional cytometry, while it provides the advantage of a high number of parameters, notably

Haematologica | 107 - June 2022

1305

D. Kuzilková et al.

Legend on following page.

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

Haematologica | 107 - June 2022

1306

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL

D. Kuzilková et al.

Figure 6. Single-cell and bulk analysis in T-cell acute lymphoblastic leukemia (T-ALL) cells of functional parameters reveals maintenance of phospho-signature at relapse in 4 of 5 T-ALL patients. (A) Vaevictis dimensionality reduction was applied on Tcell acute lymphoblastic leukemia (T-ALL) cells obtained from 5 pediatric T-ALL patients at diagnosis and relapse (top row). Signaling, proliferation and internal regulators (PTEN, MYC, BCL-2, p-4E-BP1, p-STAT5, p-AKT, p-p38, p-S6, p-LCK, p-CREB, p-RB, Ki-67, p-ERK1/2) were considered for the projection calculation. Two to 4 subsets of T-ALL cells were identified in each diagnostic sample. The same gating strategy was applied on the paired relapse samples. Histograms show phosphorylation of RB (middle row) and 4E-BP1 (bottom row) of particular subsets of T-ALL cells. The color of the histogram line corresponds with the color of the gate in the Vaevictis projection. The y-axis reflects the count of the cells. The redistribution of particular subsets from diagnosis to relapse is visualized using fish plot. The color of each subset corresponds with the color of the gate in the Vaevicts projection. Time points (diagnosis, relapse) are indicated on x-axis, percentage of particular cell subset is reflected on y-axis. (B) Heatmap represents fold change in protein phosphorylation in ex vivo IL-7 stimulated cells compared to unstimulated state, samples were obtained from T-ALL patients (in columns) at diagnosis (black label) or relapse (orange label). Only T-ALL cells from the diagnostic samples are included in the analysis. Paired diagnostic and relapse samples are connected. Heat in squares represents the fold change in phosphorylation of proteins listed in rows. Hierarchical clustering with Euclidean distance metrics and an average linkage dissects two clusters of T-ALL samples (separated with horizontal line): IL-7 responder and IL-7 non-responder. (C) Heatmap represents fold change in protein phosphorylation of PI3K pathway members in ex vivo BEZ-235 stimulated cells compared to unstimulated state, samples were obtained from T-ALL patients (in columns) at diagnosis (black label) or relapse (orange label). Only T-ALL cells from the diagnostic samples are included in the analysis. Paired diagnostic and relapse samples are connected. Heat in squares represents the fold change in phosphorylation of proteins listed in rows. Hierarchical clustering with Euclidean distance metrics and an average linkage was used. The 2 clusters of samples (separated with horizontal line) were dissected manually according to Figure 4: BEZ-235 responder and BEZ-235 non-responder.

phosphoproteins, that can be reliably measured. Our workflow overcomes some limitations of mass cytometry (high equipment cost and demanding reagent stock production) by shipping fixed samples after in vitro treatment to a collaborator. The Vaevictis projection shows recurrent patterns in patients with particular mutations. Notably, PTENmut cells displayed high MYC expression. MYC is a master transcription factor regulating several critical cell functions, such as metabolism, proliferation and survival,52 and it functions as a potent oncogene in a large number of cancers.53 In line with our findings, Bonnet et al.54 showed that modulation of MYC protein was maintained through downregulation of PTEN via the PI3K-AKT axis in T-ALL. This explains the significant inverse correlation between MYC and PTEN expression. Furthermore we analyzed the correlation between NOTCH1, PTEN status and MYC levels. We observed that MYC was low in all PTENwt cases regardless of NOTCH1 status, while all the PTENmut cases displayed high MYC in line with Bonnet et al.54 These authors showed that high MYC expression levels can be observed in the absence of (known) NOTCH1 and/or FBXW7 mutations and that modulation of MYC protein can be observed in the presence of downregulation of PTEN as a major alternative pathway of MYC activation in T-ALL independently from NOTCH1 mutations. GC resistance remains a major challenge in the treatment of pediatric ALL, and although several mechanisms have been suggested,6,55,56 our understanding of the molecular basis remains incomplete. We found that all IL-7 responding samples were from patients with a good response to GC in vivo whereas all BEZ235 responding (but IL-7 non-responsive) samples were from patients with poor response to GC in vivo. However, due to the limited number of patients studied we cannot

state that IL-7R pathway activation in vitro reflects the sensitivity to PDN therapy in vivo. Nevertheless, our preliminary observations are also supported by GC cytotoxicity experiments that assessed ex vivo prednisolone response in T-ALL primary cells in the absence or presence of IL-7. Firstly, we observed that in samples that responded to IL7 stimulation by increased phosphorylation of STAT5 the number of viable cells significantly increased upon ex vivo administration of IL-7. In contrast, the IL-7 non-responsive samples did not show any significant increase. We then exposed both kinds of samples to prednisolone in the absence or presence of IL-7. In the absence of IL-7, we observed a significant reduction in the number of viable cells, whereas when combined, IL-7 and prednisolone mutually canceled out their effects. Similar investigations were carried out by Delgado-Martin et al.57 who assessed p-STAT5 responsiveness to IL-7 with regards to GC sensitivity, and he found that the responder subsets displayed resistance to GC-induced death while the IL-7 non-responder subset was mostly GC sensitive. However these authors did not formally assess the pro-survival effect of IL-7 alone, but only as a protective factor from chemotherapy-induced death. In our series of primary T-ALL samples, the IL-7 response was assessed as both p-STAT5 response and the pro-survival response. These characteristics correlated with good GC response in vivo. Notably, the IL-7 responder cells in our series displayed a significantly higher expression of both Ki-67 and p-RB indicating a higher proliferative functional state as compared to IL-7 non-responder blasts. This is consistent with a more effective action of GC and other cytotoxic drugs on highly proliferating cells.58,59 Furthermore, we assessed whether the observed fold change in phospho-proteins after BEZ-235 treatment cor-

Haematologica | 107 - June 2022

1307

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL related with drug sensitivity, and if its addition could increase cell sensitivity to GC. Indeed, BEZ-235 treatment did not increase the cell sensitivity to prednisolone in the four BEZ-235 phospho-responder samples. Hall and collaborators60 examined the cytotoxic activity of BEZ-235 and dexamethasone, as single agents and in combination in both T-ALL cell lines and primary T-ALL samples. BEZ235 alone did not show any significant activity, in line with our data. However, these authors observed a synergistic cytotoxic effect of BEZ-235 and dexamethasone in most T-ALL cell lines tested, and in primary T-ALL lymphoblast. Li and colleagues19 studied a large cohort of primary TALL samples and demonstrated that genes involved in the IL7R–RAS–MAPK–AKT signaling pathway were significantly associated with steroid resistance and poor clinical outcome. They subsequently evaluated the drug synergism between PDN and several agents including a PI3K-AKTmTOR inhibitor in 11 primary T-ALL samples collected at diagnosis and they observed an enhanced steroid response in most samples in line with the data by Hall et al.60 We tested the synergism between prednisolone and BEZ-235 in only four patients and we did not perform any extended screening of the mutational state of the IL-7R pathway, thus our observation of the lack of additive effect of BEZ-235 and PDN activity should be taken with caution. Recently, Liu et al.22 performed an integrated genomic analysis of a large cohort of T-ALL patients and identified ten functional pathways recurrently mutated in T-ALL, including the PI3K-AKT-mTOR and JAK-STAT pathways, in 29% and 25% of cases, respectively. Mutual exclusivity was observed between the PI3K-AKT pathway and JAKSTAT or Ras pathway alterations. Our data are in line with this hypothesis at least at the phospho-proteomic level as we found two subsets of patients with opposite response to IL-7 and to BEZ-235. Notably, the signature associated with either IL-7 or BEZ235 responsiveness was maintained at relapse, with the exception of a single case in which the loss of the highly activated dominant clone was observed. In this regard the importance of the conservation of the functional dominance suggests that functional studies at an early stage (i.e., at diagnosis) may be of relevance for treatment of relapse. Therefore cases of refractory T-ALL patients warrant further in-depth studies. We are well aware that any clinical correlation of our biology-oriented study in a limited non-representative

D. Kuzilková et al.

series of childhood T-ALL (for example no IL-7R mutated patients are included) should be corroborated in further extensive investigations, nevertheless we do believe that our data should be considered in view of a better understanding of the mechanisms of GC resistance in childhood T-ALL. In summary, we demonstrated the translational potential of the mass cytometry-based T-ALL profiling at the single-cell level. Further, our analysis allowed us to investigate the heterogeneity within a single T-ALL sample, revealing a rare case of subclonal PTEN mutation28. Because of this advanced approach, we were able to identify distinct functional clusters of IL-7 and BEZ-235 responsive T-ALL cells with different sensitivity to GC, which can persist at relapse. Collectively, our observations can contribute to a better understanding of the complex signaling network governing T-ALL behavior and its correlation with the response to therapy. Disclosures No conflicts of interest to disclose. Contributions DK and CB performed the experiments, analyzed the data, wrote the manuscript; KR performed the experiments; ARS performed the conjugations experiments; HEM designed and performed conjugations experiments; MP performed part of the mutation screening; ABif and Abio revised the manuscript; TK and GG designed the research, analyzed the data and wrote themanuscript. All authors have read and approved the final submitted version of the manuscript Acknowledgments We wish to thank to Ondrej Hrusak for fruitful discussions and to Daniel Thürner and Pavel Semerak for superb handling of the Helios instrument. Funding This project was supported by the Fondazione Alessandro Maria Zancan ONLUS “Grande Ale ONLUS”, Fondazione M. Tettamanti De Marchi and NU20-05-00282 from the Ministry of Health, Czech Republic. It was also partially funded by the following grants: AIRC IG 2017 ref. id 20564, AIRC 5x1000 ref. id 21147 to AB, AIRC Accelerator Award 2018 ref. id 22791.

References 1. Pui CH. Acute lymphoblastic leukemia: introduction. Semin Hematol. 2009;46(1):1-2. 2. Girardi T, Vicente C, Cools J, Keersmaecker K De. The genetics

and molecular biology of T-ALL. Blood. 2017;129(9):1113-1123. 3. Klumper E, Pieters R, Veerman A, et al. In vitro cellular drug resistance in children with relapsed/refractory acute

Haematologica | 107 - June 2022

1308

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL lymphoblastic leukemia. Blood. 1995;86(10):3861-3868. 4. Gao J, Liu WJ. Prognostic value of the response to prednisone for children with acute lymphoblastic leukemia: a metaanalysis. Eur Rev Med Pharmacol Sci. 2018;22(22):7858-7866. 5. Lauten M, Moricke A, Beier R, et al. Prediction of outcome by early bone marrow response in childhood acute lymphoblastic leukemia treated in the ALL-BFM 95 trial: differential effects in precursor B-cell and T-cell leukemia. Haematologica. 2012;97(7):1048-1056. 6. Follini E, Marchesini M, Roti G. Strategies to overcome resistance mechanisms in T-cell acute lymphoblastic leukemia. Int J Mol Sci. 2019;20(12):3021. 7. Scheijen B. Molecular mechanisms contributing to glucocorticoid resistance in lymphoid malignancies. Cancer Drug Resist. 2019;2:647-664. 8. De Smedt R, Morscio J, Goossens S, Van Vlierberghe P. Targeting steroid resistance in T-cell acute lymphoblastic leukemia. Blood Rev. 2019;38:100591. 9. Rich BE, Campos-Torres J, Tepper RI, Moreadith RW, Leder P. Cutaneous lymphoproliferation and lymphomas in interleukin 7 transgenic mice. J Exp Med. 1993;177(2):305-316. 10. Lindqvist CM, Lundmark A, Nordlund J, et al. Deep targeted sequencing in pediatric acute lymphoblastic leukemia unveils distinct mutational patterns between genetic subtypes and novel relapse-associated genes. Oncotarget. 2016;7(39):6407164088. 11. Abraham N, Ma MC, Snow JW, Miners MJ, Herndier BG, Goldsmith MA. Haploinsufficiency identifies STAT5 as a modifier of IL-7-induced lymphomas. Oncogene. 2005;24(33):5252-5257. 12. Silva A, Laranjeira ABA, Martins LR, et al. IL-7 contributes to the progression of human T-cell acute lymphoblastic leukemias. Cancer Res. 2011;71(14):4780-4789. 13. Dibirdik I, Langlie M, Ledbetter J, et al. Engagement of interleukin-7 receptor stimulates tyrosine phosphorylation, phosphoinositide turnover, and clonal proliferation of human Tlineage acute lymphoblastic leukemia cells. Blood. 1991;78(3):564-570. 14. Treanor LM, Zhou S, Janke L, et al. Interleukin-7 receptor mutants initiate early T cell precursor leukemia in murine thymocyte progenitors with multipotent potential. J Exp Med. 2014;211(4):701-713. 15. Ribeiro D, Melão A, van Boxtel R, et al. STAT5 is essential for IL7–mediated viability, growth, and proliferation of T-cell acute lymphoblastic leukemia cells. Blood Adv. 2018;2(17):2199-2213. 16. Shochat C, Tal N, Bandapalli OR, et al. Gain-of-function mutations in interleukin-7 receptor-α (IL7R) in childhood acute lymphoblastic leukemias. J Exp Med. 2011;208(5):901-908. 17. Zenatti PP, Ribeiro D, Li W, et al. Oncogenic IL7R gain-offunction mutations in childhood T-cell acute lymphoblastic leukemia. Nat Genet. 2011;43(10):932-939. 18. Canté-Barrett K, Spijkers-Hagelstein JAP, Buijs-Gladdines JGCAM, et al. MEK and PI3K-AKT inhibitors synergistically block activated IL7 receptor signaling in T-cell acute lymphoblastic leukemia. Leukemia. 2016;30(9):1832-1843. 19. Li Y, Buijs-Gladdines JGCAM, Canté-Barrett K, et al. IL-7 receptor mutations and steroid resistance in pediatric T cell acute lymphoblastic leukemia: a genome sequencing study. PLoS Med. 2016;13(12):e1002200. 20. Oliveira ML, Akkapeddi P, Alcobia I, et al. From the outside, from within: biological and therapeutic relevance of signal transduction in T-cell acute lymphoblastic leukemia. Cell Signal. 2017;38:10-25. 21. Kim R, Boissel N, Touzart A, et al. Adult T-cell acute lymphoblastic leukemias with IL7R pathway mutations are

D. Kuzilková et al.

slow-responders who do not benefit from allogeneic stem-cell transplantation. Leukemia. 2020;34(7):1730-1740. 22. Liu Y, Easton J, Shao Y, et al. The genomic landscape of pediatric and young adult T-lineage acute lymphoblastic leukemia. Nat Genet. 2017;49(8):1211-1218. 23. Goossens S, Radaelli E, Blanchet O, et al. ZEB2 drives immature T-cell lymphoblastic leukaemia development via enhanced tumour-initiating potential and IL-7 receptor signalling. Nat Commun. 2015;6(1):5794. 24. Zhang J, Ding L, Holmfeldt L, et al. The genetic basis of early Tcell precursor acute lymphoblastic leukaemia. Nature. 2012;481(7380):157-163. 25. Vicente C, Schwab C, Broux M, et al. Targeted sequencing identifies associations between IL7R-JAK mutations and epigenetic modulators in T-cell acute lymphoblastic leukemia. Haematologica. 2015;100(10):1301-1310. 26. Meyer LK, Huang BJ, Delgado-Martin C, et al. Glucocorticoids paradoxically facilitate steroid resistance in T cell acute lymphoblastic leukemias and thymocytes. J Clin Invest. 2020;130(2):863-876. 27. Okkenhaug K, Vanhaesebroeck B. PI3K in lymphocyte development, differentiation and activation. Nat Rev Immunol. 2003;3(4):317-330. 28. Mendes RD, Sarmento LM, Canté-Barrett K, et al. PTEN microdeletions in T-cell acute lymphoblastic leukemia are caused by illegitimate RAG-mediated recombination events. Blood. 2014;124(4):567-578. 29. Milella M, Falcone I, Conciatori F, et al. PTEN: multiple functions in human malignant tumors. Front Oncol. 2015;5:24. 30. Zuurbier L, Petricoin EF, Vuerhard MJ, et al. The significance of PTEN and AKT aberrations in pediatric T-cell acute lymphoblastic leukemia. Haematologica. 2012;97(9):1405-1413. 31. Homminga I, Pieters R, Langerak AW, et al. Integrated transcript and genome analyses reveal NKX2-1 and MEF2C as potential oncogenes in T cell acute lymphoblastic leukemia. Cancer Cell. 2011;19(4):484-497. 32. Ferrando AA, Neuberg DS, Staunton J, et al. Gene expression signatures define novel oncogenic pathways in T cell acute lymphoblastic leukemia. Cancer Cell. 2002;1(1):75-87. 33. Soulier J, Clappier E, Cayuela JM, et al. HOXA genes are included in genetic and biologic networks defining human acute T-cell leukemia (T-ALL). Blood. 2005;106(1):274-286. 34. Bhatla T, Jones CL, Meyer JA, Vitanza NA, Raetz EA, Carroll WL. The biology of relapsed acute lymphoblastic leukemia. J Pediatr Hematol Oncol. 2014;36(6):413-418. 35. Loosveld M, Castellano R, Gon S, et al. Therapeutic targeting of c-Myc in T-cell acute lymphoblastic leukemia (T-ALL). Oncotarget. 2014;5(10):3168-3172. 36. Bodenmiller B, Zunder ER, Finck R, et al. Multiplexed mass cytometry profiling of cellular states perturbed by smallmolecule regulators. Nat Biotechnol. 2012;30(9):858-867. 37. Bendall SC, Simonds EF, Qiu P, et al. Single-cell mass cytometry of differential immune and drug responses across a human hematopoietic continuum. Science. 2011;332(6030):687-696. 38. Bonaccorso P, Bugarin C, Buracchi C, et al. Single-cell profiling of pediatric T-cell acute lymphoblastic leukemia: impact of PTEN exon 7 mutation on PI3K/Akt and JAK–STAT signaling pathways. Cytometry B Clin Cytom. 2020;98(6):491-503. 39. Dworzak MN, Buldini B, Gaipa G, et al. AIEOP-BFM consensus guidelines 2016 for flow cytometric immunophenotyping of pediatric acute lymphoblastic leukemia. Cytometry B Clin Cytom. 2018;94(1):82-93. 40. Watson M, Chow S, Barsyte D, et al. The study of epigenetic mechanisms based on the analysis of histone modification

Haematologica | 107 - June 2022

1309

ARTICLE - Single-cell phospho-kinase signatures in pediatric T-ALL patterns by flow cytoametry. Cytometry A. 2014;85(1):78-87. 41. Mei HE, Leipold MD, Maecker HT. Platinum-conjugated antibodies for application in mass cytometry. Cytometry A. 2016;89(3):292-300. 42. Mei HE, Leipold MD, Schulz AR, Chester C, Maecker HT. Barcoding of live human peripheral blood mononuclear cells for multiplexed mass cytometry. J Immunol. 2015;194(4):2022-2031. 43. Schulz AR, Baumgart S, Schulze J, Urbicht M, Grützkau A, Mei HE. Stabilizing antibody cocktails for mass cytometry. Cytometry A. 2019;95(8):910-916. 44. Chevrier S, Crowell HL, Zanotelli VRT, Engler S, Robinson MD, Bodenmiller B. Compensation of signal spillover in suspension and imaging mass cytometry. Cell Syst. 2018;6(5):612-620. 45. Saeed AI, Sharov V, White J, et al. TM4: a free, open-source system for microarray data management and analysis. Biotechniques. 2003;34(2):374-378. 46. Szubert B, Cole JE, Monaco C, Drozdov I. Structure-preserving visualisation of high dimensional single-cell datasets. Sci Rep. 2019;9(1):8914. 47. Ding J, Condon A, Shah SP. Interpretable dimensionality reduction of single cell transcriptome data with deep generative models. Nat Commun. 2018;9(1):2002. 48. Sulis ML, Williams O, Palomero T, et al. NOTCH1 extracellular juxtamembrane expansion mutations in T-ALL. Blood. 2008;112(3):733-740. 49. Silva A, Yunes JA, Cardoso BA, et al. PTEN posttranslational inactivation and hyperactivation of the PI3K/Akt pathway sustain primary T cell leukemia viability. J Clin Invest. 2008;118(11):3762-3774. 50. Palomero T, Sulis ML, Cortina M, et al. Mutational loss of PTEN induces resistance to NOTCH1 inhibition in T-cell leukemia. Nat Med. 2007;13(10):1203-1210. 51. Schrappe M, Valsecchi MG, Bartram CR, et al. Late MRD response determines relapse risk overall and in subsets of childhood T-cell ALL: results of the AIEOP-BFM-ALL 2000 study. Blood. 2011;118(8):2077-2084. 52. Dang C V., O’Donnell KA, Zeller KI, Nguyen T, Osthus RC, Li F.

D. Kuzilková et al.

The c-Myc target gene network. Semin Cancer Biol. 2006;16(4):253-264. 53. Meyer N, Penn LZ. Reflecting on 25 years with MYC. Nat Rev Cancer. 2008;8(12):976-990. 54. Bonnet M, Loosveld M, Montpellier B, et al. Posttranscriptional deregulation of MYC via PTEN constitutes a major alternative pathway of MYC activation in T-cell acute lymphoblastic leukemia. Blood. 2011;117(24):6650-6659. 55. Piovan E, Yu J, Tosello V, et al. Direct reversal of glucocorticoid resistance by AKT inhibition in acute lymphoblastic leukemia. Cancer Cell. 2013;24(6):766-776. 56. Serafin V, Capuzzo G, Milani G, et al. Glucocorticoid resistance is reverted by LCK inhibition in pediatric T-cell acute lymphoblastic leukemia. Blood. 2017;130(25):2750-2761. 57. Delgado-Martin C, Meyer LK, Huang BJ, et al. JAK/STAT pathway inhibition overcomes IL7-induced glucocorticoid resistance in a subset of human T-cell acute lymphoblastic leukemias. Leukemia. 2017;31(12):2568-2576. 58. Ebinger S, Özdemir EZ, Ziegenhain C, et al. Characterization of rare, dormant, and therapy-resistant cells in acute lymphoblastic leukemia. Cancer Cell. 2016;30(6):849-862. 59. Martelli AM, Lonetti A, Buontempo F, et al. Targeting signaling pathways in T-cell acute lymphoblastic leukemia initiating cells. Adv Biol Regul. 2014;56:6-21. 60. Hall CP, Reynolds CP, Kang MH. Modulation of glucocorticoid resistance in pediatric T-cell acute lymphoblastic leukemia by increasing BIM expression with the PI3K/mTOR inhibitor BEZ235. Clin Cancer Res. 2016;22(3):621-632. 61. Coustan-Smith E, Mullighan CG, Onciu M, et al. Early T-cell precursor leukaemia: a subtype of very high-risk acute lymphoblastic leukaemia indentified in two independent cohorts. Lancet Oncol. 2009;10(2):147-156. 62. Basso G, Veltroni M, Valsecchi MG, et al. Risk of relapse of childhood acute lymphoblastic leukemia is predicted by flow cytometric measurement of residual disease on day 15 bone marrow. J Clin Oncol. 2009;27(31):5168-5174.

Haematologica | 107 - June 2022

1310

ARTICLE - Acute Myeloid Leukemia

AXL/MERTK inhibitor ONO-7475 potently synergizes with venetoclax and overcomes venetoclax resistance to kill FLT3-ITD acute myeloid leukemia Sean M. Post,1 Huaxian Ma,1,2 Prerna Malaney,1 Xiaorui Zhang,1 Marisa J.L. Aitken,1 Po Yee Mak,1,2 Vivian R. Ruvolo,1,2 Tomoko Yasuhiro,3 Ryohei Kozaki,3 Lauren E. Chan,1 Lauren B. Ostermann,1,2 Marina Konopleva,1 Bing Z. Carter,1,2 Courtney DiNardo,1 Michael D. Andreeff,1,2 Joseph D. Khoury,4# and Peter P. Ruvolo1,2# Department of Leukemia, 2Section of Molecular Hematology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 3Ono Pharmaceutical Co. Ltd., Research Center of Oncology, Osaka, Japan and 4Department of Hematopathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA 1

JDK and PPR contributed equally as co-senior authors.

Correspondence: Joseph D. Khoury JKhoury@mdanderson.org Sean M. Post SPost@mdanderson.org Received: January 13, 2021. Accepted: October 28, 2021. Prepublished: November 4, 2021. https://doi.org/10.3324 haematol.2021.278369 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract FMS-like Tyrosine Kinase 3 (FLT3) mutation is associated with poor survival in acute myeloid leukemia (AML). The specific Anexelekto/MER Tyrosine Kinase (AXL) inhibitor, ONO-7475, kills FLT3-mutant AML cells with targets including Extracellular-signal Regulated Kinase (ERK) and Myeloid Cell Leukemia 1 (MCL1). ERK and MCL1 are known resistance factors for Venetoclax (ABT-199), a popular drug for AML therapy, prompting the investigation of the efficacy of ONO-7475 in combination with ABT-199 in vitro and in vivo. ONO-7475 synergizes with ABT-199 to potently kill FLT3-mutant acute myeloid leukemia cell lines and primary cells. ONO-7475 is effective against ABT-199-resistant cells including cells that overexpress MCL1. Proteomic analyses revealed that ABT-199-resistant cells expressed elevated levels of pro-growth and anti-apoptotic proteins compared to parental cells, and that ONO-7475 reduced the expression of these proteins in both the parental and ABT-199-resistant cells. ONO-7475 treatment significantly extended survival as a single in vivo agent using acute myeloid leukemia cell lines and PDX models. Compared to ONO-7474 monotherapy, the combination of ONO-7475/ABT-199 was even more potent in reducing leukemic burden and prolonging the survival of mice in both model systems. These results suggest that the ONO-7475/ABT-199 combination may be effective for AML therapy.

Introduction Internal tandem duplication (ITD) mutation in FMS-like Tyrosine Kinase 3 (FLT3) occurs in ~25% of newly diagnosed acute myeloid leukemia (AML) patients and is associated with poor survival.1-4 Therapeutic strategies for FLT3-ITD AML patients include the use of broad tyrosine kinase inhibitors (TKIs) such as Midostaurin and Gilteritinib, whose selectivity includes FLT3.1 TKIs have limited efficacy and resistance mechanisms involving FLT3 alterations, such as D835, mutations make their use challenging.1-4 The B-cell lymphoma 2 (BCL2) inhibitor ABT-199 (venetoclax) is being evaluated in numerous clinical trials for AML and is typically combined with other agents as ABT-199 has limited effectiveness as a single agent.5,6 Similar to other BH3-mimetic drugs, ABT-199 induces cross-activation of extracellular-signal regulated kinase 1/2 (ERK1/2) and

downstream induction of myeloid cell leukemia-1 (MCL1); a survival factor for AML cells with FLT3-ITD mutations.7-13 Thus, agents that ablate FLT3 and ERK1/2 signaling and block BCL2 are critical therapies for FLT3ITD AML. A recent study using FLT3-ITD AML models demonstrated that Midostaurin or Gilteritinib synergize with ABT-199.14 However, Gilteritinib induces ERK activation and promotes RAS mutations while Midostaurin promotes the development of FLT3 point mutations.2,3,14,15 Therefore, novel treatment options that do not activate the RAS/ERK pathway or select for FLT3 alterations are needed. Various studies support a role for the tyrosine kinase receptor, Anexelekto (AXL), as a critical component in FLT3ITD signaling.16-19 ONO-7475 is a specific inhibitor of AXL and MER Tyrosine Kinase (MERTK), and our group has shown it potently kills FLT3-ITD AML cells.19 Notably, ONO7475 suppresses ERK1/2 phosphorylation and reduces

Haematologica | 107 - June 2022

1311

ARTICLE - TAM kinase inhibitor synergizes with venetoclax MCL-1 expression. Therefore, we hypothesized that ONO7475 can simultaneously target BCL2 and the FLT3 signaling pathways while mitigating the deleterious alterations reported with other TKIs.2,3,13,14 In this report, we examine the efficacy of the ONO-7475/ABT-199 combination using in vitro and in vivo models of AML and test the ability of ONO-7475 to overcome ABT-199 resistance or MCL-1 overexpression (either overexpression of wild-type (WT) or gain-of-function (GOF) mutant MCL-1) in FLT3-ITD cells.

Methods Cell lines MOLM13 and OCI-AML3 cells were purchased from DSMZ (Braunschweig, Germany). MV4;11 and THP-1 cells were purchased from ATCC (Manassas, VA, USA). MOLM13 luc/gfp cells were generated as previously described.19 MV4;11 cells rendered resistant to ABT-199 (MV4;11 VenR) were developed by long-term exposure to ABT-199, as previously described.10 MV4;11 stable cell lines overexpressing MCL-1 (wild-type and GOF mutant S159A) and MOLM13 cell lines overexpressing BCL2 (wildtype and phospho-mutants S70A and S70E) with corresponding control plasmids were created, as previously described.7 Cells were cultured at 21% O2 (normoxia) or 1% O2 (hypoxia) and 5% CO2 at 37°C. Reagents ONO-7475 was supplied by Ono Pharmaceutical Co. Ltd. (Osaka, Japan); ABT-199 was purchased from LC Laboratories (Woburn, MA, USA); UMI-77 was purchased from Selleck Chem (Houston, TX, USA); stock solutions for in vitro studies were prepared with DMSO (Sigma-Aldrich, St. Louis, MO, USA). For animal studies, ONO-7475 was prepared in 0.1% Tween80 and ABT-199 was prepared in 10% ethanol/30% phosphal 50/60% PEG 400. Cell viability Cells were incubated with vehicle or varying doses of ONO-7475 and/or ABT-199, then stained with Annexin V APC (BD Biosciences, San Jose, CA, USA) and DAPI (BD Biosciences, San Jose, CA, USA). Flow cytometry was performed using the Galios561 (Beckman Coulter, Brea, CA, USA). Further, cell viability was analyzed using the Vi-Cell Cell Viability Analyzer (Beckman Coulter). Immunoblot analysis Cells were incubated with vehicle ONO-7475 and/or ABT199 and then lysed. Total protein was fractionated by SDS/PAGE. Immunoblot analysis was performed with antibodies listed in the Online Supplementary Methods and imaged as previously described.20 Tubulin and b-actin were used as loading controls.

S.M. Post et al. qRT-PCR RNA was isolated and qRT-PCR performed as described in Online Supplementary Methods. Patient samples Primary FLT3-ITD AML samples were acquired in accordance with regulations and protocols approved by the Institutional Review Board of the MD Anderson Cancer Center (MDACC). Informed consent was obtained in accordance with the Declaration of Helsinki. Patient characteristics are provided in Online Supplementary Table S2. Human AML xenograft in vivo model Human xenograft experiments were approved by the Institutional Animal Care and Use Committee at MDACC. The efficacy of the ONO-7475/ABT-199 combination in an in vivo FLT3-ITD AML xenograft model was tested using MOLM13 cells expressing luciferase/GFP and the AML PDX model 3028566 in NSG mice. The AML PDX also harbored NPM1 and DNMT3A mutations. For both models, drugs were given five days a week by oral gavage (ONO-7475 at 10mg/kg and ABT-199 at 100mg/kg). Leukemia burden was assessed by IVIS imaging for MOLM13 and flow cytometry to measure human CD45+ cells for the PDX model. Statistical analyses All in vitro experiments were conducted in triplicate. The combination index (CI), based on the Chou–Talalay method21 and determined by CalcuSyn software (BIOSOFT), was expressed as CI values obtained at the effective doses (ED) of 50%, 75%, and 90% in the population exposed to the different agents. CI<1 was considered synergistic, CI=1 additive, and CI>1 antagonistic. Statistical differences between groups were determined using either a Student’s t-test or one-way ANOVA with Dunnett’s posttest. P≤0.05 was considered statistically significant.

Results Combination of ONO-7475 with ABT-199 is effective in vitro against FLT3-ITD AML cells MV4;11 cells were treated with vehicle or varying doses of ONO-7475 +/- ABT-199. After 72 hours, cell viability and cell number were determined by flow cytometry using Annexin V antibody and DAPI. Both ONO-7475 (10 nM) and ABT-199 (30 nM) showed similar efficacy as single agents with a ~60% reduction in cells (Figure 1A). While 10 nM ONO-7475 did not induce apoptosis, 30 nM ABT-199 resulted in ~35% apoptosis (Online Supplementary Figure 1A). Interestingly, the combination of 10 nM ONO-7475 with 30 nM ABT-199 nearly eliminated MV4;11 cells (~98% reduction of viable cells; Figure 1A) with >84% of cells undergoing apoptosis (Online Supplementary Figure 1A).

Haematologica | 107 - June 2022

1312

ARTICLE - TAM kinase inhibitor synergizes with venetoclax The combination of 50 nM ONO-7475 with 30 nM ABT- 199 reduced the viability of MV4;11 cells >99% (Figure 1A) and induced ~98% apoptosis (Online Supplementary Figure 1A). MOLM13 cells showed a 60% reduction in viable cells with 10 nM ONO-7475 with limited induction (<15%) of apoptosis (Figure 1B and Online Supplementary Figure 1B). Likewise, the combination of 50 nM ONO-7475 with 30 nM ABT-199 in MOLM13 cells resulted in >99% reduction in vi-

S.M. Post et al. ability (Figure 1B) with ~77% of cells undergoing apoptosis (Online Supplementary Figure 1B). For FLT3 WT THP-1 and OCI-AML3 cells, ONO-7475 did not significantly augment cell reduction or apoptotic induction by ABT-199 (Online Supplementary Figures S2A-2D). To determine whether ONO-7475 synergizes with ABT-199, MOLM13 and MV4;11 cells were treated with vehicle or varying doses of ONO-7475 and/or ABT-199 for 72 hours.

Figure 1. ONO-7475 synergizes with ABT-199 to potently kill FLT3-ITD AML cell lines and primary cells. MV4;11 and MOLM13 cells were treated with vehicle (0.2% DMSO), 10 nM or 50 nM ONO-7475, 10 nM or 30 nM ABT-199, or combinations of the two agents for 72 hours (MV4;11) or 48 hours (MOLM13). Viable total cells (A) for MV4;11; (B) for MOLM13 were determined by flow cytometry using counting beads, Annexin V, and DAPI. One-way ANOVA with Dunnett’s post-test was performed to determine significance (*P<0.033; **P<0.002; ***P<0.001). (C) MV4;11 cells and MOLM13 were treated with vehicle (0.2% DMSO), 10 nM, 25 nM, or 50 nM, ONO-7475 or 10 nM, 25 nM, or 50 nM ABT-199, or combinations of the two agents in a 1:1 ratio for 72 hours. Apoptosis was determined by flow cytometry using Annexin V and DAPI stains. CI index was determined using CalcuSyn software for combinations at effective dose (ED) 50, ED75, and ED90. CI < 1 was considered synergistic, CI = 1 additive, and CI > 1 antagonistic. (D) Peripheral blood cells from AML patients #28 and #32 were treated with vehicle (0.2% DMSO), various doses of ONO-7475 or various doses of ABT-199, or combinations of the two agents in a ratio of 1 nM ABT-199: 2.5 nM ONO-7475 ratio for 42 hours. Apoptosis was determined by flow cytometry using Annexin V and DAPI stains. Cells were gated for CD45+ cells using CD45 antibody. CI index was determined using CalcuSyn software for combinations at effective dose (ED) 50, ED75, and ED90. CI < 1 was considered synergistic, CI = 1 additive, and CI > 1 antagonistic. Haematologica | 107 - June 2022

1313

ARTICLE - TAM kinase inhibitor synergizes with venetoclax Synergy was determined by measuring CI, based on the Chou–Talalay method.21 CI values were obtained at the ED 50%, 75%, and 90% in the population exposed to the different agents. We observed that ONO-7475 synergizes with ABT-199, resulting in a CI index below 1.0 at ED50, ED75, and ED90 (Figure 1C). In fact, CI was <0.1 in both cell lines at ED75 and ED90. To determine the efficacy of ONO-7475/ABT-199 combination in primary FLT3-ITD AML cells, fresh cells from the peripheral blood of two primary AML samples were tested. Cells were treated for 48 hours with varying doses of ONO7475+/-ABT-199. As a single agent, ONO-7475 was slightly more effective at inducing apoptosis compared to ABT-199 in the sample from Patient #28 (Figure 1D). Importantly, ONO-7475 and ABT-199 acted synergistically to enhance apoptosis (CI=0.12, ED75 and 0.07, ED90, Figure 1D). Interestingly, ONO-7475 was ineffective as a single agent in cells derived from Patient #32 while ABT-199 alone promoted apoptosis (Figure 1D). However, when combined, ONO7475/ABT-199 synergized to activate apoptosis CI=0.25, ED75 and 0.16, ED90, Figure 1D). Given that AXL expression is, in part, regulated by hypoxia,22 we next evaluated the efficacy of these agents in 1% O2. Here, we observed that these combinations remained effective and synergistic in hypoxic conditions (Online Supplementary Figure S3A-B). We next interrogated alterations in MCL-1 expression and ERK phosphorylation following treatment with vehicle, single agent, or combinations of each agent (Online Supplementary Figure S4A). We observed that low-dose ONO7475 had minimal effects on MCL-1 expression in either cell line, although the ONO-7475/ABT- 199 combination did reduce MCL-1 in MV4;11 cells (Online Supplementary Figure S4A). While 10 nM ONO-7475 in combination with 30 nM ABT-199 effectively killed MOLM13 cells (Figure 1B and Online Supplementary Figure 1B), this combination did not affect MCL-1 expression (Online Supplementary Figure 4A), suggesting that suppression of MCL-1 is not required for ONO-7475/ABT-199-mediated killing. Interestingly, ERK phosphorylation was suppressed by 10 nM ONO-7475 alone and in combination with ABT-199 in both cell lines (Online Supplementary Figure 4A), suggesting ONO-7475 may synergize with ABT-199 to impede cell proliferation. ONO-7475 is equally effective against parental and ABT-199-resistant MV4;11 cell growth We previously developed MV4;11 cells that are >100 fold more resistant to ABT-199 than parental cells.10 MV4;11 parental and ABT-199-resistant cells were treated with vehicle or varying amounts of ONO-7475 for 72 hours. Both 10 nM and 50 nM ONO-7475 were equally efficient in reducing parental and ABT-199-resistant cell growth (Figure 2A) and inducing apoptosis (Figure 2B). To determine whether ONO-7475 synergizes with ABT-199 in MV4;11 ABT-199-resistant cells, cells were treated with

S.M. Post et al. vehicle, ONO-7475, and/or ABT-199 for 72 hours. Apoptosis was determined by flow cytometry. The highest tested dose of ABT-199 (5000 nM) had little effect on apoptosis in resistant cells while the highest tested dose of ONO7475 (50 nM) resulted in only modest apoptosis (Figure 2C). However, the combination of ONO-7475 and ABT-199 was able to induce greater apoptosis with lower doses: 50% apoptotic induction with 25 nM ONO- 7475/2500 nM ABT-199 and ~90% apoptosis with 50 nM ONO-7475/5000 nM ABT-199, (CI ED50, ED75, and ED90 below 0.25, Figure 2C). Increased MCL-1 expression is a known resistance mechanism for ABT-199 and interestingly, the treatment of ABT-199-resistant cells with UMI-77, an MCL-1 inhibitor, (and/or ONO-7475) also significantly reduced cell viability (Online Supplementary Figure S4B). Proteomic analyses reveal similar ONO-7475-dependent suppression of pro-growth and pro-survival targets in parental and ABT-199-resistant cells To determine ONO-7475’s effect on growth and survival signaling pathways in parental and ABT-199-resistant cells, we initially utilized RPPA to profile proteomes of MV4;11 parental and ABT-199-resistant cells following treatment. These analyses revealed alterations in protein expression between untreated MV4;11 parental and ABT199-resistant cells (Online Supplementary Figure S5A). Critically, untreated MV4;11 ABT-199-resistant cells had a significant increase in the expression of pro-growth and survival signaling compared to parental cells. In fact, BRAF levels, an upstream positive regulator of ERK pathways,23 were markedly higher in the resistant cells, suggesting this increase may stimulate survival pathways in these cells (Figure 3A). Further, we observed elevated expression of the pro-growth translational proteins, S6 and p-S6 (Figure 3C). We once again used RPPA to evaluate ONO-7475-mediated expression changes between resistant and parental cell lines. Following treatment with 100 nM ONO-7475 for 24 hours, most proteins were similarly altered in both the resistant and parental cell lines (Online Supplementary Figure S5B). Critically, we observed that ONO-7475 treatment significantly reduced the expression of B-RAF, p-S6, pERK, and MCL1 (Figure 3A-D). Our observations that p-S6, p-ERK, and MCL1 were suppressed in both resistant and parental lines following ONO-7475 treatment, suggest that ABT-199 resistances may be overcome by ONO-7475/AXL inhibition. Overexpression of BCL2 but not MCL-1 protects cells from ONO-7475 MCL-1 overexpression protects cells from ABT-199 and other BH3 mimetic agents.7,8,10 To evaluate the relationship between MCL-1 levels and sensitivity to ONO-7475, we used MV4;11 cells containing either an EV control plasmid, WT

Haematologica | 107 - June 2022

1314

ARTICLE - TAM kinase inhibitor synergizes with venetoclax

S.M. Post et al.

Figure 2. ONO-7475 impacts growth of ABT199-resistant MV4;11 cells and parental cells. MV4;11 parental and MV4;11 ABT-199-resistant cells were treated with vehicle (0.1% DMSO), 10 nM, 50 nM, or 100 nM ONO-7475 for 72 hours. Viable total cells (A) and percentage of apoptotic cells (B) were determined by flow cytometry using counting beads, Annexin V, and DAPI. Student’s t-test was performed to determine significance (*P<0.05; **P<0.001; ***P<0.0001). (C) MV4;11 ABT-199-resistant cells were treated with vehicle (0.2% DMSO), 10 nM, 25 nM, or 50 nM ONO-7475 or 1000 nM, 2500 nM, or 5000 nM ABT-199, or combinations of the two agents in ratio of 1 nM ONO7475: 100 nM ABT-199 for 72 hours. Apoptosis was determined by flow cytometry using Annexin V and DAPI stains. CI index was determined using CalcuSyn software for combinations at effective dose (ED) 50, ED75, and ED90. CI < 1 was considered synergistic, CI = 1 additive, and CI > 1 antagonistic.

MCL-1 or GOF MCL-1 mutant (S159A), which promotes MCL1 protein stability (Figure 4A).24,25 These cells were treated with varying doses of ABT-199 or ONO-7475 for 72 hours: EV-control cells were sensitive to 30 nM ABT-199 whereas both WT and GOF mutant MCL-1 OE cells were resistant to ABT-199 (Figures 4B and 4C). While, on average, 87% of EVcontrol cells were killed by ABT-199, only 32% of WT MCL1 OE cells and 19% of GOF mutant MCL-1 OE cells were. On the other hand, EV-control, WT MCL-1 OE, and GOF mutant MCL-1 OE cells were all similarly sensitive to ONO-7475 (Figure 4B). On average, induction of apoptosis in EV cells, WT MCL-1 cells and GOF mutant MCL-1 OE cells by 50 nM ONO-7475 is ~85% in all variants. The failure of MCL-1, including the GOF mutant, to protect these cells from ONO7475 or alter cellular AXL levels suggests that the ONO-7475 mechanism of action is independent of MCL-1 expression levels (Figure 4D). We previously observed that BCL2 overexpression protects cells from the BH3-mimetic ABT-737 and that BCL2 serine 70 phosphorylation augments this protection.26

Others have demonstrated that the BCL2 mutant S70A and S70E proteins more effectively protect cells from chemotherapeutic agents compared to WT BCL2.27 In our study, we tested the ability of exogenous overexpression of WT BCL2, BCL2S70A, and BCL2S70E to protect MOLM13 cells from ONO-7475. Total viable EV control MOLM13 cells treated with 30 nM ABT-199 were reduced by ~35% compared to vehicle-treated cells (Figure 5A). MOLM13 cells overexpressing exogenous WT or mutant BCL2 were slightly more resistant, demonstrating viable cell reductions of ~22%, 14%, and 8% for WT, S70A, and S70E, respectively. Additionally, WT BCL2 OE slightly protected cells from 10 nM ONO-7475, while mutant BCL2 OE rendered MOLM13 cells more resistant to this inhibitor (Figure 5A). Exogenous WT and mutant BCL2 similarly protected cells from apoptosis (Figure 5B). We next examined the effect of BCL2 overexpression on AXL expression. In MOLM13 cells, increased BCL2 expression resulted in AXL induction (Figure 5C and D). In fact, WT BCL2 OE or mutant BCL2 OE (both S70A and S70E) each mediated a

Haematologica | 107 - June 2022

1315

ARTICLE - TAM kinase inhibitor synergizes with venetoclax

S.M. Post et al.

Figure 3. ABT ONO-7475 suppresses pro-growth and pro-survival targets in parental and ABT-199-resistant cells. MV4;11 parental and ABT-199-resistant cells were treated with 100 nM ONO-7475 for 24 hours. Expression changes in B-RAF (A), ERK and p-ERK (B), S6 and p-S6 (C), and MCL-1 (D) in MV4;11 parental and resistant cells were determined via western blotting. Tubulin serves as a control. ABT-199-mediated and ONO-7475 dependent expression changes are represented in accompanying bar graphs. Densitometry was performed using ImageJ. One-way ANOVA with Dunnett’s post-test was performed to determine significance (*P<0.033; **P<0.002; ***P<0.001). Haematologica | 107 - June 2022

1316

ARTICLE - TAM kinase inhibitor synergizes with venetoclax

S.M. Post et al.

C Figure 4. MV4;11 cells overexpressing WT or gain of function S159A mutant MCL-1 are resistant to ABT-199 but are similarly sensitive to ONO7475 compared to control cells. MV4;11 control (Empty Vector; EV) and MV4;11 cells overexpressing WT MCL-1 or S159A mutant MCL-1 were treated with vehicle (0.1% DMSO), 10 nM, or 50 nM ONO7475 for 72 hours. One-way ANOVA with Dunnett’s posttest was performed to determine significance (*P<0.033; **P<0.002; ***P<0.001) In (A) western blot of MCL-1 expression levels in EV control, WT MCL-1 OE, and S159A mutant MCL-1 OE are depicted. Viable total cells (B) and percentage of apoptotic cells (C) were determined by flow cytometry using counting beads, Annexin V, and DAPI. (D) AXL expression in MV4;11 cells based on MCL-1 expression levels. Densitometry was performed by normalizing to tubulin using ImageJ.

Haematologica | 107 - June 2022

1317

ARTICLE - TAM kinase inhibitor synergizes with venetoclax significant increase in AXL expression that was alleviated by treatment with ABT-199 (Figure 5D). The combination of ONO-7475 and ABT-199 is effective in FLT3-ITD AML MOLM13 cell line xenograft and the FLT3-ITD AML PDX model To examine in vivo efficacy of this combination, we injected 6x105 MOLM13 luc/gfp cells into NSG mice (n=35). Once leukemia burden was established by IVIS imaging, the mice were split into four groups: control (0.1% Tween80; n=10), ONO-7475 (10 mg/kg; n=8); ABT-199 (100 mg/kg; n=8); and combination (n=9). The average survival of mice in the control group was 16 days; ABT-199 only extended survival by two days although this difference was statistically significant (P=0.001). ONO-7475 was effective alone and extended survival to 22 days (P<0.0001), however, the combination therapy was most efficacious (average survival 35 days, P<0.0001, Figure 6A). Throughout the treatment course, leukemia burden was monitored in five mice from each group via IVIS imaging (Online Supplementary Figure S6). As shown in Figure 6B, the combination of ONO-7475 and ABT-199 was effective in reducing the leukemia burden. After nine days, mean fold change in radiance increased 83-fold in vehicle-treated mice, 45fold with ONO-7475 alone, 102-fold with ABT-199 alone, and only 10-fold with the combination. After 16 days, the vehicle mice had all died. However, at this time, IVIS imaging revealed that the ONO-7475/ABT-199 combination group displayed a 27-fold increase in radiance (leukemic burden) while the ONO-7475 and ABT-199 cohorts had 258-fold and 610-fold increases, respectively (Figure 6B). We next tested the efficacy of the ONO-7475/ABT-199 combination in a FLT3-ITD AML PDX model. AML PDX cells (1 million) were injected into 20 NSG mice. After verification of leukemia burden by analytical flow cytometry for human CD45 (on average 10% human CD45+ cells), the mice were divided into four cohorts (n=5 each): vehicle (0.1% Tween80), 10 mg/kg ONO- 7475, 100 mg ABT-199, and combined therapy. ONO-7475 showed efficacy as a single agent, as the average survival was 60 days in the vehicletreated cohort compared to 86 days in the ONO-7475treated cohort (P=0.0026; Figure 6C). In the PDX model, ABT-199 did not significantly extend overall survival (P=0.0751). However, the combination of ONO-7475 with ABT-199 was especially effective as it extended survival by 39 days (P=0.0026). Throughout the course of treatment, the leukemia burden was assessed by measuring human CD45+ cells in the blood of mice at various intervals. As shown in Figure 6D, after six weeks, human AML cells in vehicle-treated mice were ~91%, while mice receiving the ONO-7475/ABT-199 combination had a leukemia burden in the blood of ~27%. After eight weeks, vehicle-treated mice had ~98% human CD45+ cells while mice receiving the ONO-7475/ABT-199 combination had

S.M. Post et al. only ~44% human CD45+ cells. We next examined the extent of tumor burden in tissues from the murine PDX transplantation model. Hematoxylin and Eosin (H&E) staining revealed significant splenic infiltration by leukemic cells characterized morphologically by moderate amounts of cytoplasm, irregular nuclear contours, open (blastic) chromatin patterns, and numerous mitotic figures (red arrowheads in Online Supplementary Figure S7A). In both the ONO-7475 and ABT-199 treated groups, mitotic figures, as well as increased numbers of apoptotic figures (relative to control), were observed. In addition, islands of native hematopoiesis were observed in the splenic parenchyma. However, in the ONO7475/ABT-199 combination group, there was a notable reduction of leukemic cells in the mouse spleen, which showed substantial predominance of native hematopoiesis, including megakaryocytes and admixed granulocytic and erythroid precursors. To confirm the origin of the malignant cells, immunohistochemistry was performed using a human-specific anti-Ku-80 antibody that specifically recognizes the human Ku-80 protein (Online Supplementary Figure S7B). Ku-80 staining demonstrated a higher proportion of human leukemic cells in the control group compared to the ONO-7475 treated group.

Discussion We recently demonstrated that ONO-7475 was effective as a single agent against FLT3-ITD AML cells.19 The mechanism of ONO-7475 action involved diverse targets including the inhibition of survival kinases such as ERK1/2, MCL-1, and many cell cycle regulators. Considering that MCL-1 and ERK1/2 are major factors in ABT-199 resistance,7-10 the ability of ONO-7475 to overcome ABT-199 resistance in FLT3-ITD AML cells is an uncharacterized and intriguing area of research. ONO-7475 as a single agent was effective against both MV4;11 and MOLM13 cells. However, its combination with ABT-199 demonstrated significant synergy and resulted in near elimination of the leukemic cells in both cell lines (Figure 1A and B and Online Supplementary Figure S1). These same synergistic effects were also observed in primary patient samples (Figure 1D). While there were some differences in apoptosis between patient samples when single agents were used (ONO-7475, more effective in Patient #28 or ABT199 more effective in Patient #32), both samples were extremely sensitive to the combination of the two agents. Additionally, the notion that ONO-7475 may be useful in FLT3-ITD (or TKI-resistant) AML is highlighted by the cellular sensitivity of patient #28 compared to patient #32; this patient harbored a FLT3 D835 mutation which render cells resistant to the majority of TKIs in the clinic.1 While treatment of mutant FLT3 is challenging in AML, new

Haematologica | 107 - June 2022

1318

ARTICLE - TAM kinase inhibitor synergizes with venetoclax

S.M. Post et al.

Figure 5. Overexpression of WT or gain of function S70A or S70E BCL2 mutants induces AXL expression and protects MOLM13 cells from ONO-7475. MOLM13 control (Empty Vector; EV) and MOLM13 cells overexpressing WT BCL2 or S70A mutant or S70E mutant BCL2 were treated with vehicle (0.1% DMSO), 10 nM, or 50 nM ONO-7475 for 72 hours. Viable total cells (A) and percentage of apoptotic cells (B) were determined by flow cytometry using counting beads, Annexin V, and DAPI. One-way ANOVA with Dunnett’s post-test was performed to determine significance (*P<0.033; **P<0.002; ***P<0.001). (C) Expression of BCL2 WT, S70A mutant, or S70E mutant was determined in MOLM13 cells. RNA and protein were isolated and BCL2 expression was monitored by qRT-PCR and western blot, respectively. (D) Expression of AXL was determined in MOLM13 cells expressing WT, S70A mutant, or S70E mutant BCL2. RNA and protein were isolated and BCL2 and AXL expression was monitored by qRT-PCR and western blot, respectively. Densitometry was performed using ImageJ. Haematologica | 107 - June 2022

1319

ARTICLE - TAM kinase inhibitor synergizes with venetoclax agents that target the FLT3/AXL axis have shown efficacy. As such, Gilteritinib, a dual FLT3/AXL inhibitor, has shown activity in FLT3-mutant AM but it has been shown to activate ERK signaling and promote RAS mutations. Thus, novel agents, such as ONO-7475s that specifically target AXL signaling, may be useful to augment (or used in a resistant setting) as it does not result in ERK activation. Previously, we demonstrated that low-dose ONO-7475 directly impacts the expression of pro-growth proteins.19 Here, we demonstrate that p-ERK levels were reduced in an ONO-7475-dependent manner (Figure 3B). This is a critical observation as ERK signaling supports MCL-1 stability via phosphorylation of threonine 92 and 163, resulting in MCL-1 stabilization.28,29 Thus, the mechanisms by

S.M. Post et al. which ERK signaling and AXL inhibitors impact the regulation of MCL-1 warrants further investigation. To better understand how ABT-199 resistance impacts cellular programs, we analyzed differences between the proteomic profile of parental and ABT-199 resistance MV4;11 cells. Several cell cycle and pro-survival proteins were altered in the untreated parental and resistant cell lines. While many of these alterations portend for poor outcomes, they may also offer potential targets in ABT-199-resistant AML (Figure 3A-D and Online Supplementary Figures S4 and S5). A comparison of proteomic profiles of MV4;11 parental and MV4;11 ABT-199-resistant cells treated with 100 nM ONO-7475 for 24 hours revealed that the agent had similar effects on targets in both cell lines. Here, we ob-

Figure 6. ONO-7475 combination is effective in reducing leukemia burden and promoting mouse survival in xenograft models using a FLT3-ITD AML cell line and a FLT3-ITD AML PDX. MOLM13 luc/gfp cells were injected into NSG mice. After leukemia engraftment, determined by IVIS imaging, mice were treated with vehicle (Tween 80), 10 mg/kg ONO-7475, 100 mg/kg ABT-199 or a combination of both agents. (A) Survival data with median survival (days) is indicated for each group. (B) Measurement of fold change radiance during the course of treatment using IVIS imaging software. (C) Survival data with median survival (days) is indicated for each group. AML PDX model 3028566 cells were injected into NSG mice. After leukemia engraftment determined by flow cytometry for human CD45-positive cells, mice were treated with vehicle (Tween 80), 10 mg/kg ONO-7475, 100 mg/kg ABT199 or a combination of both agents. (D) Measurement of leukemia burden by measuring CD45-positive cells by flow cytometry during the course of treatment. Haematologica | 107 - June 2022

1320

ARTICLE - TAM kinase inhibitor synergizes with venetoclax served that ONO-7475 similarly suppressed cell cycle regulators (e.g., p-ERK, S6, p-S6, and B-RAF) and pro-survival proteins (e.g., MCL1) in parental and ABT-199-resistant cells (Figure 3 and Online Supplementary Figures S4 and S5). These results suggest that ONO-7475 remains effective in inhibiting potential ABT-199-resistant pathways. MCL-1 upregulation is a known resistance mechanism for ABT-199. Consistently, we observed that MV4;11 cells overexpressing WT or GOF mutant MCL-1 were resistant to ABT-199 compared to EV-control cells. However, ONO7475 was equally as effective reducing viable cell numbers and inducing apoptosis in EV-control cells and cells that overexpress either WT or GOF mutant MCL-1 (Figure 4B and 4C). Interestingly, overexpression of WT BCL2 or GOF mutants did impart some protection against ONO-7475 (Figure 5A and 5B), as cells overexpressing BCL2 have elevated AXL levels compared to EV-control cells (Figure 5C and 5D). Our observation that BCL2 overexpression causes increased levels of AXL (Figure 5C and 5D) seems to indicate that BCL2 may regulate AXL levels at the transcriptional level. While the mechanism for this upregulation is poorly understood, it is plausible that AXL expression is potentially regulated through the BCL2/RAS/RAF axis. This notion is supported by the findings that RAF is a target for BCL2 binding.30,31 and that RAS regulates AXL in drug resistance studies in solid tumors.32-34 These observations warrant further investigation. To investigate the in vivo efficacy of the ONO-7475/ABT199 combination, we utilized both cell line xenograft and FLT3-ITD AML PDX models. While ONO-7475 alone was efficacious in every setting, the combination was significantly more potent in reducing leukemia burden and prolonging the survival of mice compared to either single agent in both models. In the MOLM13 model, IVIS imaging revealed that mice treated with both agents had a much lower leukemia burden compared to vehicle-treated mice or mice treated with either single agent (Figure 6B and Online Supplementary Figure S6). Importantly, the combination of ONO-7475 and ABT-199 significantly prolonged survival (Figure 6A). Similar results were observed in an AML FLT3-ITD PDX model, where mice given both ONO7475 and ABT-199 survived longer compared to vehicletreated or either single agent and displayed reduced leukemia burden as observed by human CD45+ cells (Figure 6C and 6D). It is worth noting that the PDX model could, possibly, represent a high-risk AML model in view of the concomitant presence of FLT-ITD, NPM1, and DNMT3A mutations, as reported previously.35 Several clinical studies have demonstrated a favorable role for ABT-199 in AML therapy, but primary resistance

S.M. Post et al. and subsequent treatment failure remain common. In a recent study, primary or subsequent resistance to venetoclax-based treatment regimens appeared to associate with the presence of activating mutations involving FLT3 or the RAS-ERK pathway.36 This result, coupled with our findings of activation of the RAS pathway, suggests a potential resistance mechanism that may be overcome by ONO-7475. Further, our results suggest AXL/MERTK inhibition effectively synergizes with ABT-199 in AML and bypasses various mechanisms of ABT-199 resistance. In summary, our results show that inhibition of the AXL/MERTK axis exerts a synergistic effect alongside BCL2 inhibition in AML and can overcome the survival advantages conferred by MCL-1 overexpression, a common mechanism of ABT-199 resistance. These results provide pre-clinical support for exploring this strategy in AML patients. Disclosures Ono Pharmaceutical supported the work performed in this study. Tomoko Yasuhiro and Ryohei Kozaki are employees of Ono Pharmaceutical. Contributions PR and SP designed experiments, performed research, analyzed data, and wrote the manuscript; PM, HM, MA, PYM, XZ, VR, and BC performed research, and analyzed data; MK, CD, MA and JK analyzed data and helped edit the manuscript; TY and RK provided compound, helped with experimental design, and helped edit the manuscript. Acknowledgments The authors would like to express their profound respect and admiration for Dr Peter P. Ruvolo who passed away as this manuscript was going through peer review. The work in this study would not have taken place were it not for Peter’s tireless work and dedication to science and discovery. He is survived by his wife, sister, and numerous family members. Funding This work was supported by Ono Pharmaceutical Co. Ltd., Research Center of Oncology, Osaka, Japan. Flow Cytometry (MA) and RPPA Cores are funded by NCI Cancer Center Support Grant P30CA16672. S.M.P. is supported by funding from the National Cancer Institute/National Institutes of Health Award (R01CA207204) and Leukemia and Lymphoma Society (6577–19); P.M. is supported by grants from the Jane Coffin Childs Medical Trust Fund and the American Society of Hematology. We would like to thank Rithica Deepak and Oscar Benitez for carefully reviewing this manuscript.

Haematologica | 107 - June 2022

1321

ARTICLE - TAM kinase inhibitor synergizes with venetoclax

S.M. Post et al.

References 1. Short NJ, Kantarjian H, Ravandi F, Daver N. Emerging treatment paradigms with FLT3 inhibitors in acute myeloid leukemia. Ther Adv Hematol. 2019;10:2040620719827310. 2. Roskoski R Jr. The role of small molecule Flt3 receptor proteintyrosine kinase inhibitors in the treatment of Flt3-positive acute myelogenous leukemias. Pharmacol Res. 2020;155:104725. 3. Kiyoi H, Kawashima N, Ishikawa Y. FLT3 mutations in acute myeloid leukemia: Therapeutic paradigm beyond inhibitor development. Cancer Sci. 2020;111(2):312-322. 4. Smith CC. The growing landscape of FLT3 inhibition in AML. Hematology Am Soc Hematol Educ Program. 2019;2019(1):539547. 5. Lachowiez C, DiNardo CD, Konopleva M. Venetoclax in acute myeloid leukemia – current and future directions. Leuk Lymphoma. 2020;61(6):1313-1322. 6. Konopleva M, Pollyea DA, Potluri J, et al. Efficacy and biological correlates of response in a Phase II study of venetoclax monotherapy in patients with acute myelogenous leukemia. Cancer Discov. 2016;6(10):1106-1117. 7. Pan R, Hogdal LJ, Benito JM, et al. Selective BCL-2 inhibition by ABT-199 causes on-target cell death in acute myeloid leukemia. Cancer Discov. 2014;4(3):362-375. 8. Pan R, Ruvolo VR, Wei J, et al. Inhibition of Mcl-1 with the panBcl-2 family inhibitor (-)BI97D6 overcomes ABT-737 resistance in acute myeloid leukemia. Blood. 2015;126(3):363-372. 9. Lin KH, Winter PS, Xie A, et al. Targeting MCL-1/BCL-XL forestalls the acquisition of resistance to ABT-199 in acute myeloid leukemia. Sci Rep. 2016;6:27696. 10. Pan R, Ruvolo V, Mu H, et al. Synthetic lethality of combined Bcl-2 inhibition and p53 activation in AML: mechanisms and superior antileukemic efficacy. Cancer Cell. 2017;32(6):748-760. 11. Ramsey HE, Fischer MA, Lee T, et al. A novel MCL1 inhibitor combined with venetoclax rescues venetoclax-resistant acute myelogenous leukemia. Cancer Discov. 2018;8(12):1566-1581. 12. Wei AH, Roberts AW, Spencer A, et al. Targeting MCL-1 in hematologic malignancies: Rationale and progress. Blood Rev. 2020;44:100672. 13. Yoshimoto G, Miyamoto T, Jabbarzadeh-Tabrizi S, et al. FLT3-ITD up-regulates MCL-1 to promote survival of stem cells in acute myeloid leukemia via FLT3-ITD-specific STAT5 activation. Blood. 2009;114(24):5034-5043. 14. Ma J, Zhao S, Qiao X, et al. Inhibition of Bcl-2 synergistically enhances the antileukemic activity of Midostaurin and Gilteritinib in preclinical models of FLT3-mutated acute myeloid leukemia. Clin Cancer Res. 2019;25(22):6815-6826. 15. McMahon CM, Ferng T, Canaani J, et al. Clonal selection with RAS pathway activation mediates secondary clinical resistance to selective FLT3 inhibition in acute myeloid leukemia. Cancer Discov. 2019;9(8):1050-1063. 16. Park IK, Mishra A, Chandler J, et al. Inhibition of the receptor tyrosine kinase Axl impedes activation of the FLT3 internal tandem duplication in human acute myeloid leukemia: implications for Axl as a potential therapeutic target. Blood. 2013;121(11):2064-2073. 17. Ben-Batalla I, Schultze A, Wroblewski M, et al. Axl, a prognostic and therapeutic target in acute myeloid leukemia mediates paracrine crosstalk of leukemia cells with bone marrow stroma. Blood. 2013;122(14):2443-2452. 18. Park IK, Mundy-Bosse B, Whitman SP, et al. Receptor tyrosine kinase Axl is required for resistance of leukemic cells to FLT3targeted therapy in acute myeloid leukemia. Leukemia.

2015;29(12):2382-2389. 19. Ruvolo PP, Ma H, Ruvolo VR, et al. Anexelekto/MER tyrosine kinase inhibitor ONO-7475 arrests growth and kills FMS-like tyrosine kinase 3-internal tandem duplication mutant acute myeloid leukemia cells by diverse mechanisms. Haematologica. 2017;102(12):2048-2057. 20. Gallardo M, Lee HJ, Zhang X, et al. hnRNP K Is a haploinsufficient tumor suppressor that regulates proliferation and differentiation programs in hematologic malignancies. Cancer Cell. 2015;28(4):486-499. 21. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul. 1984;22:27-55. 22. Dumas P-Y, Naudin C, Martin-Lannerée S, et al. Hematopoietic niche drives FLT3-ITD acute myeloid leukemia resistance to quizartinib via STAT5-and hypoxia-dependent upregulation of AXL. Haematologica. 2019;104(10):2017-2027. 23. McCubrey JA, Steelman LS, Chappell WH, et al. Ras/Raf/MEK/ERK and PI3K/PTEN/Akt/mTOR cascade inhibitors: how mutations can result in therapy resistance and how to overcome resistance. Oncotarget. 2012;3(10):1068-1111. 24. Domina AM, Vrana JA, Gregory MA, et al. MCL1 is phosphorylated in the PEST region and stabilized upon ERK activation in viable cells, and at additional sites with cytotoxic okadaic acid or taxol. Oncogene. 2004;23(31):5301-5315. 25. Ruvolo PP. GSK-3 as a novel prognostic indicator in leukemia. Adv Biol Regul. 2017;65:26-35. 26. Konopleva M, Contractor R, Tsao T, et al. Mechanisms of apoptosis sensitivity and resistance to the BH3 mimetic ABT-737 in acute myeloid leukemia. Cancer Cell. 2006;10(5):375-388. 27. Dai H, Ding H, Meng XW, et al. Contribution of Bcl-2 phosphorylation to Bak binding and drug resistance. Cancer Res. 2013;73(23):6998-7008. 28. Ding Q, Huo L, Yang J-Y, et al. Down-regulation of myeloid cell leukemia-1 through inhibiting Erk/Pin 1 pathway by sorafenib facilitates chemosensitization in breast cancer. Cancer Res. 2008;68(15):6109-6117. 29. Thomas LW, Lam C, Edwards SW. Mcl-1; the molecular regulation of protein function. FEBS Lett. 584(14);2981-2989. 30. Deng X, Kornblau SM, Ruvolo PP, et al. Regulation of Bcl2 phosphorylation and potential significance for leukemic cell chemoresistance. J Natl Cancer Inst Monogr. 2001;(28):30-37. 31. Wang HG, Rapp UR, Reed JC. Bcl-2 targets the protein kinase Raf-1 to mitochondria. Cell. 1996;87(4):629-638. 32. Nelson-Taylor SK, Le AT, Yoo M, et al. Resistance to RETinhibition in RET-rearranged NSCLC is mediated by reactivation of RAS/MAPK signaling. Mol Cancer Ther. 2017;16(8):1623-1633. 33. Bockorny B, Rusan M, Chen W, et al. RAS-MAPK reactivation facilitates acquired resistance in FGFR1-amplified lung cancer and underlies a rationale for upfront FGFR-MEK blockade. Mol Cancer Ther. 2018;17(7):1526-1539. 34. Brand TM, Iida M, Stein AP, et al. AXL mediates resistance to cetuximab therapy. Cancer Res. 2014;74(18):5152-5164. 35. Wang YW, Tsai CH, Lin CC, et al. Cytogenetics and mutations could predict outcome in relapsed and refractory acute myeloid leukemia patients receiving BCL-2 inhibitor venetoclax. Ann Hematol. 2020;99(3):501-511. 36. DiNardo CD, Tiong IS, Quaglieri A, et al. Molecular patterns of response and treatment failure after frontline venetoclax combinations in older patients with AML. Blood. 2020;135(11):791-803.

Haematologica | 107 - June 2022

1322

ARTICLE - Bone Marrow Failure

Transforming growth factor-b signaling modifies the hematopoietic acute inflammatory response to drive bone marrow failure Jose Javier,1,2 Ashwini Hinge,1† James Bartram,1,2 Juying Xu1 and Marie-Dominique Filippi1

Correspondence:

Division of Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Research Foundation, Department of Pediatrics, University of Cincinnati College of Medicine and 2College of Medicine, University of Cincinnati, Cancer and Cell Biology graduate program, Cincinnati, OH, USA

Marie-Dominique Filippi Marie-Dominique.Filippi@cchmc.org

†

In memoriam

Received: October 6, 2020. Accepted: October 15, 2021. Prepublished: October 28, 2021. https://doi.org/10.3324 haematol.2020.273292 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract Bone marrow failure syndromes are characterized by ineffective hematopoiesis due to impaired fitness of hematopoietic stem cells. They can be acquired during bone marrow stress or innate and are associated with driver genetic mutations. Patients with a bone marrow failure syndrome are at higher risk of developing secondary neoplasms, including myelodysplastic syndromes and leukemia. Despite the identification of genetic driver mutations, the hematopoietic presentation of the disease is quite heterogeneous, raising the possibility that non-genetic factors contribute to the pathogenesis of the disease. The role of inflammation has emerged as an important contributing factor, but remains to be understood in detail. In this study, we examined the effect of increased transforming growth factor-b (TGFb) signaling, in combination or not with an acute innate immune challenge using polyinosinc:polycytidilic acid (pIC), on the hematopoietic system without genetic mutations. We show that acute rounds of pIC alone drive a benign age-related myeloid cell expansion and increased TGFb signaling alone causes a modest anemia in old mice. In sharp contrast, increased TGFb signaling plus acute pIC challenge result in chronic pancytopenia, expanded hematopoietic stem and progenitor cell pools, and increased bone marrow dysplasia 3-4 months after stress, which are phenotypes similar to human bone marrow failure syndromes. Mechanistically, this disease phenotype is uniquely associated with increased mitochondrial content, increased reactive oxygen species and enhanced caspase-1 activity. Our results suggest that chronic increased TGFb signaling modifies the memory of an acute immune response to drive bone marrow failure without the need for a preexisting genetic insult. Hence, non-genetic factors in combination are sufficient to drive bone marrow failure.

Introduction Bone marrow failure (BMF) syndromes are rare hematologic diseases characterized by impaired fitness of hematopoietic stem cells (HSC) and ineffective hematopoiesis, resulting in the absence of one or more hematopoietic lineages in the peripheral blood.1 BMF syndromes can be inherited2 or induced by inflammatory stress, allogenic or autologous HSC transplantation, myeloablative chemotherapy, or abnormal activation of the auto-immune T-cell system.3,4 Most BMF syndromes and myelodysplastic syndromes (MDS) are associated with genetic mutations, in particular

in epigenetic regulators. Nevertheless, BMF syndromes are quite heterogeneous disorders raising interest in understanding what factors contribute to the disease development, in addition to driver mutations. Substantial clinical data show that hyperactivity of inflammatory cytokines, including tumor necrosis factor-a, interleukin-6, and transforming growth factor–b (TGFb), as well as innate immune signaling pathways contribute to the pathogenesis of BMF.5,6 In particular, the TGFb signaling pathway is known to be hyperactive in Fanconi anemia,7 MDS,8-11 Shwachman-Diamond syndrome,12 and myelofibrosis.13 TGFb1 is a myelosuppressive cytokine: it is secreted as an inactive protein complex bound to latency-associated

Haematologica | Vol 7 - June 2022

1323

ARTICLE - TGFb and inflammation in bone marrow failure peptide (LAP), and the latent TGF-beta1 binding protein-1 (LTBP1). Dissociation from the complex is necessary for biological activity. Active TGFb1 proteins then signal through two serine/threonine kinases, type I and type II receptors and trigger several signaling pathways.14 The functions of TGFb1 are complex and highly context-dependent.15 In the hematopoietic system, TGFb1 inhibits or promotes cell growth, varying from cell to cell and in a dose-dependent manner. TGFb controls HSC quiescence,16,17 homing, and survival.18 During aging, TGFb1 can promote the expansion of a myeloid-biased HSC population.16 TGFb can also suppress erythropoiesis and myelopoiesis. Work in our laboratory previously showed that TGFb contributes to a functional decline in HSC in murine transplant models by promoting HSC differentiation at the expense of self-renewal.19 We found that TGFb signaling, including canonical p-Smad2 and non-canonical pp38MAPK, remains high in HSC after bone marrow transplantation in mice due to increased expression of the active form of TGFb in HSC. Interestingly, pharmacological inhibition of TGFb signaling during bone marrow reconstitution following transplantation improved HSC functions, suggesting that this increased TGFb signaling causes HSC functional decline after bone marrow transplantation.19 In mouse models of BMF or Fanconi anemia, pharmacological inhibition of TGFb signaling also restores effective hematopoiesis in vivo.7 Moreover, inhibitors of TGFb signaling have shown promising results in improving hematopoiesis in MDS patients.20 However, only 30% of patients respond to the treatment, indicating that other factors contribute to BMF.8,10,20 Dysregulation of several innate immune pathways are also known factors contributing to BMF or MDS. Toll-like receptors or their signaling effectors are often overexpressed in MDS samples compared to healthy controls, enhancing a type I interferon response through NFkB, MAPK, and IRF3.5 Other innate immune pathways, including the inflammasome and the necrosome, are also elevated in patients with BMF or MDS and contribute to ineffective hematopoiesis.21,22 Although inflammatory pathways are strongly implicated in human BMF syndromes and MDS, their exact contribution to the pathogenesis of the diseases remains to be understood. It is still unclear whether deregulated TGFb signaling and/or inflammation are secondary events that contribute to the pathogenesis of the disease or can initiate the disease, and if so in which context.23 In this study, we used a transgenic conditional mouse model overexpressing constitutively active TGFb1 (aTGFb1)19 to further investigate the role of TGFb1 in BMF/MDS. We show that a physiological increase in aTGFb1 production in the bone marrow only produces mild neutropenia and anemia in mice during aging. However, the combination of increased TGFb signaling plus polyinosinic:polycytidilic acid (pIC)driven acute inflammatory stress drives chronic BMF with

J. Javier et al. phenotypes similar to those of the human disorders associated with ineffective hematopoiesis, including BMF syndromes and MDS. Mechanistically, TGFb prevents the termination of an acute pIC response causing permanent alteration in mitochondrial functions and increased caspase-1 activity. Our findings therefore suggest that BMF syndromes can be initiated solely by multiple inflammatory hits in the context of increased TGFb signaling, and that disease outcome is dependent on the inflammatory context. Increased TGFb signaling plus pIC thus represents a novel non-genetic-driven mouse model of human BMF-like diseases.

Methods Mouse model Transgenic Tg-b1glo+/Flox mice (Jackson Labs, Stock 018393) were crossed with Mx1-Cre mice to generate MxCre+; Tgb1glo+/Flox (TgCre+) and MxCre−; Tg-b1glo+/Flox mice (TgCre–). Cre recombinase expression was induced with three injections of 10 mg/kg/mouse pIC (GE Healthcare), every other day. pIC-stressed mice were allowed to recover for at least 4 weeks prior to reinjection with the same pIC regimen. All animals were bred at an in-house, pathogenfree facility, and all studies were conducted with protocols approved by the Animal Care Committee of Cincinnati Children’s Hospital Medical Center. Flow cytometry Peripheral blood samples were stained with CD45 PerCPCy5.5/APC Cy7, B220 APC/PE Cy7, Gr1 Alexa Fluor 700, Mac1/CD11b, CD3e APC/PE, CD4 PE, and CD8a APC/PE. Whole bone marrow cells were stained as above for mature lineages. Cells were also stained for Lineage– Sca1+ Kit+ CD48– CD150+ (LSK SLAM) with biotin-conjugated anti-mouse lineage antibodies (Ter119, B220, Gr1, CD11b, CD3e) followed by staining for streptavidin V500/eF450 (eBioscience [eF450]), c-Kit APC eF780/APC (eBioscience [APC eF780]), Sca1 PE Cy7, CD48 AF700/BV605 (Biolegend [AF700]) CD150 APC/PE (eBioscience [APC]; Fisher Scientific [PE]). Bone marrow cells were further stained with CD16/32 PE (eBioscience) and CD34 eF450 (eBioscience) to immunostain for committed progenitor populations. For mitochondrial function, cells immunostained for LSK SLAM were incubated at 37°C in 5% CO2 for 30 min with either MitoSOX Deep Red Reagent (1 mM, Invitrogen) or tetramethylrhodamine ester (0.1 mM, Sigma Aldrich). Caspase 1 activity was determined using the FLICA assay (Corning), according to the manufacturer’s recommendations. Samples were then analyzed using a BD LSR II, BD LSR Fortessa, or BD Canto III (BD Biosciences). All antibodies were obtained from BD Biosciences, unless otherwise noted.

Haematologica | Vol 7 - June 2022

1324

ARTICLE - TGFb and inflammation in bone marrow failure Enzyme-linked immunosorbent assays Active TGFb1 was assessed in bone marrow fluid using the Mouse TGF-beta 1 DuoSet ELISA Kit (R&D Systems) and DuoSet ELISA Ancillary Reagent Kit 1 (R&D Systems). Bone marrow and spleen histology Tissues were fixed with 10% formalin, and stained with hematoxylin and eosin. Whole bone marrow cells were also prepared by cytospin and stained using Kwik-Diff (Fisher Scientific). RNA sequencing cDNA from 500 SLAM HSC was made using the Smart-seq v4 Ultra Low Input RNA Kit (Takara/Clontech). A barcoded DNA library was then made using the Nextera XT DNA Library Preparation Kit (Illumina). Sequencing was done by the Cincinnati Children’s Hospital Medical Center core and Alt-analyze was used for the analyses,24 as previously described.25,26 Differentially expressed genes were then analyzed using the ENRICHR database.27 Immunofluorescence assays SLAM HSC were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton-X 100, then immunostained for mitochondria using rabbit anti-Tomm20 conjugated to Alexa Fluor 555 (Abcam). The cells were then mounted with Slowfade Glass with DAPI (Invitrogen). Images were taken using the Nyquist limit setting (0.1 mm XY pixel size) at 100X magnification. Images were analyzed using the surface building Matlab extension in Imaris software for at least 30 cells in each group. Statistical analyses Experiments were done in two or three replicates unless specified otherwise. Statistical analyses were performed using an unpaired Student t-test, unless specified otherwise.

Results Increased TGFb1 causes mild neutropenia with age To further investigate the role of increased TGFb signaling in hematopoiesis, we crossed a transgenic mouse model containing a transgene of aTGFb1 with an Mx1-Cre mouse. The transgene contains two point mutations (C223S and C225S) to prevent the inhibitory latent associated peptide (LAP) from binding upon expression, permitting the produced ligand to initiate downstream signaling immediately after expression (Figure 1A).28 Offspring containing both Mx1-Cre and aTGFb1 constructs (TgCre+) were then used for further experiments, with mice containing the transgene but without Cre (TgCre–) as controls – thus preventing confounding issues associated with the leakiness of Mx1-Cre

J. Javier et al. (Online Supplementary Figure S1A).29 Mice were injected with pIC 10 mg/kg/mouse three times, once every 48 h, to cause aTGFb1 overexpression (Figure 1B). We previously confirmed that aTGFb1 levels are higher in bone marrow LSK cells (Lin-Sca1+c-Kit+) from TgCre+ mice.19 Because TGFb ligands are secreted in the extracellular matrix of tissues, we here examined levels of aTGFb1 that were released in the bone marrow microenvironment. Bone marrow fluid of TgCre+ mice contained higher levels of aTGFb1 compared to control mice, up to 6 months following aTGFb overexpression (Figure 1C). In this model, aTGFb1 levels were in the range of 50 to 100 pg/mL. Levels of the activated form of the canonical TGFb signaling transcription factor Smad2 in phenotypically identified SLAM HSC (Lin-Sca1+cKit+CD48-CD150+) 3-4 weeks after inducing overexpression were higher in TgCre+ SLAM HSC (Figure 1D). No signs of fibrosis were observed in the bone marrow of these mice up to 6 months after inducing overexpression (Online Supplementary Figure S1B). Hence, in this model, the increased TGFb1 levels remain physiological, not higher than in some clinical MDS samples and much lower than in other aTGFb1-overexpressing mouse models associated with myelofibrosis.9,30-32 To understand the impact of increased TGFb signaling on the hematopoietic system, mice were analyzed for up to 12 months after overexpression had been induced (Figure 1B). Differential blood count analysis indicated that mice overexpressing aTGFb1 developed modest neutropenia and anemia by 6 to 12 months of age compared to control animals (Figure 1E). Examination of the peripheral blood by flow cytometry analysis confirmed a mild reduction in myeloid cell frequency but increased T-cell frequency (Online Supplementary Figures S1C and S2A). We next examined the consequence of aTGFb1 overexpression in the bone marrow and spleen 3 months after TGFb overexpression. Bone marrow cellularity was not significantly different between the groups (Online Supplementary Figure S1D). However, TgCre+ mice had splenomegaly (Online Supplementary Figure S1E). Hematoxylin and eosin staining of sections of the femur and spleen taken from TgCre+ and TgCre– mice showed no gross abnormalities (Online Supplementary Figure S1F). Further analysis indicated that the total count of each hematopoietic stem or progenitor cell population in bone marrow was similar between TgCre+ and TgCre– mice during aging (Online Supplementary Figures S1G, H and S2). At 6 months, a modest dysplasia in bone marrow neutrophils and myeloid progenitors of TgCre+ mice was observed (Figure 1F-H, Online Supplementary Figure S3). A small percentage of neutrophils was hyper-nucleated; some promyelocyte/myelocytes had increased size with a higher cytoplasm to nucleus ratio (Figure 1F-H, Online Supplementary Figure S3). This effect did not persist and was no longer observed at 12 months, perhaps due to

Haematologica | Vol 7 - June 2022

1325

ARTICLE - TGFb and inflammation in bone marrow failure

J. Javier et al.

Figure 1. Mice overexpressing active TGFb1 do not develop overt hematopoietic phenotypes during steady-state hematopoiesis. (A) Schema of the mouse model of active TGFb1 (aTGFb1) overexpression. (B) Schema of the experiments. (C) Levels of aTGFb1 protein in bone marrow fluid from mice 3 weeks (3 mice/group) and 6 months (7 mice/group) after aTGFb1 overexpression determined using an enzyme-linked immunosorbent assay. (D) Analysis of the TGFb signaling effector phospho-Smad2 using immunofluorescence in LSK SLAM cells 4 weeks after aTGFb1 overexpression, 50 cells/mouse, three mice/group. (E) Differential peripheral blood counts at indicated time points following aTGFb1 overexpression. Five mice/group (3 months), seven TgCre- mice, eight TgCre+ mice (6 months), seven mice/group (12 months). (F) Wright-Giemsa staining of bone marrow cytospins 6 and 12 months after aTGFb1 overexpression, three mice/group. (G) Frequency of indicated cell types in bone marrow, three mice/group. (H) Frequency of dysplastic cells within each population, three mice/group. All experiments were conducted at least twice, with data shown as the mean ± standard error of the mean (SEM). Statistical significance was assessed using an independent Student t-test. ***P<0.001, **P<0.01, *P<0.05. ELISA: enzyme-linked immunosorbent assay; IF: immnofluorescence; WBC: white blood cells; RBC: red blood cells. Haematologica | Vol 7 - June 2022

1326

ARTICLE - TGFb and inflammation in bone marrow failure confounding aging effects. Hence, aTGFb1 overexpression causes mild hematopoietic defects. Mice overexpressing active TGFb1 show significant long-term bone marrow failure after acute polyinosinic:polycytidilic stress Several studies have previously demonstrated the role of innate immune signaling in the development of BMF and MDS.5,6 The effects of multiple inflammatory stressors are however poorly defined. We examined the effect of acute innate immune stress in the context of enhanced TGFb signaling. To this end, mice were challenged again with three injections of pIC, 1 month following aTGFb overexpression (Figure 2A) and then analyzed 2 and 90 days later (Figure 2A). Mice re-challenged with pIC were termed TgCre– pIC+ and TgCre+ pIC+, respectively. Remarkably, compared to TgCre– pIC+ mice, TgCre+ pIC+ mice developed significant peripheral blood pancytopenia, including neutropenia, lymphocytopenia and thrombocytopenia, beginning 3 months after pIC stress (Figure 2B). Persistent anemia was also noted in TgCre+ pIC+ mice and this was more pronounced than without pIC challenge (Figure 2B). TgCre+ pIC+ mice had larger spleens with an expansion of the white pulp (Online Supplementary Figure S4A, B), whereas total bone marrow cell count and density remained unchanged compared to those of control mice (Online Supplementary Figure S4C, D). We then examined bone marrow parameters in response to pIC stress. The acute response to pIC was comparable between the groups. Each group showed an increase in multipotent progenitors (MPP, Lineage– Sca1+ Kit+ CD48+) whereas SLAM numbers were unchanged 2 days after pIC challenge (Figure 2C). Granulocyte/macrophage progenitors (GMP, Lineage+ Sca1– Kit+ CD34+ CD16/32+) increased; common myeloid progenitors (CMP, Lineage+ Sca1– Kit+ CD34+ CD16/32–) and megakaryocyte/ erythrocyte progenitors (MEP, Lineage+ Sca1–Kit+ CD34-CD16/32-) decreased (Online Supplementary Figure S4E). Interestingly, 90 days after pIC stress, TgCre+ mice had higher numbers of MPP and SLAM HSC compared to control mice, whereas MEP numbers remained lower (Figure 2C, Online Supplementary Figure S3E). Furthermore, TgCre+ pIC+ mice had an increased frequency of myeloblasts and promyelocytes but reduced frequency of mature neutrophils in the bone marrow (Figure 2D,E), which was associated with a significant degree of myeloid cell dysplasia, including hypersegmented neutrophils, and increased cell size and cytoplasm to nuclear ratio in myeloblasts and promyelocytes (Figure 2D,F, Online Supplementary Figure S4F). This myeloid dysplasia is reminiscent of bone marrow cytology present in mouse models of MDS.33 Taken together, our data suggest that TgCre+ pIC+ mice develop ineffective hematopoiesis which is characterized by an expanded hematopoietic stem or progenitor cell pool,

J. Javier et al. pancytopenia and myeloid cell dysplasia. Thus, multiple inflammatory hits – increased TGFb signaling plus acute pIC challenge – together cause a disease that recapitulates features of human BMF/MDS-like diseases, suggesting that non-genetic factors can initiate the onset of long-lasting BMF/MDS disorders. Double-stranded RNA and enhanced TGFb signaling cause permanent changes in the gene expression profile of hematopoietic stem cells To understand how an acute pIC challenge in the context of enhanced TGFb signaling causes long-lasting ineffective hematopoiesis, we analyzed the global transcriptome profile of four groups of SLAM HSC: (i) from 3-month-old TgCre- mice, (ii) from 3-month-old TgCre+ mice, (iii) from TgCre- mice 3 months after pIC re-challenge and (iv) from TgCre+ mice 3 months after pIC re-challenge. For the pIC re-challenge groups, we chose to collect the SLAM HSC when pancytopenia began to manifest, i.e., 3 months after pIC re-challenge (Figure 3A). Data were analyzed using unsupervised principle component analysis and supervised hierarchical clustering in AltAnalyze®.25,26 Principal component analysis separated the four groups of cells into four distinct clusters, indicating that each group possesses a unique transcriptional signature (Figure 3B). Hierarchical clustering indicated that pIC re-challenge profoundly altered the transcriptional landscape in SLAM HSC from both TgCre- and TgCre+ mice in comparison to controls (Figure 3C), even 3 months following the transient pIC challenge. A large number of these differentially expressed genes were downregulated by pIC challenge and belonged to chromosome organization, mitochondrion, and cell cycle (Figure 3C, D - cluster 1). Cluster 2 represents genes that were upregulated by pIC; these genes mostly relate to mitochondrion and the respiratory chain complex (Figure 3C, E). Genes highlighted in cluster 3 relate to signal transduction, and were downregulated by pIC challenge, but more so in TgCre+ SLAM HSC. Finally, specific differences in gene expression between TgCre- or TgCre+ SLAM HSC after pIC challenge were noted, and are underscored by white boxes (Figure 3C). Thus, a transient pIC challenge caused long-lasting transcriptional changes in HSC. A more detailed examination of which genes are differentially expressed in TgCre+ SLAM HSC specifically after pIC challenge revealed that they belong to two main categories, interferon response genes and nuclear-encoded genes related to mitochondrial regulation. The MarkerFinder algorithm in AltAnalze identified that myeloid and innate immune genes, including TLR2/4/6 co-receptor cd14, cxcl10, anxa3, olfm4, and s100a8/s100a9 were upregulated in TgCre– pIC+ SLAM HSC but not in TgCre+ pIC+ SLAM HSC (Figure 3D, F, and Online Supplementary Figure S5A), thus correlating with the increase in peripheral blood myeloid cells in these

Haematologica | Vol 7 - June 2022

1327

ARTICLE - TGFb and inflammation in bone marrow failure

J. Javier et al.

Figure 2. Mice overexpressing active TGFb1 develop ineffective hematopoiesis with cell dysplasia after acute pIC stress. (A) Schema of the experiment. Mice were challenged with three injections of pIC 10 mg/kg/mouse every other day, performed at least 4 weeks after the induction of active TGFb1 (aTGFb1) overexpression. (B) Differential peripheral blood counts. Six mice/group. (C) Bone marrow cell counts of LSK, LSK CD48+ and LSK SLAM. Five mice/group (No stress), seven mice (2 days), eight TgCre- mice, and nine TgCre+ mice (90 days). (D) Wright-Giemsa staining of bone marrow cells 3-4 months after pIC stress. Normal arrows denote hyper-lobulated neutrophils; block arrows denote dysplastic myeloblasts; arrow heads denote dysplastic erythroblasts. Three mice/group. (E) Frequency of indicated cell types in bone marrow, three mice/group. (F) Frequency of dysplastic cells within each population. Three mice/group. Experiments were performed at least twice, with data shown as the mean ± standard error of the mean (SEM). Statistical significance was assessed using an independent Student t-test. ***P<0.001, **P<0.01, *P<0.05. WBC: white blood cells; RBC: red blood cells.

Haematologica | Vol 7 - June 2022

1328

ARTICLE - TGFb and inflammation in bone marrow failure

J. Javier et al.

Figure 3. Active TGFb1-overexpressing SLAM hematopoietic stem cells display a unique transcriptional signature long-term following acute pIC stress. (A) Schematic of the workflow of the transcriptomic analysis from SLAM hematopoietic stem cells (HSC) before and 3 months after pIC stress. (B) Principle component analysis visualization. (C) Hierarchical clustering of differentially regulated genes using pairwise comparative analysis. Columns represent cell populations. Rows represent genes. Three mice/group. (D-F) Top gene ontology categories of differentially expressed genes identified in cluster 1 (D), cluster 2 (E) and cluster 3 (F). Haematologica | Vol 7 - June 2022

1329

ARTICLE - TGFb and inflammation in bone marrow failure

J. Javier et al.

mice (Figure 2). We specifically interrogated differential gene expression of the interferon-a and -b signaling pathway. These genes were more downregulated in TgCre+ SLAM HSC (Online Supplementary Figure S5B). We also interrogated genes related to mitochondrial regulation. Genes important for the regulation of mitochondrial translation such as mrpl46 were upregulated only in TgCre+ SLAM HSC after pIC stress, (Figure 3D, F and Online Supplementary Figure S5C). On the other hand, genes encoding regulators of mitophagy were downregulated in TgCre+ SLAM HSC after pIC stress (Online Supplementary Figure S5C). These findings strongly suggest that pIC causes significant and permanent transcriptional changes in SLAM HSC, some of which are modified only by aTGFb1 overexpression.

lations from both TgCre– pIC+ and TgCre+ pIC+ mice (Figure 4C). In the longer term (3 months), mitochondrial membrane potential returned to baseline in TgCre– pIC+ SLAM HSC, MPP and CP populations. Interestingly, mitochondrial membrane potential remained high in TgCre+ pIC+ SLAM HSC but was reduced in TgCre+ pIC+ CP. Finally, we examined total cellular levels of ROS using CellROX staining. All hematopoietic cell populations, i.e, CP, MPP and SLAM HSC, from TgCre+ pIC+ mice displayed increased ROS levels compared to those from TgCre– pIC+ mice (Figure 4D). Increased ROS in TgCre– pIC+ HSC SLAM did not necessarily come from mitochondria since mitochondrial-driven superoxide levels, as assessed using MitoSOX Red dye, were not different between the groups (data not shown). Caspase-1 activation can be induced by mitochondrial activation or intracellular ROS.39,40 In other cell types, pIC can trigger the activation of caspase 1. We thus examined the effect of aTGFb overexpression on pIC-induced caspase-1 activity using the FAM-FLICA caspase-1 assay. TgCre+ pIC+ SLAM HSC had sustained caspase-1 activity compared to TgCre– pIC+ SLAM HSC, as indicated by increased caspase1 in TgCre+ pIC+ SLAM HSC up to 3 months after pIC challenge, compared to controls (Figure 4E). These data suggest that aTGFb1-overexpressing SLAM HSC maintain more active mitochondria, and have increased ROS levels and caspase-1 activity compared to control cells in the long term following acute pIC stress.

Active TGFb1-overexpressing SLAM hematopoietic stem cells show aberrant mitochondrial polarization and increased caspase 1 activity long-term following pIC stress To functionally validate the gene expression findings, we first examined nuclear localization of IRF3, representing the active form of IRF3, and confirmed that IRF3 was not chronically activated in HSC from TgCre+ pIC+ mice (Online Supplementary Figure S5D). We then focused on mitochondria, as suggested by the transcriptional profile of TgCre+ pIC+ SLAM HSC. Mitochondria have emerged as a central platform for the activation of intracellular innate immune responses, including the inflammasome, which can be activated in response to pIC.34,35 These immune responses are known to depend on and subsequently to alter mitochondrial metabolism. Interestingly, several groups have shown that samples from MDS patients exhibit abnormal mitochondrial functions, including increased cellular reactive oxygen species (ROS) and hyperpolarized mitochondria.36-38 Alteration in nuclear-encoding mitochondrial gene expression was also found to be predictive of the development of secondary MDS after chemotherapy or bone marrow transplantation. We first examined mitochondrial content by immunostaining for the mitochondrial outer membrane protein Tomm20 and performing high resolution z-stacked imaging and three-dimensional reconstruction analyses. Mitochondrial content was similar in SLAM HSC from both TgCre– and TgCre+ mice before pIC stress (Figure 4A, B). However, 3 months after pIC stress, TgCre+ pIC+ SLAM HSC had higher mitochondrial content compared to control cells. This is consistent with a gene expression signature of elevated regulators of mitochondrial biogenesis and reduced regulators of mitophagy. Mitochondrial membrane potential, analyzed using tetramethylrhodamine ester, showed several differences between the groups. Shortly after pIC re-challenge (2 days and 7 days), mitochondrial membrane potential increased in SLAM HSC, MPP and the committed progenitor pool (CP, Lineage- Sca1- Kit+) popu-

Discussion In this study, we found that while a physiological and chronic increase in TGFb signaling alone has little impact on the hematopoietic system, an additional but acute insult with pIC leads to long-lasting ineffective hematopoiesis that closely resembles the chronic BMF associated with myelodysplasia. Mechanistically, acute pIC imposes permanent transcriptional changes in HSC, which, in the context of increased TGFb signaling, are associated with a long-lasting increase in mitochondrial content, hyperpolarized mitochondria, increased intracellular ROS and caspase-1 activity. These results imply that inflammatory stresses are sufficient to cause long-lasting BMF/MDS-like disorders without the need for driver mutations. These findings may also provide insights into the causes of BMF/MDS-like disease heterogeneity in which disease outcome may vary with specific combinations of inflammatory insults and dosage of insult. Finally, these findings also have long-term implications for the use of combinatorial therapies for treating human BMF syndromes. Inflammation has long been associated with acquired BMF syndromes.5,6 There is also a strong correlation between inflammation (regardless of cause, duration and frequency) and the development of MDS. Independent

Haematologica | Vol 7 - June 2022

1330

ARTICLE - TGFb and inflammation in bone marrow failure

J. Javier et al.

Figure 4. Active TGFb1-overexpressing SLAM hematopoietic stem cells display sustained mitochondrial activity and caspase-1 activity long-term following acute pIC stress. (A,B) Mitochondrial content assessed in SLAM hematopoietic stem cells (HSC) using Tomm20 immunostaining. (A) Representative immunofluorescence images (Tomm20 in red, DAPI in blue). (B) Quantification of Tomm20 mean fluorescent intensity (MFI). Fifty cells from each mouse, six mice/group. (C) Mitochondrial membrane potential was assessed using tetra-methyl rhodamine ester dye (TMRE) staining at the indicated time after pIC stress. Four mice/group (0 days after pIC stress/no stress), Five TgCre-, six TgCre+ (2 days post stress), six TgCre-and seven TgCre+ (7 days after stress). Six mice/group (90 days after pIC stress). (D) Intracellular reactive oxygen species were measured using the CellROX Deep Red Reagent. Five mice/group (no stress), six mice/group (3 months after pIC stress). (E) Active caspase 1 was measured using the FAM-FLICA 660 kit. Five mice/group (no stress), eight TgCre-, nine TgCre+ mice (3 months after pIC stress). Data are from at least two independent experiments and statistics were performed using an independent Student t-test. ***P<0.001, **P<0.01, *P<0,05. MFI: mean fluorescent intensity; AU: arbitrary units. Haematologica | Vol 7 - June 2022

1331

ARTICLE - TGFb and inflammation in bone marrow failure

J. Javier et al.

studies have demonstrated that innate immune signaling is responsible for phenotypes of some MDS subtypes, including del5q MDS. TGFb signaling is a key driver of MDS and has been implicated in aplastic anemia, Fanconi anemia and Shwachman-Diamond syndrome.7,9,11,12,41 Previous studies exploring the relationship between TGFb signaling and BMF have used a similar aTGFb1-overexpressing construct but under the control of the albumin promoter.20,30 In the albumin-aTGFb1 mouse model, aTGFb1 overexpression produces concentrations much higher than in the mouse model used in our study, and higher than those found in MDS patients.30 In this model, mice developed severe anemia, megakaryocyte dysplasia and marrow reticulin fibrosis within 3 weeks postpartum. Although informative, the acute presentation of the hematopoietic defects of this model prevented long-term assessment of the effects of a chronic increase in aTGFb1 on disease development. TGFb functions in a dose-dependent manner. In the hematopoietic system, low aTGFb1 concentrations (pg/mL) stimulate HSC proliferation, whereas higher concentrations (ng/mL) are inhibitory.42,43 Our model suggests that a modest increase in TGFb signaling alone is not sufficient to drive severe BMF during steady-state hematopoiesis. Interestingly, an added acute innate immune signal allows a persistent HSC response leading to BMF. In the WT context, acute pIC challenge seems to cause an accelerated aging phenotype, at least related to myeloid expansion. In the context of enhanced TGFb signaling, acute pIC challenge causes a BMF/MDS-like syndrome. These findings mean that disease outcome depends on a specific combination of inflammatory stressors, supporting the emerging hypothesis of the multiple inflammatory hit hypothesis to explain heterogeneity in BMF syndromes and MDS. Crosstalk between pIC and TGFb signaling in the development of BMF was previously described in the context of Fanconi anemia, which is caused by mutations in DNA repair proteins via homologous recombination. Milsom’s group dissected the response of HSC to pIC. They showed that in response to pIC, HSC exiting from quiescence sustain DNA damage that can be resolved by the Fanconi anemia-mediated DNA repair response. As such, WT mice recover from pIC stress. In contrast, Fanca-deficient mice had reduced numbers of HSC, unresolved DNA damage and developed severe BMF.44 Interestingly, Zhang et al. showed that enhanced TGFb signaling, known to be upregulated in patients with Fanconi anemia, contributes to pIC-induced BMF in Fanca-deficient mice by modifying the DNA repair response to pIC-induced DNA damage. When TGFb signaling is high, HSC use error-prone nonhomologous end-joining instead of homologous recombination, thus favoring DNA mutations. In this model, inhibition of TGFb signaling rescued hematopoiesis in pICtreated Fanca-deficient mice.7 Collectively, these findings support the idea that TGFb is a ‘modifier’ of HSC functions,

which predisposes to the development of BMF/MDS. It also raises the interesting possibility that TGFb-modified pIC-induced DNS damage could contribute to our phenotype. This will be interesting to examine further. The fact that an acute pIC challenge causes long-lasting effects in HSC is also notable. It means that pIC can induce longlasting transcriptional memory in HSC. It will be interesting to examine whether this resembles the recently described trained immunity phenomenon,45-47 whether other innate immune insults similarly synergize with TGFb1 in disease development and what are the exact mechanisms behind this synergy. TGFb signaling is mostly known to act though canonical Smad signaling and non-canonical p38 MAPK signaling. Our data suggest that increased TGFb signaling alters mitochondrial function and caspase-1 activity after pIC stress. The association between altered mitochondria and BMF/MDS is not unprecedented. Studies have demonstrated that Fanconi anemia patients have altered mitochondria48 and respiratory chain defects.49 Mitochondrial diseases themselves have hematologic phenotypes of varying degrees, such as Pearson syndrome, which presents with pancytopenia, and Barth syndrome, which presents with neutropenia. A study of patients with MDS or acute myeloid leukemia revealed transcriptional dysregulation of their mitochondria.50 Several studies have also implicated mitochondrial dysfunction and increased intracellular ROS in driving the refractory anemia associated with MDS.36-38 We recently reported that aberrant mitochondrial function is one source of abnormal HSC function.26 Thus, our study supports the current knowledge that impaired or altered mitochondrial function contributes to ineffective hematopoiesis. Our study further suggests that there may be direct links between mitochondrial dysfunction and altered TGFb signaling. It remains to be seen how TGFb causes mitochondrial defects and how those defects contribute to ineffective hematopoiesis. One possibility is enhancing mitochondrial biogenesis, perhaps via Myc, which has been involved in BMF.51 Another possibility would be abnormal activation of caspase-1, which can contribute to BMF.21,52 Intracellular ROS, perhaps as a result of abnormal mitochondria, may be responsible for sustained caspase-1 activation.39,40 The elevation of intracellular ROS in aTGFb1-overexpressing SLAM HSC after pIC stress in our mouse model may therefore not only have direct genotoxic effects, but may also synergize with and amplify pIC-mediated caspase-1 activation to drive BMF. The functional outcome of enhanced caspase-1 activation in HSC remains unclear. In our model, it is unlikely that the outcome is only cell death since the SLAM HSC pool is in fact expanded in TgCre+ pIC+ mice. While caspase-1 is known to cause cell death by pyroptosis, other studies have shown that caspase-1 also controls glycolysis during Salmonella typhimurium infection by tar-

Haematologica | Vol 7 - June 2022

1332

ARTICLE - TGFb and inflammation in bone marrow failure geting and cleaving key glycolytic enzymes such as aldolase, triose-phosphate isomerase and α-enolase.53 Activated caspase-1 can also induce the activation of sterol regulatory element binding proteins, responsible for regulating lipid membrane biogenesis, to favor cell survival instead of causing cell death.54 A careful examination of the role of caspase-1 activity in HSC is therefore needed. In conclusion, we have described a mouse model of BMF that results from an acute inflammatory challenge in the context of increased TGFb signaling. This model recapitulates phenotypes of human BMF syndromes that are linked to TGFb signaling. This mouse model will help not only to further our understanding of the pathogenesis of BMF syndromes and MDS associated with TGFb signaling but also provides an in vivo model to test effects of combinatorial therapy to cure these disorders.

J. Javier et al. conducted the investigations; M-DF was responsible for the formal analysis; JJ and M-DF wrote the original draft, reviewed and edited it; JJ and M-DF acquired funding for the study; M-DF provided supervision. Acknowledgments We thank the mouse core staff, Jeff Bailey and Victoria Summey, for bone marrow transplants, the flow cytometry core for assistance in cell sorting at Cincinnati Children’s Hospital Medical Center, and the confocal imaging core for assistance in immunofluorescence imaging analyses. We also thank Ellen Javier for technical assistance. Funding The work was supported by the NIH (DK102890 [to MDF]); a Pilot Innovative Project (CCHMC [to MDF]), DOD (BM 190093 [to MDF]); and a Pelotonia fellowship [to JJ].

Disclosures No conflicts of interest to disclose.

Data availability All data are available, including scRNA-sequencing accesContributions sion codes and a Gene Expression Omnibus accession JJ, AH and M-DF conceived the study; JJ, AH, JB, and JX number. Figures 1-4 have associated raw data. Raw images were responsible for the methodology; JJ, AH, JB, and JX are also available upon request

References 1. Zeidan AM, Shallis RM, Wang R, Davidoff A, Ma X. Epidemiology of myelodysplastic syndromes: why characterizing the beast is a prerequisite to taming it. Blood Rev. 2019;34:1-15. 2. Kallen ME, Dulau-Florea A, Wang W, Calvo KR. Acquired and germline predisposition to bone marrow failure: diagnostic features and clinical implications. Semin Hematol. 2019;56(1):69-82. 3. Maciejewski JP, O'Keefe C, Gondek L, Tiu R. Immune-mediated bone marrow failure syndromes of progenitor and stem cells: molecular analysis of cytotoxic T cell clones. Folia Histochem Cytobiol. 2007;45(1):5-14. 4. Glenthoj A, Orskov AD, Hansen JW, Hadrup SR, O'Connell C, Gronbaek K. Immune mechanisms in myelodysplastic syndrome. Int J Mol Sci. 2016;17(6):944. 5. Barreyro L, Chlon TM, Starczynowski DT. Chronic immune response dysregulation in MDS pathogenesis. Blood. 2018;132(15):1553-1560. 6. Demerdash Y, Kain B, Essers MAG, King KY. Yin and yang: the dual effects of interferons on hematopoiesis. Exp Hematol. 2021;96:1-12. 7. Zhang H, Kozono DE, O'Connor KW, et al. TGF-beta inhibition rescues hematopoietic stem cell defects and bone marrow failure in Fanconi anemia. Cell Stem Cell. 2016;18(5):668-681. 8. Bewersdorf JP, Zeidan AM. Transforming growth factor (TGF)beta pathway as a therapeutic target in lower risk myelodysplastic syndromes. Leukemia. 2019;33(6):1303-1312. 9. Bhagat TD, Zhou L, Sokol L, et al. miR-21 mediates hematopoietic suppression in MDS by activating TGF-beta signaling. Blood. 2013;121(15):2875-2881. 10. Geyh S, Rodriguez-Paredes M, Jager P, et al. Transforming growth factor beta1-mediated functional inhibition of mesenchymal stromal cells in myelodysplastic syndromes and

acute myeloid leukemia. Haematologica. 2018;103(9):1462-1471. 11. Zhou L, McMahon C, Bhagat T, et al. Reduced SMAD7 leads to overactivation of TGF-beta signaling in MDS that can be reversed by a specific inhibitor of TGF-beta receptor I kinase. Cancer Res. 2011;71(3):955-963. 12. Joyce CE, Saadatpour A, Ruiz-Gutierrez M, et al. TGFbeta signaling underlies hematopoietic dysfunction and bone marrow failure in Shwachman-Diamond syndrome. J Clin Invest. 2019;130(9):3821-3826. 13. Chagraoui H, Komura E, Tulliez M, Giraudier S, Vainchenker W, Wendling F. Prominent role of TGF-beta 1 in thrombopoietininduced myelofibrosis in mice. Blood. 2002;100(10):3495-3503. 14. Massague J, Gomis RR. The logic of TGFbeta signaling. FEBS Letters. 2006;580(12):2811-2820. 15. Hinge A, Filippi MD. Deconstructing the complexity of TGFbeta signaling in hematopoietic stem cells: quiescence and beyond. Curr Stem Cell Rep. 2016;2(4):388-397. 16. Brenet F, Kermani P, Spektor R, Rafii S, Scandura JM. TGFbeta restores hematopoietic homeostasis after myelosuppressive chemotherapy. J Exp Med. 2013;210(3):623-639. 17. Yamazaki S, Ema H, Karlsson G, et al. Nonmyelinating Schwann cells maintain hematopoietic stem cell hibernation in the bone marrow niche. Cell. 2011;147(5):1146-1158. 18. Capron C, Lacout C, Lecluse Y, et al. A major role of TGF-beta1 in the homing capacities of murine hematopoietic stem cell/progenitors. Blood. 2010;116(8):1244-1253. 19. Hinge A, Xu J, Javier J, et al. p190-B RhoGAP and intracellular cytokine signals balance hematopoietic stem and progenitor cell self-renewal and differentiation. Nat Commun. 2017;8:14382. 20. Zhou L, Nguyen AN, Sohal D, et al. Inhibition of the TGF-beta receptor I kinase promotes hematopoiesis in MDS. Blood. 2008;112(8):3434-3443.

Haematologica | Vol 7 - June 2022

1333

ARTICLE - TGFb and inflammation in bone marrow failure 21. Basiorka AA, McGraw KL, Eksioglu EA, et al. The NLRP3 inflammasome functions as a driver of the myelodysplastic syndrome phenotype. Blood. 2016;128(25):2960-2975. 22. Wagner PN, Shi Q, Salisbury-Ruf CT, et al. Increased Ripk1mediated bone marrow necroptosis leads to myelodysplasia and bone marrow failure in mice. Blood. 2019;133(2):107-120. 23. Migliaccio AR. A vicious interplay between genetic and environmental insults in the etiology of blood cancers. Exp Hematol. 2018;59:9-13. 24. Emig D, Salomonis N, Baumbach J, Lengauer T, Conklin BR, Albrecht M. AltAnalyze and DomainGraph: analyzing and visualizing exon expression data. Nucleic Acids Res. 2010;38(Web Server issue):W755-762. 25. Olsson A, Venkatasubramanian M, Chaudhri VK, et al. Single-cell analysis of mixed-lineage states leading to a binary cell fate choice. Nature. 2016;537(7622):698-702. 26. Hinge A, He J, Bartram J, et al. Asymmetrically segregated mitochondria provide cellular memory of hematopoietic stem cell replicative history and drive HSC attrition. Cell Stem Cell. 2020;26(3):420-430. 27. Chen EY, Tan CM, Kou Y, et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics. 2013;14:128. 28. Hall BE, Zheng C, Swaim WD, et al. Conditional overexpression of TGF-beta1 disrupts mouse salivary gland development and function. Lab Invest. 2010;90(4):543-555. 29. Velasco-Hernandez T, Sawen P, Bryder D, Cammenga J. Potential pitfalls of the Mx1-Cre system: implications for experimental modeling of normal and malignant hematopoiesis. Stem Cell Rep. 2016;7(1):11-18. 30. Sanderson N, Factor V, Nagy P, et al. Hepatic expression of mature transforming growth factor beta 1 in transgenic mice results in multiple tissue lesions. Proc Natl Acad Sci U S A. 1995;92(7):2572-2576. 31. Shehata M, Schwarzmeier JD, Hilgarth M, Hubmann R, Duechler M, Gisslinger H. TGF-beta1 induces bone marrow reticulin fibrosis in hairy cell leukemia. J Clin Invest. 2004;113(5):676-685. 32. Akiyama T, Matsunaga T, Terui T, et al. Involvement of transforming growth factor-beta and thrombopoietin in the pathogenesis of myelodysplastic syndrome with myelofibrosis. Leukemia. 2005;19(9):1558-1566. 33. Zhou T, Kinney MC, Scott LM, Zinkel SS, Rebel VI. Revisiting the case for genetically engineered mouse models in human myelodysplastic syndrome research. Blood. 2015;126(9):10571068. 34. Breda CNS, Davanzo GG, Basso PJ, Saraiva Camara NO, MoraesVieira PMM. Mitochondria as central hub of the immune system. Redox Biol. 2019;26:101255. 35. Kim SJ, Ahn DG, Syed GH, Siddiqui A. The essential role of mitochondrial dynamics in antiviral immunity. Mitochondrion. 2018;41:21-27. 36. Goncalves AC, Alves V, Silva T, Carvalho C, Oliveira CR, SarmentoRibeiro AB. Oxidative stress mediates apoptotic effects of ascorbate and dehydroascorbate in human myelodysplasia cells in vitro. Toxicol In Vitro. 2013;27(5):1542-1549. 37. Greenberg PL, Young NS, Gattermann N. Myelodysplastic syndromes. Hematology Am Soc Hematol Educ Program. 2002:136-161. 38. Picou F, Vignon C, Debeissat C, et al. Bone marrow oxidative

J. Javier et al. stress and specific antioxidant signatures in myelodysplastic syndromes. Blood Adv. 2019;3(24):4271-4279. 39. Franchi L, Eigenbrod T, Munoz-Planillo R, Nunez G. The inflammasome: a caspase-1-activation platform that regulates immune responses and disease pathogenesis. Nat Immunol. 2009;10(3):241-247. 40. Harijith A, Ebenezer DL, Natarajan V. Reactive oxygen species at the crossroads of inflammasome and inflammation. Front Physiol. 2014;5:352. 41. Youn M, Huang H, Chen C, et al. MMP9 inhibition increases erythropoiesis in RPS14-deficient del(5q) MDS models through suppression of TGF-beta pathways. Blood Adv. 2019;3(18):27512763. 42. Kale VP. Differential activation of MAPK signaling pathways by TGF-beta1 forms the molecular mechanism behind its dosedependent bidirectional effects on hematopoiesis. Stem Cells Dev. 2004;13(1):27-38. 43. Kale VP, Vaidya AA. Molecular mechanisms behind the dosedependent differential activation of MAPK pathways induced by transforming growth factor-beta1 in hematopoietic cells. Stem Cells Dev. 2004;13(5):536-547. 44. Walter D, Lier A, Geiselhart A, et al. Exit from dormancy provokes DNA-damage-induced attrition in haematopoietic stem cells. Nature. 2015;520(7548):549-552. 45. Saeed S, Quintin J, Kerstens HH, et al. Epigenetic programming of monocyte-to-macrophage differentiation and trained innate immunity. Science. 2014;345(6204):1251086. 46. Cheng SC, Quintin J, Cramer RA, et al. mTOR- and HIF-1alphamediated aerobic glycolysis as metabolic basis for trained immunity. Science. 2014;345(6204):1250684. 47. Netea MG, Joosten LA, Latz E, et al. Trained immunity: a program of innate immune memory in health and disease. Science. 2016;352(6284):aaf1098. 48. Cappelli E, Ravera S, Vaccaro D, et al. Mitochondrial respiratory complex I defects in Fanconi anemia. Trends Mol Med. 2013;19(9):513-514. 49. Bottega R, Nicchia E, Cappelli E, et al. Hypomorphic FANCA mutations correlate with mild mitochondrial and clinical phenotype in Fanconi anemia. Haematologica. 2018;103(3):417426. 50. Schildgen V, Wulfert M, Gattermann N. Impaired mitochondrial gene transcription in myelodysplastic syndromes and acute myeloid leukemia with myelodysplasia-related changes. Exp Hematol. 2011;39(6):666-675. 51. Rodriguez A, Zhang K, Farkkila A, et al. MYC promotes bone marrow stem cell dysfunction in Fanconi anemia. Cell Stem Cell. 2021;28(1):33-47. 52. Garbati MR, Hays LE, Keeble W, Yates JE, Rathbun RK, Bagby GC. FANCA and FANCC modulate TLR and p38 MAPK-dependent expression of IL-1beta in macrophages. Blood. 2013;122(18):31973205. 53. Shao W, Yeretssian G, Doiron K, Hussain SN, Saleh M. The caspase-1 digestome identifies the glycolysis pathway as a target during infection and septic shock. J Biol Chem. 2007;282(50):36321-36329. 54. Gurcel L, Abrami L, Girardin S, Tschopp J, van der Goot FG. Caspase-1 activation of lipid metabolic pathways in response to bacterial pore-forming toxins promotes cell survival. Cell. 2006;126(6):1135-1145.

Haematologica | Vol 7 - June 2022

1334

ARTICLE - Chronic Lymphocytic Leukemia

Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients Jennifer R. Brown,1 John C. Byrd,2 Paolo Ghia,3 Jeff P. Sharman,4 Peter Hillmen,5 Deborah M. Stephens,6 Clare Sun,7 Wojciech Jurczak,8 John M. Pagel,9 Alessandra Ferrajoli,10 Priti Patel,11 Lin Tao,11 Nataliya Kuptsova-Clarkson,12 Javid Moslehi13 and Richard R. Furman14

Correspondence:

Dana-Farber Cancer Institute, Boston, MA, USA; 2The Ohio State University Comprehensive Cancer Center, Columbus, OH, USA; 3Università Vita-Salute San Raffaele and IRCCS Ospedale San Raffaele, Milano, Italy; 4Willamette Valley Cancer Institute/US Oncology, Eugene, OR, USA; 5St. James’s University Hospital, Leeds, UK; 6University of Utah Huntsman Cancer Institute, Salt Lake City, UT, USA; 7National Heart, Lung, and Blood Institute, Bethesda, MD, USA; 8Maria Sklodowska-Curie National Research Institute of Oncology, Krakow, Poland; 9 Swedish Cancer Institute, Seattle, WA, USA; 10University of Texas MD Anderson Cancer Center, Houston, TX, USA; 11AstraZeneca, South San Francisco, CA, USA; 12AstraZeneca, Gaithersburg, MD, USA; 13Section of Cardio-Oncology & Immunology, Division of Cardiology and the Cardiovascular Research Institute, University of California San Francisco, San Francisco, CA, USA and 14Weill Cornell Medicine, New York Presbyterian Hospital, New York, NY, USA

Received: April 6, 2021. Accepted: September 21, 2021. Prepublished: September 30, 2021. https://doi.org/10.3324/haematol.2021.278901

Jennifer R. Brown jennifer_brown@dfci.harvard.edu

Abstract Cardiovascular (CV) toxicities of the Bruton tyrosine kinase (BTK) inhibitor ibrutinib may limit use of this effective therapy in patients with chronic lymphocytic leukemia (CLL). Acalabrutinib is a second-generation BTK inhibitor with greater BTK selectivity. This analysis characterizes pooled CV adverse events (AE) data in patients with CLL who received acalabrutinib monotherapy in clinical trials (clinicaltrials gov. Identifier: NCT02029443, NCT02475681, NCT02970318 and NCT02337829). Acalabrutinib was given orally at total daily doses of 100–400 mg, later switched to 100 mg twice daily, and continued until disease progression or toxicity. Data from 762 patients (median age: 67 years [range, 32–89]; median follow-up: 25.9 months [range, 0–58.5]) were analyzed. Cardiac AE of any grade were reported in 129 patients (17%; grade ≥3, n=37 [5%]) and led to treatment discontinuation in seven patients (1%). The most common any-grade cardiac AE were atrial fibrillation/flutter (5%), palpitations (3%), and tachycardia (2%). Overall, 91% of patients with cardiac AE had CV risk factors before acalabrutinib treatment. Among 38 patients with atrial fibrillation/flutter events, seven (18%) had prior history of arrhythmia or atrial fibrillation/flutter. Hypertension AE were reported in 67 patients (9%), 43 (64%) of whom had a preexisting history of hypertension; no patients discontinued treatment due to hypertension. No sudden cardiac deaths were reported. Overall, these data demonstrate a low incidence of new-onset cardiac AE with acalabrutinib in patients with CLL. Findings from the head-to-head, randomized trial of ibrutinib and acalabrutinib in patients with highrisk CLL (clinicaltrials gov. Identifier: NCT02477696) prospectively assess differences in CV toxicity between the two agents.

Introduction Bruton tyrosine kinase (BTK) inhibitors, such as ibrutinib and acalabrutinib, are preferred treatment regimens for chronic lymphocytic leukemia (CLL).1 BTK is involved in downstream amplification of B-cell receptor signaling and is a key anticancer target for CLL and other B-cell malignancies such as mantle cell lymphoma and Waldenström’s macroglobulinemia.2-6 Ibrutinib was the first BTK inhibitor approved for the treatment of CLL.2 In addition to irreversibly binding to the

cysteine residue (C481) and blocking the adenosine triphosphate (ATP)-binding pocket of BTK,7,8 ibrutinib binds to analogous cysteine residues in other kinases such as interleukin-2–inducible T-cell kinase (ITK), tyrosine kinase expressed in hepatocellular carcinoma (TEC), epidermal growth factor receptor (EGFR), and T-cell X chromosome kinase (TXK) as well as to several Src family kinases.9-11 Cardiovascular (CV) toxicities, particularly hypertension and atrial fibrillation, have been associated with ibrutinib, which may be attributed in part to off-target kinase inhibition.12-15 In the RESONATE study in patients with pre-

Haematologica | 107 - June 2022

1335

ARTICLE - Cardiovascular adverse events with acalabrutinib viously treated CLL, atrial fibrillation of any grade was reported in 5% of patients receiving ibrutinib monotherapy at a median follow-up of 9.4 months, which increased to 11% with 44 months of follow-up; in the RESONATE-2 study in patients with treatment-naïve CLL, any-grade atrial fibrillation was reported in 6% and 11% of patients at median follow-up durations of 18.4 and 60 months, respectively.16-19 Additionally, in the CLL-12 trial in patients with asymptomatic, treatment-naïve, early stage CLL, any-grade atrial fibrillation was reported in 18% of patients receiving ibrutinib versus 7% of patients on placebo after a median follow-up of 31 months, with grade 3 to 4 events reported in 6% and 2% of patients, respectively.20 With longer follow-up in the phase Ib/II PCYC-1102 ibrutinib trial, atrial fibrillation of grade 3 or higher was reported in 10% of patients with relapsed/refractory CLL at a median of 6.8 years of follow-up and in 6% of patients with treatment-naive CLL at a median of 7.3 years of follow-up.19 A review of 16 ibrutinib clinical studies found the rate of any-grade atrial fibrillation (4–16%) to be higher than that in a general population (1–2%).21 In a singlecenter, retrospective analysis in ibrutinib-treated patients (n=562), new or worsening hypertension was associated with an increased risk of atrial fibrillation.14,22 Acalabrutinib, a second-generation, potent, highly selective, covalent BTK inhibitor, was approved for the treatment of CLL based on the results of two randomized, controlled phase III studies, ELEVATE-TN and ASCEND.3,23,24 Compared with ibrutinib, acalabrutinib has greater selectivity for BTK in vitro,7,9 which has been hypothesized to explain differences in the tolerability profiles of the two agents.6,9,16,25 At median follow-up duration of 16.1 months in the ASCEND studies, the incidence of atrial fibrillation of any grade was 5% in the acalabrutinib monotherapy arm compared with 3% in the comparator arm.23,24 Similarly, at a median follow-up of 28.3 months in the ELEVATE-TN study, the incidence of any-grade atrial fibrillation was 4% in the acalabrutinib monotherapy arm versus 1%

J.R. Brown et al.

in the comparator arm.23 In order to further explore cardiac and hypertension-related effects of acalabrutinib in patients with CLL, we conducted a retrospective pooled analysis of data from all clinical studies of acalabrutinib (from phase I to III) in CLL containing an acalabrutinib monotherapy arm and characterized CV adverse events (AE).

Methods Data were from a pooled population of patients with CLL treated with at least one dose of acalabrutinib monotherapy in all sponsored clinical trials containing an acalabrutinib monotherapy arm (Table 1).9,23,24,26-29 Because of the inclusion of a phase I study and an initial phase II study, acalabrutinib was given orally at total daily doses of 100– 400 mg (although the majority of patients in this analysis [78%] received 100 mg twice daily [BID]). Treatment continued until progressive disease or toxicity. Each study’s protocol was approved by an Institutional Review Board and each study was conducted in accordance with the general principles set forth in the International Conference on Harmonization Guidelines for Good Clinical Practice and the Declaration of Helsinki, and in accordance with all applicable legal and regulatory requirements. All participants gave written informed consent. AE were coded using Medical Dictionary for Regulatory Activities v21.1. Severity was graded according to National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) v4.03. The protocols defined AE as those occurring or worsening on or after the first dose of acalabrutinib, through the treatment phase, and within 30 days of the last dose. For this analysis, cardiac AE were those categorized under the system organ class “cardiac disorders”; hypertension AE, which were considered an AE of clinical interest, included the following preferred terms: hypertension, blood pressure increased, essential hyper-

Table 1. Study summaries. Study name

Study description

Number of patientsa

ACE-CL-001 (NCT02029443)

Ph I/II study of acalabrutinib in patients with CLL

301; TN/RR: 112/189

ACE-CL-007 (NCT02475681; ELEVATE-TN)b

Ph III study of acalabrutinib ± O vs C+O in TN CLL

224; all TN

ACE-CL-309 (NCT02970318; ASCEND)

Ph III study of acalabrutinib vs IdR or BR in RR CLL

189; all RR

15-H-0016 (NCT02337829)

Ph II study of acalabrutinib in patients with RR or TN with del(17p) CLL

48; TN/RR: 16/32

596 patients started on the acalabrutinib 100 mg twice daily (BID) dose; 166 patients started on a different dose, and 106 of these patients were later switched to the 100 mg BID dose. bStudy has acalabrutinib monotherapy and combination therapy arms; only monotherapy patients were included. BR: bendamustine plus rituximab; C: chlorambucil; CLL: chronic lymphocytic leukemia; IdR: idelalisib plus rituximab; O: obinutuzumab; PD: progressive disease; Ph: phase; PO: orally; RR: relapsed/refractory; TN: treatment-naïve. a

Haematologica | 107 - June 2022

1336

ARTICLE - Cardiovascular adverse events with acalabrutinib tension, hypertensive crisis, malignant hypertension, hypertensive heart disease, and orthostatic hypertension. Assessments of cardiac AE and hypertension AE included incidence, rate per patient exposure years, time to onset, incidence of events leading to dose delays and discontinuations, and event management. Event duration was also assessed for overall cardiac AE and atrial fibrillation/flutter events. Cardiac AE severity and resolution of grade ≥3 cardiac AE also were assessed. Incidence of cardiac AE in the first 6 months was assessed based on a predominance of atrial fibrillation events during this time period in published reports with ibrutinib.13 CV and hypertension risk factors, based on patient medical histories, were also assessed (methodology described in the Online Supplementary Appendix). Prior history of arrhythmia, atrial fibrillation, and atrial flutter events was assessed in patients who experienced atrial fibrillation events, and prior history of hypertension was assessed in patients who experienced hypertension events. Occurrence of major hemorrhage events (defined as any hemorrhagic event that was serious or grade ≥3 in severity, or any central nervous system hemorrhage) after receipt of anticoagulant therapy on study was assessed. Atrial fibrillation risk scores (range, 0–7) were computed in patients without a prior history of atrial fibrillation using the Shanafelt predictive model, calculated as the sum of the risk values assigned to the specific factors independently associated with atrial fibrillation: older age (2 points for age 65–74 years; 3 points for age ≥75 years), male sex (1 point), valvular heart disease (2 points), and hypertension (1 point).13,30,31 Statistical analysis Data were summarized using descriptive statistics. Time to first cardiac AE overall, atrial fibrillation/flutter AE, and hypertension AE were analyzed using the Kaplan-Meier method. Incidence of atrial fibrillation was analyzed by age group (<65 years, 65 to <75 years, ≥75 years) and Shanafelt risk category (0–1, 2–3, 4, and ≥5). The incidence of cardiac AE also was compared between the pooled acalabrutinib monotherapy arms and pooled comparator arms from the randomized, controlled, phase III trials (ASCEND and ELEVATE-TN) included in this analysis; comparator treatments included investigator’s choice of idelalisib plus rituximab or bendamustine plus rituximab (ASCEND) and obinutuzumab plus chlorambucil (ELEVATE-TN). Statistical comparison of cardiac AE incidence between acalabrutinib monotherapy and comparator arms was performed using the Chi-squared test.

Results Patients In total, 762 patients who received acalabrutinib monotherapy across four clinical trials were included in this analysis

J.R. Brown et al.

(Table 1). Patient demographics and baseline characteristics were similar among the total acalabrutinib monotherapy CLL population and the subgroup of patients who experienced a cardiac AE (Table 2). At data cut-off, the median duration of follow-up was 25.9 months (range, 0.0–58.5; n=761) and the median duration of acalabrutinib exposure was 24.9 months (range, 0.0–58.5; n=760). Five hundred fifty-three patients (73%) remained on acalabrutinib at data cut-off. A total of 208 patients (27%) discontinued acalabrutinib treatment; the most common reasons were progressive disease (n=82; 11%) and AE (n=70; 9%). When comparing the treatment-naïve (n=352) and relapsed/refractory (n=410) populations, 68 (19%) and 140 (34%) discontinued acalabrutinib treatment, respectively; the most common reasons were progressive disease (n=19 [5%] and n=63 [15%], respectively) and AE (n=27 [8%] and n=43 [10%], respectively). Most patients (78%; n=596/762) received only the 100 mg BID dose of acalabrutinib; 166 patients (22%) received a different initial acalabrutinib dose (55 received doses greater than 100 mg BID [I.e., 200 mg BID, 250 mg daily [QD], and 400 mg QD]). Most of the patients who received an initial dose of acalabrutinib that differed from the standard dose of 100 mg BID later switched to 100 mg BID (n=106 [64%]). Patient demographics and baseline characteristics were similar in the pooled acalabrutinib monotherapy and comparator arms from ASCEND and ELEVATE-TN (Online Supplementary Table S1). Safety Cardiac adverse events

Across studies and exposure durations, there were 199 cardiac AE of any grade reported among 129 (n=58 with treatment naïve CLL; n=71 with relapsed/refractory CLL) of 762 patients (17%) (Table 3). Among the 55 patients who received initial acalabrutinib doses greater than 100 mg BID, 12 (22%) experienced cardiac AE, whereas 84 (14%) reported cardiac AE among the 596 who had taken acalabrutinib only at the 100 mg BID dose. In the overall population of 762 patients, the most common cardiac AE of any grade (incidence ≥2%) were atrial fibrillation (4% [n=34]; grade ≥3, 1% [n=10]), palpitations (3% [n=23]; grade ≥3, 0), and tachycardia (2% [n=17]; grade ≥3, 0). The median time to onset of cardiac AE was 10.1 months (range: 0.1–49.7) (Figure 1A). Overall, 7% (n=53/760) of patients reported any-grade cardiac AE in the first 6 months on acalabrutinib, 3% (n=18/653) >6 to 12 months after treatment initiation, and 8% (n=46/585) >12 to 24 months after starting treatment. Cardiac AE led to acalabrutinib discontinuation in seven patients (1%), six of whom discontinued due to grade ≥3 cardiac events (described below) and one patient who discontinued due to grade 2 atrial fibrillation. Comparatively, 91% (n=117/129) of patients with cardiac AE versus 80% (n=503/633) without cardiac AE had at least one CV risk factor before starting acalabrutinib. The most prevalent CV risk factors (defined in the Online Supplementary Appendix) for patients with versus

Haematologica | 107 - June 2022

1337

ARTICLE - Cardiovascular adverse events with acalabrutinib

J.R. Brown et al.

Table 2. Patient demographics and baseline characteristics. All patients (N=762)

Patients with cardiac events (N=129)

67.0 (32–89)

69.0 (42–84)

508 (66.7)

83 (64.3)

698 (91.6)

113 (87.6)

Black or African American

26 (3.4)

9 (7.0)

Asian

14 (1.8)

3 (2.3)

American Indian or Alaska Native

1 (0.1)

Native Hawaiian or Other Pacific Islander

1 (0.1)

Other

11 (1.4)

4 (3.1)

Missing

11 (1.4)

26.7 (16–49)a

27.1 (18–47)b

292 (38.3)

41 (31.8)

415 (54.5)

81 (62.8)

54 (7.1)

7 (5.4)

1 (0.1)

1 (0–13)

1 (0–10)

Treatment-naïve disease, n (%)

352 (46.2)

58 (45.0)

Relapsed/refractory disease, n (%)

410 (53.8)

71 (55.0)

Demographic/baseline characteristic Age, median (range), years Male, n (%) Race, n (%) White

BMI (kg/m2), median (range) ECOG PS score, n (%)

Number of prior regimens, median (range)

n=746. bn=126. BMI: body mass index; ECOG PS: Eastern Cooperative Oncology Group performance status.

without cardiac AE were hypertension (65% [n=84] vs. 53% [n=333], respectively), diabetes mellitus (19% [n=24] vs. 12% [n=79]), hyperlipidemia (16% [n=21] vs. 11% [n=67]), hypothyroidism (16% [n=21] vs. 12% [n=73]), hypercholesterolemia (14% [n=18] vs. 16% [n=104]), and atrial fibrillation (12% [n=16] vs. 6% [n=40]). The most common (≥15%) types of concomitant medications reported among patients who experienced cardiac events were antithrombotic agents (n=36 [28%]) and beta-blocking agents (n=30 [23%]) (Online Supplementary Table S2); 16 patients (12%) received both antithrombotic and beta-blocking agents, and a total of 17 patients (13%) received aspirin (13 [10%] received aspirin along with other antithrombotic agents and 4 [3%] received aspirin without additional antithrombotic agents). Fifty-one grade ≥3 cardiac AE were reported among 37 patients (5%). Overall, nine of 760 patients (1%) experienced 13 grade ≥3 cardiac AE (representing 25% of total grade ≥3 events) during the first 6 months on acalabrutinib (Online Supplementary Table S3); six of 653 (1%) and 15 of 585 patients (3%) experienced grade ≥3 cardiac AE >6 to 12 months and >12 to 24 months after starting acalabrutinib, respectively. Among the 37 patients with grade ≥3 cardiac events, 12 patients experienced a grade 4 cardiac event

and two patients experienced a grade 5 cardiac event. The grade 5 AE included acute myocardial infarction in one patient with preexisting hypertension, type 2 diabetes mellitus, CV disorder, and chronic kidney disease; and congestive cardiac failure in another patient with a history of congestive cardiac failure, aortic stenosis, and atrial fibrillation. Among the 37 patients with grade ≥3 cardiac AE, 18 patients (49%) were continuing acalabrutinib at data cut-off. Six of the 37 patients (16%) with grade ≥3 cardiac AE discontinued acalabrutinib due to those grade ≥3 cardiac AE, including acute myocardial infarction (n=2; grade 3 and grade 5), cardiac failure congestive (n=2; grade 3 and 5), cardiac failure (n=1; grade 3), and cardiac tamponade (n=1; grade 4). Thirteen of the 37 patients discontinued acalabrutinib due to other AE (n=4), progressive disease (n=5), death (n=3), and other reasons (n=1). Sixteen of the 51 grade ≥3 cardiac AE (31%) led to dose delay and 36 (71%) were managed with concomitant medications. In total, 43 of the grade ≥3 cardiac AE (84%) resolved, of which 15 were associated with dose delays, four resulted in drug withdrawal, and 24 had no associated dose modifications or changes. No sudden cardiac deaths were reported. In the analysis comparing acalabrutinib monotherapy with

Haematologica | 107 - June 2022

1338

ARTICLE - Cardiovascular adverse events with acalabrutinib

J.R. Brown et al.

Table 3. Incidence of cardiac adverse events. All patients (N=762)

Event

Any grade

Grade ≥3c

129 (17), [199], {0.081}

37 (5), [51], {0.023}

38 (5), [48], {0.024}

11 (1.4), [12], {0.007}

Atrial fibrillation

34 (4), [44], {0.021}

10 (1), [11], {0.006}

Atrial flutter

4 (0.5), [4], {0.003}

1 (0.1), [1], {0.001}

Palpitations

23 (3), [27], {0.014}

Tachycardia

17e (2), [18], {0.011}

Sinus tachycardia

11e (1), [13], {0.007}

1 (0.1), [1], {0.001}

Angina pectoris

10 (1), [11], {0.006}

2 (0.3), [2], {0.001}

Bradycardia

9 (1), [10], {0.006}

2 (0.3), [2], {0.001}

Cardiac failure

6 (0.8), [6], {0.004}

3 (0.4), [3], {0.002}

Acute myocardial infarction

5 (0.7), [6], {0.003}

Supraventricular tachycardia

4e (0.5), [4], {0.003}

1 (0.1), [1], {0.001}

Any cardiac AE,a n (%),b [number of individual events], {rate per PEY} Most common cardiac AE (preferred terms; occurring in ≥4 patients), n (%), [number of individual events], {rate per PEY} Atrial fibrillation/flutterd

Adverse events (AE) categorized under the system organ class cardiac disorders. b199 AE were reported in 129 patients (17%). No events under the preferred terms sudden death or sudden cardiac death were reported. cOther grade ≥3 AE of interest occurring in <4 patients each included complete atrioventricular (AV) block (n=2; 0.3%), acute coronary syndrome (n=1; 0.1%), second-degree AV block (n=1; 0.1%), and ventricular fibrillation (n=1; 0.1%). dPatients with “atrial fibrillation” or “atrial flutter” preferred terms combined. There was no overlap between patients with “atrial fibrillation” and “atrial flutter” events. eOne patient had both “tachycardia” and “sinus tachycardia” events. Another patient had both “sinus tachycardia” and “supraventricular tachycardia” events. A third patient had both “tachycardia” and “supraventricular tachycardia” events. All other reports of “tachycardia,” “sinus tachycardia,” and “supraventricular tachycardia” occurred in unique patients. PEY: patient exposure years. a

the comparator arms in ASCEND and ELEVATE-TN, 45 of 333 patients (14%) in the pooled acalabrutinib monotherapy group and 25 of 322 (8%) in the pooled comparator group experienced cardiac AE (odds ratio, 1.86; 95% confidence interval [CI]: 1.11-3.11; P=0.02). Grade ≥3 cardiac AE were reported in 4% (n=14) and 3% (n=10) of patients in the acalabrutinib and comparator groups, respectively. Atrial fibrillation was the most common cardiac AE in both groups, with incidence rates of 4% (n=14) and 2% (n=5), respectively. The proportions of patients with CV risk factors were similar between the pooled acalabrutinib monotherapy and pooled comparator groups among those with cardiac AE (n=39 [87%] and n=22 [88%], respectively) and those without cardiac AE (n=220 [76%] and n=234 [79%]). Atrial fibrillation/flutter adverse events

Atrial fibrillation/flutter events of any grade were reported in 38 (n=15 with treatment-naïve CLL; n=23 with relapsed/refractory CLL) of 762 patients (5%). Atrial fibrillation was the most common cardiac AE based on individual

preferred terms and occurred in 34 patients (4%). Median time to an event of atrial fibrillation/flutter was 17.1 months (Figure 1B) and atrial fibrillation/flutter events had a median duration of 0.1 month (range, 0.0–12.4). Rates of atrial fibrillation were consistently low over time, with approximately 1% (n=10/760) of acalabrutinib-treated patients having atrial fibrillation/flutter events in the first 6 months on treatment, 0.2% (n=1/653) >6 to 12 months after treatment initiation, and 3% (n=16/585) >12 to 24 months after starting acalabrutinib treatment. Among 38 patients with atrial fibrillation/flutter events, seven (18%) had a prior history of arrhythmia, atrial fibrillation, or flutter. Twelve (32%) patients with treatment-emergent atrial fibrillation/flutter were subsequently initiated on at least 1 anticoagulant due to atrial fibrillation, one of whom experienced a major hemorrhage event after starting the anticoagulant. In the full analysis population of 762 patients, 299 patients started on-study anticoagulant treatment for any reason, among whom 12 (4%) subsequently experienced a major hemorrhage event. The major hemor-

Haematologica | 107 - June 2022

1339

ARTICLE - Cardiovascular adverse events with acalabrutinib rhage events had resolved in nine of the 12 patients; events were ongoing in three patients at the time of data cut-off. Twenty-nine patients had de novo atrial fibrillation/flutter (i.e., no prior history), among whom one patient experienced a subsequent stroke event (grade 3 ischemic cerebral infarction occurring 682 days after starting treatment and 2 days after the atrial fibrillation event). This patient was female and 67 years of age with a medical history of grade 2 arterial hypertension, grade 1 heart failure, and recurrent transient cerebral ischemic attacks 7 to 15 years before study entry. Notably, the patient de-

J.R. Brown et al.

veloped acute myocardial infarction on treatment day 221 and had received relevant concomitant medications prior to atrial fibrillation onset (clopidogrel, fraxiparin, aspirin). Atrial fibrillation/flutter events occurred relatively more frequently with increasing age (age <65 years: 2% [n=7/288]; 65 to <75 years: 6% [n=19/318]; ≥75 years: 8% [n=12/156]). Among patients with no history of atrial fibrillation before study enrollment (n=706), the number (%) of patients in each Shanafelt risk score category was 171 (24%; score 0–1), 297 (42%; score 2–3), 190 (27%; score 4), and 48 (7%; score ≥5); the rate of treatment-emergent atrial fibrillation among these patients increased with in-

Figure 1. Time to onset of cardiac, atrial fibrillation/flutter, and hypertension adverse events. Time to onset of (A) cardiac events and (B) atrial fibrillation/flutter and hypertension adverse events (AE). AEa categorized under the system organ class cardiac disorders; mo: months. Haematologica | 107 - June 2022

1340

ARTICLE - Cardiovascular adverse events with acalabrutinib creasing Shanafelt risk score category (2% [n=3], 5% [n=14], 3% [n=6], and 13% [n=6], respectively; Online Supplementary Table S4). The cumulative incidence of atrial fibrillation events was higher in patients with Shanafelt risk scores ≥5 and in those with a previous history of atrial fibrillation compared with patients with Shanafelt risks scores ≤4 (Figure 2). Infections occurred concurrently in seven (18%) of the 38 patients who experienced an atrial fibrillation/flutter event, including three patients with concurrent pneumonia. Hypertension adverse events

In total, hypertension events were reported in 67 patients (9%). The majority of patients with hypertension events had relapsed/refractory CLL (n=39 [58%]) compared with treatment-naïve CLL (n=28 [42%]). Median time to an event of hypertension was 6.5 months (Figure 1B). Patients who experienced hypertension events more commonly had risk factors for hypertension (defined in the Online Supplementary Appendix) compared with patients without hypertension events (79% [n=53/67] vs. 59% [n=412/695], respectively). Among the 67 patients with hypertension events, 43 (64%) had preexisting hypertension. Other than preexisting hypertension, the most common risk factors (incidence >5%) included hyperlipidemia (19% [n=13]), hypercholesterolemia (19% [n=13]), diabetes mellitus (10% [n=7]), coronary artery disease (9% [n=6]), chronic obstructive pulmonary disease (7% [n=5]), and coronary ar-

J.R. Brown et al.

tery bypass (6% [n=4]). In total, 13 of 43 (30%) patients with preexisting hypertension and 0 of 24 patients without preexisting hypertension received at least one concomitant medication for hypertension management during the hypertension event. Thirty-five grade ≥3 hypertension events (grade 3, n=34; grade 4, n=1) were reported in 30 patients, among whom 24 (80%) had a prior history of hypertension. Ten patients (1%) experienced 10 grade ≥3 hypertension events (29% of total grade ≥3 hypertension events) during the first 6 months of acalabrutinib treatment; 0.6% (n=4) experienced grade ≥3 hypertension events >6 to 12 months after treatment initiation and 1% (n=7) >12 to 24 months after starting treatment. Among the 30 patients with grade ≥3 hypertension events, no patient discontinued acalabrutinib treatment due to these events, and 21 (70%) were continuing acalabrutinib at data cut-off. Three of the grade ≥3 hypertension AE (9%) led to dose delay and 18 (51%), including the three resulting in dose delay, were managed with concomitant medications. Among the 17 events that were not managed with concomitant medications or dose delay, no additional actions were documented. Two patients experienced grade 3 hypertension that required hospitalization or prolonged hospitalization; both events resolved with drug interruption/dosing delay and concomitant or additional medication. One patient had grade 4 hypertension and experienced the event while hospitalized for a grade 3 serious hypercalcemia event.

Figure 2. Time to onset of atrial fibrillation by Shanafelt risk score category and previous history of atrial fibrillation (AF).

Haematologica | 107 - June 2022

1341

ARTICLE - Cardiovascular adverse events with acalabrutinib This patient concurrently experienced other adverse events including fall, encephalopathy, respiratory failure, and seizure and had notable medical history of severe hypertension and Guillian-Barre syndrome. In total, 20 (57%) of the grade ≥3 hypertension AE resolved, including two that resulted in dose delay and 18 that had no associated dose modifications or changes; 15 (43%) events were ongoing at data cut-off.

Discussion In this pooled analysis of four clinical studies of acalabrutinib monotherapy in patients with CLL, the incidence of cardiac AE was relatively low. Overall, cardiac AE of any grade occurred in 17% of patients, with only a 5% incidence of atrial fibrillation/flutter events of any grade over a median 25.9 months of follow-up. Among the 129 patients who experienced cardiac AE, 91% had preexisting cardiac risk factors. Any-grade cardiac AE occurred at a consistent rate over time. While it is unknown whether some cardiac AE resolve with discontinuation of drug, there were relatively few discontinuations due to cardiac AE in this analysis, reported in only 1% of patients. BTK inhibitors have become a preferred therapy in CLL; however, cardiac toxicity has been a safety concern for ibrutinib, particularly due to increased incidence of atrial fibrillation (reported incidence up to 18%)12,13,20,21,32-34 and sudden death.12,35 While the mechanism behind ibrutinib’s cardiac toxicity profile is not well understood, a preclinical study in rats suggested ibrutinib may result in inhibition of the PI3K-Akt signaling pathway, which under normal conditions provides a cardioprotective effect on the heart under stress.36 A separate recent study in mice found that 4 weeks of ibrutinib treatment led to inducible atrial fibrillation, left atrial enlargement, myocardial fibrosis, and inflammation, while comparable treatment with acalabrutinib did not.37 Ibrutinib treatment in mice with BTKinactivating mutations still led to inducible atrial fibrillation, indicating that BTK in mice is not the relevant target.37 Chemoproteomic profiling and knockout experiments in mice suggested that C-terminal Src kinase (CSK), an off target of ibrutinib, may be responsible for ibrutinibinduced atrial fibrillation.37 While CSK inhibition was not tested directly, a biochemical and cellular profiling study of BTK inhibitors showed that acalabrutinib did not inhibit any of the Src kinases at physiologically relevant concentrations.11 A study of peripheral blood samples from patients with CLL also showed that ibrutinib inhibited kinases downstream of CSK whereas acalabrutinib did not.38 These experiments suggest that acalabrutinib would be expected to have a significantly lower rate of atrial fibrillation than ibrutinib. The relevance of these findings to humans, however, should be based on prospective studies

J.R. Brown et al.

comparing ibrutinib with acalabrutinib.39 Atrial fibrillation poses a significant burden to patients with CLL, with 6.1% of patients presenting with a history of atrial fibrillation at the time of CLL diagnosis.31 Patients with atrial fibrillation have a 5-fold increased risk of stroke and a 9-fold increased risk of developing cardiac failure.4043 Management of atrial fibrillation is challenging in patients with CLL who are oftentimes older with other comorbidities such as hypertension, which also must be adequately managed to reduce the risks of stroke and bleeding in the context of anticoagulant therapy.40 Patients receiving BTK inhibitor treatment should be assessed for prior atrial fibrillation history and risk factors for stroke and bleeding in alignment with current atrial fibrillation management guidelines.44-46 It is important to note that the focus of our analysis of atrial fibrillation/flutter was on any-grade events. While grade ≥3 atrial fibrillation events were observed among patients in this pooled analysis, CTCAE grading in atrial fibrillation primarily distinguishes the presence or absence of symptoms, with grades 1 and 2 representing “asymptomatic” and “non-urgent” cases of atrial fibrillation, respectively, and grades 3 and 4 representing “symptomatic” and “life-threatening” atrial fibrillation, respectively. However, evidence suggests patients with asymptomatic atrial fibrillation are at higher risk of cerebrovascular events and CV mortality than those with more typical presentation, suggesting that CTCAE grading does not adequately assess medical risk of atrial fibrillation.47 In the current analysis, atrial fibrillation/flutter of any grade was reported in 5% of patients (n=38) over a median of 25.9 months of follow-up. The rate of atrial fibrillation/flutter is reasonably consistent with the incidence of new-onset atrial fibrillation after a median follow-up of 7.3 years in an untreated CLL population with no prior history of atrial fibrillation (6%),31 and with the incidence of any-grade atrial fibrillation among placebotreated patients with asymptomatic, treatment-naïve, early stage CLL after a median follow-up of 31 months in the CLL-12 study (7%).20 In our comparative analysis of pooled data from the ASCEND and ELEVATE-TN studies, the incidence of any-grade atrial fibrillation was 4% with acalabrutinib monotherapy compared with 2% in the pooled comparator arm. By contrast, in a previously reported analysis of the RESONATE-2 study with a similar median follow-up of 29 months, the rate of any-grade atrial fibrillation in patients with CLL receiving ibrutinib was 10% (4% for grade ≥3 atrial fibrillation), suggesting a higher risk with ibrutinib.48 In a pooled analysis of data from four randomized controlled trials of ibrutinib (median follow-up 16.6 months), approximately 63% of atrial fibrillation events reported in ibrutinib-treated patients over 24 months of treatment occurred in the first 6 months,13 whereas 29% of atrial fibrillation events in our

Haematologica | 107 - June 2022

1342

ARTICLE - Cardiovascular adverse events with acalabrutinib analysis occurred in the first 6 months of acalabrutinib monotherapy, with a more constant rate over time. In the same pooled analysis of four ibrutinib studies, de novo atrial fibrillation events increased with increasing Shanafelt risk score category, with incidence rates ranging from 4% to 9% for categories 0 to 4 and the highest incidence of atrial fibrillation in those with Shanafelt risk score ≥5 (15%).13 In our analysis, the incidence of atrial fibrillation also was highest in those with the highest Shanafelt risk scores, but was generally lower across all categories compared with those reported for ibrutinib,13 with incidence rates ranging from 2% to 5% for categories 0 to 4 and 13% for categories ≥5. A notable increase in the rate of atrial fibrillation was seen only in patients with the highest Shanafelt risk category, suggesting that lower risk scores may not be as informative in acalabrutinib-treated patients. Ibrutinib treatment is associated with an increased risk of hypertension and an associated increased risk of atrial fibrillation.14 Overall, hypertension-related events were infrequent in the present analysis (any-grade incidence of 9%; grade ≥3 incidence of 4%) and only 13 of the 67 patients (19%) with events received concurrent medication for hypertension management during the event. In addition, a majority of patients with hypertension events (64%) had preexisting hypertension. By comparison, in a retrospective analysis of 301 patients with CLL, 71% of patients without hypertension at baseline developed new hypertension, and 56% of patients with hypertension at baseline had worsening hypertension with ibrutinib treatment.22 In a pooled analysis of four clinical trials of ibrutinib, the rates of any-grade and grade 3/4 hypertension were 10% and 4%, respectively, similar to the rates reported in our study; however, the median follow-up duration was shorter in that analysis (~16 months).49 In a separate integrated analysis of two ibrutinib clinical trials with a median treatment duration of 29 months, anygrade and grade 3/4 hypertension were reported in 21% and 7% of ibrutinib-treated patients, respectively.50 Comparatively, in a phase Ib/II study of ibrutinib in 132 patients with treatment-naïve (median follow-up, 87 months) or relapsed/refractory (median follow-up, 82 months) CLL, the incidence of grade ≥3 hypertension was 28% overall, with yearly incidence rates of 9% (≤1 year), 8% (>1–2 years), 19% (>2–3 years), 15% (>3–4 years), 16% (>4–5 years), 16% (>5–6 years), and 5% (>6–7 years).19 The results of our analysis suggest a lower risk of hypertension events with acalabrutinib, although longer follow-up is required in patients on both ibrutinib and acalabrutinib. There are some limitations of our analysis. Although we conducted an analysis comparing pooled acalabrutinib monotherapy with the pooled comparator arms from the two phase III studies, the individual trials were not designed to detect statistical differences in AE between the treatment arms. However, baseline characteristics and

J.R. Brown et al.

risk factors were similar between the comparator groups, mitigating concerns regarding underlying differences between these patient subgroups. Furthermore, any comparisons to ibrutinib rely on cross-trial comparisons, which are fraught with confounding factors, particularly with respect to the patient population enrolled. The Shanafelt risk score used in some of our analyses is a simplified scoring system used previously in patients with CLL and therefore is included here for comparison,13,30,31 whereas guideline-recommended risk assessments for atrial fibrillation take a more comprehensive approach and may be more appropriate for use in clinical practice.44-46 In addition, our results should be considered in the context of the differing dosing schedules (though most patients were treated with 100 mg twice daily) and treatment exposure times (0.0–58.5 months) across studies, as well as the exclusion of patients with significant CV disease. Finally, CV issues in cancer patients can take several years to manifest; longer follow-up of CV sequelae may be necessary.51,52 In conclusion, based on the results of this pooled analysis, the incidence of cardiac AE with acalabrutinib treatment is relatively low in patients with CLL. The results from our analysis also suggest an intriguing difference in atrial fibrillation and hypertension rates with acalabrutinib compared with ibrutinib. Our findings are supported by the recently reported results from the head-to-head ELEVATE-RR trial (ACE-CL-006; clinicaltrials gov. Identifier: NCT02477696), which demonstrated a statistically lower incidence of any-grade atrial fibrillation and hypertension with acalabrutinib versus ibrutinib in patients with previously treated CLL.39 Disclosures JRB has received grants or contracts from Gilead, Loxo/Lilly, Sun, TG Therapeutics, and Verastem/SecuraBio, is a consultant for AbbVie, Acerta/AstraZeneca, BeiGene, Bristol Myers Squibb/Juno/Celgene, Catapult, Dynamo, Eli Lilly, Genentech/Roche, Gilead, Kite, Loxo, MEI Pharma, Morphosys AG, Nextcea, Octapharma, Pfizer, Pharmacyclics, Rigel, Sunesis, TG Therapeutics, and Verastem, has received payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing or educational events from Janssen and Teva, and has participated on data safety monitoring or advisory boards for Invectys and Morphosys. JCB is a consultant for AstraZeneca, Trillium, Syndax, Novartis, Kartos, and has ownership in Vercerx. PG has received consulting/advisory fees/honoraria from AbbVie, Acerta/AstraZeneca, Adaptive Bio, ArQule/MSD, BeiGene, Gilead, Janssen, Juno/Celgene/Bristol Myers Squibb, and Loxo/Lilly, and has received research funding from AbbVie, Gilead, Janssen, and Sunesis. JPS is an employee of the US Oncology Network, is a consultant for AbbVie, Acerta/AstraZeneca, BeiGene, Bristol Myers Squibb, Cel-

Haematologica | 107 - June 2022

1343

ARTICLE - Cardiovascular adverse events with acalabrutinib gene, Genentech, Pharmacyclics, Pfizer, and TG Therapeutics, has ownership in VelosBio, and receives research funding from Acerta, Celgene, Genentech, Gilead, Merck, Pharmacyclics, Seattle Genetics, Takeda, and TG Therapeutics. PH has received travel, accommodations, and expenses from AbbVie and Janssen, research funding from AbbVie and Janssen, honoraria and research funding from F. Hoffmann-LaRoche, honoraria from AstraZeneca, and research funding from Pharmacyclics and Gilead, and is employed by the University of Leeds. DMS has participated in advisory boards for Adaptive Bio, BeiGene, Epizyme, Karyopharm, and TG Therapeutics, and has received clinical trial funding from Acerta, ArQule, Gilead, Juno, Mingsight, Novartis, Karyopharm, and Verastem. CS has received research funding from Genmab. WJ is a consultant for and currently employed by Maria Sklodowska-Curie National Research Institute of Oncology, has received research funding from and was previously employed by Jagiellonian University, and has received research funding from Janssen, Mei Pharma, Merck, Pharmacyclics, Roche, Takeda, and TG Therapeutics. JMP is a consultant for Actinium, AstraZeneca, BeiGene, Gilead, Loxo, and MEI Pharma. AF has nothing to disclose. PP is employed by and a stockholder in AstraZeneca. LT is an employee of AstraZeneca. NK-C is an employee of AstraZeneca. JM is a consultant for AstraZeneca, Janssen, Bristol Myers Squibb, Boston Biomedical, Immunocure, Myovant, and Deciphera, and is supported by National Institutes of Health grants (R01HL141466, R01HL155990, and R01HL156021). RRF is a consultant for AbbVie, AstraZeneca, BeiGene, Janssen, Loxo, and Pharmacyclics, has received payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing, or educational events from AbbVie, AstraZeneca, and Janssen,

J.R. Brown et al.

has received payment for expert testimony from AbbVie, and has participated on a data safety monitoring board or advisory board for Incyte. Contributions All authors contributed to data interpretation. Statistical and safety analyses were performed by MB and NK-C. All authors reviewed and provided important intellectual contributions to the manuscript; all authors approved the final version for publication. Acknowledgements The authors thank the investigators and coordinators at each of the clinical sites, and the patients who participated in the trials included in this pooled analysis and their families. This project was supported by Acerta Pharma, a member of the AstraZeneca Group. Medical writing assistance, funded by AstraZeneca, was provided by Robert J. Schoen, PharmD, and Cindy Gobbel, PhD, of Peloton Advantage, LLC, an OPEN Health company. Funding Support to JCB was provided by the National Cancer Institute R35 CA197734, Four Winds Foundation, and the D. Warren Brown Foundation. CS was supported by the Intramural Research Program of the National Heart, Lung, and Blood Institute. Data-sharing statement Data underlying the findings described in this manuscript may be obtained in accordance with AstraZeneca’s data sharing policy described at https://astrazenecagrouptrials. pharmacm.com/ST/Submission/Disclosure.

References 1. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Chronic Lymphocytic Leukemia/Small Lymphocytic Lymphoma version 4.2021. Available from: https://www.nccn.org/professionals/physician_gls/default.aspx. Accessed July 20, 2021. 2. Imbruvica [package insert]. Sunnyvale, CA, Horsham, PA: Pharmacyclics, Janssen Biotech, Inc., 2020. 3. Calquence [package insert]. Wilmington, DE: AstraZeneca Pharmaceuticals, 2019. 4. Brukinsa [package insert]. San Mateo, CA: BeiGene USA, Inc, 2019. 5. Vitale C, Burger JA. Chronic lymphocytic leukemia therapy: new targeted therapies on the way. Expert Opin Pharmacother. 2016;17(8):1077-1089. 6. Pal Singh S, Dammeijer F, Hendriks RW. Role of Bruton's tyrosine kinase in B cells and malignancies. Mol Cancer. 2018;17(1):57. 7. Barf T, Covey T, Izumi R, et al. Acalabrutinib (ACP-196): a covalent Bruton tyrosine kinase inhibitor with a differentiated

selectivity and in vivo potency profile. J Pharmacol Exp Ther. 2017;363(2):240-252. 8. Tam CS, Trotman J, Opat S, et al. Phase 1 study of the selective BTK inhibitor zanubrutinib in B-cell malignancies and safety and efficacy evaluation in CLL. Blood. 2019;134(11):851-859. 9. Byrd JC, Harrington B, O'Brien S, et al. Acalabrutinib (ACP-196) in relapsed chronic lymphocytic leukemia. N Engl J Med. 2016;374(4):323-332. 10. Bond DA, Woyach JA. Targeting BTK in CLL: beyond Ibrutinib. Curr Hematol Malig Rep. 2019;14(3):197-205. 11. Kaptein A, de Bruin G, Emmelot-van Hoek M, et al. Potency and selectivity of BTK inhibitors in clinical development for B-cell malignancies. Blood. 2018;132(Supplement 1):1871. 12. Salem JE, Manouchehri A, Bretagne M, et al. Cardiovascular toxicities associated with ibrutinib. J Am Coll Cardiol. 2019;74(13):1667-1678. 13. Brown JR, Moslehi J, O'Brien S, et al. Characterization of atrial fibrillation adverse events reported in ibrutinib randomized controlled registration trials. Haematologica. 2017;102(10):17961805.

Haematologica | 107 - June 2022

1344

ARTICLE - Cardiovascular adverse events with acalabrutinib 14. Dickerson T, Wiczer T, Waller A, et al. Hypertension and incident cardiovascular events following ibrutinib initiation. Blood. 2019;134(22):1919-1928. 15. Caldeira D, Alves D, Costa J, Ferreira JJ, Pinto FJ. Ibrutinib increases the risk of hypertension and atrial fibrillation: systematic review and meta-analysis. PLoS One. 2019;14(2):e0211228. 16. Byrd JC, Brown JR, O'Brien S, et al. Ibrutinib versus ofatumumab in previously treated chronic lymphoid leukemia. N Engl J Med. 2014;371(3):213-223. 17. Byrd JC, Hillmen P, O'Brien S, et al. Long-term follow-up of the RESONATE phase 3 trial of ibrutinib vs ofatumumab. Blood. 2019;133(19):2031-2042. 18. Burger JA, Tedeschi A, Barr PM, et al. Ibrutinib as initial therapy for patients with chronic lymphocytic leukemia. N Engl J Med. 2015;373(25):2425-2437. 19. Byrd JC, Furman RR, Coutre SE, et al. Ibrutinib treatment for first-line and relapsed/refractory chronic lymphocytic leukemia: final analysis of the pivotal phase 1b/2 PCYC-1102 study. Clin Cancer Res. 2020;26(15):3918-3927. 20. Langerbeins P, Bahlo J, Rhein C, et al. Ibrutinib versus placebo in patients with asymptomatic, treatment-naïve early stage CLL: primary endpoint results of the phase 3 double-blind randomized CLL12 trial. Hematol Oncol. 2019;37(suppl 2):38-40. 21. Ganatra S, Sharma A, Shah S, et al. Ibrutinib-associated atrial fibrillation. JACC Clin Electrophysiol. 2018;4(12):1491-1500. 22. Jones J, George BS, Peterson CB, et al. Hypertension, cardiovascular and renal complications observed in patients with chronic lymphocytic leukemia receiving long-term therapy with ibrutinib. Blood. 2020;136(Supplement 1):27-28. 23. Sharman JP, Egyd M, Jurczak W, et al. Acalabrutinib with or without obinutuzumab versus chlorambucil and obinutuzmab for treatment-naive chronic lymphocytic leukaemia (ELEVATE TN): a randomised, controlled, phase 3 trial. Lancet. 2020;395(10232):1278-1291. 24. Ghia P, Pluta A, Wach M, et al. ASCEND: Phase III, randomized trial of acalabrutinib versus idelalisib plus rituximab or bendamustine plus rituximab in relapsed or refractory chronic lymphocytic leukemia. J Clin Oncol. 2020;38(25):2849-2861. 25. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-42. 26. Byrd JC, Wierda WG, Schuh A, et al. Acalabrutinib monotherapy in patients with relapsed/refractory chronic lymphocytic leukemia: updated phase 2 results. Blood. 2020;135(15):12041213. 27. Awan FT, Schuh A, Brown JR, et al. Acalabrutinib monotherapy in patients with chronic lymphocytic leukemia who are intolerant to ibrutinib. Blood Adv. 2019;3(9):1553-1562. 28. Byrd JC, Woyach JA, Furman RR, et al. Acalabrutinib in treatment-naive (TN) chronic lymphocytic leukemia (CLL): updated results from the phase 1/2 ACE-CL-001 study. Blood. 2018; 132(Suppl 1):S692. 29. Sun CCL, Nierman PK, Kendall EK, et al. Clinical and biological implications of target occupancy in CLL treated with the BTK inhibitor acalabrutinib. Blood. 2020;136(1):93-105. 30. Archibald WJ, Rabe KG, Kabat BF, et al. Atrial fibrillation in patients with chronic lymphocytic leukemia (CLL) treated with ibrutinib: risk prediction, management, and clinical outcomes. Ann Hematol. 2021;100(1):143-155. 31. Shanafelt TD, Parikh SA, Noseworthy PA, et al. Atrial fibrillation in patients with chronic lymphocytic leukemia (CLL). Leuk Lymphoma. 2017;58(7):1630-1639. 32. Farooqui M, Valdez J, Soto S, Bray A, Tian X, Wiestner A. Atrial

J.R. Brown et al. fibrillation in CLL/SLL patients on ibrutinib. Blood. 2015;126(23):2933. 33. Burger JA, Barr PM, Robak T, et al. Long-term efficacy and safety of first-line ibrutinib treatment for patients with CLL/SLL: 5 years of follow-up from the phase 3 RESONATE-2 study. Leukemia. 2020;34(3):787-798. 34. Moslehi JJ. Cardiovascular toxic effects of targeted cancer therapies. N Engl J Med. 2016;375(15):1457-1467. 35. Lampson BL, Yu L, Glynn RJ, et al. Ventricular arrhythmias and sudden death in patients taking ibrutinib. Blood. 2017;129(18):2581-2584. 36. McMullen JR, Boey EJ, Ooi JY, Seymour JF, Keating MJ, Tam CS. Ibrutinib increases the risk of atrial fibrillation, potentially through inhibition of cardiac PI3K-Akt signaling. Blood. 2014;124(25):3829-3830. 37. Xiao L, Salem JE, Clauss S, et al. Ibrutinib-mediated atrial fibrillation attributable to inhibition of C-terminal Src kinase. Circulation. 2020;142(25):2443-2455. 38. Patel V, Balakrishnan K, Bibikova E, et al. Comparison of acalabrutinib, a selective Bruton tyrosine kinase inhibitor, with ibrutinib in chronic lymphocytic leukemia cells. Clin Cancer Res. 2017;23(14):3734-3743. 39. Byrd JC, Hillmen P, Ghia P, et al. Acalabrutinib versus ibrutinib in previously treated chronic lymphocytic leukemia: results of the first randomized phase 3 trial. J Clin Oncol. 2021 Jul 26. 2021;39(31):3441-3452. 40. Win NT, Teo SP. Atrial fibrillation in older patients-reducing stroke risk is not only about anticoagulation. J Geriatr Cardiol. 2016;13(10):880-882. 41. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22(8):983-988. 42. Verma A, Kalman JM, Callans DJ. Treatment of patients with atrial fibrillation and heart failure with reduced ejection fraction. Circulation. 2017;135(16):1547-1563. 43. Chatterjee NA, Chae CU, Kim E, et al. Modifiable risk factors for incident heart failure in atrial fibrillation. JACC Heart Fail. 2017;5(8):552-560. 44. January CT, Wann LS, Calkins H, et al. 2019 AHA/ACC/HRS focused update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society in collaboration with the Society of Thoracic Surgeons. Circulation. 2019;140(2):e125-e151. 45. January CT, Wann LS, Alpert JS, et al. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. Circulation. 2014;130(23):2071-2104. 46. Hindricks G, Potpara T, Dagres N, et al. 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): the Task Force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) Developed with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC. Eur Heart J. 2021;42(5):373-498. 47. Siontis KC, Gersh BJ, Killian JM, et al. Typical, atypical, and asymptomatic presentations of new-onset atrial fibrillation in the community: Characteristics and prognostic implications. Heart Rhythm. 2016;13(7):1418-1424. 48. Barr PM, Robak T, Owen C, et al. Sustained efficacy and detailed

Haematologica | 107 - June 2022

1345

ARTICLE - Cardiovascular adverse events with acalabrutinib clinical follow-up of first-line ibrutinib treatment in older patients with chronic lymphocytic leukemia: extended phase 3 results from RESONATE-2. Haematologica. 2018;103(9):15021510. 49. O'Brien S, Hillmen P, Coutre S, et al. Safety analysis of four randomized controlled studies of ibrutinib in patients with chronic lymphocytic leukemia/small lymphocytic lymphoma or mantle cell lymphoma. Clin Lymphoma Myeloma Leuk. 2018;18(10):648-657.

J.R. Brown et al. 50. Coutre SE, Byrd JC, Hillmen P, et al. Long-term safety of singleagent ibrutinib in patients with chronic lymphocytic leukemia in 3 pivotal studies. Blood Adv. 2019;3(12):1799-1807. 51. Moslehi J. The cardiovascular perils of cancer survivorship. N Engl J Med. 2013;368(11):1055-1056. 52. Groarke JD, Cheng S, Moslehi J. Cancer-drug discovery and cardiovascular surveillance. N Engl J Med. 2013;369(19):17791781.

Haematologica | 107 - June 2022

1346

ARTICLE - Hematopoiesis

Profound systemic alteration of the immune phenotype and an immunoglobulin switch in Erdheim-Chester disease in 78 patients from a single center Fleur Cohen Aubart,1* Lucie Poupel,2* Flora Saint-Charles,2 Frédéric Charlotte,3 Youssef Arsafi,2 Eric Frisdal,2 Damien Roos-Weil,3 Jean-François Emile,4 Zahir Amoura,1 Maryse Guerin,2 Philippe Lesnik,2 Julien Haroche1 and Wilfried Le Goff2 Sorbonne Université, Assistance Publique-Hôpitaux de Paris, Service de Médecine Interne 2, Centre National de Référence Maladies Systémiques Rares et Histiocytoses, Hôpital Pitié-Salpêtrière, Paris; 2Sorbonne Université, INSERM, Institute of Cardiometabolism and Nutrition (ICAN), UMR_S1166, Paris; 3Sorbonne Université, Assistance Publique-Hôpitaux de Paris, Service d'anatomopathologie, Hôpital Pitié-Salpêtrière, Paris and 4EA4340, Université Versailles-Saint Quentin, Assistance Publique-Hôpitaux de Paris, Hôpital Ambroise Paré, Département de Pathologie, Boulogne, France. 1

*FCA and LP contributed equally as co-first authors.

Correspondence: Julien Haroche Julien.haroche@psl.aphp.fr Wilfried Le Goff wilfried.le_goff@sorbonne-universite.fr Received: April 30, 2021. Accepted: October 6, 2021. Prepublished: October 14, 2021. https://doi.org/10.3324/haematol.2021.279118 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract Erdheim-Chester disease (ECD) is a rare, systemic, non-Langerhans cell histiocytosis neoplasm, which is characterized by the infiltration of CD63+ CD1a- histiocytes in multiple tissues. The BRAFV600E mutation is frequently present in individuals with ECD and has been detected in hematopoietic stem cells and immune cells from the myeloid and systemic compartments. Immune cells and pro-inflammatory cytokines are present in lesions, suggesting that ECD involves immune cell recruitment. Although a systemic cytokine T-helper-1-oriented signature has been reported in ECD, the immune cell network orchestrating the immune response in ECD has yet to be described. To address this issue, the phenotypes of circulating leukocytes were investigated in a large, single-center cohort of 78 patients with ECD and compared with those of a group of 21 control individuals. Major perturbations in the abundance of systemic immune cells were detected in patients with ECD, with decreases in circulating plasmacytoid, myeloid 1, and myeloid 2 dendritic cells, mostly in BRAFV600E carriers, in comparison with individuals in the control group. Similarly, marked decreases in blood Thelper, cytotoxic, and B-lymphocyte numbers were observed in patients with ECD, relative to the control group. Measurement of circulating immunoglobulin concentrations revealed an immunoglobulin G switch, from IgG1 to IgG4 subclasses, which are more frequently associated with the BRAF mutation. First-line therapies, including pegylated interferon-a and vemurafenib, were able to correct most of these alterations. This study reveals a profound disturbance in the systemic immune phenotype in patients with ECD, providing important new information, helping to understand the physiopathological mechanisms involved in this rare disease and improving the therapeutic management of patients.

Introduction Erdheim-Chester disease (ECD) is a rare, systemic, nonLangerhans cell histiocytosis neoplasm, frequently caused by mutations in the MEK-extracellular signal-regulated kinase (ERK) signaling pathway; these are mostly BRAF mutations.1 ECD is characterized by the infiltration of tissues by foamy histiocytes expressing markers of the monocyte/macrophage lineage, including CD45, CD68, CD163, and CD14, whereas ECD histiocytes are negative for CD1a and CD207 dendritic cell (DC) markers. It is proposed that in ECD, histiocytes originate from myeloid CD34+ and

CD14+ progenitor cells.2,3 The BRAFV600E mutation has been detected in hematopoietic stem cells, including common myeloid progenitors and granulocyte-macrophage progenitors, in the bone marrow of patients with ECD,2 supporting a model in which BRAF-mutated myeloid cells disseminate from bone marrow to the periphery for tissue infiltration. Consistent with this model, the BRAFV600E mutation was also found in circulating leukocytes, including classical (CD14+) and nonclassical (CD16+) monocytes and CD1c+ myeloid DC in individuals with ECD.2 The accumulation of histiocytes within lesions in cases of ECD is accompanied by the expression of a chemokine

Haematologica | 107 - June 2022

1347

ARTICLE - Alteration of the systemic immune cell phenotype in ECD and cytokine network favoring immune cell recruitment.4,5 Indeed, pro-inflammatory cytokines are highly expressed in ECD lesions, together with the infiltration of pro-inflammatory T-cell helper 1 (Th-1) lymphocytes. In addition, immunohistological examination of ECD biopsies revealed that infiltrated histiocytes express a large set of chemokines and chemokine receptors.4 Consistent with these observations, patients with ECD exhibited a systemic immune Th1-oriented cytokine profile,6 thereby providing important clues for the therapeutic management of these patients. However, the therapeutic management of patients with ECD remains difficult. First-line therapies are mostly determined by the severity of the disease. Thus, pegylated interferon-a (pegIFNa) is used to treat mild disease and non-refractory ECD,7 whereas drugs targeting the mutated BRAF, such as vemurafenib, are used in multisystemic and refractory ECD.8 The underlying mechanisms that orchestrate the immune response in ECD remain largely unknown, and a comprehensive characterization of systemic immune cells in ECD patients is lacking. Therefore, the goal of our study was to determine whether patients with ECD exhibit abnormalities in their systemic immune phenotype and whether these are affected by the presence of the BRAF mutation and therapeutic agents. We demonstrated that patients with ECD had a profound alteration in their systemic immune cell phenotype, characterized by a low abundance of DC subsets and by specific lymphocyte populations, together with a switch in immunoglobulin (Ig) G subclasses, which may be partially corrected by first-line therapies.

Methods Patients Fasting blood samples were obtained from 17 healthy individuals who formed the control group (13 males and 4 females; mean age; 53±25 years, range, 21–90 years) and 78 patients with ECD (60 males and 18 females; mean age; 60±14 years, range, 18–84 years) who were followed at the Pitié-Salpêtrière Hospital, Paris, France, between December 2012 and July 2015 (Online Supplementary Table S1). For all patients, ECD was diagnosed based on the consensus guidelines for the diagnosis and clinical management of ECD.9 The BRAFV600E mutation was detected using multiplex picodroplet digital polymerase chain reaction analysis (Raindance Technologies), as previously described.10 The prevalence of the BRAFV600E mutation was 64% in the ECD group (50/71 patients, indeterminate for 7 patients). The absence of the BRAFV600E mutation in ECD patients was referred to as the wild-type (WT) in this study. At the time of the blood sample collection, the patients were free of any treatment (n=42) or were receiving treatment with either

F.C. Aubart et al.

pegIFNa (n=31), vemurafenib (n=13), or other drugs (n=17). It is worth noting that we cannot determine which of the cases reported in the present study were previously reported. Blood samples were collected from 25 patients at several time points (free of any treatment and upon treatment). This study was approved by the ethics committee of Ile de France III (#2011-A00447-34) and was conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from all patients. Analysis of blood immune cells by flow cytometry Fresh blood samples were collected in EDTA tubes at the same time of the day for all patients and control individuals; the samples were used immediately for the flow cytometry analysis. Analysis of blood immune cells was carried out simultaneously in both patients and control individuals throughout the study (2013–2015). Similar blood immune cell counts were obtained when flow cytometry analysis was performed for the same control individual at different times of the study. Quantification of the immune cell subsets is described in detail in the Online Supplementary Materials. Circulating chemokines, cytokines and immunoglobulins The quantification of circulating chemokines, cytokines and immunoglobulins is described in the Online Supplementary Materials. Statistical analyses Values are given as medians and interquartile ranges (Q1– Q3). Comparisons of two groups were performed using the Mann-Whitney test. Comparisons of more than two groups were performed using the Kruskal-Wallis test followed by a Dunn comparison test. The impacts of the BRAFV600E mutation and treatment with first-line therapies on blood leukocyte counts were tested using the Jonckheere-Terpstra trend test. Correlations were calculated using the Spearman rank-order test. A χ2 test was performed to analyze the distribution of individuals with ECD around the median value of the indicated parameter, according to BRAF status. For skewed variables, the raw data were logarithmically transformed prior to conducting the analyses. Statistical analyses were performed using R statistical software version 3.3.2 (R foundation for Statistical Computing) and Prism software from GraphPad (San Diego, CA USA). Principal component analysis was performed using the public MetaboAnalyst web server (https://www.metaboanalyst.ca/).

Results Profound alteration of the systemic immune cell phenotype in patients with Erdheim-Chester disease Flow cytometry analysis of blood immune cells in patients

Haematologica | 107 - June 2022

1348

ARTICLE - Alteration of the systemic immune cell phenotype in ECD with ECD allowed the identification of the complete set of monocytes (classical, intermediate, and nonclassical), DC (plasmacytoid [pDC], myeloid 1 [mDC1], and myeloid 2 [mDC2]), and lymphocytes (T-helper [Th], cytotoxic [CT], T regulatory [Treg], natural killer [NK], and B) independently of the BRAFV600E mutation as observed in control individuals (Online Supplementary Figure S1). As shown in Online Supplementary Figure S1, no atypical population was detected in ECD patients, in comparison with control individuals, irrespective of their BRAF status. Although the number of total blood monocytes was higher in patients with ECD who had the BRAF-mutation than in the controls (+58.9%, P<0.05), none of the monocyte subsets was found to be significantly increased in those individuals (Table 1). Instead, a trend for a decrease in nonclassical CD14+CD16++ monocytes was observed in ECD patients carrying the BRAFV600E mutation (-73.8%, P<0.08). More strikingly, a marked decrease in the absolute count of DC, 1349including pDC (-63.6%, P<0.0005), mDC1 (-62.0%, P<0.05), and mDC2 (-72.6%, P<0.005), was observed in patients with ECD when compared with the values in healthy individuals; this effect mostly reflected the reduction in all DC subsets in patients with the BRAF-mutation. Such effects were independent of a patient’s sex (data not shown). Although the number of blood neutrophils, NK, NKT, and Treg cells was not altered in patients with ECD, we noticed a large decrease in CT (-80.8%, P<0.0005) and B (-66.5%, P<0.005) lymphocytes in ECD patients relative to the counts for these cells in controls. Finally, a substantial reduction in the absolute count of Th lymphocytes (-84.5%, P<0.05) was observed in patients with ECD who lacked the BRAF mutation. Principal component analysis of the blood immune cell populations of individuals in the control and ECD groups illustrated the peculiar systemic immune signature that characterized ECD (Figure 1A), as well as a potential effect of the BRAFV600E mutation, as was suggested by the analysis of individual cell populations (Table 1). Assessment of the impact of the BRAFV600E mutation on populations of blood immune cells in ECD supports an enhancing effect of the mutation on the reduction in blood nonclassical CD14+CD16++ monocyte (P<0.03) and DC (pDC, P<0.0002; mDC1, P<0.05, and mDC2, P<0.0009) numbers and on the increase in blood total monocytes (P<0.04) in ECD patients compared with control individuals (Figure 1B-I). Analysis of the effect of first-line therapies on this disturbed systemic immune cell signature indicated that patients with ECD who also had the BRAF mutation and who were treated with first-line therapies, including pegIFNa and vemurafenib, did not exhibit such a massive alteration of the systemic immune cell phenotype when compared with control individuals (Table 1). As an illustration, the absolute counts of mDC1 and mDC2 populations in treated ECD patients carrying the BRAFV600E mutation were

F.C. Aubart et al.

not significantly different from those of control individuals. Assessment of the impact of first-line therapies in ECD patients with the BRAF mutation highlighted the capacity of treatments to partially correct or restore the circulating numbers of several altered leukocyte populations in ECD; this was observed for nonclassical CD14+CD16++ monocyte (P<0.03), mDC1 (P<0.03), and mDC2 (P<0.0006) populations (Figure 2). However, treatments taken individually or as a whole were unable to restore the decrease in pDC populations in patients with ECD (Table 1 and Figure 2). Taken together, these findings highlight a major perturbation of the systemic immune cell phenotype in ECD cases, characterized by a deficit of DC and lymphocytes, which could be partially restored by first-line treatments in patients with the BRAF mutation. Impact of first-line therapies on the systemic cytokine and chemokine network in patients with ErdheimChester disease To provide clues about the mechanism underlying the alteration of the systemic immune cell phenotype of patients with ECD according to the BRAF status, a comprehensive quantification of circulating chemokines and cytokines was performed on this single-center group of 78 patients with ECD (Online Supplementary Table S2). As previously reported,6 the levels of many circulating cytokines and chemokines are highly heterogeneous among ECD patients (Online Supplementary Table S2). However, when we investigated the impact of the BRAFV600E mutation on ECD patients’ cytokine and chemokine profiles, we observed that the proportion of individuals with high levels of numerous circulating cytokines driving the Th1 response (IL-6, IL-8, IL-12p40, and TNFa) and chemokines (IP-10, CCL2, MIP-1a, and CCL22) was higher in carriers of the BRAFV600E mutation than in noncarriers (Figure 3). It is noteworthy that patients with the BRAF-mutation also exhibited high levels of the anti-inflammatory cytokine, IL-10. In contrast, a higher proportion of patients with elevated circulating levels of eotaxin, EGF, and IL-15 was detected among patients lacking the BRAF mutation than in their counterparts with the BRAF mutation. Because of the specific mode of action of the treatments, i.e., vemurafenib and pegIFNa, no difference in the circulating concentrations of cytokines and chemokines was detected between ECD patients with the BRAF mutation who were treated or not treated with first-line therapies when taken as a whole (Online Supplementary Table S2). Conversely, compared with their non-treated counterparts, ECD patients with the BRAF mutation who were treated with pegIFNa exhibited higher circulating levels of cytokines that drive either a pro-inflammatory Th1 (IFNa, 5.6-fold, P<0.0001; and IL-15, 1.8-fold, P<0.05) or an anti-inflammatory Th2 (IL-10, 1.9-fold, P<0.05) re-

Haematologica | 107 - June 2022

1349

ARTICLE - Alteration of the systemic immune cell phenotype in ECD

F.C. Aubart et al.

Figure 1. Patients with Erdheim-Chester disease are characterized by a peculiar systemic immune cell signature (A-I) Principal component analysis (A) and blood counts of total (B) and nonclassical (C) monocytes, T helper (D), cytotoxic (E) and B (F) lymphocytes, and plasmacytoid (G), and myeloid 1 (H) and 2 (I) dendritic cells in patients with untreated Erdheim-Chester disease (ECD) according to their BRAF status in comparison with individuals in the control group. Controls, n=17; ECD patients without the BRAF mutation (WT), n=11; and ECD patients carrying the BRAFV600E mutation (V600E), n=23. P-value for the trend was assessed using the Jonckheere-Terpstra trend test. Haematologica | 107 - June 2022

1350

1351

Haematologica | 107 - June 2022

131.7 (57.52-204.2) 4.56 (2.07-11.13) 23.20 (12.75-35.01) 95.93 (52.14-174.30) 200.9 (71.28-440.9) 235.8 (63.89-557.8)

11.93 (9.10-19.60)

4044 (2266-5531) 303.4 (188.0-564.0) 632.0 (377.5-2000)

277.3 (131.9-441.7) 196.8 (99.65-372.1) 11.50 (4.27-18.46) 18.57 (5.05-36.59)

Control (n=17)

104.3 (22.05-214.3) 9.00 (4.23-15.01) 10.39 (5.45-27.19) 31.73 (15.88-90.11)* 29.13 (12.92-76.26)*** 36.44 (11.55-203.3)*

133.3 (39.61-267.8) 7.42 (2.83-13.85) 16.57 (7.65-43.81) 30.64 (13.30-84.51)* 45.69 (18.19-82.11)** 211.7 (51.47-506.0)†

66.9 (27.72-187.3) 3.67 (1.44-5.80) 20.33 (17.51-37.90) 67.34 (10.90-169-24) 48.23 (13.47-114.9)** 174.5 (57.15-289.6)

53.1 (4.90-101.7)*,† 4.27 (1.67-8.52) 29.74 (4.82-41.13) 9.87 (5.17-30.21)***,† 46.36 (9.01-67.01)** 95.12 (10.40-215.0)*

55.2 (27.09-162.5) 3.71 (1.49-8.31) 24.05 (11.39-40.64) 21.65 (7.51-89.54)* 48.23 (10.36-75.44) *** 130.6 (33.74-256.6)

3.55 (5.19-23.82) 11.23 (1.19-24.13) 1.04 (0.31-23.82)

20.12 (1.25-40.54)

17.53 (4.02-25.46)

20.12 (1.58-32.28)

106.8 (25.67-228.4) 7.52 (3.41-14.77) 15.94 (7.48-30.94) 32.17 (14.98-84.32)** 38.63 (17.24-84.66)*** 88.77 (31.40-278.8)

1030 (281.2-1997) *** 229.7 (93.33-415.8) 327.1 (135.7-960.8)

413.3 (268.6-567.7) 241.2 (144.8-357.1) 13.75 (7.27-26.80) 13.82 (2.27-20.25)

All (n=29)

1243 (590.0-1976)** 250.5 (140.9-467.1) 523.0 (285.3-863.2)†

399.2 (280.1-446.3) 260.6 (182.0-372.9) 10.19 (7.32-20.05) 9.15 (5.00-16.77)

Vemurafenib (n=12)

663.1 (180.5-2081)*** 186.3 (25.62-405.3) 233.0 (67.38-1114)

416.3 (174.9-594.2) 241.2 (60.89-344.2) 18.41 (7.26-32.54) 15.33 (0.64-32.84)

PegIFNa (n=17)

1167 (186.1-2004)** 100.3 (41.13-595.5)* 135.9 (63.60-399.9)**

440.6 (248.5-984.9)* 288.2 (152.1-686.9) 13.22 (3.85-24.12) 4.86 (1.16-16.69)

V600E (n=23)

2092 (542.9-2710)** 276.2 (42.51-890.0) 359.8 (91.27-872.2)

183.8 (106.7-965.5) 203.2 (99.51-485.4) 13.22 (3.85-24.12) 11.10 (1.99-23.09)

WT (n=11)

1472 (238.2-2479)*** 115.2 (42.51-589.1)* 172.9 (81.17-594.4)**

384.0 (181.6-921.9) 281.3 (144.8-553.3) 11.62 (4.03-23.48) 6.72 (1.58-20.97)

All (n=38)

Treated ECD (V600E)

EDC: Erdheim-Chester disease; WT: wild-type; pegIFNa: pegylated interferon alpha; pDC: plasmacytoid dendritic cells; mDC: myeloid dendritic cells; NK: natural killer cells; NKT: natural killer T cells; Treg: regulatory T cells; B: B lymphocytes; CTL: cytotoxic T lymphocytes; Th, T helper cells. *P<0.05, **P<0.005 and ***P<0.005 versus the control individuals. †P<0.05 versus untreated ECD patients carrying the BRAFV600E mutation. Expressed as median (quartile 1- quartile 3). WT and V600E, absence and presence of the BRAFV600E mutation.

Treg

NKT

Lymphocytes (x103)

Neutrophils (x106)

mDC2

mDC1

pDC

Dendritic cells

CD14+CD16++

CD14++CD16+

CD14++CD16-

Total

Monocytes (x103)

Circulating Leukocytes (absolute count/mL)

BRAF mutation status

Untreated ECD

Table 1. Blood leukocyte phenotyping of Erdheim-Chester patients according to BRAFV600E mutation.

ARTICLE - Alteration of the systemic immune cell phenotype in ECD F.C. Aubart et al.

ARTICLE - Alteration of the systemic immune cell phenotype in ECD sponse, as well as higher levels of chemokines (IP-10, 1.5fold, P<0.05; and CCL2, 1.3-fold, P<0.05) and a cytokine involved in hematopoiesis (GCSF; 2.2-fold, P<0.05) (Online Supplementary Table S2). However, it is worth mentioning that there was a reduction in plasma CCL22 levels (-34%, P<0.05) in ECD patients carrying the BRAFV600E mutation, after they received pegIFNa treatment. A similar pegIFNa signature was observed when all patients with ECD were

F.C. Aubart et al.

considered, irrespective of their BRAF status (data not shown). Finally, a significant reduction in plasma CCL2 levels was only observed in ECD patients upon vemurafenib therapy compared with levels in untreated ECD patients carrying the BRAF mutation (-49.5%, P<0.05) (Online Supplementary Table S2). Taken together, our results show that patients with ECD who carried the BRAFV600E mutation had an overall more

Figure 2. Impact of first-line therapies on the systemic immune cell signature in patients with Erdheim-Chester disease carrying the BRAF mutation. (A-I) Principal component analysis (A) and blood counts of total (B) and nonclassical (C) monocytes, T helper (D), cytotoxic (E) and B (F) lymphocytes, and plasmacytoid (G), and myeloid 1 (H) and 2 (I) dendritic cells in untreated or treated patients with Erdheim-Chester disease (ECD) carrying the BRAFV600E mutation in comparison with individuals in the control group. Controls (n=17), untreated (n=23), and treated (n=29) ECD patients carrying the BRAFV600E mutation (V600E). Treatments included pegylated interferon a and vemurafenib. P-value for the trend was assessed using the Jonckheere-Terpstra trend test. Haematologica | 107 - June 2022

1352

ARTICLE - Alteration of the systemic immune cell phenotype in ECD pro-inflammatory cytokine and chemokine signature than ECD patients who did not carry this mutation, and this signature appears to have been further exacerbated by pegIFNa treatment. Interrelationship between blood immune cell phenotype and cytokine and chemokine network in patients with Erdheim-Chester disease We next investigated whether the modifications to the circulating cytokine and chemokine concentrations may translate into the major perturbation of the systemic immune cell phenotype, as well as its partial restoration following first-line therapy, in ECD patients carrying the BRAF mutation. For this purpose, correlations were explored between circulating immune cell numbers and the concentrations of cytokines and chemokines, in the entire ECD cohort As shown in Online Supplementary Table S3, although none of these biomolecules was found to be

correlated with blood B or Th lymphocyte levels, the results indicated that the absolute count of nonclassical monocytes (CD14+CD16++) was positively correlated with the plasma concentrations of IFNa2 (r=0.31, P<0.005), IL6 (r=0.30, P<0.05), IL-8 (r=0.23, P<0.05), and IL-5 (r=0.27, P<0.05). Interestingly, plasma IP-10 levels were positively correlated with the abundance of both nonclassical monocytes (r=0.31, P<0.05) and mDC2 cells (r=0.28, P<0.05) in the blood, while a correlation was detected between TNFa and MIP-1b levels with the number of mDC1 cells (r=0.40, P<0.0005) and CT lymphocytes (r=-0.24, P<0.05). As a whole, these findings have led to the identification of a set of cytokines and chemokines that might account for the abundance of nonclassical monocytes and myeloid DC following first-line therapies in the blood of ECD patients who carry the BRAF mutation.

F.C. Aubart et al.

Haematologica | 107 - June 2022

1353

Figure 3. Impact of the BRAFV600E mutation on the systemic chemokine and cytokine network in patients with Erdheim-Chester disease.(A-L) Analysis of the repartition of untreated patients with Erdheim-Chester disease (ECD) according to their BRAF status around the median values of systemic concentrations of IL-6 (A), IL-12p40 (B), IL-15 (C), TNFa (D), IL-10 (E), CCL2 (F), CCL22 (G), eotaxin (H), IL-8 (I), IP-10 (J), MIP-1a (K), and EGF (L). ECD patients without the BRAF mutation (WT), n=9 and ECD patients carrying the BRAFV600E mutation (V600E), n=21. Statistical significance was tested using a χ2 test.

ARTICLE - Alteration of the systemic immune cell phenotype in ECD Immunoglobulin switch toward immunoglobulin G4 in patients with Erdheim-Chester disease Finally, to determine whether the disturbance of the systemic immune cell phenotype translates into a defect in immunoglobulin production, plasma concentrations of immunoglobulin isotypes (IgA, IgM, IgG1, IgG2, IgG3, and IgG4) were quantified in patients with untreated ECD. Although the quantities of IgA and IgM were within the reference ranges for adults,12 those of IgG were more elevated, which mostly reflected the high abundance of IgG4 and, to a lesser extent, high concentrations of IgG2 13 (Online Supplementary Table S4). As a result, in patients with ECD, the proportion of IgG1 (IgG1/IgGs) was low, whereas that of IgG4 (IgG4/IgGs) was high; this effect appeared to be more pronounced in patients carrying the BRAF mutation. Analysis of the distribution of patients according to their IgG4 levels (normal <135 and high ≥135 mg/dL)14 indicated that whereas a roughly similar proportion of patients without the BRAF mutation had either normal or high levels of IgG4, the level of IgG4 in patients carrying the BRAFV600E mutation was 1.7-fold higher than that of patients who lacked the mutation in the normal

F.C. Aubart et al.

IgG4 group and up to 3-fold higher in the high IgG4 group (Figure 4A). Overall, 64.5% of patients with ECD had a high IgG4 immune phenotype and this was predominantly among those patients who carried the BRAF mutation. Strikingly, first-line therapies corrected the IgG switch in this latter group, with a significant increase in IgG1 being observed upon pegIFNa therapy (untreated, 30.9 [25.3– 43.8] versus pegIFNa, 50.3 [36.1–58.3], P<0.005), while IgG4 returned to normal values following vemurafenib treatment (untreated, 16.5 [6.51–37.4] versus vemurafenib, 4.69 [2.02–7.94], P<0.05) (Figure 4). A similar correction of the IgG profile was also observed for the entire ECD cohort treated with first-line therapies (Online Supplementary Figure S2). These findings revealed that patients with ECD exhibited an IgG switch, from IgG1 to IgG4, which was corrected by first-line therapies.

Discussion The present study, involving a single-center series of 78 patients with ECD, revealed a profound perturbation of

Figure 4. Correction of the IgG1/IgG4 switch by first-line therapies in patients carrying the BRAF mutation. (A) Prevalence of the high-level IgG4 phenotype in untreated patients with ErdheimChester disease (ECD) according to the presence of the BRAFV600E mutation. Normal IgG4 <135 mg/dL, high IgG4 ≥135 mg/dL. ECD patients without the BRAF mutation (WT), n=9 and ECD patients carrying the BRAFV600E mutation (V600E), n=22. (B-E) Impact of first-line therapies on the percentage of IgG1 (B), IgG2 (C), IgG3 (D), and IgG4 (E). Untreated (n=22) and treated (n=27; pegIFNa=16 and vemurafenib=11) ECD patients carrying the BRAFV600E mutation (V600E). Differences between groups were tested using the Kruskal-Wallis test. *P<0.05 and **P<0.005 versus untreated ECD patients carrying the BRAFV600E mutation.

Haematologica | 107 - June 2022

1354

ARTICLE - Alteration of the systemic immune cell phenotype in ECD the blood immune phenotype in these patients, characterized by a decrease in the DC and lymphocyte populations accompanied by a switch in IgG subclasses. This perturbation was exacerbated in patients carrying the BRAFV600E mutation, who also exhibited a higher pro-inflammatory status than patients who lacked this mutation. First-line therapies were able to partially correct the altered immune cell phenotype and restore the IgG pattern. This first comprehensive analysis of systemic immune cell populations in patients with ECD revealed a peculiar immune cell ECD signature, characterized by a very low abundance of DC, including pDC, mDC1, and mDC2, in comparison with the abundance of these cells in matched control individuals. Although there is limited information available about the levels of immune cells in the blood of patients with histiocytosis, this observation contrasts with the increased quantity of DC precursors detected in the blood of patients with Langerhans cell histiocytosis, a histiocytic neoplasm that arises from the dendritic lineage.15 Although a trend for such a decrease in DC was observed in patients who lacked the BRAF mutation, a much stronger effect was detected in patients who did carry this mutation, suggesting that the activation of the ERK signaling pathway could underlie this phenotype. The perturbation in blood DC levels was unlikely to have resulted from the increased infiltration of these cells into tissues, as no CD123+ cells (pDC) have previously been detected in ECD lesions.6 Rather, activation of the MEK/ERK signaling pathway was reported to inhibit the maturation of monocyte-derived DC.16,17 More recently, Hogstad et al. elegantly demonstrated that the MAPK pathway, including the BRAFV600E mutation, suppresses DC migration and traps DC in Langerhans cell histiocytosis lesions.18 BRAF mutations have been detected in myeloid progenitors in bone marrow from ECD patients;2 therefore, our findings lead us to propose that the presence of the BRAFV600E mutation in myeloid DC precursors might cause these cells to be retained in the bone marrow compartment and impede their migration to the blood circulation. This mechanism could explain the paradoxical elevated systemic IFNa concentrations reported in patients with ECD,6 despite the low abundance of blood DC described here. Additional investigations are needed to determine whether an increase in myeloid DC precursors can be detected in the bone marrow of patients with ECD. Antigen-presenting cells such as DC interact with lymphocytes and contribute to their proliferation and maturation and the establishment of an immune response. Together with the decrease in blood DC, the systemic concentrations of helper, cytotoxic, and B lymphocytes were markedly reduced in patients with ECD in the present study. Moreover, decreased systemic levels of IL-7, a cytokine involved in B- and T-lymphocyte differentiation,

F.C. Aubart et al.

have been reported in patients with ECD.6 The infiltration of Th1 cells into ECD lesions4 could also contribute to the reduction in the abundance of circulating T lymphocytes. Indeed, CCL19/MIP-3b, a chemo-attractant for B and T lymphocytes and DC, was reported to be expressed in ECD lesions that were analyzed by immunohistochemistry.4 In contrast, the expansion of Treg lymphocytes alone in both the blood compartment and lesions has been reported in Langerhans cell histiocytosis, while monocyte and DC populations were not altered. 19 Despite the low abundance of circulating B cells, a recent study pointed out the high prevalence (42%) of autoimmunity in patients with ECD.20 Here, we brought to light perturbations in the IgG profile characterized by high IgG4 levels and leading to an IgG1/IgG4 switch. A few case reports have documented high IgG4 levels in ECD patients, suggesting ECD mimics IgG4-related disease.14,21 In a review of a single-center cohort, Gianfreda et al. observed that high levels of IgG4 were present in 26.7% (4/15) of patients with ECD.14 In the present study, involving 78 patients, high levels of IgG4 (≥135 mg/dL) were observed more frequently, affecting 64.5% of patients. However, while ECD and IgG4-related disease share some physiopathological characteristics, these diseases have distinctive clinical features, suggesting they are distinct disorders. The increased production of IgG4 is frequently driven through a Th2 response to IL-4, IL-5, or Il-13 and by anti-inflammatory IL-10 and TGFb cytokines.22 Although ECD patients exhibit a Th1 immune response,4,6 the present study indicated that patients who carry the BRAF mutation have higher circulating IL-10 concentrations and are more likely to exhibit high IgG4 levels than patients who lack the mutation, suggesting that IL-10 might contribute to the IgG4 immune response in ECD. Moreover, IFNa, which is secreted by pDC and initiates the Th1 response and whose systemic concentrations are elevated in ECD,6 has been reported to increase IgG4 production by B lymphocytes.23 Thus, infiltration of pDC in pancreatic lesions of patients with IgG4-related autoimmune pancreatitis has been proposed to induce IgG4 production by plasma cells via IFNa.23 Similar to what is observed in cases of IgG4-related disease, IgG4-positive plasma cell infiltrates were observed in ECD lesions at perirenal and subcutaneous sites.14 Whether or not the reduction in circulating B cells in ECD detected in the present study reflects the infiltration of these cells into lesions or impaired B cell differentiation deserves further investigation. The quantification of serum cytokines in a single-center series of 37 patients with ECD was previously reported; it included the identification of an ECD signature based on the concentrations of IFNa2, IL-12, MCP-1, IL-4, and IL-7, which allowed ECD patients to be distinguished from control individuals.6 The present study provides new information regarding the effect of the BRAFV600E mutation on this

Haematologica | 107 - June 2022

1355

ARTICLE - Alteration of the systemic immune cell phenotype in ECD

ECD signature, as well as on the systemic immune Th1 phenotype that characterizes ECD. We found an exacerbated Th1-mediated systemic immune response in patients carrying the BRAF mutation, characterized by a higher proportion of participants carrying the BRAFV600E mutation with elevated circulating concentrations of proinflammatory cytokines (IL-12p40, IL-6, IL-8, and TNFa) and chemokines (IP-10, CCL2, CCL22, and MIP-1a) than participants who lacked this mutation. However, first-line therapies were unable to dampen this pro-inflammatory phenotype. Elevated levels of circulating chemokines in patients carrying the BRAF mutation are consistent with those in previous studies that found that the presence of the BRAFV600E mutation is a major determinant in histiocyte infiltration24 and that vemurafenib shows a high efficacy in multisystemic and refractory ECD.8 The decrease in CCL2 concentrations in ECD patients treated with vemurafenib in comparison with their untreated counterparts carrying the BRAF mutation might account, at least in part, for the reduced infiltration upon receiving this therapy. An analysis of the systemic chemokine and cytokine network in 52 patients with Langerhans cell histiocytosis versus 34 control individuals revealed that patients carrying the BRAFV600E mutation only showed elevated levels of MCP-3 in serum, with no other abnormalities detected.25 Such an elevation of MCP-3 was not observed in patients carrying the BRAF mutation in the present study. In the present study, we found that first-line therapies, although having a modest impact on the systemic chemokine and cytokine concentrations, were able to correct most of the alterations in blood immune cell counts, especially those of nonclassical CD14+CD16++ monocytes,

F.C. Aubart et al.

mDC1, mDC2, and B lymphocytes, whereas they failed to restore those of pDC. While disturbances in the immune response were more frequent in ECD patients carrying the BRAF mutation, targeted therapy appeared to be less effective than pegIFNa in these patients. Nevertheless, vemurafenib, similarly to pegIFNa, corrected the IgG1/IgG4 switch. The major findings of the present study are summarized in Figure 5. The limitations of this study include the relatively low number of patients with ECD on the different therapies (vemurafenib versus pegIFNa). Another limitation is the absence of data from ECD patients before and after treatment, which would be useful to investigate in more detail the impact of first-line therapies on the systemic disturbance of the immune cell phenotype and the IgG switch. Finally, the inclusion of a control group comprising patients with Langerhans cell histiocytosis would have helped as a comparison with the specific inflammatory patterns in ECD. In conclusion, our study is the first to document the marked alteration of the systemic immune response in ECD and brings to light the involvement of DC in this non- Langerhans cell histiocytosis neoplasm. This new information will help our understanding of the mechanisms taking place in ECD physiopathology and provides additional clues to the best approach to the therapeutic management of patients with ECD. Disclosures No conflicts of interest to disclose. Contributions FCA, LP, FS-C, YA, EF, and WLG performed the research.

Figure 5. Major alterations of the systemic immune cell phenotype in patients with Erdheim-Chester disease. Flow cytometry analysis of blood leukocytes in patients with Erdheim-Chester disease revealed a marked decrease in dendritic cells (pDC, mDC1, and mDC2) and lymphocytes (CTL and BL), as well as a reduction in nonclassical monocytes in comparison with levels of these cells in individuals in the control group. Such a reduction in antigen-presenting cells might impair the activation of cytotoxic T lymphocytes and B lymphocytes and the production of IgG, leading to an IgG switch toward IgG4. These alterations were mostly observed in ECD patients carrying the BRAFV600E mutation (in red), who exhibited a more pronounced systemic inflammation. First-line therapies partially corrected the systemic immune cell phenotype and normalized blood IgG concentrations. BL: B lymphocytes; CTL: cytotoxic T lymphocytes; IgG: immunoglobulin G; NC: nonclassical; mDC: myeloid dendritic cells; pDC: plasmacytoid dendritic cells. Haematologica | 107 - June 2022

1356

ARTICLE - Alteration of the systemic immune cell phenotype in ECD FCA, PL, JH, and WLG designed the research. FCA and JH recruited the patients. MG and WLG analyzed the data. JH and WLG obtained funding for the research. All authors gave critical comments on the manuscript. WLG supervised the study and wrote the manuscript.

F.C. Aubart et al.

Funding INSERM, Sorbonne Université, French National Agency (ANR-10-IAHU-05), and the Erdheim-Chester Disease Global Alliance.

Acknowledgments The authors are indebted to all the participants for their cooperation.

References 1. Durham BH. Molecular characterization of the histiocytoses: neoplasia of dendritic cells and macrophages. Semin Cell Dev Biol. 2019;86:62-76. 2. Milne P, Bigley V, Bacon CM, et al. Hematopoietic origin of Langerhans cell histiocytosis and Erdheim-Chester disease in adults. Blood. 2017;130(2):167-175. 3. Durham BH, Roos-Weil D, Baillou C, et al. Functional evidence for derivation of systemic histiocytic neoplasms from hematopoietic stem/progenitor cells. Blood. 2017;130(2):176-180. 4. Stoppacciaro A, Ferrarini M, Salmaggi C, et al. Immunohistochemical evidence of a cytokine and chemokine network in three patients with Erdheim-Chester disease: implications for pathogenesis. Arthritis Rheum. 2006;54(12):4018-4022. 5. Dagna L, Girlanda S, Langheim S, et al. Erdheim-Chester disease: report on a case and new insights on its immunopathogenesis. Rheumatology. 2010;49(6):1203-1206. 6. Arnaud L, Gorochov G, Charlotte F, et al. Systemic perturbation of cytokine and chemokine networks in Erdheim-Chester disease: a single-center series of 37 patients. Blood. 2011;117(10):2783-2790. 7. Cohen-Aubart F, Maksud P, Emile J-F, et al. Efficacy of infliximab in the treatment of Erdheim-Chester disease. Ann Rheum Dis. 2018;77(9):1387-1390. 8. Haroche J, Cohen-Aubart F, Emile J-F, et al. Dramatic efficacy of vemurafenib in both multisystemic and refractory ErdheimChester disease and Langerhans cell histiocytosis harboring the BRAF V600E mutation. Blood. 2013;121(9):1495-1500. 9. Goyal G, Heaney ML, Collin M, et al. Erdheim-Chester disease: consensus recommendations for evaluation, diagnosis, and treatment in the molecular era. Blood. 2020;135(22):1929-1945. 10. Haroche J, Charlotte F, Arnaud L, et al. High prevalence of BRAF V600E mutations in Erdheim-Chester disease but not in other non-Langerhans cell histiocytoses. Blood. 2012;120(13):27002703. 11. Szalat R, Pirault J, Fermand J-P, et al. Physiopathology of necrobiotic xanthogranuloma with monoclonal gammopathy. J Intern Med. 2014;276(3):269-284. 12. Gonzalez-Quintela A, Alende R, Gude F, et al. Serum levels of immunoglobulins (IgG, IgA, IgM) in a general adult population and their relationship with alcohol consumption, smoking and common metabolic abnormalities: serum immunoglobulin levels in adults. Clin Exp Immunol. 2008;151(1):42-50. 13. Schauer U, Stemberg F, Rieger CHL, et al. IgG subclass concentrations in certified reference material 470 and reference

values for children and adults determined with the binding site reagents. Clin Chem. 2003;49(11):1924-1929. 14. Gianfreda D, Musetti C, Nicastro M, et al. Erdheim-Chester disease as a mimic of IgG4-related disease: a case report and a review of a single-center cohort. Medicine (Baltimore). 2016;95(21):e3625. 15. Rolland A, Guyon L, Gill M, et al. Increased blood myeloid dendritic cells and dendritic cell-poietins in Langerhans cell histiocytosis. J Immunol. 2005;174(5):3067-3071. 16. Puig-Kröger A, Relloso M, Fernández-Capetillo O, et al. Extracellular signal-regulated protein kinase signaling pathway negatively regulates the phenotypic and functional maturation of monocyte-derived human dendritic cells. Blood. 2001;98(7):2175-2182. 17. Aguilera-Montilla N, Chamorro S, Nieto C, et al. Aryl hydrocarbon receptor contributes to the MEK/ERK-dependent maintenance of the immature state of human dendritic cells. Blood. 2013;121(15):e108-117. 18. Hogstad B, Berres M-L, Chakraborty R, et al. RAF/MEK/extracellular signal-related kinase pathway suppresses dendritic cell migration and traps dendritic cells in Langerhans cell histiocytosis lesions. J Exp Med. 2018;215(1):319-336. 19. Senechal B, Elain G, Jeziorski E, et al. Expansion of regulatory T cells in patients with Langerhans cell histiocytosis. PLoS Med. 2007;4(8):e253. 20. Roeser A, Cohen-Aubart F, Breillat P, et al. Autoimmunity associated with Erdheim-Chester disease improves with BRAF/MEK inhibitors. Haematologica. 2019;104(11):e502-e505. 21. Miron G, Karni A, Faust-Soher A, Giladi N, Alroy H, Gadoth A. Erdheim-Chester disease presenting with chorea and mimicking IgG4-related disorder. Neurol Clin Pract. 2019;9(6):524-526. 22. Liu C, Zhang P, Zhang W. Immunological mechanism of IgG4related disease. J Transl Autoimmun. 2020;3:100047. 23. Arai Y, Yamashita K, Kuriyama K, et al. Plasmacytoid dendritic cell activation and IFN- production are prominent features of murine autoimmune pancreatitis and human IgG4-related autoimmune pancreatitis. J Immunol. 2015;195(7):3033-3044. 24. Cohen-Aubart F, Guerin M, Poupel L, et al. Hypoalphalipoproteinemia and BRAF V600E mutation are major predictors of aortic infiltration in the Erdheim-Chester disease. Arterioscler Thromb Vasc Biol. 2018;38(8):1913-1925. 25. Morimoto A, Oh Y, Nakamura S, et al. Inflammatory serum cytokines and chemokines increase associated with the disease extent in pediatric Langerhans cell histiocytosis. Cytokine. 2017;97:73-79.

Haematologica | 107 - June 2022

1357

ARTICLE - Hemostasis

Critical role of peroxisome proliferator-activated receptor a in promoting platelet hyperreactivity and thrombosis under hyperlipidemia Li Li,1,2* Jiawei Zhou,3* Shuai Wang,1,2* Lei Jiang,4 Xiaoyan Chen,1,2 Yangfan Zhou,1,2 Jingke Li,1,2 Jingqi Shi,1,2 Pu Liu,5 Zheyue Shu,6,7 Frank J. Gonzalez,8 Aiming Liu9# and Hu Hu1,2,10# Department of Pathology and Pathophysiology and Bone Marrow Transplantation Center of the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China; 2 Institute of Hematology, Zhejiang Engineering Laboratory for Stem Cell and Immunotherapy, Zhejiang University, Hangzhou, China; 3State Key Laboratory of Diagnosis and Treatment for Infectious Diseases, The First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China;4Department of Pathology, Zhejiang Provincial Key Laboratory of Pathophysiology, Ningbo University School of Medicine, Ningbo, China; 5 Department of Pathology of the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China; 6Division of Hepatobiliary and Pancreatic Surgery, Department of Surgery, First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China; 7Key Laboratory of Combined Multi-Organ Transplantation, Ministry of Public Health, Hangzhou, China; 8Laboratory of Metabolism, Center for Cancer Research, National Cancer Institute, NIH, Bethesda, MD, USA; 9Ningbo University School of Medicine, Ningbo, China; 10 Key Laboratory of Disease Proteomics of Zhejiang Province, Hangzhou, China. 1

Correspondence: Hu Hu huhu@zju.edu.cn Aiming Liu liuaiming@nbu.edu.cn Received: August 5, 2021. Accepted: September 29, 2021. Prepublished: October 7, 2021. https://doi.org/10.3324 haematol.2021.279770 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

LL, JZ and SW contributed equally as co-first authors. AML and HH contributed equally as co-senior authors.

Abstract Platelet hyperreactivity and increased atherothrombotic risk are specifically associated with dyslipidemia. Peroxisome proliferator-activated receptor alpha (PPARa) is an important regulator of lipid metabolism. It has been suggested to affect both thrombosis and hemostasis, yet the underlying mechanisms are not well understood. In this study, the role and mechanism of PPARa in platelet activation and thrombosis related to dyslipidemia were examined. Employing mice with deletion of PPARa (Ppara-/-), we demonstrated that PPARa is required for platelet activation and thrombus formation. The effect of PPARa is critically dependent on platelet dense granule secretion, and is contributed by p38MAPK/Akt, fatty acid b-oxidation, and NAD(P)H oxidase pathways. Importantly, PPARa and the associated pathways mediated a prothrombotic state induced by a high-fat diet and platelet hyperactivity provoked by oxidized low density lipoproteins. Platelet reactivity was positively correlated with the levels of expression of PPARa, as revealed by data from wild-type, chimeric (Ppara+/-), and Ppara-/- mice. This positive correlation was recapitulated in platelets from hyperlipidemic patients. In a lipid-treated megakaryocytic cell line, the lipid-induced reactive oxygen species-NF-kB pathway was revealed to upregulate platelet PPARa in hyperlipidemia. These data suggest that platelet PPARa critically mediates platelet activation and contributes to the prothrombotic status under hyperlipidemia.

Introduction Underpinned by platelet hyperactivity, atherothrombotic disease is the leading cause of mortality and morbidity worldwide. Dyslipidemia has been firmly established as a risk factor for atherothrombotic disease.1,2 Despite the vigorous efforts that have been devoted to establishing the pathways leading to platelet hyperactivity in dyslipidemia, 3,4 the mechanisms responsible are still unclear. Identification of key targets by which dyslipidemia regulates platelet activity is im-

perative for the prevention and management of atherothrombotic disease. Oxidized low density lipoproteins (oxLDL), the product of dysfunctional lipid metabolism, are major promoters of a prothrombotic state in both animal models and human patients.5,6 Scavenger receptor CD36 and signaling pathways such as Src family kinases (SFK), mitogen-activated protein kinases and reactive oxygen species (ROS) are involved in oxLDL-induced platelet activation.7-9 Molecules involved in lipid metabolism such as the transcription factors farnesoid X receptor,10 liver X receptor,11 and PPAR12-14

Haematologica | 107 - June 2022

1358

ARTICLE - PPARa promotes platelet hyperreactivity. are also expressed in platelets. How these molecules interact with the established platelet activation network is ill-defined. PPARa is a major regulator of lipid metabolism in nucleated cells by upregulating the transcription of lipid-metabolizing enzymes, such as carnitine palmitoyl-CoA transferase-I (CPT-I) and acyl-CoA oxidase.15-17 It is expressed in anucleate platelets and was reported to play roles in thrombosis and hemostasis.18,19 Although PPARa may interact with protein kinase C during platelet activation,18 the underlying signaling mechanism has not been elucidated. A variety of metabolic and pathological conditions are related to PPARa expression,20-22 but it has not been explored whether or how platelet activation is regulated by PPARa. In this study, we investigated the role of PPARa in dyslipidemia-related prothrombotic potential and platelet hyperactivity. PPARa expression in platelets was enhanced in both a hyperlipidemic mouse model and patients, which correlated well with platelet hyperactivity. The mechanism increasing PPARa expression in platelets and the platelet functions targeted by PPARa were explored.

Methods Subjects The procedures in human subjects were approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University, and informed consent was obtained from the study participants. Blood from 36 patients with hypertriglyceridemia, 16 patients with hypercholesterolemia and 31 healthy subjects was obtained. None of the participants had taken any antiplatelet or other nonsteroidal anti-inflammatory drugs for at least 14 days before blood collection. None of the patients had clinical evidence of cardiovascular disease (according to their clinical history, physical examination, and electrocardiogram). Moreover, exclusion criteria for all subjects included renal insufficiency, proteinuria, altered hepatic function and alcohol abuse. Patients with diabetes mellitus (fasting blood glucose level >115 mg/dL or treatment with a hypoglycemic agent), hypertension (systolic blood pressure >140 mmHg, diastolic blood pressure >90 mmHg) and smokers were also excluded. Animals Previously reported PPARa-deficient mice (Ppara-/-) were used in these experiments.23 All mice were 8 to 14 weeks old, and matched for weight and sex. Male ApoE-deficient (Apoe-/-) mice (6 weeks old) were purchased from the Model Animal Research Center of Nanjing University (Nanjing, China). Ppara-/-/Apoe-/- mice were generated by cros-

L. Li et al. sing Ppara-/- and Apoe-/- mice in the animal facilities of Zhejiang University. The animals were fed a normal chow diet until 8 weeks. Their diet was then switched to a highfat diet, containing 40% fat and 1.25% cholesterol (Trophic Animal Feed High-Tech Co., Ltd, China), for 8 weeks. All animals were maintained under standard conditions of room temperature and humidity with a 12-hour dark-light cycle. All animal protocols were approved by Zhejiang University Laboratory Animal Welfare and Ethics Committee. Preparation of washed human platelets All blood donors had antecubital veins allowing a clean venipuncture. Blood was drawn without stasis into siliconized vacutainers containing 1/9 v/v 3.8% sodium citrate, then washed platelets were re-suspended as previously described.24 Preparation of washed mice platelets Whole blood was collected from the inferior vena cava into a 0.2 volume of ACD buffer (75 mM sodium citrate, 39 mM citric acid, and 135 mM dextrose, pH 6.5), and was diluted 1:3 with modified Tyrode buffer (20 mM HEPES, 137 mM NaCl, 13.8 mM NaHCO3, 2.5 mM KCl, 0.36 mM NaH2PO4, 5.5 mM glucose, pH 7.4). Diluted whole blood was centrifuged at 180 g for 10 min at room temperature. The platelet-rich plasma was collected into a fresh tube containing 500 mL ACD, and centrifuged at 700 g for 10 min. The platelet pellet was then re-suspended in modified Tyrode buffer. Statistical analysis Results are expressed as mean ± standard error of the mean (SEM). Statistical significance was evaluated with a paired t-test, two-tailed Mann-Whitney U tests and twoway analysis of variance (ANOVA) using the statistical software GraphPad Prism (GraphPad Software, La Jolla, CA, USA).

Results Ppara-/- mice display impaired hemostasis and thrombosis Ppara-/- mice were genotyped by polymerase chain reaction and ablation of Ppara was confirmed in PPARa-deficient platelets by western blot analysis (Figure 1A). The levels of expression of PPARb and PPARg were similar in heterozygous (Ppara+/-) platelets, Ppara-/- platelets and wild-type (WT) platelets (Figure 1A). Moreover, Ppara-/mice were viable and fertile, and did not exhibit any evident bleeding tendency or thrombotic events over their lifespan. Ppara-/- mice did not differ significantly from their WT littermates with regard to platelet count, red blood cell count, white blood cell count, hematocrit, or

Haematologica | 107 - June 2022

1359

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Figure 1. PPARa-deficient mice display impaired hemostasis and thrombosis. (A) Genotyping results of wild-type (WT), Ppara+/and Ppara-/- mice using polymerase chain reaction. Immunoblot analysis of PPARa, PPARb and PPARg expression in platelets from WT, Ppara+/-, Ppara-/- mice and humans. (B) Bleeding times for WT (●), Ppara+/- (■) and Ppara-/- (▲) mice. Means are indicated by horizontal lines. Statistical significance was evaluated with a paired t test (*P<0.05; ***P<0.001; ns: not significant). Percentages of WT, Ppara+/- and Ppara-/- mice bleeding times that exceeded 15 min ( ) or were within 15 min (■). Results were obtained from 26 WT, 26 Ppara+/- and 26 Ppara-/- mice. (C) An injury to the carotid artery was induced by FeCl3. The dot plot shows occlusion times for carotid arterioles as a result of FeCl3-induced thrombosis in WT (●, n=6), Ppara+/- (■, n=9) and Ppara-/- mice (▲, n=6). Means are indicated by horizontal lines. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; Legend on following page. Haematologica | 107 - June 2022

1360

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

**P<0.01; ns: not significant). (D) Representative images and time courses of thrombus formation induced by FeCl3 injury to mesenteric arterioles in WT (top row), Ppara+/- (middle row) and Ppara-/- (bottom row) mice. a: arteriole; v: venule. Scale bars, 100 mm (left panel). Dot plot showing occlusion times for arterioles as a result of FeCl3-induced thrombosis in WT (●, n=22), Ppara+/- (■, n=33) and Ppara-/- mice (▲, n=26). Means are indicated by horizontal lines. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; ns: not significant). (E) Photomicrographs showing the progression of adhesion of platelets from WT, Ppara+/- and Ppara-/- mice on collagen. Whole blood from WT, Ppara+/- and Ppara-/- mice, collected in heparin (7.5 U/mL), was fluorescently labeled by incubation with mepacrine (100 mM) for 30 min, and then perfused through fibrillar collagen-coated bioflux plates at a shear rate of 40 dynes/cm2 for 5 min. Original magnification, ×10. Scale bar, 100 mm (left panel). Dot plot showing area coverage of platelets from WT (●), Ppara+/- (■) and Ppara-/- (▲) mice (n=3 for each group; two-way analysis of variance test, *P<0.05; ***P<0.001).

hemoglobin concentration (Online Supplementary Table S1). Electron microscopy showed normal discoid morphology of Ppara-/- platelets with unaltered numbers of a granules and dense granules, compared to those in the WT platelets (Online Supplementary Figure S1A). No significant differences in the surface expression of platelet CD41 (aIIb subunit), and CD42b (GPIbα subunit) were found between WT and Ppara-/- platelets (Online Supplementary Figure S1B). With a tail-bleeding assay, Ppara-/- mice showed significantly prolonged tail bleeding time (712.80 ± 58.11 seconds vs. 333.40 ± 64.76 seconds; P<0.001) (Figure 1B), agreeing with findings from a previous study.18 Moreover, 73% of the Ppara-/- mice had a bleeding time exceeding 15 min, while the percentage in WT littermates was 19% (Figure 1B). In a FeCl3-induced model of carotid artery thrombosis, the time to formation of a stable occlusive thrombus in the carotid artery was significantly longer in Ppara-/- mice than in WT mice (10.88 ± 1.51 min vs. 3.53 ± 0.59 min, P<0.01) (Figure 1C). In a FeCl3-induced model of mesenteric arteriole thrombosis, the time to formation of stable occlusive thrombi was significantly longer in Ppara-/- mice than in WT mice (38.35 ± 3.10 min vs. 27.55 ± 2.81 min, P<0.05) (Figure 1D). Interestingly, heterozygous Ppara+/mice also had a significant perturbance of thrombotic and hemostatic functions, with significantly increased tail bleeding time, rate of non-stoppable bleeding, and time to the formation of stable occlusive thrombi compared to those of WT mice (Figure 1B-D). In a model of deep vein thrombosis, Ppara+/- and Ppara-/- mice developed thrombi similar in weight and length to those observed in WT mice (Online Supplementary Figure S2A, B), indicating a complex multi-cellular interaction upon PPARa deficiency in thrombo-inflammation. These in vivo data indicate that PPARa is essential for hemostasis and thrombosis, functions governed by platelets; nevertheless, the role of PPARa is complicated in thrombo-inflammation because the outcome is dictated by the interaction of platelets and inflammatory cells. In a microfluidic perfusion assay, when whole-blood was perfused over an immobilized collagen surface at the shear stress of 1000 s-1 for 5 min, the areas covered by Ppara+/- and Ppara-/- platelets were 27.5% and 44.9% smaller in average than those by WT platelets (Figure 1E). A recombinant whole-blood system with diluted washed

platelets (2 × 107/mL) was used in the same assay to assess the collagen-adhesion ability of platelets. In the absence of platelet aggregation, the areas covered by collagen-adhered Ppara-/- platelets were similar to those of WT platelets (Online Supplementary Figure S3). These findings indicate that PPARa functions in regulating the growth of platelet thrombi, not the initial adhesion. Ppara-/- platelets show functional defects due to an impaired ATP secretion Next, platelet aggregation in response to common platelet stimuli was analyzed. Compared to WT platelets, Ppara-/platelets displayed an average 30% reduction of aggregation rates in response to thrombin (0.025 U/mL) and 57% reduction in response to collagen (0.8 mg/mL) (Figure 2A). However, ADP and TXA2 analog U46619-induced platelet aggregation was not affected by PPARa deficiency (Online Supplementary Figure S4A). Although dense granule content was normal in Ppara-/- platelets (Online Supplementary Figure S4B), ATP release induced by low doses of thrombin (0.025 U/mL), collagen (0.8 mg/mL) and U46619 (0.3 mM) was largely inhibited in Ppara-/- platelets (Figure 2A and Online Supplementary Figure S4A). Again, Ppara+/platelets exhibited intermediate rates of aggregation and dense granule secretion (Figure 2A). Higher concentrations of thrombin (0.05 U/mL) and collagen (2 mg/mL) overcame the defective aggregation and dense granule secretion in Ppara-/- platelets (Online Supplementary Figure S4A). The aggregation differences between WT and Ppara-/- platelets were abolished when apyrase was applied to hydrolyze dense granule-secreted ATP and ADP (Figure 2B and Online Supplementary Figure S4C). Conversely, supplementation with a low concentration of ADP (1 mM), which was insufficient to induce aggregation on its own, rescued the defective aggregation of Ppara+/- and Ppara-/- platelets stimulated by thrombin or collagen (Figure 2C and Online Supplementary Figure S4D). As indicated by the measurement of TXB2, collagen- or thrombin-induced TXA2 production was comparable between WT and Ppara-/platelets (Online Supplementary Figure S4E). Moreover, the secretion of a-granules and activation of aIIbb3, measured respectively by P-selectin expression and the binding of Jon/A antibody, were not influenced by PPARa deficiency in response to thrombin and convulxin (Online Supplementary Figure S4F). These data suggest that the impaired

Haematologica | 107 - June 2022

1361

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Legend on following page. Haematologica | 107 - June 2022

1362

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Figure 2. Ppara-/- platelets show impaired aggregation, ATP secretion, spreading and delayed clot retraction. (A) Aggregation and ATP release of washed wild-type (WT), Ppara+/- and Ppara-/- platelets stimulated with thrombin (0.025 U/mL) or collagen (0.8 mg/mL). Aggregation and ATP release were assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments. Statistical significance was evaluated with a two-tailed Mann-Whitney test and a paired t test (*P<0.05; **P<0.01; ***P<0.001). (B) Aggregation and ATP release of washed WT, Ppara+/- and Ppara-/- platelets stimulated with thrombin (0.025 U/mL) or collagen (0.8 mg/mL) in the presence of vehicle or apyrase (1 U/mL) incubated for 5 min. Traces are representative of at least three independent experiments. (C) Aggregation of washed WT, Ppara+/- and Ppara-/- platelets stimulated with thrombin (0.025 U/mL) or (0.8 mg/mL) in the presence of a low concentration of ADP (1 mM). Traces are representative of at least three independent experiments. (D) Spreading of WT, Ppara+/- and Ppara-/- platelets on immobilized fibrinogen in the presence or absence of apyrase (1 U/mL) or ADP (1 mM). Images are representative of three independent experiments with similar results. Original magnification, ×100. Scale bar, 10 mm (left panel). Statistical significance was evaluated with a two-tailed Mann-Whitney test (**P<0.01; ***P<0.001; ns: not significant). (E) Platelets from WT, Ppara+/and Ppara-/- mice were resuspended with human platelet-poor plasma at a concentration of 4×108/mL, and recombined plasma was stimulated to coagulate with thrombin (0.4 U/mL), then photographed at different time points. Statistical significance was evaluated with a paired t test (*P<0.05; **P<0.01).

aggregation in Ppara-/- platelets is caused by the reduced ADP secretion. Consistent with the role of ADP in thrombus amplification, Ppara-/- platelets formed smaller aggregates than WT platelets when stimulated with low doses of thrombin and collagen (Online Supplementary Figure S4G). Platelet spreading on immobilized fibrinogen and clot retraction, two processes controlled by early and late integrin aIIbb3-mediated outside-in signaling, respectively, were then measured. Platelet spreading on immobilized fibrinogen (Figure 2D) and clot retraction (Figure 2E) were also inhibited by PPARa deficiency. Apyrase eliminated the spreading difference between WT and Ppara-/- platelets, and exogenous ADP (1 mM) rescued the defective spreading of Ppara-/- platelets (Figure 2D). Clot retraction mediated by Ppara-/- platelets showed a significant delay compared to that by WT platelets (Figure 2E). These data demonstrate an important role for PPARa in platelet dense granule secretion and its activation. PPARa promotes platelet activation through a p38/ROS/Akt signal axis When platelet signaling events were analyzed, both collagen and thrombin induced a significantly reduced phosphorylation of Akt (Thr308/Ser473) and p38 (Thr180/Tyr182) in Ppara-/- platelets (Figure 3A and Online Supplementary Figure S5A), while phosphorylation of ERK1/2 (Thr202/Tyr204) and JNK (Thr183/Tyr185) remained unaltered (Online Supplementary Figure S5A). Moreover, both SB203580 and SH-6, inhibitors of p38 and Akt, strongly inhibited platelet aggregation and ATP release induced by thrombin and collagen, but no additive effects with PPARa deficiency were observed (Figure 3B and Online Supplementary Figure S5B). Therefore, PPARa is functionally coupled to p38 and Akt activation. As p38 regulates the production of ROS,25,26 thrombin and convulxin induced significantly less ROS production in Ppara+/- and Ppara-/- platelets than in WT platelets (Figure 3C), while ROS scavenging by N-acetylcysteine (NAC) essentially eliminated the aggregation and ATP release difference among WT, Ppara+/-, and Ppara-/platelets (Figure 3D and Online Supplementary Figure S5C).

Intriguingly, NAC treatment abolished the difference of Akt phosphorylation but left intact the difference of p38 phosphorylation among WT, Ppara+/- and Ppara-/- (Figure 3E and Online Supplementary Figure S6A). Furthermore, SB203580 and SH-6 eliminated the differences of Akt phosphorylation (Figure 3F and Online Supplementary Figure S6B) and ROS production (Figure 3G) between WT and Ppara-/- platelets; whereas SH-6 did not change the phosphorylation of p38 (Figure 3F and Online Supplementary Figure S6B). These data suggest a sequential relay of PPARa, p38, ROS production, and Akt during platelet activation. Sources of ROS were also investigated using the NADPH oxidase inhibitor VAS2780, CPT-I inhibitor etomoxir, and xanthine oxidase inhibitor allopurinol (Online Supplementary Figure S6C). VAS2780 and etomoxir, but not allopurinol, eliminated the differences of platelet aggregation and ATP release between WT and Ppara-/- platelets (Figure 3H and Online Supplementary Figure S6D). Both VAS2780 and etomoxir eliminated the phosphorylation difference of Akt, but not that of p38 between WT and Ppara-/- platelets (Online Supplementary Figure S6E). These data suggest that ROS from NADPH oxidase and mitochondrial fatty acid b-oxidation constitute the important sources of ROS for PPARa-regulated platelet activation. Consistent with previous reports,18,19 a synthetic PPARa agonist WY14643 inhibited aggregation and ATP release induced by low doses of thrombin and collagen in WT platelets (Online Supplementary Figure S7A), and essentially abolished the differences between WT and Ppara-/- platelets (Online Supplementary Figure S7A). Unexpectedly, a PPARa antagonist GW6471 also inhibited aggregation and ATP release in WT platelets and eliminated the difference between WT and Ppara-/- platelets in response to low doses of thrombin and collagen (Online Supplementary Figure S7B). WY14643 and GW6471 both inhibited phosphorylation of Akt Ser473 upon platelet activation by collagen, indicating that these compounds act by interrupting PPARa signaling in platelets (Online Supplementary Figure S7C). Moreover, GW6471 per se but not WY14643 induced phosphorylation of Akt Ser473 (Online Supplementary Figure

Haematologica | 107 - June 2022

1363

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Legend on following page. Haematologica | 107 - June 2022

1364

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Figure 3. PPARa promotes platelet activation through a p38/ROS/Akt signal axis. (A) Immunoblot analysis of wild-type (WT), Ppara+/and Ppara-/- platelets, stimulated with thrombin (0.025 U/mL) and collagen (0.8 mg/mL) for 5 min, with antibodies recognizing phosphorylated Akt Thr308, phosphorylated Akt Ser473, total Akt, phosphorylated p38 Thr180/Tyr182 (T180/Y182), and total p38. Representative immunoblots from at least three independent experiments. (B) Washed WT, Ppara+/- and Ppara-/- platelets (2×108/mL) were incubated with dimethyl sulfoxide (DMSO), SH-6 (10 mM), SB203580 (10 mM) for 10 min, then stimulated with thrombin (0.025 U/mL) or collagen (0.8 mg/mL), respectively. Aggregation and ATP release were assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments. (C) Generation of reactive oxygen species (ROS) analyzed by flow cytometry. H2DCFDA-loaded (50 mM) mice platelets were stimulated with thrombin (0.025 and 0.05 U/mL) or convulxin (50 and 100 ng/mL) for 5 min. Samples were analyzed immediately. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; **P<0.01; ns: not significant). (D) Washed WT, Ppara+/- and Ppara-/- platelets (2×108/mL) were incubated with or without N-acetylcysteine (NAC, 2 mM) for 5 min, then stimulated with thrombin (0.025 U/mL) or collagen (0.8 mg/mL), respectively. Aggregation and ATP release were assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments. (E) Immunoblot analysis of WT, Ppara+/- and Ppara-/- platelets stimulated with thrombin (0.025 U/mL) and collagen (0.8 mg/mL) for 5 min in the absence or presence of NAC, with antibodies recognizing phosphorylated Akt Thr308, phosphorylated Akt Ser473, total Akt, phosphorylated p38 Thr180/Tyr182 (T180/Y182), and total p38. Representative immunoblots from at least three independent experiments. (F) Immunoblot analysis of WT, Ppara+/- and Ppara-/platelets, stimulated with thrombin (0.025 U/mL) or collagen (0.8 mg/mL) for 5 min in the presence of DMSO, SH-6 (10 mM), and SB203580 (10 mM), with antibodies recognizing phosphorylated Akt Thr308, phosphorylated Akt Ser473, total Akt, phosphorylated p38 Thr180/Tyr182 (T180/Y182), and total p38. Representative immunoblots from at least three independent experiments. (G) H2DCFDA-loaded (50 mM) mice platelets were incubated with DMSO, SH-6 (10 mM) or SB203580 (10 mM), stimulated with thrombin (0.05 U/mL) or convulxin (100 ng/mL) for 5 min. Samples were analyzed immediately. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; **P<0.01; ***P<0.001; ns: not significant). (H) Washed WT, Ppara+/- and Ppara-/- platelets (2×108/mL) were incubated with DMSO, etomoxir (25 mM), VAS2870 (10 mM), or ALP (200 mM) for 10 min, then stimulated with thrombin (0.025 U/mL) or collagen (0.8 mg/mL), respectively. Aggregation and ATP release were assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments.

S7D). Therefore, GW6471 may be a partial agonist in the context of platelet activation. Hence, the agonist and antagonist of PPARa both appear to inhibit platelet function, possibly due to the disruption of signal transduction mediated by PPARa. PPARa mediates hyperlipidemia-associated prothrombotic status and oxidized low-density lipoprotein-evoked platelet activation Consistent with the previous study,27 after 8 weeks of a high-fat diet, total plasma levels of cholesterol and triglycerides in Ppara+/+/Apoe-/- mice were significantly increased, while they did not undergo further change in Ppara-/-/Apoe-/- mice (Online Supplementary Table S2). The occlusion time in FeCl3-induced mesenteric arteriole thrombosis in Ppara+/+/Apoe-/- mice was significantly shortened by a high-fat diet when compared with the control diet. But it was comparable between Ppara-/-/Apoe-/- mice fed with a high-fat diet and Ppara+/+/Apoe-/- mice fed the control diet (Figure 4A). KODiA-PC, 1-(palmitoyl)-2-(5-keto-6-octene-dioyl) phosphatidylcholine, one of the most potent CD36 ligands in the oxLDL species, caused direct platelet aggregation, which was largely decreased in Ppara-/- platelets (Figure 4B and Online Supplementary Figure S8A). Consistently, KODiA-PC induced a significantly reduced phosphorylation of Src (Tyr418), p40phox and ERK5 (Thr218/Tyr220), Akt (Ser473) and p38 (Thr180/Tyr182) in Ppara-/- platelets, compared with WT platelets (Figure 4C). Moreover, the NADPH oxidase inhibitor VAS2780, the ROS scavenger NAC, the ERK5 inhibitor BIX02188, the p38 inhibitor SB203580 and the Akt inhibitor SH-6, but not the CPT-I inhibitor etomoxir, elimi-

nated the difference in aggregation between WT and Ppara-/- platelets (Figure 4D and Online Supplementary Figure S8B). Robust ROS production induced by KODiA-PC was also reduced by PPARa deficiency (Figure 4E). Interestingly, ROS production by KODiA-PC was inhibited by VAS2780 and NAC, but not etomoxir (Figure 4E). These data indicate that PPARα mediates oxLDL-induced platelet activation, which is associated with an altered ROS generation pathway. Platelet PPARa expression correlates with platelet hyperreactivity Having established the importance of PPARa in hyperlipidemic-induced platelet activation, we further found that PPARa expression in platelets was significantly increased in mice fed a high-fat diet, compared to that in mice fed with a control diet (Figure 5A). The expression of platelet PPARb and PPARg was however not significantly influenced by the high-fat diet (Figure 5A). In mice fed a high-fat diet, the increase of PPARa expression was accompanied by an increase in platelet aggregation induced by thrombin and collagen (Figure 5B and Online Supplementary Figure S9). These data indicate that increased platelet PPARa expression in hyperlipidemic mice is responsible for platelet hyperactivity. Platelet PPARa expression was also determined in healthy volunteers and patients with hyperlipidemia (Online Supplementary Table S3). Platelet PPARa protein and mRNA levels were significantly increased in patients with hypertriglyceridemia and hypercholesterolemia (Figure 5C, D), although the expression of PPARb and PPARg proteins was similar in healthy subjects and hyperlipidemic patients (Fi-

Haematologica | 107 - June 2022

1365

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Legend on following page. Haematologica | 107 - June 2022

1366

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Figure 4. PPARa mediates hyperlipidemia-associated prothrombotic status and oxLDL-evoked platelet activation. (A) Ppara+/+/Apoe-/- and Ppara-/-/Apoe-/- mice were fed a high-fat diet (HFD) or control diet (CD) for 8 weeks before undergoing in vivo thrombosis experiments. Platelets were labeled by direct tail vein injection of DiOC6 (10 mM, 100 mL). Dot plot showing occlusion times for arterioles as a result of FeCl3-induced thrombosis in Ppara+/+/Apoe-/- and Ppara-/-/Apoe-/- mice. Means are indicated by horizontal lines. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; **P<0.01). (B) Platelets were stimulated with 1-(palmitoyl)-2-(5-keto-6-octenedioyl) phosphatidylcholine (KODiA-PC, 15 mM). Aggregation was assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments. (C) Immunoblot analysis of wild-type (WT), Ppara+/- and Ppara-/- platelets stimulated with KODiA-PC for 5 min, with antibodies recognizing phosphorylated Src Tyr418 (Y418), total Src, phosphorylated p40phox, b-actin, phosphorylated ERK5 Thr218/Tyr220 (T218/Y220), total ERK5, phosphorylated p38, total p38, phosphorylated Akt Ser473, and total Akt (top panel). Statistical significance was evaluated with a paired Student t test (*P<0.05; **P<0.01; ***P<0.001) (bottom panel). (D) Washed WT, Ppara+/- and Ppara-/- platelets were incubated with N-acetylcysteine (NAC, 2 mM), etomoxir (Eto, 25 mM), dimethylsulfoxide (DMSO), VAS2870 (10 mM), BIX02188 (10 mM), SB203580 (10 mM), or SH-6 (10 mM) for 10 mins, then stimulated with KODiA-PC (15 mM). Aggregation was assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments. (E) H2DCFDAloaded (50 mM) mice platelets were incubated with DMSO, NAC (2 mM), VAS2870 (10 mM), or etomoxir (25 mM), stimulated with KODiA-PC (15 mM) for 10 min. PAPC was used as a negative control. Samples were analyzed immediately. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; ns: not significant).

gure 5C). As expected, platelet aggregation in response to thrombin was enhanced in patients with hypertriglyceridemia or hypercholesterolemia compared to that in healthy subjects (Figure 5E). The level of platelet PPARa expression was closely correlated with platelet aggregation in response to thrombin (Figure 5F). These findings demonstrate that the increased expression of platelet PPARa in patients with hyperlipidemia is closely related to platelet activity.

ked binding of p65 to region 1 of the Ppara promoter of megakaryocytes when treated with fatty acid, cholesterol or oxLDL (Figure 6E), indicating that NF-kB directly regulates Ppara transcription in Meg-01 cells. Thus, fatty acids, cholesterol or oxLDL upregulate PPARa expression in Meg01 cells through ROS and subsequent NF-kB signaling.

Oxidized low-density lipoproteins and lipids upregulate megakaryocyte- but not platelet- PPARa Compared with platelets treated with normal medium, platelets incubated with a fatty acid (oleic acid or palmitic acid), cholesterol or oxLDL for 12 h or 24 h did not alter the PPARa levels (Online Supplementary Figure S10A). In contrast, PPARa protein and mRNA levels in Meg-01 cells were significantly increased after 24 h incubation with the fatty acid, cholesterol or oxLDL (Figure 6A, B and Online Supplementary Figure S10B), without a concomitant change of the expression of PPARb and PPARg (Figure 6A and Online Supplementary Figure S10B). These data suggest that the increased PPARa in hyperlipidemic platelets may derive from megakaryocytes. It was reported that hyperlipidemia induces ROS generation and activation of the NF-kB signaling pathway.28 Indeed, treating Meg-01 cells with the NF-kB inhibitor BAY11-7082, antioxidants NAC or DTT, abolished the PPARa upregulation by fatty acids, cholesterol or oxLDL (Figure 6C, D and Online Supplementary Figure S10C). Moreover, BAY11-7082 and NAC or DTT inhibited fatty acid-, cholesterol- or oxLDL-induced phosphorylation of IκBa (Figure 6C, D and Online Supplementary Figure S10C). An in silico promoter analysis (Jaspar and ensemble Genome Brower) identified the possible NF-kB-binding sites on the Ppara promoter. Four potential NF-kB sites in the sense strand of the region -105/-114 bp (region 1), -168/-177 bp (region 2), -1588/-1597 bp (region 3) and -1878/-1887 bp (region 4). Chromatin immunoprecipitation analyses revealed a mar-

The present study investigated the role of PPARa in platelet activation and the impact of PPARa in the prothrombotic potential caused by hyperlipidemia. The results demonstrated that PPARa is an indispensable signaling molecule supporting platelet activation and thrombosis. Importantly, increased PPARa expression in platelets is responsible for enhanced platelet activity by hyperlipidemia. Hyperlipidemia does not trigger PPARa expression in platelets directly, but rather does so in megakaryocytes through ROS and NF-kB pathways. Our study not only elucidated the signaling function of PPARa in supporting platelet activation, but also revealed a key role for PPARa in bridging the genetic effect of hyperlipidemia on megakaryocytes with the prothrombotic potential operated by platelets. This study clearly demonstrated the positive role that PPARa serves in supporting platelet activation and thrombosis. However, previous studies showed some synthetic18,19 or endogenous molecules19 inhibited platelet activation and thrombosis in a PPARa-dependent manner. This discrepancy may suggest the existence of endogenous PPARa ligands which serve as a positive regulator of platelet function and thrombosis. Although such ligands of PPARa have yet to be defined, given the phenotype of platelets upon PPARa deficiency, it is possible that these stimulatory ligands are the ones playing dominant roles in platelet activation. While lipids with cardioprotective effects (e.g., polyunsaturated fatty acids)29 are able to produce endogenous PPARa ligand with platelet-inhibitory

Discussion

Haematologica | 107 - June 2022

1367

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Legend on following page. Haematologica | 107 - June 2022

1368

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Figure 5. Increased platelet PPARa expression correlates with platelet hyperreactivity in hyperlipidemic mice and patients with hyperlipidemia. (A) Immunoblot analysis of PPARa, PPARb and PPARg expression in platelets from Apoe-/- mice fed a high-fat diet (HFD) or control diet (CD) for 8 weeks with PPARa, PPARb and PPARg antibodies. Statistical significance was evaluated with a two-tailed Mann-Whitney test (**P<0.01; ns: not significant). (B) Aggregation and ATP release of platelets from Apoe-/- mice fed with a HFD or CD were stimulated with thrombin (0.015 U/mL) or collagen (0.6 mg/mL). Aggregation was assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Traces are representative of at least three independent experiments. (C) Immunoblot analysis of PPARa, PPARb and PPARg expression in platelets from patients with hypertriglyceridemia (HTG) and hypercholesterolemia (HTC) with PPARa, PPARb and PPARg antibodies. Representative immunoblots of platelet PPARa and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) from four healthy subjects, six patients with HTG and three patients with HTC. Representative immunoblots of platelet PPARb, PPARg and GAPDH from four healthy subjects, three patients with HTG and three patients with HTC. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; ns: no significance). (D) PPARA mRNA expression in platelets from healthy subjects (n=10) and patients with HTG (n=12) or healthy subjects (n=9) and patients with HTC (n=10) was analyzed by quantitative real-time polymerase chain reaction. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05). (E) Aggregation of platelets from healthy subjects (n=25) and patients with HTG (n=36) or healthy subjects (n=16) and patients with HTC (n=16) in response to thrombin (0.025 U/mL). Aggregation was assessed with a Chrono-log lumiaggregometer under stirring at 1,200 rpm. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05). (F) Platelet aggregation induced by thrombin is well correlated with protein level of platelet PPARa expression in healthy subjects (n=25) and patients with HTG (n=36) or healthy subjects (n=16) and patients with HTC (n=16). Each solid circle represents a different individual (Pearson correlation, GraphPad Prism 5).

properties (e.g., DPAn-6),19 lipid species with cardiovascular disease-promoting properties, such as saturated fatty acids,29 may generate derivatives which serve as stimulatory PPARa ligands to promote platelet activation. Identification of the stimulatory ligands may thus provide novel targets for the intervention of thrombosis and constitutes an important theme in its own right. Moreover, based on the fact that ligands mainly target the ligand-binding domain of PPARa,30,31 it is tempting to hypothesize that a conformational change induced by occupancy of the region of the ligand-binding domain may be key to the nongenomic function of PPARa. Future structural studies are therefore imperative to provide further insight into this promising anti-thrombotic target. This study not only confirmed that the key contribution of PPARa to thrombosis and hemostasis is through the regulation of platelet dense granule secretion, but also revealed the pathway on which PPARa relies to perform its function. Hence, p38 and Akt were identified as the sequential signals downstream of PPARa to regulate platelet dense granule secretion. These findings are in agreement with the previously reported roles of p38 and Akt in platelets. For example, p38 has been shown to positively regulate dense granule secretion,25 and Akt have also been reported to be important in dense granule secretion through promoting nitric oxide/cGMP signaling32 or inhibiting GSK3b.33 Moreover, our results suggest that p38 activation is relayed by Akt phosphorylation, which is also consistent with a previous report.34 Given the intact Jon/A binding and the significantly reduced spreading on immobilized fibrinogen upon PPARa deficiency, it seems that PPARa mainly participates in outside-in signaling rather than inside-out signaling. This observation is also consistent with previous reports, which suggested that outside-in signaling and dense granule secretion are coupled and both are regulated by p38 and Akt signaling, evidenced by the studies employing either inhibition or genetic

ablation of p38 or Akt.25,32 Downstream of p38, ROS generation has been found to be a pivotal link between PPARa and platelet activation, echoing a previous finding in macrophages.35 It seems that hemostatic stimuli36-38 and oxLDL9 may differentially employ the ROS generation pathways. Notably, our data indicate that PPARa critically contributes to the generation of ROS controlled by both mitochondrial fatty acid b-oxidation and NADPH oxidases. Therefore, the PPARa/p38/ROS/Akt axis may function as a central gatekeeper for platelet activation and is employed by major platelet receptors, such as immunoreceptor tyrosine-based activation motif receptor, G-protein-coupled receptors, and possibly CD36. A key finding of the present study is the correlation between the level of expression of PPARa and the extent of platelet reactivity. The subsequent data from the hyperlipidemic Ppara-/-/Apoe-/- mice model further revealed a causative relationship between PPARa and platelet hyperactivity under the condition of hyperlipidemia. The correlation is found in both humans and mice, which indicates that it might be an evolutionary conservative mechanism. Unlike in macrophages, where oxLDL elevates the expression of both PPARa21 and PPARg,39 megakaryocytes seems to respond to hyperlipidemia specifically with upregulated expression of PPARa, but not of PPARb or PPARg. Considering the importance of fatty acids in supporting both megakaryocyte maturation40 and platelet production,41 and the central roles of PPARa in fatty acid metabolism,17,42 it may not be surprising that megakaryocyte PPARa is a sensitive and specific target regulated by hyperlipidemia. Although it is commonly accepted that nongenomic mechanisms direct PPAR family-mediated platelet function,13,18,43,44 our findings suggest that in the case of PPARa, a genomic regulation in megakaryocytes and a nongenomic signaling role in platelets are concerted to underscore the platelet hyperreactivity under hyperlipidemia.

Haematologica | 107 - June 2022

1369

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Legend on following page. Haematologica | 107 - June 2022

1370

ARTICLE - PPARa promotes platelet hyperreactivity.

L. Li et al.

Figure 6. Oxidized low-density lipoproteins and lipids upregulate megakaryocyte but not platelet PPARa. (A) Immunoblot analysis of PPARa, PPARb and PPARg expression in Meg-01 cells cultured with fatty acids (oleic acid [OA], 400 mM and palmitic acid [PA], 200 mM), cholesterol (CHO, 2.5 mg/mL, 5.0 mg/mL, 7.5 mg/mL) or oxidized low-density lipoproteins (oxLDL, 10 mg/mL, 50 mg/mL) for 24 h with PPARa, PPARb and PPARg antibodies. Representative immunoblots from at least three independent experiments. (B) PPARA mRNA expression in Meg01 cells cultured with fatty acids (OA, 400 mM and PA, 200 mM), cholesterol (CHO, 2.5 mg/mL, 5.0 mg/mL, 7.5 mg/mL) or oxLDL (10 mg/mL, 50 mg/mL) for 24 h was analyzed by quantitative real-time polymerase chain reaction. Statistical significance was evaluated with a two-tailed Mann-Whitney test (*P<0.05; ns: not significant). (C) Immunoblot analysis of PPARα and phosphorylated IκBα level in Meg-01 cells cultured with fatty acids (OA, 400 mM and PA, 200 mM), cholesterol (CHO, 2.5 mg/mL, 5.0 mg/mL, 7.5 mg/mL) or oxLDL (10 mg/mL, 50 mg/mL) in the absence or presence of BAY11-7082 (10 mM) for 24 h. Representative immunoblots from at least three independent experiments. (D) Immunoblot analysis of PPARa and phosphorylated IκBa level in Meg-01 cells cultured with fatty acids (OA, 400 mM and PA, 200 mM), CHO (2.5 mg/mL, 5.0 mg/mL, 7.5 mg/mL) or oxLDL (10 mg/mL, 50 mg/mL) in the absence or presence of N-acetylcysteine (NAC, 1 mM) or DTT (1 mM) for 24 h. Representative immunoblots from at least three independent experiments. (E) NF-kB binding to the Ppara promoter of Meg-01 cells as determined by chromatin immunoprecipitation. Schematic diagram showing the NF-kB-binding site in the Ppara promoter (top panel). Amplification of the Ppara promoter region containing the NF-kB-binding motif in Meg-01 cells cultured with fatty acids (OA, 400 mM and PA, 200 mM), CHO (2.5 mg/mL, 5.0 mg/mL, 7.5 mg/mL) or oxLDL (10 mg/mL, 50 mg/mL) for 24 h. GAPDH was used as a control to show precipitation specificity (bottom panel). Results shown are representative of three or more separate experiments run on different days.

In the current study, NF-kB, activated by ROS, was revealed as the possible mechanism of the upregulation of PPARa in megakaryocytes by hyperlipidemia. Our data suggest that NF-kB upregulates PPARa expression possibly through direct binding to the Ppara promoter region in lipid-treated megakaryocytes. It is worth mentioning that hyperglycemia elicits a similar NF-κB activation which subsequently upregulates the expression of P2Y12, a key receptor in mediating platelet activation and thrombosis.45 It is therefore possible that metabolic disorders activate common inflammatory pathways to promote thrombosis. Interestingly, it was reported that PPARa acts as a feedback to negatively regulate NF-kB activation in smooth muscle cells46 and the metabolism of the arachidonic acid metabolite leukotriene B4 in hepatocytes.47 It is yet to be determined whether PPARa acts similarly in megakaryocytes to negatively regulate inflammation. However, while upregulation of PPARa may suppress inflammation, it may exert a prothrombotic effect through enhanced platelet activation. PPARa may thus be a key molecule maintaining the balance of thrombosis and inflammation, as suggested by the thrombo-inflammatory model of deep vein thrombosis in the current study. The study has several limitations. First, expression of PPARa in vascular cells has been reported;48-50 in the absence of a cell-specific knockout model, a possible contribution to hemostasis from vasculature PPARa could not be excluded. Second, functional validation of the signaling involved was performed with inhibitors, which inevitably have possible off-target effects. Although it is beyond the scope of the study to chase every off-target effect, care must be taken when interpreting the results. To summarize, we found that platelet PPARa positively

mediates platelet activation through promoting dense granule secretion. Hyperlipidemia may contribute to the prothrombotic status through upregulated expression of PPARa in megakaryocytes/platelets and enhanced activation signaling mediated by PPARa in platelets. This work suggested that coupled genomic and nongenomic interventions targeting PPARa may be necessary for the prevention of thrombosis under hyperlipidemia. Disclosures No conflicts of interest to disclose. Contributions LL, JZ and SW performed experiments, analyzed data, and wrote and revised the manuscript. LJ, XC and YZ performed experiments, analyzed data, and revised the manuscript. JL, JS, PL, ZS and FJG contributed intellectually and revised the manuscript. AL and HH designed the study and revised the manuscript. Acknowledgments The authors would like to thank the core facilities and the Center of Cryo-Electron Microscopy of Zhejiang University School of Medicine for technical support. The authors would also like to thank Ms. Xueping Zhou for excellent assistance in animal facilities. Funding This study was supported by grants from the National Natural Science Foundation of China (81870106, 82070138 to HH and 82000139 to PL) and the Zhejiang Provincial Natural Science Foundation (LZ18H080001 to HH and LQ21H030003 to ZS).

References 1. Skulas-Ray AC, Wilson PWF, Harris WS, et al. Omega-3 fatty acids for the management of hypertriglyceridemia: a science

advisory from the American Heart Association. Circulation. 2019;140(12):e673-e691.

Haematologica | 107 - June 2022

1371

ARTICLE - PPARa promotes platelet hyperreactivity. 2. Korporaal SJ, Meurs I, Hauer AD, et al. Deletion of the highdensity lipoprotein receptor scavenger receptor BI in mice modulates thrombosis susceptibility and indirectly affects platelet function by elevation of plasma free cholesterol. Arterioscler Thromb Vasc Biol. 2011;31(1):34-42. 3. Maschberger P, Bauer M, Baumann-Siemons J, et al. Mildly oxidized low density lipoprotein rapidly stimulates via activation of the lysophosphatidic acid receptor Src family and Syk tyrosine kinases and Ca2+ influx in human platelets. J Biol Chem. 2000;275(25):19159-19166. 4. Cipollone F, Mezzetti A, Porreca E, et al. Association between enhanced soluble CD40L and prothrombotic state in hypercholesterolemia: effects of statin therapy. Circulation. 2002;106(4):399-402. 5. Korporaal SJ, Gorter G, van Rijn HJ, Akkerman JW. Effect of oxidation on the platelet-activating properties of low-density lipoprotein. Arterioscler Thromb Vasc Biol. 2005;25(4):867-872. 6. Wraith KS, Magwenzi S, Aburima A, Wen Y, Leake D, Naseem KM. Oxidized low-density lipoproteins induce rapid platelet activation and shape change through tyrosine kinase and Rho kinase-signaling pathways. Blood. 2013;122(4):580-589. 7. Chen K, Febbraio M, Li W, Silverstein RL. A specific CD36dependent signaling pathway is required for platelet activation by oxidized low-density lipoprotein. Circ Res. 2008;102(12):15121519. 8. Yang M, Cooley BC, Li W, et al. Platelet CD36 promotes thrombosis by activating redox sensor ERK5 in hyperlipidemic conditions. Blood. 2017;129(21):2917-2927. 9. Magwenzi S, Woodward C, Wraith KS, et al. Oxidized LDL activates blood platelets through CD36/NOX2-mediated inhibition of the cGMP/protein kinase G signaling cascade. Blood. 2015;125(17):2693-2703. 10. Moraes LA, Unsworth AJ, Vaiyapuri S, et al. Farnesoid X receptor and its ligands inhibit the function of platelets. Arterioscler Thromb Vasc Biol. 2016;36(12):2324-2333. 11. Unsworth AJ, Bye AP, Tannetta DS, et al. Farnesoid X receptor and liver X receptor ligands initiate formation of coated platelets. Arterioscler Thromb Vasc Biol. 2017;37(8):1482-1493. 12. Hashizume S, Akaike M, Azuma H, et al. Activation of peroxisome proliferator-activated receptor alpha in megakaryocytes reduces platelet-derived growth factor-BB in platelets. J Atheroscler Thromb. 2011;18(2):138-147. 13. Akbiyik F, Ray DM, Gettings KF, Blumberg N, Francis CW, Phipps RP. Human bone marrow megakaryocytes and platelets express PPARgamma, and PPARgamma agonists blunt platelet release of CD40 ligand and thromboxanes. Blood. 2004;104(5):1361-1368. 14. Ali FY, Davidson SJ, Moraes LA, et al. Role of nuclear receptor signaling in platelets: antithrombotic effects of PPARbeta. FASEB J. 2006;20(2):326-328. 15. Gilde AJ, van der Lee KA, Willemsen PH, et al. Peroxisome proliferator-activated receptor (PPAR) alpha and PPARbeta/delta, but not PPARgamma, modulate the expression of genes involved in cardiac lipid metabolism. Circ Res. 2003;92(5):518-524. 16. Lalloyer F, Wouters K, Baron M, et al. Peroxisome proliferatoractivated receptor-alpha gene level differently affects lipid metabolism and inflammation in apolipoprotein E2 knock-in mice. Arterioscler Thromb Vasc Biol. 2011;31(7):1573-1579. 17. Badman MK, Pissios P, Kennedy AR, Koukos G, Flier JS, MaratosFlier E. Hepatic fibroblast growth factor 21 is regulated by PPARalpha and is a key mediator of hepatic lipid metabolism in ketotic states. Cell Metab. 2007;5(6):426-437. 18. Ali FY, Armstrong PC, Dhanji AR, et al. Antiplatelet actions of statins and fibrates are mediated by PPARs. Arterioscler Thromb

L. Li et al. Vasc Biol. 2009;29(5):706-711. 19. Yeung J, Adili R, Yamaguchi A, et al. Omega-6 DPA and its 12lipoxygenase-oxidized lipids regulate platelet reactivity in a nongenomic PPARalpha-dependent manner. Blood Adv. 2020;4(18):4522-4537. 20. Drosatos K, Pollak NM, Pol CJ, et al. Cardiac myocyte KLF5 regulates Ppara expression and cardiac function. Circ Res. 2016;118(2):241-253. 21. Yu X, Li X, Zhao G, et al. OxLDL up-regulates Niemann-Pick type C1 expression through ERK1/2/COX-2/PPARalpha-signaling pathway in macrophages. Acta Biochim Biophys Sin (Shanghai). 2012;44(2):119-128. 22. Sartippour MR, Renier G. Differential regulation of macrophage peroxisome proliferator-activated receptor expression by glucose : role of peroxisome proliferator-activated receptors in lipoprotein lipase gene expression. Arterioscler Thromb Vasc Biol. 2000;20(1):104-110. 23. Lee SS, Pineau T, Drago J, et al. Targeted disruption of the alpha isoform of the peroxisome proliferator-activated receptor gene in mice results in abolishment of the pleiotropic effects of peroxisome proliferators. Mol Cell Biol. 1995;15(6):3012-3022. 24. Jiang L, Xu C, Yu S, et al. A critical role of thrombin/PAR-1 in ADP-induced platelet secretion and the second wave of aggregation. J Thromb Haemost. 2013;11(5):930-940. 25. Shi P, Zhang L, Zhang M, et al. Platelet-specific p38alpha deficiency improved cardiac function after myocardial infarction in mice. Arterioscler Thromb Vasc Biol. 2017;37(12):e185-e196. 26. Arthur JF, Qiao J, Shen Y, et al. ITAM receptor-mediated generation of reactive oxygen species in human platelets occurs via Syk-dependent and Syk-independent pathways. J Thromb Haemost. 2012;10(6):1133-1141. 27. Lu Y, Harada M, Kamijo Y, et al. Peroxisome proliferator-activated receptor alpha attenuates high-cholesterol diet-induced toxicity and pro-thrombotic effects in mice. Arch Toxicol. 2019;93(1):149161. 28. Cominacini L, Pasini AF, Garbin U, et al. Oxidized low density lipoprotein (ox-LDL) binding to ox-LDL receptor-1 in endothelial cells induces the activation of NF-kappaB through an increased production of intracellular reactive oxygen species. J Biol Chem. 2000;275(17):12633-12638. 29. Zhuang P, Zhang Y, He W, et al. Dietary fats in relation to total and cause-specific mortality in a prospective cohort of 521 120 individuals with 16 years of follow-up. Circ Res. 2019;124(5):757768. 30. Roy A, Jana M, Kundu M, et al. HMG-CoA reductase inhibitors bind to PPARalpha to upregulate neurotrophin expression in the brain and improve memory in mice. Cell Metab. 2015;22(2):253265. 31. Xu HE, Stanley TB, Montana VG, et al. Structural basis for antagonist-mediated recruitment of nuclear co-repressors by PPARalpha. Nature. 2002;415(6873):813-817. 32. Stojanovic A, Marjanovic JA, Brovkovych VM, et al. A phosphoinositide 3-kinase-AKT-nitric oxide-cGMP signaling pathway in stimulating platelet secretion and aggregation. J Biol Chem. 2006;281(24):16333-16339. 33. O'Brien KA, Stojanovic-Terpo A, Hay N, Du X. An important role for Akt3 in platelet activation and thrombosis. Blood. 2011;118(15):4215-4223. 34. Kamiyama M, Shirai T, Tamura S, et al. ASK1 facilitates tumor metastasis through phosphorylation of an ADP receptor P2Y12 in platelets. Cell Death Differ. 2017;24(12):2066-2076. 35. Teissier E, Nohara A, Chinetti G, et al. Peroxisome proliferatoractivated receptor alpha induces NADPH oxidase activity in macrophages, leading to the generation of LDL with PPAR-alpha

Haematologica | 107 - June 2022

1372

ARTICLE - PPARa promotes platelet hyperreactivity. activation properties. Circ Res. 2004;95(12):1174-1182. 36. Choo HJ, Saafir TB, Mkumba L, Wagner MB, Jobe SM. Mitochondrial calcium and reactive oxygen species regulate agonist-initiated platelet phosphatidylserine exposure. Arterioscler Thromb Vasc Biol. 2012;32(12):2946-2955. 37. Bakdash N, Williams MS. Spatially distinct production of reactive oxygen species regulates platelet activation. Free Radic Biol Med. 2008;45(2):158-166. 38. Liu Y, Hu M, Luo D, et al. Class III PI3K positively regulates platelet activation and thrombosis via PI(3)P-directed function of NADPH oxidase. Arterioscler Thromb Vasc Biol. 2017;37(11):2075-2086. 39. Ricote M, Huang J, Fajas L, et al. Expression of the peroxisome proliferator-activated receptor gamma (PPARgamma) in human atherosclerosis and regulation in macrophages by colony stimulating factors and oxidized low density lipoprotein. Proc Natl Acad Sci U S A. 1998;95(13):7614-7619. 40. Valet C, Batut A, Vauclard A, et al. Adipocyte fatty acid transfer supports megakaryocyte maturation. Cell Rep. 2020;32(1):107875. 41. Kelly KL, Reagan WJ, Sonnenberg GE, et al. De novo lipogenesis is essential for platelet production in humans. Nat Metab. 2020;2(10):1163-1178. 42. Ide T, Shimano H, Yoshikawa T, et al. Cross-talk between peroxisome proliferator-activated receptor (PPAR) alpha and liver X receptor (LXR) in nutritional regulation of fatty acid metabolism. II. LXRs suppress lipid degradation gene promoters through inhibition of PPAR signaling. Mol Endocrinol. 2003;17(7):1255-1267. 43. Moraes LA, Spyridon M, Kaiser WJ, et al. Non-genomic effects of

L. Li et al. PPARgamma ligands: inhibition of GPVI-stimulated platelet activation. J Thromb Haemost. 2010;8(3):577-587. 44. Ali FY, Hall MG, Desvergne B, Warner TD, Mitchell JA. PPARbeta/delta agonists modulate platelet function via a mechanism involving PPAR receptors and specific association/repression of PKCalpha--brief report. Arterioscler Thromb Vasc Biol. 2009;29(11):1871-1873. 45. Hu L, Chang L, Zhang Y, et al. Platelets express activated P2Y12 receptor in patients with diabetes mellitus. Circulation. 2017;136(9):817-833. 46. Delerive P, De Bosscher K, Besnard S, et al. Peroxisome proliferator-activated receptor alpha negatively regulates the vascular inflammatory gene response by negative cross-talk with transcription factors NF-kappaB and AP-1. J Biol Chem. 1999;274(45):32048-32054. 47. Devchand PR, Keller H, Peters JM, Vazquez M, Gonzalez FJ, Wahli W. The PPARalpha-leukotriene B4 pathway to inflammation control. Nature. 1996;384(6604):39-43. 48. Staels B, Koenig W, Habib A, et al. Activation of human aortic smooth-muscle cells is inhibited by PPARalpha but not by PPARgamma activators. Nature. 1998;393(6687):790-793. 49. Inoue I, Shino K, Noji S, Awata T, Katayama S. Expression of peroxisome proliferator-activated receptor alpha (PPAR alpha) in primary cultures of human vascular endothelial cells. Biochem Biophys Res Commun. 1998;246(2):370-374. 50. Chinetti G, Gbaguidi FG, Griglio S, et al. CLA-1/SR-BI is expressed in atherosclerotic lesion macrophages and regulated by activators of peroxisome proliferator-activated receptors. Circulation. 2000;101(20):2411-2417.

Haematologica | 107 - June 2022

1373

ARTICLE - Hemostasis

Proline-rich tyrosine kinase Pyk2 regulates deep vein thrombosis Stefania Momi,1 Jessica Canino,2 Mauro Vismara,2 Luca Galgano,2,3 Emanuela Falcinelli,1 Giuseppe Guglielmini,1 Giulia Ciarrocca Taranta,1 Gianni Francesco Guidetti,2 Paolo Gresele,1 Mauro Torti2 and Ilaria Canobbio2 Department of Medicine and Surgery, Division of Internal and Cardiovascular Medicine, University of Perugia, Perugia; 2Department of Biology and Biotechnology, University of Pavia, Pavia and 3Scuola Universitaria Superiore, IUSS, Pavia, Italy 1

Correspondence: Mauro Torti mtorti@unipv.it Received: July 26, 2021. Accepted: February 1, 2022. Prepublished: February 10, 2022. https://doi.org/10.3324/haematol.2021.279703 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract Deep vein thrombosis results from the cooperative action of leukocytes, platelets, and endothelial cells. The proline-rich tyrosine kinase Pyk2 regulates platelet activation and supports arterial thrombosis. In this study, we combined pharmacological and genetic approaches to unravel the role of Pyk2 in venous thrombosis. We found that mice lacking Pyk2 almost completely failed to develop deep venous thrombi upon partial ligation of the inferior vena cava. Pyk2-deficient platelets displayed impaired exposure of phosphatidylserine and tissue factor expression by endothelial cells and monocytes was completely prevented by inhibition of Pyk2. In human umbilical vein endothelial cells (HUVEC), inhibition of Pyk2 hampered IL-1b-induced expression of VCAM and P-selectin, and von Willebrand factor release. Pyk2-deficient platelets showed defective adhesion on von Willebrand factor and reduced ability to bind activated HUVEC under flow. Moreover, inhibition of Pyk2 in HUVEC strongly reduced platelet adhesion. Similarly, Pyk2-deficient neutrophils were unable to efficiently roll and adhere to immobilized endothelial cells under venous flow conditions. Moreover, platelets and neutrophils from Pyk2knockout mice showed defective ability to form heterogeneous aggregates upon stimulation, while platelet monocyte interaction occurred normally. Consequently, platelet neutrophil aggregates, abundant in blood of wild-type mice upon inferior vena cava ligation, were virtually undetectable in Pyk2-knockout mice. Finally, we found that expression of Pyk2 was required for NETosis induced by activated platelets. Altogether our results demonstrate a critical role of Pyk2 in the regulation of the coordinated thromboinflammatory responses of endothelial cells, leukocytes and platelets leading to venous thrombosis. Pyk2 may represent a novel promising target in the treatment of deep vein thrombosis.

Introduction Deep vein thrombosis (DVT) is a consequence of impaired hemostasis and increased inflammation.1 Thrombi formed in the lower limbs frequently embolize and reach the lungs causing pulmonary thromboembolism.2 DVT results from the cooperative action of leukocytes and platelets that interact with activated endothelial cells (EC).3 The vascular endothelium can be activated by multiple insults, such as hypercholesterolemia, high blood pressure, and reactive oxygen species (ROS).4 Stasis of venous blood is a crucial determinant for DVT and facilitates the development of a pro-inflammatory phenotype in EC characterized by the expression of cell adhesion molecules (CAM) and P-selectin and associated to the release of von Willebrand factor (VWF).5,6 EC recruit blood borne leuko-

cytes. Tissue factor (TF) expression by monocytes initiates the extrinsic pathway of the coagulation cascade,7 whereas thrombus-resident neutrophils propagate DVT through the activation of coagulation factor XII. Neutrophils also release neutrophil extracellular traps (NET) and engage circulating platelets8 which, in turn, recruit additional neutrophils forming platelet neutrophil aggregates in a vicious circle reinforcing thrombosis and inflammation.3 The activation of different vascular cells involved in the propagation of DVT is regulated by signaling molecules and key enzymes. Among them, protein tyrosine kinases play a pivotal role in cellular activation and are often directly linked to membrane receptors.9 In particular, Src family kinases mediate neutrophil adhesion to adherent platelets and modulate platelet neutrophil aggregate

Haematologica | 107 - June 2022

1374

ARTICLE - The multifaced role of Pyk2 in venous thrombosis formation.10,11 Additional tyrosine kinases act as second line signal transducers and, among them, focal adhesion kinases (FAK) have attracted increasing attention over the past few years. Proline-rich tyrosine kinase Pyk2 is a cytosolic non-receptor tyrosine kinase of 1,009 aminoacids that belongs to the FAK family.12 Pyk2 is highly expressed in hematopoietic cells.13 In macrophages, Pyk2 mediates adhesion, migration, and complement-mediated phagocytosis14,15 and in neutrophils it controls adhesion, migration, degranulation, ROS production and host responses to bacterial infection.16,17 These observations highlight a critical role of Pyk2 in innate immunity and inflammation. We have previously shown that Pyk2 is also an important regulator of platelet activation.18 Pyk2 gets phosphorylated and thus activated in platelets stimulated by thrombin, collagen, ADP and VWF19-21 and it is also recruited in platelets adherent to collagen and fibrinogen.22,23 Pyk2 is required for TxA2 formation, integrin aIIbb3 activation and platelet aggregation.18 Mice lacking Pyk2 show delayed formation of occlusive thrombi in a model of photochemical induced arterial thrombosis and increased bleeding time, showing that Pyk2 is a critical regulator of arterial thrombosis and primary hemostasis. Interestingly, Pyk2knockout (Pyk2-KO) mice are also protected against platelet pulmonary embolism.18 The role of Pyk2 in the development of DVT however remains unexplored. In this study we show that Pyk2 is an important regulator of DVT and orchestrates the interplay between platelets, neutrophils, and EC. Our results disclose a novel potential therapeutic target for the pharmacological prevention of venous thrombosis by targeting Pyk2.

Methods Materials Pyk2-KO mice generation was described previously.14 All the procedures involving the use of mice were approved by the Ethics Committee (University of Pavia), by the Committee on Ethics of Animal Experiments (University of Perugia), and by the Italian Ministry of Public Health (authorization numbers 561/2015 PR; 8/2018 PR; 259/2018 PR). HUVEC were from Promocell, and THP-1 cells were a gift from Prof. Lanni, University of Pavia. PE-conjugated anti-VCAM-1 (CD106) (PE), FITC-conjugated anti-P-selectin and anti-mouse immunoglobulin G (IgG), as well as FITClabeled Annexin V kit were from Beckam Coulter. Anti-tissue factor antibody (CD142), anti-Ly6G and anti-CD115 were from eBioscience. FITC-labeled anti-CD41/CD61 (Leo.D2) was from Emfret Analytics. Peridin chlorophyl protein (PerCP)-conjugated anti-mouse CD45 (clone 30F11) was from Beckton Dickinson. Pyk2 inhibitor PF-

S. Momi et al.

4594755 was from Pfizer. Asserachrom VWF:Ag Kit was from Diagnostica Stag; IL-1b from PeproTech: CellTracker™ Red CMTPX Dye from Thermo Fisher Scientific. VWF was from Calbiochem. Experimental procedures Experimental procedures were performed as previously described.24-31 Additional details are included in the Online Supplementary Appendix.

Results Pyk2 promotes venous thrombosis We have previously demonstrated that Pyk2 deficiency protects mice from platelet pulmonary embolism triggered by the injection of collagen plus epinephrine.18 In order to investigate the role of Pyk2 in the development of DVT in mice, we applied a model of partial inferior vena cava (IVC) ligation, in which thrombosis is driven by a flow disturbance without physical disruption of the endothelial lining.5,32 Mice were sacrificed 24- and 48-hours post IVC ligation and formed thrombi were extracted and analyzed. All the WT mice developed an evident vena cava thrombus both at 24 and 48 hours upon IVC ligation (Figure 1A). By contrast, thrombus formation was observed only in four of 15 Pyk2-KO mice at 24 hours, and in six of 16 Pyk2-KO mice 48 hours after IVC ligation (Figure 1A). When present, thrombi formed in Pyk2-KO mice were very small compared to those formed in WT mice or were, in some animals, nearly undetectable (Figure 1B). Thrombus weight and length were strikingly and significantly reduced in Pyk2-KO compared to WT mice (Figure 1C and D). At 24 hours, the mean weight of thrombi in WT mice was 6.92±4.89 mg, while that of thrombi observed in Pyk2-KO was 1.09±2.4 mg. The thrombi length was about 3.06±1.97 mm and 0.63±1.14 mm in WT and Pyk2-KO mice, respectively. At 48 hours, the average weight of the thrombi further increased in WT mice to 10.45±5 mg and the size to 5.69±2.2 mm but remained unchanged in Pyk2-deficient mice (0.63± 1.1 mg and 0.67±1.0 mm). These data clearly show that the absence of Pyk2 protects mice from venous thrombosis upon IVC ligation. In order to discriminate the contribution of Pyk2 expressed in platelets versus endothelial cells or leukocytes we performed cross-transfusion experiments in which platelets from Pyk2-KO mice were reinfused into recipient thrombocytopenic WT mice and vice versa. Figure 1E shows that WT mice receiving Pyk2-deficient platelets were still able to form thrombi upon partial IVC ligation. However, both the length and the weight of these thrombi were slightly reduced when compared to those formed in WT mice (P<0.05). By contrast, transfusion of WT platelets into Pyk2-KO mice did not restore at all the ability to form ve-

Haematologica | 107 - June 2022

1375

ARTICLE - The multifaced role of Pyk2 in venous thrombosis

S. Momi et al.

Figure 1. IVC-ligation-induced deep vein thrombosis in wild-type and Pyk2-knockout mice. Thrombus formation in the inferior vena cava (IVC) was induced by partial vein ligation. Thrombi were extracted 24 or 48 hours (h) after surgery. (A) Incidence of thrombus formation in wild-type (WT) vs. Pyk2-knockout (Pyk2-KO) mice is reported as percentage of the analyzed animals (black bars: mice developing thrombus; white bars: mice with no detectable thrombus). (B) Comparative representative images showing the different size of thrombi from WT and Pyk2-KO mice after 24 hours of IVC ligation. (C and D) Thrombus weight and length at 24 and 48 hours, respectively. **P≤0.01; ***P≤0.005; ****P≤0.001 (17 WT and 31 Pyk2-KO mice were analyzed). (E) Thrombus formation in cross-transfused mice. Pyk2-deficient platelets were re-infused in recipient thrombocytopenic WT mice and WT platelets were re-infused in recipient Pyk2-KO mice (4 animals/group), as indicated. Thrombi were isolated and measured 48 hours upon IVC ligation.

nous thrombi, as thrombus formation was never observed in all the analysed animals. These results indicate that Pyk2 in endothelial cells and/or leukocytes is essential for initiation of DVT and that Pyk2 in platelets has a secondary but evident role in thrombus growth. Pyk2 is required for endothelial cell activation DVT results from a complex interplay between enzymatic and cellular processes in which EC, platelets and leukocytes orchestrate a proinflammatory condition that culminates in clot formation. FAK have been implicated in vascular inflammation induced by TNFa and IL-1b in human aortic EC.33 We investigated the involvement of Pyk2 in EC activation using HUVEC and a pharmacological Pyk2 inhibitor, PF-4594755, whose selectivity and potency have been previously characterized.34 HUVEC were pre-incubated with PF-4594755 and then stimulated with IL-1b

for 4 hours. Surface expression of the adhesive molecules P-selectin and VCAM was significantly increased by IL-1b in stimulated HUVEC, and this response was completely suppressed by the Pyk2 inhibitor PF-4594755 (Figure 2A). We also analyzed the contribution of Pyk2 to the secretion of VWF, which plays a critical role in hemostasis and inflammation, as well as in DVT.35 We found that the constitutive release of VWF was unchanged upon pre-incubation with PF-4594755. However, VWF secretion stimulated by IL-1b was significantly affected by inhibition of Pyk2 activity (Figure 2B). Clot formation is initiated by TF that can be expressed by stimulated EC as well as by monocytes. Figure 2C(i) shows that stimulation of HUVEC with IL-1b induced a marked expression of TF, which was completely prevented upon Pyk2 inhibition by PF-4594755. By measuring the increase of TF mRNA in IL-1b-stimulated HUVEC treated with PF-

Haematologica | 107 - June 2022

1376

ARTICLE - The multifaced role of Pyk2 in venous thrombosis 4594755, we obtained clear evidence that Pyk2 regulates TF expression mainly at the transcriptional level (Figure 2Cii). Similarly, we found that also in monocytes TF exposure stimulated by treatment with LPS was completely prevented upon Pyk2 inhibition (Figure 2D). Altogether, these results indicate that TF exposure by pro-inflammatory stimulation of EC and monocytes is strictly regulated by Pyk2. Pyk2 regulates neutrophil rolling over endothelial cells In order to unravel the contribution of Pyk2 expressed by neutrophils in the early steps of DVT, we compared the ability of neutrophils isolated from the bone marrow of WT and Pyk2-KO mice to adhere over immobilized EC under flow. Isolated WT and Pyk2-deficient neutrophils were perfused over a layer of HUVEC stimulated with IL-

S. Momi et al.

1b at a shear stress of 250 sec-1, which mimics venous blood flow. Neutrophils from WT mice adhered to stimulated HUVEC in a time-dependent manner and maximal, firm adhesion was observed after 300 seconds. Conversely, Pyk2-deficient neutrophils almost completely failed to stop on endothelial cells and rolled away in few seconds (Figure 3A and B). Consequently, the number of firmly adherent neutrophils remained very low over time. In a separated set of experiments, we analyzed whether inhibition of Pyk2 in HUVEC could affect neutrophil rolling. Figure 3C shows that WT neutrophils almost completely failed to roll and adhere over IL-1b-stimulated HUVEC in which Pyk2 is inhibited by PF-4594755. Thus, Pyk2 activity in neutrophil as well as in EC is necessary for a correct recruitment of neutrophils on the activated endothelium.

Figure 2. Pyk2 supports a pro-inflammatory and procoagulant phenotype in endothelial cells. Human umbilical vein endothelial cells (HUVEC) were pre-incubated with dimethyl sulfoxide (DMSO) as vehicle (black bars) or with 10 µM Pyk2 inhibitor PF-4594755 (white bars) for 30 minutes and then stimulated with IL-1b for 4 hours (IL-1b) or left untreated (none). (A) Surface expression of P-selectin and VCAM-1, as assessed by flow cytometry, expressed as % of positive cells. (B) Measurement of von Willebrand factor (VWF) release in the supernatant of HUVEC. (C) Surface expression of tissue factor (TF), analyzed by flow cytometry, is reported in panel (i), while reverse transcription polymerase chain reaction (RT-PCR) analysis of TF mRNA expression in vehicle (none, black bars) or PF-4594755 (white bars)-treated HUVEC stimulated with IL-1b is reported in panel (ii), as fold increase over unstimulated cells. (D) THP-1 cells were pre-incubated with DMSO as vehicle (black bars) or with 10 µM Pyk2 inhibitor PF-4594755 (white bars) for 30 minutes and stimulated with 10 µg/ml lipopolysaccharide (LPS) for 3 hours (LPS) or left untreated (none) and surface expression of TF was analyzed by flow cytometry. Results are the mean ± standard error of the mean of 3 different experiments. *P≤0.05; **P≤0.01; ***P≤0.005; ****P≤0.001. Haematologica | 107 - June 2022

1377

ARTICLE - The multifaced role of Pyk2 in venous thrombosis

S. Momi et al. Figure 3. Pyk2 mediates neutrophil rolling over human umbilical vein endothelial cells. Neutrophils were isolated from the bone marrow of wild-type (WT) and Pyk2-knockout (Pyk2-KO) mice and perfused over IL-1b-activated human umbilical vein endothelial cells (HUVEC) at 250 sec-1, a shear stress that mimics venous blood flow. Number of adherent neutrophils was acquired at a rate of 1 frame/second. (A) Representative phase contrast microscopy images of rolling neutrophils (bright spots), obtained by overlapping of all the acquired frames are shown in panels (i) for WT neutrophils, and (ii) for Pyk2-deficient neutrophils, respectively. Arrows indicate the trajectory of rolling neutrophils over time. (B) The number of adherent neutrophils/field acquired at the different time points over a period of 10 minutes is reported in panel (i), while the percentage of surface coverage is shown in panel (ii). ****P≤0.001. (C) Rolling and adhesion of WT neutrophils perfused at 250 sec-1 over IL-1b-activated HUVEC pretreated with dimethyl sulfoxide (DMSO) or with this is 10 µM Pyk2 inhibitor PF-4594755 for 30 minutes (panel i) and the percentage of surface coverage (panel ii). The reported data are representative of 3 different experiments. ***P≤0.005.

Pyk2 modulates the interactions between platelets and neutrophils and neutrophil extracellular trap formation Circulating platelet leukocyte aggregates (PLA) in WT and Pyk2-deficient mice were similar in basal conditions, but they were significantly reduced in Pyk2-KO mice upon stimulation with thrombin (Figure 4A). Similar results were also obtained upon stimulation with the GPVI agonist convulxin (data not shown). Using antibodies against Ly6G or CD115, we found that Pyk2 deficiency selectively affected interaction of platelets with neutrophils, but not with monocytes (Figures 4B and C). The number of circulating PLA in whole blood collected 48 hours after DVT induced by IVC partial ligation was dramatically reduced in Pyk2KO mice compare to WT littermates (Figure 4D). We next evaluated in vitro platelet-stimulated NET extrusion. Neutrophils were left untreated or mixed with thrombin-activated platelets for 30 minutes, and NET formation was evaluated by DNA staining with Hoechst. Intact neutrophils showed the typical nucleus with multilobulated shape, that lost its compactness and shape and extruded DNA and histones upon NET formation (Figure 4E). In the absence of added platelets, no differences in NET formation were observed between WT and Pyk2-deficient neutrophils. However, upon addition of thrombinstimulated WT platelets a significant increase of NET

formation was observed with WT but not with Pyk2-deficient neutrophils, indicating that, in neutrophils Pyk2 is required to elaborate a functional response to activated platelets. Interestingly, addition of thrombin-stimulated platelets from Pyk2-KO mice failed to induce NET extrusion by both WT and Pyk2-deficient neutrophils (Figures 4E and F), demonstrating a role also for platelet Pyk2 in the development of a neutrophil NET phenotype. Importantly, upon neutrophil stimulation with PMA, that bypasses early signal transduction events, NET extrusion was comparable in WT and Pyk2-deficient cells (Figure 4G). Pyk2 modulates platelet adhesion to von Willebrand factor During DVT, platelets also bind to the activated endothelium and to the released VWF.36 Figure 5A shows that, in the presence of botrocetin, platelet adhesion to VWF under static conditions, was significantly impaired in the absence of Pyk2 (Figure 5A). Adhesion of platelets to activated EC in function of time was investigated under flow conditions over IL-1b activated HUVEC. Figure 5B shows that, when compared to WT platelets, Pyk2-KO platelets displayed a modestly reduced ability to adhere to a monolayer of activated

Haematologica | 107 - June 2022

1378

ARTICLE - The multifaced role of Pyk2 in venous thrombosis

S. Momi et al.

Figure 4. Pyk2 regulates platelets-neutrophils interaction and is required for neutrophil extracellular traps formation. (A) Platelet-leukocyte, (B) platelet-monocyte and (C) platelet-neutrophil aggregates in whole blood from wild-type (WT) (black bars) and Pyk2-knockout (Pyk2-KO) mice (white bars) before (basal) of upon stimulation with 0.1 U/mL thrombin were measured by flow cytometry using specific antibodies: CD45 for leukocytes, CD115 for monocytes, Ly6G for neutrophils and CD42 for platelets. Results are the mean ± standard error of the mean (SEM) of 3 different experiments. *P≤0.05; ***P≤0.005. (D) Platelet-neutrophil aggregates in whole blood of WT (black bar) or Pyk2-KO mice (white bar) measured 48 hours after partial inferior vena cava (IVC). ***P≤0.005. (E to G) Neutrophils were isolated from bone marrow and platelets from whole blood of WT and Pyk2-KO mice. Neutrophils (from either genotypes) were left untreated (none) or incubated with thrombin-activated platelets from WT mice or from Pyk2-KO mice, as indicated in (E and F), or with 100 nM phorbol 12-myristate 13-acetate (PMA) (G). Neutrophil extracellular trap (NET) formation was analyzed by fluorescent microscopy upon staining of nuclei with Hoechst. (E) Representative images at 400X magnification, while quantification of NET formation, expressed as % of NET-extruding neutrophils, is reported in panels (F and G). Results are the mean ± SEM of 3 different experiments. **P≤0.01

HUVEC. However, quantitative analysis of collected images showed that the percentage of total area covered by adherent platelets was slightly, but significantly higher for WT than for Pyk2-KO platelets (0.17±0.01% vs. 0.15±0.02%, respectively, P<0.05) (Figure 5Bii). Importantly, however, the ability of platelets to adhere over IL-1b-activated HUVEC was almost completely prevented when endothelial Pyk2 was pharmacological inhibited by PF-4594755 (Figure 5C, panels i and ii). These results indicate that Pyk2 activity in the endothelium, rather than in platelets, plays a predominant role in the regulation of platelet adhesion. Finally, we investigated the role of Pyk2 in the expression of platelet procoagulant activity, by measuring the surface exposure of PS, through annexin V binding. Figure 5D

shows that platelet PS exposure increased in both genotypes when platelets were stimulated with low or high doses of thrombin plus convulxin, but the percentage of annexin V-positive cells remained significantly lower in Pyk2-deficient platelets.

Discussion In this study we have shown that the proline-rich tyrosine kinase Pyk2 is a critical regulator of venous thrombosis in mice. This novel function of Pyk2 results from its multifaceted role in the pro-inflammatory responses of several blood cells involved in DVT development, including platelets, neutrophils and EC. Pyk2-KO mice failed to develop

Haematologica | 107 - June 2022

1379

ARTICLE - The multifaced role of Pyk2 in venous thrombosis

S. Momi et al.

Figure 5. Pyk2 regulates platelet adhesion to von Willebrand factor, rolling over human umbilical vein endothelial cells and phosphatidylserine expression. (A) Washed platelets from wild-type (WT) or Pyk2-knockout (Pyk2-KO) mice were let to adhere over immobilized von Willebrand afcator (VWF) in the presence of 0.5 mg/mL botrocetin for 60 minutes. Adherent platelets were stained with TRITC-conjugated phalloidin and counted with the software Image J. A representative image is reported in (i), while quantification of the results is reported in the histograms (ii) (black bar: WT platelets; white bar: Pyk2-deficient platelets). Results are the mean ± standard error of the mean (SEM) of 4 different experiments. ***P≤0.005. (B) Fluorescent-labeled platelets were perfused at a shear rate of 200 sec-1 over a monolayer of IL-1b (1 ng/mL)-activated human umbilical vein endothelial cells (HUVEC) and rolling platelets were evaluated by measuring fluorescent spots moving on the HUVEC surface. The relative area covered by platelets was calculated with ImageJ at 1 FPS steps. The plots in (i) show the covered area and quantitative analysis of are presented in (ii) as mean ± SEM of 3 different experiments. *P≤0.05; ***P≤0.005. (C) Fluorescent-labeled platelets were perfused over PF-4594755-treated HUVEC and the number of adherent cells was acquired at a rate of 1 frame/second (panel i). Quantification of the percentage of surface coverage is reported in panel (ii). ****P≤0.001. (D) Phosphatidylserine (PS) exposure on activated WT and Pyk2-KO mouse platelets. Murine platelets were left untreated (basal) or were treated with 0.5 U/mL thrombin plus 250 ng/mL convulxin (low dose) or with 1 U/mL thrombin plus 500 ng/mL convulxin (high dose) for 10 minutes. Cells were then stained with annexin V-FITC and analyzed by flow cytometry. Data are expressed as the percentage of positive cells and are the mean ± SEM of 3 different experiments. *P≤0.05; ****P≤0.001.

thrombi in the lower limbs upon IVC ligation or formed, occasionally, thrombi very small in size and weight. Specifically, we found that Pyk2: i) regulates the expression of adhesive molecules by activated EC; ii) modulates the expression of TF by activated monocytes and EC and the expression of pro-coagulant activity by platelets; iii) mediates neutrophil and platelet adhesion to EC under flow; iv) supports the formation of platelet neutrophil aggregates and stimulates NETosis. Previous studies have recognized the functional relevance of Pyk2 in vascular cells, but none have assessed its role in the mechanisms triggering venous thrombosis. We have

previously shown that Pyk2 is essential for platelet activation and adhesion,22,23 granule secretion, and arterial thrombosis.18 In neutrophils, Pyk2 regulates adhesion, motility and inflammatory responses,15,17,37,38 while in EC, Pyk2 was shown to regulate adhesion dynamics and to promote vascular inflammation.33 Platelets, neutrophils, and EC cooperate in the development of venous thrombosis.39 The present study documents for the first time that Pyk2 coordinates the activation of these vascular cells and regulates multiple events which converge toward the development of a pro-inflammatory, hypercoagulable, and prothrombotic phenotype. This articulated contribution of

Haematologica | 107 - June 2022

1380

ARTICLE - The multifaced role of Pyk2 in venous thrombosis Pyk2 to EC, platelet, and leukocyte functions provides a molecular explanation for its essential role in DVT. Importantly, cross-transfusion experiments indicate that the role of Pyk2 in the different vascular cells implicated in DVT is not redundant but may regulate specific steps of thrombus formation. In particular, the inability of transfused WT platelets to restore DVT in Pyk2-KO mice indicates that Pyk2 in EC or leukocytes is essential for the initiation of thrombus formation. By contrast, transfusion of Pyk2-deficient platelets into WT mice only caused a reduction of thrombus length and weight, suggesting that Pyk2 in platelets is not required for the initial formation of venous thrombi, but contributes to its development and growth. This is in line with the notion that EC and leukocytes act in an early phase during DVT while platelets may then contribute to recruit leukocytes and support fibrin formation by enhancing neutrophil-dependent coagulation.3 The contribution of Pyk2 on DVT is demonstrated using both genetic and pharmacologic approaches. Pyk2-KO mice have been used for in vivo studies and for the investigation of the molecular mechanisms involved at the level of isolated cells. Some experiments have been performed with human cells and the role of Pyk2 has been determined using the pharmacological inhibitor PF-4594755. The use of this inhibitor was necessary under all those circumstances in which it was impossible to obtain a significant number of primary cells from Pyk2-KO mice, mainly in the case of the studies performed on EC and monocytes. The combined use of murine and human cells does not represent a limitation of the study, but rather provides reliability to the translation of the in vivo observations made in mice to humans. In this regard, it is noteworthy that the inhibitor used in our study, PF-4594755, is selective and specific for Pyk2, as previously demonstrated by the analysis of its effects on Pyk2-deficient cells.34 We also found that PF-4594755 was able to prevent Pyk2 phosphorylation in activated platelets34 and well as in LPS-stimulated monocytes (data not shown). In the present study, the use of PF-4594755 was limited to in vitro experiments, because previous analysis of the pharmacokinetic properties of this compound highlighted its low bioavailability and unsuitability to be used in vivo probably because of its high hydrophobicity.40,41 A pro-inflammatory phenotype deriving in part from the prolonged disturbance of blood flow represents a common first trigger of DVT. Our results highlight a precise role for Pyk2 in the development of such a pro-inflammatory phenotype and are in line with previous works, based on the use of less selective inhibitors of FAK. For instance, Murphy and collaborators showed that dual FAK and Pyk2 inhibition by PF-271 reduced cytokine-induced expression of inflammatory adhesion molecules by HUVEC.33 Here, we found that the selective inhibition of Pyk2 by PF-4594755

S. Momi et al.

is sufficient to completely prevent IL-1b-induced expression of P-selectin, VCAM and TF, as well as VWF release by HUVEC, suggesting that Pyk2 is an important regulator of the pro-inflammatory phenotype of human EC. Accordingly, we found that, inhibition of Pyk2 in EC strongly prevented neutrophil rolling and adhesion under venous flow conditions. Moreover, also Pyk2-deficient neutrophils rolled on and adhered to HUVEC less efficiently than Pyk2-expressing cells. These data show that Pyk2 is an important regulator of the cell activation pathways supporting the leukocyte endothelium interactions. We also collected evidence that Pyk2 is required for the initiation of blood clotting. In fact, we have documented that expression of TF in stimulated monocytes and EC was completely prevented by the inhibition of Pyk2 by PF4594755. We also suggest that, at least in ECs, Pyk2 regulate TF expression mainly at a transcriptional level. Moreover, our results indicate that also Pyk2 expressed in platelets may contribute to clot formation, as the expression of procoagulant PS was significantly, albeit not completely, reduced in Pyk2-deficient platelets. These results point to a role for Pyk2 in the regulation of the extrinsic blood coagulation pathway. Actually, we did measure a prolonged aPTT in Pyk2-KO mice compared to WT littermates (data not shown). However, the specific mechanism underlying this observation remains to be elucidated, as the current aPTT assay implies the addition of a certain amount of exogenous TF, thus making difficult to appreciate the precise contribution of endogenous TF. Neutrophil and platelet activation are closely intertwined in the development of thromboinflammation. Stimulated neutrophils release NET which activate platelets promoting the release of pro-inflammatory cytokines and triggering clot formation.8 Platelet neutrophil interactions thus support the development of DVT. We found that platelet neutrophil aggregates were abundant in blood of WT mice upon induction of DVT by IVC ligation but were almost undetectable in Pyk2-KO mice which failed to develop DVT. Although it cannot be excluded that the reduced number of platelet neutrophil aggregates in Pyk2KO mice may be consequence of the failure to develop DVT, it is reasonable that this represents one of the causes of the lack of thrombus formation. Indeed, we found that the expression of Pyk2 both in platelets and neutrophils is required for the formation in vitro of heterologous aggregates upon stimulation with thrombin. These observations integrate and extend previous findings by Evangelista et al. showing that Src family kinases mediate neutrophil adhesion to platelets through the recruitment of Pyk2.10 P-selectin mediates platelet interaction with neutrophils, and it is noteworthy that the lack of Pyk2 impairs P-selectin exposure on activated platelets.18 The reduced ability of platelets and neutrophils to form heterologous aggregates may provide an additional mech-

Haematologica | 107 - June 2022

1381

ARTICLE - The multifaced role of Pyk2 in venous thrombosis anistic explanation for the markedly reduced DVT in Pyk2KO mice, since circulating platelet neutrophil aggregates have been recognized as a risk factor for DVT.42 NET formation by neutrophils is stimulated by the interaction with activated platelets and represents a triggering event for DVT. Src and MAP kinases which are commonly indicated as upstream and downstream effectors of Pyk2, respectively, are regulators of NETosis.43 In our study we show that Pyk2-deficient neutrophils form less NET in vitro when stimulated with either Pyk2-expressing or Pyk2-deficient platelets. We also verified that the lack of Pyk2 in platelets hampers NET formation by WT neutrophils in manner comparable to that produced by the lack of Pyk2 in neutrophils. Thus, Pyk2 elicits a biological effect by controlling the contribution of both cell types involved in this response. In conclusion, our study highlights a novel multifaceted role of Pyk2 in the cross-talk between platelets, neutrophils and EC and identifies this kinase as a novel player in the different steps involved in the development of DVT. Pyk2, therefore, could represent a novel potential therapeutic target for the prevention of DVT associated with inflammatory conditions.

S. Momi et al.

Disclosures No conflicts of interest to disclose. Contributions SM designed and performed experiments, analyzed data, and wrote the manuscript; JC, MV, LG, EF, GG, GCT and GFG performed experiments and analyzed data; PG analyzed data and edited the manuscript; MT and IC designed research, analyzed data, and wrote the manuscript. Acknowledgments The authors thank Prof. Cristina Lanni for providing THP-1. Funding This research was supported by the Italian Ministry of Education, University and Research (MIUR): Dipartimenti di Eccellenza Program (2018–2022), Department of Biology and Biotechnology “L. Spallanzani,” University of Pavia, and by Fondazione CARIPLO (2018-0483 to MT and PG). Data sharing statement Questions regarding data sharing should be addressed to the corresponding author.

References 1. Schulz C, Engelmann B, Massberg S. Crossroads of coagulation and innate immunity: the case of deep vein thrombosis. J Thromb Haemost. 2013;11(1):233-241. 2. Di Nisio M, van Es N, Büller HR. Deep vein thrombosis and pulmonary embolism. Lancet. 2016;388(10063):3060-3073. 3. von Brühl ML, Stark K, Steinhart A, et al. Monocytes, neutrophils, and platelets cooperate to initiate and propagate venous thrombosis in mice in vivo. J Exp Med. 2012;209(4):819835. 4. Scioli MG, Storti G, D'Amico F, et al. Oxidative stress and new pathogenetic mechanisms in endothelial dysfunction: potential diagnostic biomarkers and therapeutic targets. J Clin Med. 2020;9(6):1995. 5. Brill A, Fuchs TA, Chauhan AK, et al. von Willebrand factormediated platelet adhesion is critical for deep vein thrombosis in mouse models. Blood. 2011;117(4):1400-1407. 6. Budnik I, Brill A. Immune factors in deep vein thrombosis initiation. Trends Immunol. 2018;39(8):610-623. 7. Cimmino G, Cirillo P. Tissue factor: newer concepts in thrombosis and its role beyond thrombosis and hemostasis. Cardiovasc Diagn Ther. 2018;8(5):581-593. 8. Fuchs TA, Brill A, Wagner DD. Neutrophil extracellular trap (NET) impact on deep vein thrombosis. Arterioscler Thromb Vasc Biol. 2012;32(8):1777-1783. 9. Senis YA, Mazharian A, Mori J. Src family kinases: at the forefront of platelet activation. Blood. 2014;124(13):2013-2024. 10. Evangelista V, Pamuklar Z, Piccoli A, et al. Src family kinases mediate neutrophil adhesion to adherent platelets. Blood. 2007;109(6):2461-2469. 11. Totani L, Amore C, Di Santo A, et al. Roflumilast inhibits

leukocyte-platelet interactions and prevents the prothrombotic functions of polymorphonuclear leukocytes and monocytes. J Thromb Haemost. 2016;14(1):191-204. 12. Guidetti GF, Torti M, Canobbio I. Focal adhesion kinases in platelet function and thrombosis. Arterioscler Thromb Vasc Biol. 2019;39(5):857-868. 13. Avraham S, Avraham H. Characterization of the novel focal adhesion kinase RAFTK in hematopoietic cells. Leuk Lymphoma. 1997;27(3-4):247-256. 14. Okigaki M, Davis C, Falasca M, et al. Pyk2 regulates multiple signaling events crucial for macrophage morphology and migration. Proc Natl Acad Sci U S A. 2003;100(19):10740-10745. 15. Paone C, Rodrigues N, Ittner E, et al. The Tyrosine kinase Pyk2 contributes to complement-mediated phagocytosis in murine macrophages. J Innate Immun. 2016;8(5):437-451. 16. Kamen LA, Schlessinger J, Lowell CA. Pyk2 is required for neutrophil degranulation and host defense responses to bacterial infection. J Immunol. 2011;186(3):1656-1665. 17. Canino J, Guidetti GF, Galgano L, et al. The proline-rich tyrosine kinase Pyk2 modulates integrin-mediated neutrophil adhesion and reactive oxygen species generation. Biochim Biophys Acta Mol Cell Res. 2020;1867(10):118799. 18. Canobbio I, Cipolla L, Consonni A, et al. Impaired thrombininduced platelet activation and thrombus formation in mice lacking the Ca(2+)-dependent tyrosine kinase Pyk2. Blood. 2013;121(4):648-657. 19. Canobbio I, Lova P, Sinigaglia F, et al. Proline-rich tyrosine kinase 2 and focal adhesion kinase are involved in different phases of platelet activation by vWF. Thromb Haemost. 2002;87(3):509-517.

Haematologica | 107 - June 2022

1382

ARTICLE - The multifaced role of Pyk2 in venous thrombosis 20. Manganaro D, Consonni A, Guidetti GF, et al. Activation of phosphatidylinositol 3-kinase β by the platelet collagen receptors integrin α2β1 and GPVI: The role of Pyk2 and c-Cbl. Biochim Biophys Acta. 2015;1853(8):1879-1888. 21. Canobbio I, Cipolla L, Guidetti GF, et al. The focal adhesion kinase Pyk2 links Ca2+ signalling to Src family kinase activation and protein tyrosine phosphorylation in thrombin-stimulated platelets. Biochem J. 2015;469(2):199-210. 22. Consonni A, Cipolla L, Guidetti G, et al. Role and regulation of phosphatidylinositol 3-kinase β in platelet integrin α2β1 signaling. Blood. 2012;119(3):847-856. 23. Cipolla L, Consonni A, Guidetti G, et al. The proline-rich tyrosine kinase Pyk2 regulates platelet integrin αIIbβ3 outside-in signaling. J Thromb Haemost. 2013;11(2):345-356. 24. Canobbio I, Visconte C, Momi S, et al. Platelet amyloid precursor protein is a modulator of venous thromboembolism in mice. Blood. 2017;130(4):527-536. 25. Pitchford SC, Momi S, Giannini S, et al. Platelet P-selectin is required for pulmonary eosinophil and lymphocyte recruitment in a murine model of allergic inflammation. Blood. 2005;105(5):2074-2081. 26. Giannini S, Falcinelli E, Bury L, et al. Interaction with damaged vessel wall in vivo in humans induces platelets to express CD40L resulting in endothelial activation with no effect of aspirin intake. Am J Physiol Heart Circ Physiol. 2011;300(6):H2072-2079. 27. Falcinelli E, Petito E, Becattini C, et al. Role of endothelial dysfunction in the thrombotic complications of COVID-19 patients. J Infect. 2020;82(5):186-230. 28. Bosshart H, Heinzelmann M. THP-1 cells as a model for human monocytes. Ann Transl Med. 2016;4(21):438. 29. Petito E, Falcinelli E, Paliani U, et al. COVIR study investigators. Association of neutrophil activation, more than platelet activation, with thrombotic complications in Coronavirus disease 2019. J Infect Dis. 2021;223(6):933-944. 30. Mócsai A, Zhang H, Jakus Z, et al. G-protein-coupled receptor signaling in Syk-deficient neutrophils and mast cells. Blood. 2003;101(10):4155-4163. 31. Canobbio I, Visconte C, Oliviero B, et al. Increased platelet adhesion and thrombus formation in a mouse model of

S. Momi et al.

Alzheimer's disease. Cell Signal. 2016;28(12):1863-1871. 32. Momi S, Caracchini R, Falcinelli E, et al. Stimulation of platelet nitric oxide production by nebivolol prevents thrombosis. Arterioscler Thromb Vasc Biol. 2014;34(4):820-829. 33. Murphy JM, Jeong K, Rodriguez YAR, et al. FAK and Pyk2 activity promote TNF-α and IL-1β-mediated pro-inflammatory gene expression and vascular inflammation. Sci Rep. 2019;9(1):7617. 34. Guidetti GF, Zarà M, Canobbio I, et al. Novel pharmacological inhibitors demonstrate the role of the tyrosine kinase Pyk2 in adhesion and aggregation of human platelets. Thromb Haemost. 2016;116(5):904-917. 35. Kawecki C, Lenting PJ, Denis CV. von Willebrand factor and inflammation. J Thromb Haemost. 2017;15(7):1285-1294. 36. Coenen DM, Mastenbroek TG, Cosemans JMEM. Platelet interaction with activated endothelium: mechanistic insights from microfluidics. Blood. 2017;130(26):2819-2828. 37. Tse KW, Lin KB, Dang-Lawson M, et al. Small molecule inhibitors of the Pyk2 and FAK kinases modulate chemoattractant-induced migration, adhesion and Akt activation in follicular and marginal zone B cells. Cell Immunol. 2012;275(1-2):47-54. 38. Cheung SM, Ostergaard HL. Pyk2 Controls integrin-dependent CTL migration through regulation of de-adhesion. J Immunol. 2016;197(5):1945-1956. 39. Schrottmaier WC, Mussbacher M, Salzmann M, et al. Plateletleukocyte interplay during vascular disease. Atherosclerosis. 2020;307:109-120. 40. Bhattacharya SK, Aspnes GE, Bagley SW, et al. Identification of novel series of pyrazole and indole-urea based DFG-out PYK2 inhibitors. Bioorg Med Chem Lett. 2012;22(24):7523-7529. 41. Grossi M, Bhattachariya A, Nordström I, et al. Pyk2 inhibition promotes contractile differentiation in arterial smooth muscle. J Cell Physiol. 2017;232(11):3088-3102 42. Zhou J, Xu E, Shao K, et al. Circulating platelet-neutrophil aggregates as risk factor for deep venous thrombosis. Clin Chem Lab Med. 2019;57(5):707-715. 43. Nie M, Yang L, Bi X, et al. Neutrophil extracellular traps induced by IL8 promote diffuse large B-cell lymphoma progression via the TLR9 Signaling. Clin Cancer Res. 2019;25(6):1867-1879.

Haematologica | 107 - June 2022

1383

ARTICLE - Non-Hodgkin Lymphoma

CCR6 activation links innate immune responses to mucosa-associated lymphoid tissue lymphoma development Boguslawa Korona,1 Dagmara Korona,2 Wanfeng Zhao,3 Andrew C. Wotherspoon4 and MingQing Du1,5 Division of Cellular and Molecular Pathology, Department of Pathology, University of Cambridge, Cambridge; 2Department of Genetics, University of Cambridge, Cambridge; 3The Human Research Tissue Bank, Cambridge University Hospitals NHS Foundation Trust, Cambridge; 4Department of Histopathology, Royal Marsden Hospital, London and 5 Department of Histopathology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK 1

Correspondence: Ming-Qing Du mqd20@cam.ac.uk Received: September 23, 2021. Accepted: January 31, 2022. Prepublished: February 10, 2022. https://doi.org/10.3324/haematol.2021.280067 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract The genesis of extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT) is driven by oncogenic co-operation among immunological stimulations and acquired genetic changes. We previously identified recurrent CCR6 mutations in MALT lymphoma, with majority predicted to result in truncated proteins lacking the phosphorylation motif important for receptor desensitization. Functional consequences of these mutational changes, the molecular mechanisms of CCR6 activation and how this receptor signaling contributes to MALT lymphoma development remain to be investigated. In the present study, we demonstrated that these mutations impaired CCR6 receptor internalization and were activating changes, being more potent in apoptosis resistance, malignant transformation, migration and intracellular signaling, particularly in the presence of the ligands CCL20, HBD2 (human b defensin 2) and HD5 (human a defensin 5). CCR6 was highly expressed in malignant B cells irrespective of the lymphoma sites. HBD2 and CCL20 were constitutively expressed by the duct epithelial cells of salivary glands, and also those involved in lymphoepithelial lesions (LEL) in salivary gland MALT lymphoma. While in the gastric setting, HBD2, and HD5, to a less extent CCL20, were highly expressed in epithelial cells of pyloric and intestinal metaplasia respectively including those involved in LEL, which are adaptive responses to chronic Helicobacter pylori infection. These findings suggest that CCR6 signaling is most likely active in MALT lymphoma, independent of its mutation status. The observations explain why the emergence of malignant B cells and their clonal expansion in MALT lymphoma are typically around LEL, linking the innate immune responses to lymphoma genesis.

Introduction Extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) is a broad group of low-grade B-cell lymphomas from various mucosal sites. Despite sharing similar histological and biological characteristics, the lymphoma at various sites shows divergence in its etiology and mutation profile.1 As the lymphoma derives from a background of chronic inflammatory disorder, the emergence and expansion of the neoplastic B-cell clone are most likely the result of oncogenic cooperation between the persistent antigenic stimulations and acquired somatic genetic changes. There is increasing evidence showing that the acquired genetic changes affect the signaling pathways critical for the

function of marginal zone B cells, and corroborate with the antigenic drive or stimuli generated by the local microenvironment in lymphoma development.1 A classic example is that overexpression of MALT lymphoma associated oncogenes such as BCL10, MALT1 or BIRC3-MALT1 alone is insufficient for lymphoma development, but capable of inducing lymphoma-like lesions in the presence of immunogenic stimulations in various animal models.2,3 Among immunological stimulations, chronic B-cell receptor (BCR) activation, and T-cell help, particularly CD40/CD40L co-stimulation, are the signaling pathways known to be critically involved in the development of gastric MALT lymphoma.4-8 This notion has now been extended to MALT lymphoma of other sites, with more receptor signaling identified to be involved in

Haematologica | 107 - June 2022

1384

ARTICLE - CCR6 links inflammatory responses to lymphoma lymphomagenesis. For example, in thyroid MALT lymphoma that invariably arises in a background of Hashimoto’s thyroiditis, CD274 (PD-L1) and TNFRSF14 are commonly inactivated by mutation and/or deletion, and their inactivation eliminates their inhibitory regulation of T-helper cells, thus may enable T cells to provide more help to malignant B cells.9 Similarly, in salivary gland MALT lymphoma that is closely associated with lymphoepithelial sialadenitis (myoepithelial sialadenitis), GPR34 is activated by somatic mutations or t(X;14)(p11;q32), and also a paracrine stimulation via ligands generated by the lymphoepithelial lesion.10-12 Apart from GPR34, CCR6 is another G protein-coupled receptor (GPCR) which is recurrently mutated in MALT lymphoma.12 But unlike GPR34, CCR6 mutation is not restricted to MALT lymphoma of the salivary glands, but also seen in those of the stomach and thyroid.12 The majority of CCR6 mutations are nonsense changes or frameshift indels that are clustered in the C-terminal region, resulting in truncated products lacking the C-terminal phosphorylation motif (Figure 1A),12 which is responsible for binding to b-arrestin and receptor desensitization. The remaining mutations are missense changes including R159S and Y352C affecting the second intracellular loop and a putative C-terminal phosphorylation site respectively.12 It is unclear how these mutations affect CCR6 function and whether CCR6 signaling is also maintained by microenvironmental stimuli in MALT lymphoma, independent of its genetic changes. CCR6 is a member of class-A GPCR superfamily, which transduces ligand stimulation to intracellular signaling through G proteins. CCR6 is expressed in a range of leukocytes including mature B cells and their derived lymphomas, with the protein expression seen in 84-100% of MALT lymphoma.13-16 The ligands for CCR6 include CCL20 and b defensins.13 CCL20, also known as macrophage inflammatory protein 3a (MIP-3a), is produced by a range of cell types including macrophages and endothelial cells, and its expression is typically upregulated by inflammatory cytokines.13 b defensins are small, cationic, antimicrobial peptides, and produced by epithelial cells and leukocytes, particularly the former.17,18 Among human b defensins (HBD), HBD1-3 have been shown to act as a ligand for CCR6, and these defensins, like CCL20, are highly upregulated during inflammatory responses, to recruit and activate leukocytes.17,18 The CCR6/CCL20 axis is critical for humoral immune responses, particularly at mucosal sites, while the role of CCR6 and HBD interaction in B-cells and their derived lymphomas remain to be investigated.17,18 In order to understand the oncogenic action of CCR6 and its mutants, we investigated their receptor signaling and transformation potential in vitro, and their responses to stimulation by CCL20, HBD2 and human a defensin 5

B. Korona et al.

(HD5). We also investigated the expression of CCR6, CCL20, HBD2 and HD5 in primary MALT lymphoma tissue specimens, and identified evidence supporting paracrine stimulation of lymphoma B cells via CCR6 by CCL20/HBD2/HD5 released by the inflamed epithelial cells.

Methods CCR6 expression constructs The CCR6 wild-type and mutants (R159S and W335X) were cloned separately into pcDNA5/FRT (Online Supplementary Figure S1) and pIRESpuro vectors. All the above constructs contained a HA-tag at the N-terminus of CCR6, and a separate set of these constructs with an additional C-terminal GFP-tag were also generated. Generation of a single copy CCR6 stable expression isogenic cell lines Flp-InTRex293 host cells (genetically engineered from HEK293 to enable targeted integration of a single copy expression vector) were used to generate isogenic cell lines that stably express a single copy of the wild-type or mutant CCR6 expression construct, thus permitting direct comparison in their expression levels and response to ligand stimulations. Expression of recombinant CCR6 was confirmed (Online Supplementary Figure S2). Generation of CCR6 enriched expression DG75 line DG75 cells overexpressing the wild-type or mutant CCR6 were generated and recombinant protein expression was confirmed by flow cytometry (Online Supplementary Figure S3). Effect of CCR6 expression on cytotoxic challenge The isogenic Flp-InTRex293 cells and the DG75 B cells that express recombinant CCR6, were subjected to staurosporine treatment in the presence of CCL20 or vehicle for 16 hours. Level of cell death was determined by flow cytometry analysis of annexin V. Soft agar colony formation assay, wound scratch healing assay and transwell migration assays These experiments were performed using the Flp-InTRex293 cells that carried a single copy of the wild-type or mutant CCR6 expression construct, together with parental cells (please refer to the Online Supplementary Appendix for details) Analysis of CCR6 internalization For time-lapse microscopy, the Flp-InTRex293 cells that carried a single copy of wild-type or various mutant CCR6

Haematologica | 107 - June 2022

1385

ARTICLE - CCR6 links inflammatory responses to lymphoma with C-terminal GFP-tag were treated with CCL20 (50 nM) in FluoroBrite medium (Gibco) while recording by timelapse microscopy for 30 minutes (Online Supplementary Figure S4). The CCR6 expression was quantified using Image J (https://theolb.readthedocs.io/en/latest/imaging/measuring-cell-fluorescence-using-imagej.html). For flow cytometry analysis, the Flp-InTRex293 cells with a single copy of the wild-type or mutant CCR6 expression construct were treated with CCL20 (50 nM), and an aliquot (100 mL) was taken out at the indicated times for measuring surface CCR6.10,19 Dual luciferase reporter assay The firefly reporter plasmids for CRE (cAMP/PKA), SRF-RE (RhoA), SRE (MAPK/ERK), NFAT-RE (Calcium/Calcineurin), TCF/LEF-RE (Wnt), ISRE (JAK/STAT1/2), AP1 (MAPK/JNK), NFκB, CSL (NOTCH) and the Renilla pRL-TK control vector were from our previous studies.10,20,21 Reporter assays were optimized before data collection (Online Supplementary Figure S5). For detailed methodology, please see the On-

B. Korona et al.

line Supplementary Appendix. Prediction of coupling probabilities of CCR6 Coupling probability of G protein a to CCR6 wild-type and its various mutants was estimated by using an online machine-learning program PRECOG (predicting coupling probabilities of G-protein coupled receptors) (http://precog.russelllab.org) (Online Supplementary Figure S6).22 Immunohistochemistry Local ethical guidelines were followed for the use of archival tissues for research with ethical approval (05Q1604-10). The expression of CCR6, CCL20, HBD2, TFF2 (trefoil factor 2) and HD5 in MALT lymphoma, lymphoepithelial sialadenitis and normal salivary glands, was investigated by immunohistochemistry on formalin-fixed paraffin-embedded tissue sections (Online Supplementary Table S1). Please refer to the Online Supplementary Appendix for details.

Figure 1. CCR6 mutants confer increased resistance to apoptosis and greater transforming capacity. Schematic presentation of CCR6 mutations seen in MALT lymphoma. (A) Effect of CCR6 mutation on cell survival induced by staurosporine. Isogenic Flp-IN293 cells (top panel) that stably expressed a single copy of CCR6 or its mutants, and DG75 cells (below) that express CCR6 or its mutants were treated with 1.5 nM staurosporine in the presence or absence of CCL20 stimulation for 16 hours, cell death was then measured by flow cytometry analysis of annexin V binding. (B) Transforming potential of CCR6 and its mutants determined by soft agar colony formation assay. Isogenic FlpINTRex293 cells that stably expressed a single copy of CCR6 or its mutants were grown on soft agar for 3 weeks and colonies were stained with a crystal violet (top) and quantified (below). The data (mean + standard deviation) presented in (B and C) are from at least 3 independent experiments. Statistical significance is analyzed by a two‐tailed unpaired t‐test, with significance indicated. WT: wild-type.

Haematologica | 107 - June 2022

1386

ARTICLE - CCR6 links inflammatory responses to lymphoma

Results CCR6 mutants confer resistance to cell death In order to investigate whether CCR6 mutations conferred more resistance to cytotoxic damages, we investigated their effect on staurosporine induced cell death. In the absence of any ligand stimulation, both isogenic Flp-InTRex293 and DG75 cells expressing the CCR6 truncation, and the Flp-InTRex293 cells expressing the R159S mutant showed a significantly reduced cell death than those expressing the wild-type respectively (Figure 1B). In the presence of CCL20 ligand stimulation, all cell lines overexpressing CCR6 showed a significant reduction of cell death, and the Flp-InTRex293 and DG75 cells expressing the CCR6 truncation mutant displayed a significantly reduced cell death than those expressing the wild-type. CCR6 mutants show enhanced transforming potential In order to test whether mutation potentiated CCR6 transforming ability, we performed soft agar colony formation assay using the isogenic Flp-InTRex293 cells expressing the wild-type or mutant CCR6. The cells expressing the CCR6 W335X truncation, to a lesser extent the R159S mutant, showed a significantly higher number of colonies than those expressing the wild-type (Figure 1C). CCR6 mutants show enhanced migration ability In the absence of ligand stimulation, the isogenic Flp-InTRex293 cells expressing the CCR6 W335X truncation mutant showed an enhanced migration capacity than those expressing the wild-type as shown by the wound healing assay (Figure 2A). In the presence of CCL20 stimulation, both CCR6 W335X and R159S mutants displayed a significantly greater response to wound healing than the wildtype. Similar results were also seen from the transwell assays. Under the CCL20 stimulation, both CCR6 mutants displayed a significantly greater migration capacity than the wild-type (Figure 2B). While under the HBD2 stimulation, both mutants showed a similar trend, but only the truncation mutant reached a significant difference in comparison with the wild-type. CCR6 mutants show delayed internalization Receptor internalization following ligand stimulation downregulates GPCR signaling. In order to investigate whether mutation affected CCR6 internalization, we monitored the protein membrane expression in isogenic Flp-InTRex293 cell lines that expressed a single copy of wild-type or mutant CCR6 following CCL20 stimulation. In comparison with the wild-type CCR6, the W335X truncation mutant showed a significant higher level of membrane retention by time lapse microscopy (Figure 3A;

B. Korona et al.

Online Supplementary Figure S4). Similarly, flow cytometry analysis of these isogenic cell lines also demonstrated a significantly higher level of surface retention of the truncation mutant than the wild-type following CCL20 stimulation (Figure 3B). In both assays, there was no significant difference in surface CCR6 retention between the R159S missense mutant and the wild-type. In line with this, the microscopic appearance also showed a higher level of CCR6 W335X expression than both CCR6 R159S and CCR6 wild-type (Figure 3A). CCR6 mutants show enhanced signaling capacity Following ligand stimulation, GPCR may activate multiple cellular signaling through interaction with G-proteins. In order to investigate the signaling pathways activated by CCR6 and whether this was potentiated by mutation, we examined the activities of nine signaling pathways in DG75 cells using dual luciferase report assays (Figure 4; Online Supplementary Figure S7). In the absence of ligand stimulation, both the wild-type and mutant CCR6, particularly the truncation mutant showed a low level of enhanced NFAT RE (calcium/calcineurin) and TCF-LEF (Wnt/b-catenin) luciferase reporter activities in comparison with the control (Figure 4). In the presence of CCL20 stimulation, both CCR6 wild-type and mutant displayed a significantly increased SRE (MAPK/ERK) than their respective controls. Additionally, the two mutants, particularly the W335X truncation change, exhibited significantly greater SRF-RE (RhoA) reporter activities than the wild-type CCR6. The CCR6 W335X mutant also showed small, but significantly enhanced CSL (Notch), NFκB and CRE (cAMP/PKA) reporter activities than the wild-type (Online Supplementary Figure S7). We also performed SRE (MAPK/ERK) and SRF-RE (RhoA) reporter assays using the isogenic Flp-InTRex293 cells that expressed a single copy of wild-type or mutant CCR6 (Figure 5A). In the presence of CCL20 or HBD2 stimulation, both CCR6 wild-type and mutants showed significantly enhanced SRE (MAPK/ERK) and SRF-RE (RhoA) reporter activities than their respective controls. Furthermore, the CCR6 W335X truncation, to a lesser extent the R159S mutant where indicated, displayed significantly greater reporter activities than the wild-type. As expected, the SRF-RE (RhoA) reporter activities could be eliminated in the presence of the Rho-associated protein kinase ROCK inhibitor Y27632 (Figure 5A).23 In order to explore whether other defensin family members may also serve as ligand for CCR6,18 we modeled CCR6 and ligand binding using molecular docking software (https://bioinfo3d.cs.tau.ac.il/PatchDock/). This revealed that HD5 had a predicted binding affinity to CCR6, similar to CCL20, and subsequent reporter assays confirmed that HD5 is capable of stimulating CCR6 and its mutants (Figure 5B).

Haematologica | 107 - June 2022

1387

ARTICLE - CCR6 links inflammatory responses to lymphoma

B. Korona et al.

Figure 2. Effect of CCR6 mutation on cell migration. (A) Isogenic Flp-IN293 cells (left) that stably expressed a single copy of CCR6 or its mutants, together with their parental cell line, were measured for their migration capacity using wound closure assay (left panel). For each cell line, at least 19 images of the wound are taken at the time points indicated (right panel), and wound closure is presented as percentage of the initial scratched area. The data presented (mean ± standard deviation) are from 3 independent experiments. (B) Similarly, the isogenic Flp-IN293 cells that stably expressed a single copy of CCR6 or its mutants, together with their parental cell line, were assessed for their migrating capacity using transwell assays. The data presented (mean + standard deviation) are from 3 independent experiments. Statistical significance is analyzed by one‐way ANOVA with significant differences indicated. WT: wild-type.

Prediction of Gα coupling probabilities of CCR6 In general, there was a similar profile of Gα coupling probabilities among the wild-type CCR6 and its mutants as shown by the machine-learning program PRECOG although a higher binding probability for Gai3 and Ga15 was found for the truncation mutant (Online Supplementary Figure S6).22,24 The predicted Ga coupling profile was in line with the observed signaling activities measured by reporter assays (Figure 4).24-27 CCR6 and its ligand expression in mucosa-associated lymphoid tissue lymphoma In order to investigate whether CCR6 signaling was operational in MALT lymphoma, we investigated CCR6 and its ligands (CCL20, HBD2, HD5) expression in primary lymphoma tissue specimens by immunohistochemistry. Among the 29 cases of MALT lymphoma examined, strong

CCR6 expression in malignant B cells was seen 26 cases, independent of the CCR6 mutation status and lymphoma sites (Figures 6 and 7), although all the mutants cases (n=3, p.C336X, p.S344X, p.Y352C) examined showed strong, more uniform staining (Online Supplementary Figure S8). Interestingly, CCR6 expression was high in the malignant B cells involving in lymphoepithelial lesions, but downregulated in those undergoing plasmacytic differentiation (Figure 6). The expression of CCR6 ligands HBD2 and CCL20 was investigated in MALT lymphoma (salivary gland: n=14; stomach: [n=29 for HBD2; n=11 for CCL20]; thyroid: n=79), lymphoepithelial sialadenitis (n=5), chronic gastritis (n=3) and normal salivary gland tissues (n=2), and HD5 only in gastrointestinal tissues in light its known expression in Paneth cells.28 These ligands were not expressed in malignant B cells, but found in epithelial cells with their ex-

Haematologica | 107 - June 2022

1388

ARTICLE - CCR6 links inflammatory responses to lymphoma

B. Korona et al.

Figure 3. Effect of CCR6 mutation on receptor internalization. (A) Analysis of sub-cellular localization of CCR6 and it representative mutants by time-lapse microscopy (n=3). Isogenic Flp-INTRex293 cell lines that stably expressed a single copy of HA-CCR6-GFP or its mutants were treated with CCL20 (50 nM), and CCR6 subcellular localization was monitored for 30 minutes. Shown are images at the indicated times following ligand stimulation (left). Please also refer to the Online Supplementary Figure S4 for the video record. The CCR6 membrane expression in individual cells was qualified at the indicated times using Image J and normalized to the 0 time point (right). Comparison between wild-type (WT) CCR6 and its various mutants was performed with Prism6 non-linear regression analyses, with significantly difference indicated. (B) The experiment is similarly carried out as above but with CCR6 surface expression analyzed by flow cytometry using an anti-HA antibody. The CCR6 surface expression is normalized to the time 0 point in each cell line. The data (mean ± standard deviation) presented are from 3 independent experiments. Statistical comparison between wild-type CCR6 and its mutants was carried out using Prism6 non-linear regression analyses (sigmoidal model fitting), with significant differences indicated.

Figure 4. Comparison of CCR6 and its representative mutants in activation of various cellular signaling using reporter assays. This was carried out in DG75 B-cell lymphoma cell line transiently co-transfected with CCR6 expression constructs and reporter plasmids. Data (mean + standard deviation) are from at least 2 independent experiments performed in triplicate. Comparisons between two groups were assessed using one‐way ANOVA with significant differences indicated. WT: wild-type. Haematologica | 107 - June 2022

1389

ARTICLE - CCR6 links inflammatory responses to lymphoma pression patterns varying among different mucosal sites where MALT lymphoma occurred (Figures 6 and 7; Online Supplementary Figures S9 and S10). In normal salivary glands including those adjacent to MALT lymphoma lesion, HBD2 was expressed moderately in the ductal, but not in acinar epithelial cells, while CCL20 was expressed moderately in the ductal, and strongly in acinar epithelial cells (Online Supplementary Figure S9). In salivary gland MALT lymphoma where acini were obliterated, and ducts involved by LEL, weak HBD2 and CCL20 expression were seen in the ductal epithelial cells of LEL (Figure 6A). In chronic gastritis including those in gastric MALT lym-

B. Korona et al.

phoma, HBD2 was expressed weakly in gastric epithelial cells, but strongly in a subset cell population of the inflamed epithelia, particularly in those of basal glands, which were also strong positive for TFF2, a marker of pyloric metaplasia 29 (Figure 7A; Online Supplementary Figure S10). HBD2 was at very low level or negative in the Paneth cells of intestinal metaplasia (Online Supplementary Figure S10). Similarly, CCL20 expression was weakly expressed in gastric epithelial cells, but moderately in a subset cell population of the inflamed epithelia (Online Supplementary Figure S10). In gastric MALT lymphoma, the HBD2 and CCL20 expression pattern in LEL was very simi-

Figure 5. Response of CCR6 and its representative mutants to HBD2 and HD5 stimulation as measured by reporter assays using isogenic Flp-InTRex293 cell lines that stably express a single copy of CCR6 or its mutant. (A) CCR6 and its mutants show similar response profile to CCL20 and HBD2 stimulation as measured by SRF-RE reporter assay, and the receptor mediated RhoA activation can be effectively abrogated by the Rho-associated protein kinase ROCK inhibitor Y27632 (Miltenyi Biotec). (B) Structural modeling of GPCR and ligand binding. HD5 is predicted to have a binding affinity to CCR6, similar to CCL20 using molecular docking software (https://bioinfo3d.cs.tau.ac.il/PatchDock/), and is capable of stimulating CCR6 and its mutants as shown by reporter assay. Data (mean + standard deviation) presented are from at least 3 independent experiments. Comparisons between various groups are assessed using one‐way ANOVA, with significant differences indicated. WT: wild-type.

Haematologica | 107 - June 2022

1390

ARTICLE - CCR6 links inflammatory responses to lymphoma lar to those of inflamed epithelia with HBD2 highly expressed in a subset of the epithelial cells in glands with high TFF2 expression, including those involved in LEL (Figure 7B). Intestinal metaplasia is a frequent finding in gastric MALT lymphoma, and as expected, HD5 was strongly expressed in the Paneth cells of intestinal metaplasia, including

B. Korona et al.

those involved in LEL (Figure 7B). This was particularly prominent in small intestinal MALT lymphoma (Online Supplementary Figure S11). There was no apparent expression of HBD2 and CCL20 in the normal thyroid epithelial cells. In thyroid MALT lymphoma, both protein expression were weak or negative in the epithelial cells of LEL.

Figure 6. HBD2, CCL20 and CCR6 expression in a representative case of salivary gland mucosa-associated lymphoid tissue lymphoma. (A) Adjacent tissue shows lymphoepithelial sialadenitis and uninvolved salivary glands. HBD2 is expressed moderately in the ductal (D), but not in acinar (A) epithelial cells, and weakly in the epithelial cells involved in lymphoepithelial lesions (LEL). CCL20 is expressed moderately in the ductal, strongly in acinar epithelial cells, and weakly in those involved in LEL. (B) Area of mucosa-associated lymphoid tissue (MALT) lymphoma shows expansion of malignant B cells surrounding LEL, which frequently display plasmacytic differentiation (PD). The neoplastic nature of the B-cell is indicated by IgΚ light chain restriction. The malignant B cells, including those forming LEL, show strong CCR6 expression, but those undergoing plasmacytic differentiation display weak to negative staining. Haematologica | 107 - June 2022

1391

ARTICLE - CCR6 links inflammatory responses to lymphoma

Discussion By using a range of in vitro functional assays, we have shown that the CCR6 C-terminal truncation and R159S

B. Korona et al.

mutations are activation changes, causing CCR6 constitutive activation and enhanced signaling upon ligand stimulation. We have also demonstrated the expression of CCR6 in malignant B cells and its ligands (HBD2, CCL20,

Figure 7. Expression of CCR6 ligands in pyloric and intestinal metaplasia, and gastric mucosa-associated lymphoid tissue lymphoma. (A) HBD2 is highly expressed in a subset of epithelial cells within glands which are consistent with pyloric metaplasia as identified by high TFF2 (trefoil factor 2) expression. While HD5, but not HBD2, is highly expressed in the Paneth cells of intestinal metaplasia identified by strong AB (Alcian blue) staining and goblet cells. (B) A representative case of gastric mucosa-associated lymphoid tissue (MALT) lymphoma showing CCL20, HBD2 and HD5 expression in the epithelial cells involved in lymphoepithelial lesion (LEL). CCR6 is highly expressed in malignant B-cells including those involved in LEL. Haematologica | 107 - June 2022

1392

ARTICLE - CCR6 links inflammatory responses to lymphoma HD5) in the inflamed epithelial cells within the lymphoma microenvironment, suggesting that CCR6 signaling is operational in MALT lymphoma. In vitro assays showed clear functional differences between the CCR6 W335X truncation and R159S mutants, with the former being more potent in apoptosis resistance, malignant transformation, migration and intracellular signaling, particularly in the presence of ligand stimulation. The CCR6 truncation mutant lacks the C-terminal phosphorylation motif,12,19 which mediates binding to b-arrestin, triggering receptor desensitization and internalization.30 As expected, the truncation mutant showed a higher level of membrane retention following ligand stimulation, and this most likely underpins its much potent oncogenic activities observed by in vitro assays. It is unclear how the R159S mutation enhances CCR6 signaling. R159 is located in the second intracellular loop, and its change to serine may destabilize the receptor as suggested by an online protein stability prediction neural network (a predictive score of -1.153, much lesser than <0 required for a destabilizing change).31 Destabilizing mutation can increase receptor flexibility and obviate the necessity for ligand binding to open G-protein binding pocket, and enhance basal signaling. 32 Interestingly, R159 is located in G-protein binding pocket based on the CCR6 structure in complex with a Go protein (Online Supplementary Figure S12).33 Thus, R159 residue change may alter CCR6 structure and provide more favorable interacting pocket for G-protein activation.32,34 Among the nine intracellular signaling pathways examined, the two CCR6 mutants, and to a lesser extent the wild-type were highly sensitive to ligand stimulation in both SRE (MAPK/ERK) and SRF-RE (RhoA) reporter assays. SRE (MAPK/ERK) activation mainly drives cell proliferation, while SRF-RE (RhoA) signaling largely promotes cell migration.35 Through activation of both cellular pathways, CCR6 signaling may help B cells to migrate to inflammatory sites, and maintain their activation, consequently

B. Korona et al.

contributing to malignant transformation and clonal expansion. In support of this, CCR6 is highly expressed in MALT lymphoma cells, independent of its mutation status. Moreover, both CCL20 and HBD2, the ligand of CCR6, are constitutively expressed in epithelial cells, including those involved in the LEL within MALT lymphoma. These findings suggest that CCR6 signaling is likely commonly involved in the pathogenesis of MALT lymphoma, and its activation may be maintained by persistent exposure to ligand stimulation, independent of CCR6 mutation. Nonetheless, CCR6 mutations can sensitize the receptor to ligand stimulation, thereby enhancing the receptor signaling and its pathogenic potential. The inflamed epithelial cells are likely the major source of HBD2 and CCL20 in MALT lymphoma, and there are important differences in their origin between salivary gland and gastric MALT lymphoma. In salivary gland MALT lymphoma, these ligands are constitutively expressed by ductal epithelial cells as shown in the present study and also supported by the literature,36,37 and are likely more actively released when these epithelial cells undergo cellular stress or damage due to LEL. Importantly, it is these ductal glands that are preferentially invaded by B cells, forming LEL, and the LEL are a dynamic structure, with epithelial cells undergoing active regeneration. This provides a perpetual microenvironment for relentless production and release of HBD2 and CCL20, thus chronic stimulation of B cells via CCR6 (Figure 8). In line with these findings, the emergence of malignant B cells and their clonal expansion are always around the LEL, and this is known histologically as a halo appearance, an important feature for recognizing early lesion of salivary gland MALT lymphoma.38,39 In gastric MALT lymphoma, HBD2 and CCL20 are mainly expressed in the inflamed epithelial glands, particularly in a subset of epithelial cells within glands consistent with pyloric metaplasia, while HD5 is strongly expressed in the Paneth cells of intestinal metaplasia.28 Defensins

Figure 8. CCR6 activation links innate immune responses to mucosa-associated lymphoid tissue lymphoma development. HBD2 and CCL20 are constitutively expressed in salivary gland epithelial cells, and together with HD5 are induced to be expressed at a high level in the inflamed gastric epithelial cells, likely through the metaplastic process. These ligands are likely actively released when the epithelial cells are under stress or involved in lymphoepithelial lesions (LEL). LEL are a dynamic structure, undergoing active regeneration, thus providing an enduring source of ligands for CCR6 activation, and linking innate immune responses to mucosa-associated lymphoid tissue (MALT) lymphoma genesis. Haematologica | 107 - June 2022

1393

ARTICLE - CCR6 links inflammatory responses to lymphoma are a group of antimicrobial peptides, better known in its high expression in the intestinal Paneth cells, and serve as innate responses to constrain bacterial colonization.40 Most gastric MALT lymphomas arise in a background of chronic H. pylori gastritis, in which both pyloric and intestinal metaplasia are a common histological finding, often being multifocal.29,41-43 These metaplasia are most likely an adaptive response to H. pylori infection, as the intestinal mucosa is evolutionarily developed to serve as a barrier to defend gut microbiota invasion. Apart from producing mucins and a high pH environment, intestinal metaplasia also harbors the Paneth cells that produce abundant antimicrobial peptides including defensins.40 In support of this, the areas of gastric mucosa, which show intestinal metaplasia with defensins 5/6 expressing Paneth cells, have reduced H. pylori colonization, and these defensins also bear antibacterial activities in vitro at a low concentration.28, 44 Similarly, HBD2, although not other family members, was induced and expressed at a high level in H. pylori associated gastritis, and HBD2 can inhibit H pylori growth in vitro in a dose-dependent manner.44,45 Our findings of high HBD2 and CCL20 expression in inflamed gastric epithelia are in keeping with the above observations, and further bridge the innate immune responses to MALT lymphoma genesis (Figure 8). Similar to salivary gland MALT lymphoma, the emergence of neoplastic B cells and their clonal expansion in the gastric form also start around the LEL, extending to the marginal zone of reactive B-cell follicles. The injured epithelial cells also undergo active regeneration. This together with broad mucosal involvement by the lymphoma cells may provide them enduring favorable environmental milieu. Apart from the survival advantage, the lymphoma cells close to LEL also frequently show histological features of plasmacytic differentiation and blast transformation, suggesting the presence of a plethora of immunological stimulations in this microenvironment. In this context, the CCR6 and CCL20/HBD2/HD5 interaction between the malignant B cells and inflamed epithelia identified in the present study may represent only one of the many crosstalks to be discovered between the two cell populations. Indeed, in salivary gland MALT lymphoma, we recently identified evidence supporting a paracrine stimulation of malignant B-cells via GPR34 by ligand generated by LEL.10 In a proportion of cases, both CCR6 and GPR34 are potentially co-expressed in the same lymphoma cells (Online Supplementary Figure S13) and the two receptor signaling may

B. Korona et al.

cooperate in their oncogenic activities in the development of MALT lymphoma. In this context, it is worth noting that the two receptors activate different signaling, with CCR6 primarily SRE (MAPK/ERK) and SRF-RE (RhoA), while GPR34 predominantly NF-kB and AP1-RE (MAPK/JNK), hence potentially complementing in their biological activities. In conclusion, our findings indicate that CCR6 mutations seen in MALT lymphoma are activation changes, and these mutants are highly sensitive to ligand stimulation, hence enhancing the receptor signaling. More importantly, CCR6 is commonly expressed in MALT lymphoma, independent of CCR6 mutation status, and its signaling is most likely operational in the lymphoma cells via relentless stimulation by HBD2 and CCL20 produced by inflamed epithelial cells, particularly those involved in LEL (Figure 8). In gastric MALT lymphoma, HBD2 and HD5 are mainly derived from pyloric and intestinal metaplasia, potentially linking innate immune responses to lymphoma genesis. These findings elegantly explain why the emergence of malignant B cells and their clonal expansion typically around the LEL at multiple mucosal sites. Disclosures No conflicts of interest to disclose. Contributions BK, DK, WZ and MQD set up the experimental design, collected and analyzed data; AW provided pathology input; BK prepared and wrote the manuscript; MQD provided research funding, set up the study design and coordination. All authors commented on the manuscript and approved its submission for publication. Acknowledgements The authors would like to thank Professor Timothy Cox for help to access their microplate reader, and Mr Ian Clark for help to use their microscopy facility. Funding This research in MQD's lab was supported by grants from the Kay Kendall Leukemia Fund (KKL1141) UK, Blood Cancer UK (13 006), Isaac Newton Trust, Addenbrookes Charitable Trust and Department of Pathology, University of Cambridge. The Human Research Tissue Bank is supported by the NIHR Cambridge Biomedical Research Center. Data sharing statement Not applicable as data generated in this study have been included in the published article files.

Haematologica | 107 - June 2022

1394

ARTICLE - CCR6 links inflammatory responses to lymphoma

B. Korona et al.

References 1. Du MQ. MALT lymphoma: a paradigm of NF-κB dysregulation. Semin Cancer Biol. 2016;39:49-60. 2. Li Z, Wang H, Xue L, et al. Emu-BCL10 mice exhibit constitutive activation of both canonical and noncanonical NF-kappaB pathways generating marginal zone (MZ) B-cell expansion as a precursor to splenic MZ lymphoma. Blood. 2009;114(19):41584168. 3. Baens M, Fevery S, Sagaert X, et al. Selective expansion of marginal zone B cells in Emicro-API2-MALT1 mice is linked to enhanced IkappaB kinase gamma polyubiquitination. Cancer Res. 2006;66(10):5270-5277. 4. Hussell T, Isaacson PG, Crabtree JE, Spencer J. The response of cells from low-grade B-cell gastric lymphomas of mucosaassociated lymphoid tissue to Helicobacter pylori. Lancet. 1993;342(8871):571-574. 5. Hussell T, Isaacson PG, Spencer J. Proliferation and differentiation of tumour cells from B-cell lymphoma of mucosa-associated lymphoid tissue in vitro. J Pathol. 1993;169(2):221-227. 6. Greiner A, Knörr C, Qin Y, et al. Low-grade B cell lymphomas of mucosa-associated lymphoid tissue (MALT-type) require CD40mediated signaling and Th2-type cytokines for in vitro growth and differentiation. Am J Pathol. 1997;150(5):1583-1593. 7. Craig VJ, Cogliatti SB, Arnold I, et al. B-cell receptor signaling and CD40 ligand-independent T cell help cooperate in Helicobacter-induced MALT lymphomagenesis. Leukemia. 2010;24(6):1186-1196. 8. Hömig-Hölzel C, Hojer C, Rastelli J, et al. Constitutive CD40 signaling in B cells selectively activates the noncanonical NFkappaB pathway and promotes lymphomagenesis. J Exp Med. 2008;205(6):1317-1329. 9. Wu F, Watanabe N, Tzioni MM, et al. Thyroid MALT lymphoma: self-harm to gain potential T-cell help. Leukemia. 2021;35(12):3497-3508. 10. Korona B, Korona D, Zhao W, Wotherspoon A, Du M-Q. GPR34 activation potentially bridges lymphoepithelial lesion to genesis of salivary gland MALT lymphoma. Blood. 2022;139(14):21862197. 11. Moody S, Escudero-Ibarz L, Wang M, et al. Significant association between TNFAIP3 inactivation and biased immunoglobulin heavy chain variable region 4-34 usage in mucosa-associated lymphoid tissue lymphoma. J Pathol. 2017;243(1):3-8. 12. Moody S, Thompson JS, Chuang SS, et al. Novel GPR34 and CCR6 mutation and distinct genetic profiles in MALT lymphomas of different sites. Haematologica. 2018;103(8):13291336. 13. Lee AYS, Körner H. The CCR6-CCL20 axis in humoral immunity and T-B cell immunobiology. Immunobiology. 2019;224(3):449454. 14. Rodig SJ, Jones D, Shahsafaei A, Dorfman DM. CCR6 is a functional chemokine receptor that serves to identify select Bcell non-Hodgkin's lymphomas. Hum Pathol. 2002;33(12):1227-1233. 15. Deutsch AJ, Aigelsreiter A, Steinbauer E, et al. Distinct signatures of B-cell homeostatic and activation-dependent chemokine receptors in the development and progression of extragastric MALT lymphomas. J Pathol. 2008;215(4):431-444. 16. Takata K, Tanino M, Ennishi D, et al. Duodenal follicular lymphoma: comprehensive gene expression analysis with insights into pathogenesis. Cancer Sci. 2014;105(5):608-615.

17. Lee AY, Phan TK, Hulett MD, Körner H. The relationship between CCR6 and its binding partners: does the CCR6-CCL20 axis have to be extended? Cytokine. 2015;72(1):97-101. 18. Fruitwala S, El-Naccache DW, Chang TL. Multifaceted immune functions of human defensins and underlying mechanisms. Semin Cell Dev Biol. 2019;88:163-172. 19. Lu MY, Lu SS, Chang SL, Liao F. The phosphorylation of CCR6 on distinct Ser/Thr residues in the carboxyl terminus differentially regulates biological function. Front Immunol. 2018;9:415. 20. Escudero-Ibarz L, Wang M, Du MQ. Significant functional difference between TNFAIP3 truncation and missense mutants. Haematologica. 2016;101(9):e382-384. 21. Clipson A, Wang M, de Leval L, et al. KLF2 mutation is the most frequent somatic change in splenic marginal zone lymphoma and identifies a subset with distinct genotype. Leukemia. 2015;29(5):1177-1185. 22. Singh G, Inoue A, Gutkind JS, Russell RB, Raimondi F. PRECOG: PREdicting COupling probabilities of G-protein coupled receptors. Nucleic Acids Res. 2019;47(W1):W395-W401. 23. Vongsa RA, Zimmerman NP, Dwinell MB. CCR6 regulation of the actin cytoskeleton orchestrates human beta defensin-2- and CCL20-mediated restitution of colonic epithelial cells. J Biol Chem. 2009;284(15):10034-10045. 24. Wu V, Yeerna H, Nohata N, et al. Illuminating the OncoGPCRome: novel G protein-coupled receptor-driven oncocrine networks and targets for cancer immunotherapy. J Biol Chem. 2019;294(29):11062-11086. 25. Cheng Z, Garvin D, Paguio A, Stecha P, Wood K, Fan F. Luciferase reporter assay system for deciphering GPCR pathways. Curr Chem Genomics. 2010;4:84-91. 26. Kamato D, Thach L, Bernard R, et al. Structure, function, pharmacology, and therapeutic potential of the G protein, Galpha/q,11. Front Cardiovasc Med. 2015;2:14. 27. Vogt S, Grosse R, Schultz G, Offermanns S. Receptor-dependent RhoA activation in G12/G13-deficient cells: genetic evidence for an involvement of Gq/G11. J Biol Chem. 2003;278(31):2874328749. 28. Tanabe H, Sato T, Watari J, et al. Functional role of metaplastic paneth cell defensins in Helicobacter pylori-infected stomach. Helicobacter. 2008;13(5):370-379. 29. Wright NA. Aspects of the biology of regeneration and repair in the human gastrointestinal tract. Philos Trans R Soc Lond B Biol Sci. 1998;353(1370):925-933. 30. Gupta MK, Mohan ML, Naga Prasad SV. G protein-coupled receptor resensitization paradigms. Int Rev Cell Mol Biol. 2018;339:63-91. 31. Cao H, Wang J, He L, Qi Y, Zhang JZ. DeepDDG: predicting the stability change of protein point mutations using neural networks. J Chem Inf Model. 2019;59(4):1508-1514. 32. Stoy H, Gurevich VV. How genetic errors in GPCRs affect their function: possible therapeutic strategies. Genes Dis. 2015;2(2):108-132. 33. Wasilko DJ, Johnson ZL, Ammirati M, et al. Structural basis for chemokine receptor CCR6 activation by the endogenous protein ligand CCL20. Nat Commun. 2020;11(1):3031. 34. Huang H, Tao YX. Functions of the DRY motif and intracellular loop 2 of human melanocortin 3 receptor. J Mol Endocrinol. 2014;53(3):319-330. 35. Lavoie H, Gagnon J, Therrien M. ERK signalling: a master regulator of cell behaviour, life and fate. Nat Rev Mol Cell Biol. 2020;21(10):607-632.

Haematologica | 107 - June 2022

1395

ARTICLE - CCR6 links inflammatory responses to lymphoma 36. Kaneda Y, Yamaai T, Mizukawa N, et al. Localization of antimicrobial peptides human beta-defensins in minor salivary glands with Sjögren's syndrome. Eur J Oral Sci. 2009;117(5):506510. 37. Wang Y, Shnyra A, Africa C, Warholic C, McArthur C. Activation of the extrinsic apoptotic pathway by TNF-alpha in human salivary gland (HSG) cells in vitro, suggests a role for the TNF receptor (TNF-R) and intercellular adhesion molecule-1 (ICAM-1) in Sjögren's syndrome-associated autoimmune sialadenitis. Arch Oral Biol. 2009;54(11):986-996. 38. Hyjek E, Smith WJ, Isaacson PG. Primary B-cell lymphoma of salivary glands and its relationship to myoepithelial sialadenitis. Hum Pathol. 1988;19(7):766-776. 39. Bacon CM, Du MQ, Dogan A. Mucosa-associated lymphoid tissue (MALT) lymphoma: a practical guide for pathologists. J Clin Pathol. 2007;60(4):361-372. 40. Singh R, Balasubramanian I, Zhang L, Gao N. Metaplastic paneth cells in extra-intestinal mucosal niche indicate a link to

B. Korona et al.

microbiome and inflammation. Front Physiol. 2020;11:280. 41. Yoshida A, Isomoto H, Hisatsune J, et al. Enhanced expression of CCL20 in human Helicobacter pylori-associated gastritis. Clin Immunol. 2009;130(3):290-297. 42. Bauer B, Wex T, Kuester D, Meyer T, Malfertheiner P. Differential expression of human beta defensin 2 and 3 in gastric mucosa of Helicobacter pylori-infected individuals. Helicobacter. 2013;18(1):6-12. 43. Pero R, Coretti L, Nigro E, et al. β-defensins in the fight against Helicobacter pylori. Molecules. 2017;22(3):424. 44. Pero R, Angrisano T, Brancaccio M, et al. Beta-defensins and analogs in Helicobacter pylori infections: mRNA expression levels, DNA methylation, and antibacterial activity. PLoS One. 2019;14(9):e0222295. 45. Hamanaka Y, Nakashima M, Wada A, et al. Expression of human beta-defensin 2 (hBD-2) in Helicobacter pylori induced gastritis: antibacterial effect of hBD-2 against Helicobacter pylori. Gut. 2001;49(4):481-487.

Haematologica | 107 - June 2022

1396

ARTICLE - Plasma Cell Disorders

Isatuximab plus carfilzomib and dexamethasone versus carfilzomib and dexamethasone in relapsed multiple myeloma patients with renal impairment: IKEMA subgroup analysis Marcelo Capra,1 Thomas Martin,2 Philippe Moreau,3 Ross Baker,4 Ludek Pour,5 Chang-Ki Min,6 Xavier Leleu,7 Mohamad Mohty,8 Marta Reinoso Segura,9 Mehmet Turgut,10 Richard LeBlanc,11 Marie-Laure Risse,12 Laure Malinge,13 Sandrine Schwab12 and Meletios Dimopoulos14 Centro Integrado de Hematologia e Oncologia, Hospital Mãe de Deus, Porto Alegre, Brazil; Department of Medicine, University of California at San Francisco, San Francisco, CA, USA; 3 Department of Hematology, University of Nantes, Nantes, France; 4Perth Blood Institute, Murdoch University, Perth, Western Australia, Australia; 5Department of Internal Medicine, Hematology and Oncology, University Hospital Brno, Brno, Czech Republic; 6Department of Hematology, Catholic Hematology Hospital and Leukemia Research Institute, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea; 7 Service d'Hématologie et Thérapie Cellulaire, CHU and CIC INSERM 1402, Poitiers Cedex, France; 8Department of Hematology, Hôpital Saint-Antoine, Sorbonne University, INSERM UMRS 938, Paris, France; 9Hospital Universitario Virgen del Rocio, Sevilla, Spain; 10 Department of Hematology, Ondokuz Mayıs University Faculty of Medicine, Samsun, Turkey; 11Hôpital Maisonneuve-Rosemont, Université de Montréal, Montréal, Quebec, Canada; 12Sanofi Research and Development, Vitry/Alfortville, France; 13Aixial, BoulogneBillancourt, France and 14Department of Clinical Therapeutics, School of Medicine, National and Kapodistrian University of Athens School of Medicine, Athens, Greece. 1

Correspondence: Marcelo Capra marcelocapra@hotmail.com Received: May 28, 2021. Accepted: October 5, 2021. Prepublished: October 14, 2021. https://doi.org/10.3324/haematol.2021.279229 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract Renal impairment (RI) is common in patients with multiple myeloma (MM) and new therapies that can improve renal function are needed. The phase III IKEMA study (clinicaltrials gov. Identifier: NCT03275285) investigated isatuximab (Isa) with carfilzomib and dexamethasone (Kd) versus Kd in relapsed MM. This subgroup analysis examined results from patients with RI, defined as estimated glomerular filtration rate <60 mL/min/1.73 m². Addition of Isa prolonged progression-free survival (PFS) in patients with RI (hazard ratio: 0.27; 95% confidence interval [CI]: 0.11–0.66; median PFS not reached for Isa-Kd versus 13.4 months for Kd [20.8-month follow-up]). Complete renal responses occurred more frequently with Isa-Kd (52.0%) versus Kd (30.8%) and were durable in 32.0% versus 7.7% of patients, respectively. Treatment exposure was longer with Isa-Kd, with median number of started cycles and median duration of exposure of 20 versus 9 cycles and 81.0 versus 35.7 weeks for Isa-Kd versus Kd, respectively. Among patients with RI, the incidence of patients with grade ≥3 treatment-emergent adverse events was similar between the two arms (79.1% in Isa-Kd vs. 77.8% in Kd). In summary, the addition of Isa to Kd improved clinical outcomes with a manageable safety profile in patients with RI, consistent with the benefit observed in the overall IKEMA study population.

Introduction Multiple myeloma (MM) is characterized by abnormal proliferation of plasma cells and production of M-protein, a monoclonal immunoglobulin (Ig). Renal impairment (RI) affects up to 50% of MM patients, depending on how RI is defined. MM-related RI is multifactorial, but mainly caused by precipitation of Ig-free light chains in the distal tubules, leading to tubule obstruction and cast nephropathy.1 RI is a major cause of morbidity and an adverse predictor of survival in MM patients.2,3 As renal

function recovery is associated with improved clinical outcomes, it is one of the main therapeutic goals in MM patients with RI. Urgent therapy is required to achieve reversal of severe RI, since renal failure established for >2 weeks would substantially compromise the possibility of recovery.4-7 Newly introduced anti-myeloma therapies such as proteasome inhibitors (i.e., bortezomib, carfilzomib)3,8-12 and immunomodulatory drugs (i.e., lenalidomide, pomalidomide)13-20 aid in the recovery of renal function.21,22 Carfilzomib is a nextgeneration proteasome inhibitor approved as monotherapy

Haematologica | 107 - June 2022

1397

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

or in combination with dexamethasone (Kd), lenalidomide/ dexamethasone, or daratumumab/dexamethasone for relapsed/refractory MM (RRMM).8,23 The phase III ENDEAVOR study demonstrated superiority of Kd versus bortezomib/ dexamethasone (Vd) in RRMM patients with 1–3 prior treatment lines.8 Median progression-free survival (PFS) was 18.7 months with Kd versus 9.4 months with Vd (hazard ratio [HR]: 0.53; 95% confidence interval [CI]: 0.44–0.65; P<0.0001]. Median overall survival (OS) was 47.6 months with Kd versus 40.0 months with Vd (HR: 0.791; 95% CI: 0.65–0.96; one-sided P=0.010). A post-hoc exploratory subgroup analysis of ENDEAVOR reported complete renal response in 15.3% of Kd-treated patients, with longer median PFS and OS in patients achieving complete renal responses.9 The ENDEAVOR study results showed activity in patients with renal function impairment, supporting Kd as a therapeutic option for MM patients with RI.7-9 However, Kd treatment in patients with RI may present challenges, as carfilzomib has been associated with renal toxicity and hypertension, and may require repeated administration of intravenous fluids compared with oral or subcutaneous alternatives.7-9,23 Based on the phase III ICARIA-MM study, isatuximab (Isa), an anti-CD38 monoclonal antibody, is approved in a number of countries in combination with pomalidomide/ dexamethasone for the treatment of RRMM patients who have received ≥2 prior therapies, including lenalidomide and a proteasome inhibitor.24-27 Based on the phase III IKEMA study, Isa to date is also approved in combination with Kd in the United States for patients with relapsed MM who have received 1–3 prior treatment lines and in the European Union for MM patients who have received ≥1 prior therapy.24,25,28 A pre-specified IKEMA interim analysis showed that PFS was prolonged by the addition of Isa (median PFS, not reached for Isa-Kd versus 19.2 months with Kd; stratified HR: 0.53; 99% CI: 0.32–0.89; one-sided log-rank test P=0.0007), crossing the pre-specified efficacy boundary (P=0.005).28 This pre-specified subgroup analysis of IKEMA examined efficacy, renal response, and safety in patients with RI, at the time of the interim analysis.

Methods Study design IKEMA (clinicaltrials gov. Identifier: NCT03275285) was a prospective, multinational, randomized, open-label, parallel-group, phase III study conducted at 69 study centers in 16 countries.29 Institutional ethics committees or independent review boards approved the study protocol for each center. The study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Guidelines for Good Clinical Practice. All patients provided written informed consent.

Patients Details of the study methodology have been reported previously.28,29 Briefly, eligible patients had relapsed MM with 1–3 prior lines of therapy. Patients were excluded if they had primary refractory MM or serum free-light chain measurable disease only, had received prior carfilzomib treatment, were refractory to anti-CD38 antibody therapy, or presented with left ventricular ejection fraction <40%. Patients with a baseline estimated glomerular filtration rate (eGFR) as low as 15 mL/min/1.73m² were eligible for enrolment.30 Patients with prior pulmonary comorbidities, including chronic obstructive pulmonary disease, could be enrolled.29 Randomization Patients were randomly assigned in a 3:2 ratio to receive Isa-Kd or Kd. Randomization was stratified by number of prior treatment lines (1 vs. >1) and Revised International Staging System (R-ISS) stage I or II versus stage III versus not classified, at study entry.28 Treatment Patients in the Isa-Kd arm received Isa intravenously at 10 mg/kg on days 1, 8, 15, and 22 in the first 28-day cycle; and days 1 and 15 in subsequent cycles. In both arms, carfilzomib was administered intravenously at 20 mg/m2 on days 1 and 2; 56 mg/m2 on days 8, 9, 15, and 16 of cycle 1; and then 56 mg/m2 on days 1, 2, 8, 9, 15, and 16 of subsequent cycles.28 Dexamethasone 20 mg was administered intravenously or orally on days 1, 2, 8, 9, 15, 16, 22, and 23. Treatment continued until unacceptable adverse event (AE), disease progression, or other discontinuation criteria. Study endpoints and measured outcomes The primary efficacy endpoint was PFS, as per blinded independent response committee (IRC). The IRC reviewed disease assessments for response and progression (central radiological evaluation, M-protein quantification from central laboratory, and local bone marrow aspiration for plasma cell infiltration when needed). Key secondary efficacy endpoints included overall response rate (ORR) according to the International Myeloma Working Group (IMWG) response criteria,31 very good partial response (VGPR) or better rate, measurable residual disease (MRD) negativity rate, complete response (CR) rate, and OS.32-34 MRD was assessed by central laboratory using next-generation sequencing (NGS) Adaptive clonoSEQ Assay (Adaptive Biotechnologies, Seattle, WA) with a minimum sensitivity of 1/105 nucleated cells in patients reaching ≥VGPR. Efficacy assessments were performed on day 1 of every cycle and at end of treatment. Safety assessments included recording of AE (graded per NCI-CTCAE v4.03), lab-

Haematologica | 107 - June 2022

1398

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

oratory parameters (including complete blood, neutrophil, and platelet counts; and hemoglobin values, graded per NCI-CTCAE version 4.03), vital signs, electrocardiograms, and Eastern Cooperative Oncology Group performance status (ECOG PS). Safety was regularly reviewed by an Independent Data Monitoring Committee. Renal response Both renal function impairment and renal response were analyzed. The eGFR was assessed using the modification of diet in renal disease (MDRD) equation on days 1, 2, 8, 9, 15, 16, and 22 of cycle 1; days 1, 8, and 15 of cycle 2; days 1 and 15 of each subsequent treatment cycle, and as clinically indicated. eGFR results were classified as RI (<60 mL/min/1.73 m²) or no RI (≥60 mL/min/1.73 m²). Based on IMWG criteria, complete renal response was defined as an increase in eGFR from <50 mL/min/1.73 m² at baseline to ≥60 mL/min/1.73 m² (no RI) in ≥1 post-baseline assessment.6,7 Responses were considered durable when lasting ≥60 days.6 A minor renal response was defined as an improvement in eGFR from ≥15 to <30 mL/min/1.73 m² at baseline to ≥30 mL/min/1.73 m² in ≥1 assessment during treatment.7 Statistical analysis Sample size calculation was based on the primary efficacy endpoint; 159 events were needed to detect a 41% lower risk of disease progression (HR: 0.59) using a log-rank test (one-sided significance level of 0.025, 90% power). An interim PFS analysis was pre-planned when 65% of the 159 PFS events (103 events) were observed to detect overwhelming efficacy. All efficacy analyses were conducted in the intent-to-treat population and summarized by randomized treatment. Extent of study treatment and treatment-emergent AE (TEAE) analyses were conducted in the safety population. Median PFS, probabilities of being progression-free, and corresponding CI were estimated using the Kaplan-Meier method. HR estimates were determined using the Cox proportional-hazard model by subgroup. Comparisons between patients with and without RI were observational only, with no formal statistical analysis performed. SAS 9.4 (SAS, Cary, NC) was used for all analyses.

Results Patients A total of 302 patients were randomized to Isa-Kd (n=179) or Kd (n=123). Baseline eGFR values could be calculated for 165 patients in the Isa-Kd arm and 111 in the Kd arm. Baseline eGFR was not evaluable for 14 patients in Isa-Kd and 12 in Kd, due to local legal restrictions on collecting racial group information. Among evaluable patients, 43

(26.1%) in the Isa-Kd arm and 18 (16.2%) in the Kd arm had RI (eGFR <60 mL/min/1.73 m2). Of these, 39 (23.6%) patients in Isa-Kd and 15 (13.5%) in Kd had moderate RI (eGFR ≥30 to <60 mL/min/1.73 m²); 2.4% of patients in IsaKd and 2.7% in Kd had severe RI (eGFR ≥15 to <30 mL/min/1.73 m²). Among patients with RI at baseline, characteristics were generally well balanced between study arms (Table 1), except for more patients aged ≥75 years in the Isa-Kd than the Kd arm (14.0% vs. 5.6%, respectively) and more patients refractory to lenalidomide (25.6% Isa-Kd vs. 50.0% Kd) or to immunomodulatory drugs and proteasome inhibitors (18.6% Isa-Kd vs. 44.4% Kd) in the control arm. Patients with RI in both the Isa-Kd and Kd arms tended to be older, had more ISS stage III disease, and received more prior therapy lines than patients without RI (Table 1). Patient flow was described previously.28 Efficacy At a median overall follow-up of 20.7 months, the PFS benefit of Isa-Kd versus Kd in patients with and without RI, according to the assessment by the IRC, was consistent with that seen for the overall IKEMA study population (Figure 1). The addition of Isa prolonged PFS in patients with RI (HR: 0.27; 95% CI: 0.11–0.66; median PFS not reached for Isa-Kd vs. 13.4 months for Kd) and in patients without RI (HR: 0.63; 95% CI: 0.39–1.00; medians not reached). Probability to be free of a PFS event at 18 months was 79% with Isa-Kd versus 41% with Kd in patients with RI and 71% with Isa-Kd versus 59% with Kd in those without RI. Multivariate analysis of PFS for patients with RI was performed to adjust the imbalance at baseline between IsaKd and Kd, including ISS stage, gain(1q21), refractory to PI or IMiD therapy, sex, and regulatory region as covariates. Adjusted HR was equal to 0.21 (95% CI: 0.07–0.68), suggesting that the imbalance did not influence the treatment effect in favor of Isa-Kd for PFS. Consistent treatment effect was also observed in patients with the most severe RI at baseline (eGFR <45 mL/min/1.73 m²), as an exploratory analysis, in favor of patients treated in Isa-Kd (HR: 0.16; 95% CI: 0.04–0.67; median PFS, not reached for Isa-Kd [n=19] versus 11.14 months for Kd [n=10]) and in patients with eGFR ≥45 mL/min/1.73 m² (HR: 0.60; 95% CI: 0.39–0.93; medians not reached, n=146 versus n=101). In the intent-to-treat population, the ORR was higher with Isa-Kd versus Kd for patients with RI (93.1% vs. 61.1%, respectively; Figure 2). Although the ORR was 83.6% with IsaKd versus 89.2% with Kd for patients without RI, the ≥VGPR rate for patients with RI was 79.1% with Isa-Kd versus 44.4% with Kd, and for patients without RI, it was 71.3% versus 59.1%, respectively. The MRD negativity rate, assessed by NGS at 10-5 sensitivity level in bone marrow aspirates from

Haematologica | 107 - June 2022

1399

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

Table 1. Baseline characteristics in patients with and without renal impairment in the isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd) and Kd arms – intent-to-treat population eGFR <60 mL/min/1.73 m2 (n = 61)

eGFR ≥60 mL/min/1.73 m2 (n = 215)

Isa-Kd (n = 43)

Kd (n = 18)

Isa-Kd (n = 122)

Kd (n = 93)

67 (39–86)

69 (49–90)

64 (37–81)

62 (33–78)

<65

15 (34.9)

4 (22.2)

64 (52.5)

55 (59.1)

65–74 ≥75

22 (51.2) 6 (14.0)

13 (72.2) 1 (5.6)

47 (38.5) 11 (9.0)

31 (33.3) 7 (7.5)

Stage I

8 (18.6)

4 (22.2)

29 (23.8)

23 (24.7)

Stage II

9 (20.9)

6 (33.3)

40 (32.8)

39 (41.9)

Stage III

18 (41.9)

7 (38.9)

30 (24.6)

16 (17.2)

Unknown

8 (18.6)

1 (5.6)

23 (18.9)

15 (16.1)

Stage I

3 (7.0)

2 (11.1)

36 (29.5)

28 (30.1)

Stage II

34 (79.1)

11 (61.1)

72 (59.0)

52 (55.9)

Stage III

5 (11.6)

3 (16.7)

8 (6.6)

5 (5.4)

1 (2.3)

2 (11.1)

6 (4.9)

8 (8.6)

High risk

9 (20.9)

5 (27.8)

30 (24.6)

24 (25.8)

Standard risk

29 (67.4)

12 (66.7)

76 (62.3)

57 (61.3)

5 (11.6)

1 (5.6)

16 (13.1)

12 (12.9)

2 (1–4)

2 (1–3)

1 (1–4)

1 line, n (%)

17 (39.5)

5 (27.8)

55 (45.1)

47 (50.5)

≥2 lines, n (%)

26 (60.5)

13 (72.2)

67 (54.9)

46 (49.5)

Lenalidomide

11 (25.6)

9 (50.0)

39 (32.0)

26 (28.0)

IMiD and PI

8 (18.6)

8 (44.4)

23 (18.9)

15 (16.1)

Median age, years (range) Age in years by category, n (%)

ISS stage at initial diagnosis, n (%)

R-ISS stage at study entry, n (%)

Not classified Cytogenetic risk , n (%) a

Missing Prior lines of therapy Median (range)

Patients refractory to, n (%)

High risk was defined as del(17p), t(4;14), or t(14;16) by fluorescence in situ hybridization. Cytogenetics was assessed by a central laboratory with a cut-off of 50% for del(17p), and 30% for t(4;14) and t(14;16). d: dexamethasone; eGFR: estimated glomerular filtration rate; IMiD: immunomodulatory drug; Isa: isatuximab; ISS: International Staging System; ITT: intent-to-treat; K: carfilzomib; PI: proteasome inhibitor; RI: renal impairment; R-ISS: revised International Staging System.

patients who achieved ≥VGPR, was 30.2% with Isa-Kd versus 11.1% with Kd for patients with RI and 29.5% with IsaKd versus 14.0% with Kd for patients without RI. In addition, the CR rate for patients with RI was 41.9% with Isa-Kd versus 22.2% with Kd, and for patients without RI, it was 40.2% versus 30.1%, respectively (Figure 2). Although OS data were not mature at the interim analysis, 17% and 20% of patients died in the Isa-Kd and Kd arms, respectively (among patients with RI: 12% in Isa-Kd versus 39% in Kd and among patients without RI: 18% in Isa-Kd versus 15% in Kd). Renal response Among the 25 and 13 patients in the Isa-Kd and Kd arms,

respectively, with eGFR <50 mL/min/1.73 m² at baseline, more patients in the Isa-Kd than the Kd arm had a complete renal response (52.0% vs. 30.8%; Figure 3A). Durable complete renal response occurred in eight of 25 (32.0%) Isa-Kd versus one of 13 (7.7%) Kd patients. In patients with severe RI at baseline (eGFR ≥15 to <30 mL/min/1.73 m²), all patients in the Isa-Kd arm achieved minor renal response compared with only one patient in the Kd arm (4/4 [100%] versus 1/3 [33.3%], respectively) (Figure 3A). Moreover, the time to first renal response and time to complete renal response were shorter in patients with baseline eGFR <50 mL/min/1.73 m² treated with Isa-Kd. Median time (95% CI) to first renal response was 1.51 (0.82–not calculable [NC]) months in the Isa-Kd arm ver-

Haematologica | 107 - June 2022

1400

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

sus 6.51 (0.69–NC) months in the Kd arm (Figure 3B). Median time (95% CI) to complete renal response was 7.82 (1.22–NC) months in the Isa-Kd arm versus NC (1.28–NC) months in the Kd arm (Figure 3C). A similar incidence of patients experienced, at least once, end-stage RI (eGFR <15 mL/min/1.73 m²) during treatment with Isa-Kd versus Kd (1.8% vs. 2.7%, respectively). In the safety population, the number of patients with ≥1 TEAE mapped in the acute renal failure Standardized MedDRA

Queries (SMQ) narrow terms was nine of 177 (5.1%, of which 1.1% were grade ≥3) in the Isa-Kd arm versus ten of 122 (8.2%, 2.5% grade ≥3). Acute kidney injury was observed in five of 177 (2.8%, 1.1% grade ≥3) Isa-Kd patients versus seven of 122 (5.7%, 1.6% grade ≥3) Kd patients. Treatment exposure Longer treatment duration was observed with Isa-Kd versus Kd in patients with and without RI (Table 2). The

Figure 1. Progression-free survival with isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd) compared with Kd. (A) Patients with renal impairment (RI) (eGFR <60 mL/min/1.73 m²) or (B) without RI (eGFR ≥60 mL/min/1.73 m²), (ITT population). Progression-free survival (PFS) as per blinded independent response committee. d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; ITT: intent to treat; K: carfilzomib; PFS: progression-free survival. Haematologica | 107 - June 2022

1401

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

median (range) number of cycles for Isa-Kd was 20 (2–25) for patients with RI and 19 (1–27) cycles for those without RI, and for Kd it was 9 (1–24) cycles for patients with RI and 17 (1–28) cycles for those without RI. Median duration of exposure for patients with and without RI was 81.0 (6– 104) and 78.6 (1–111) weeks with Isa-Kd versus 35.7 (1–97) and 68.5 (4–114) weeks with Kd. More patients were still on treatment at the cut-off date in the Isa-Kd arm (55.8% with RI and 54.1% without RI) versus the Kd arm (16.7% with RI and 36.6% without RI). The reasons for definitive treatment discontinuation in patients with RI were progressive disease (27.9% in the Isa-Kd vs. 33.3% in the Kd arm) and AE (7.0% in the Isa-Kd vs. 27.8% in the Kd arm). In patients without RI, 26.2% in the Isa-Kd versus 37.6% in the Kd arm discontinued treatment due to progressive disease and 9.8% in the Isa-Kd versus 9.7% in the Kd arm due to AE.

The median relative dose intensity of Isa was similar in patients with or without RI; thus, RI did not impact Isa administration. The relative dose intensity of carfilzomib in patients with RI was lower in the Kd arm (84.6%) than in the Isa-Kd arm (93.1%), but similar in patients without RI (90.1% in the Isa-Kd vs. 91.4% in the Kd arm), indicating that more carfilzomib doses were delayed, reduced, or omitted in patients with RI who received Kd (Table 2). Safety TEAE were experienced in 97.7% of Isa-Kd versus 100% of Kd patients with RI, whereas 93.7% versus 94.6% of patients without RI experienced TEAE in the Isa-Kd versus Kd arms, respectively (Table 3). In patients with RI, grade ≥3 TEAE were reported in 79.1% of Isa-Kd versus 77.8% of Kd patients and serious TEAE in 62.8% of Isa-Kd versus 77.8% of Kd pa-

Figure 2. Response rates with isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd) compared with Kd. (A) Patients with renal impairment (RI) (eGFR <60 mL/min/1.73 m²) or (B) without RI (eGFR ≥60 mL/min/1.73 m²), (ITT population). CR: complete response; d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; ITT: intent to treat; K: carfilzomib; MRD neg: minimal residual disease negativity; ORR: overall response rate; PR: partial response; VGPR: very good partial response. MRD was assessed by next-generation sequencing with a sensitivity level 10-5. Haematologica | 107 - June 2022

1402

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

Figure 3. Renal response in the isatuximab (Isa) carfilzomib (K) dexamethasone (d) (IsaKd) compared with Kd arms. (A) Complete and durable (≥60 days) renal responses in patients with eGFR <50 mL/min/1.73 m² at baseline and minor renal responses in patients with eGFR ≥15 and <30 mL/min/1.73 m² at baseline, (ITT population). (B) Time to first renal response and (C) time to first complete renal response in patients with eGFR <50 mL/min/1.73 m² at baseline. CrR: complete renal response; d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; ITT: intent to treat; K: carfilzomib; rR: renal response.

Haematologica | 107 - June 2022

1403

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

Table 2. Overall extent of exposure in patients with and without renal impairment in the isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd) and Kd arms – safety population

Exposure

eGFR <60 mL/min/1.73 m2 (n = 61)

eGFR ≥60 mL/min/1.73 m2 (n = 212)

Isa-Kd (n = 43)

Kd (n = 18)

Isa-Kd (n = 120)

Kd (n = 92)

20 (2–25)

9 (1–24)

19 (1–27)

17 (1–28)

81.0 (6–104)

35.7 (1–97)

78.6 (1–111)

68.5 (4–114)

Isatuximab

94.3 (79.2–105.0)

–

94.3 (66.7–108.2)

–

Carfilzomib

93.1 (47.7–108.7)

84.6 (44.7–100.2)

90.1 (18.2–107.5)

91.4 (41.8–108.6)

Dexamethasone

85.8 (26.8–101.1)

90.2 (31.2–100.0)

85.4 (24.5–100.2)

88.1 (27.4–101.1)

Median number of cycles started, (range) Median treatment duration, weeks (range) Median relative dose intensity, % (range)

d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; K: carfilzomib; RI: renal impairment.

Table 3. Safety summary in patients with and without renal impairment in the isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd) and Kd arms – safety population

n (%)

eGFR <60 mL/min/1.73 m2 (n = 61) Isa-Kd Kd (n = 43) (n = 18)

eGFR ≥60 mL/min/1.73 m2 (n = 212) Isa-Kd Kd (n = 120) (n = 92)

Patients with any TEAE

42 (97.7)

18 (100)

116 (96.7)

87 (94.6)

Patients with any grade ≥3 TEAE

34 (79.1)

14 (77.8)

93 (77.5)

60 (65.2)

Patients with any grade 5 TEAEa

2 (11.1)

5 (4.2)

1 (1.1)

27 (62.8)

14 (77.8)

71 (59.2)

50 (54.3)

3 (7.0)

5 (27.8)

12 (10.0)

9 (9.8)

Patients with any serious TEAE Patients with any TEAE leading to definitive discontinuation a

TEAE with fatal outcome during the treatment period. d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; K: carfilzomib; RI: renal impairment; TEAE: treatment-emergent adverse event.

tients. In patients with RI, treatment with Isa-Kd did not increase the incidence of TEAE with fatal outcome during treatment (Isa-Kd, 0% vs. Kd, 11.1% [2/18]) nor of TEAE leading to treatment discontinuation (Isa-Kd, 7.0% vs. Kd, 27.8%, Table 3). TEAE occurring in ≥15% of patients treated with Isa-Kd are shown in Table 4, by renal function group and treatment arm. In patients with RI, the most common TEAE with IsaKd versus Kd were diarrhea (41.9% vs. 22.2%), upper respiratory tract infection (39.5% vs. 27.8%), hypertension (34.9% vs. 27.8%), and fatigue (34.9% vs. 22.2%). The most common TEAE with Isa-Kd versus Kd in patients without RI were hypertension (40.0% vs. 32.6%), upper respiratory tract infections (39.2% vs. 26.1%), and diarrhea (36.7% vs. 31.5%). Infusion reactions were observed in 37.2% of IsaKd versus 5.6% of Kd patients with RI and 45.8% of Isa-Kd

versus 3.3% of Kd patients without RI, but no grade ≥3 infusion reactions were reported. Hypertension was the most common grade ≥3 TEAE independently of renal function: 20.9% with Isa-Kd versus 22.2% with Kd in patients with RI and 20.8% with Isa-Kd versus 18.5% with Kd in patients without RI (Table 4). Carfilzomib has been reported to cause cardiac complications.35 Cardiac failure (by standardized MedDRA query) was observed in 11.6% (2.3% grade ≥3) of Isa-Kd patients with RI versus 5.6% (5.6% grade ≥3) of Kd patients with RI. In patients without RI, cardiac failure was observed in 5.8% (4.2% grade ≥3) of Isa-Kd patients versus 6.5% (3.3% grade ≥3) of Kd patients. The most common hematologic abnormalities based on laboratory results in treated patients with RI were anemia (all patients in both arms) and thrombocytopenia

Haematologica | 107 - June 2022

1404

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

(93.0% with Isa-Kd versus 83.3% with Kd, Table 5). In patients without RI, incidence of anemia was 99.2% with Isa-Kd versus 100% with Kd, whereas incidence of thrombocytopenia was 94.2% with Isa-Kd versus 89.1% with Kd.

Importantly, the incidence of grade 3–4 anemia and thrombocytopenia were comparable in all subgroups, while the incidence of grade 3 neutropenia was higher in the Isa-Kd arm in patients with or without RI (Table 5).

Table 4. Treatment-emergent adverse events occurring in ≥15% of patients treated with isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd), according to the renal impairment status – safety population.

Most common TEAE (in ≥15% of patients treated with Isa-Kd, worst grade) by preferred term, n (%)

eGFR <60 mL/min/1.73 m2 (n = 61) Isa-Kd (n = 43)

eGFR ≥60 mL/min/1.73 m2 (n = 212)

Kd (n = 18)

Isa-Kd (n = 120)

Kd (n = 92)

Any grade

Grade ≥3

Any grade

Grade ≥3

Any grade

Grade ≥3

Any grade

Grade ≥3

Diarrhea

18 (41.9)

4 (22.2)

1 (5.6)

44 (36.7)

5 (4.2)

29 (31.5)

1 (1.1)

Upper respiratory tract infection

17 (39.5)

5 (27.8)

1 (5.6)

47 (39.2)

6 (5.0)

24 (26.1)

1 (1.1)

Infusion reaction

16 (37.2)

1 (5.6)

55 (45.8)

3 (3.3)

Hypertensiona

15 (34.9)

9 (20.9)

5 (27.8)

4 (22.2)

48 (40.0)

25 (20.8)

30 (32.6)

17 (18.5)

Fatigue

15 (34.9)

4 (9.3)

4 (22.2)

1 (5.6)

35 (29.2)

2 (1.7)

19 (20.7)

Dyspnea

14 (32.6)

1 (2.3)

2 (11.1)

31 (25.8)

8 (6.7)

24 (26.1)

1 (1.1)

Back pain

13 (30.2)

1 (2.3)

2 (11.1)

24 (20.0)

2 (1.7)

19 (20.7)

1 (1.1)

Cough

10 (23.3)

1 (5.6)

22 (18.3)

16 (17.4)

Pneumonia

8 (18.6)

5 (11.6)

5 (27.8)

4 (22.2)

31 (25.8)

22 (18.3)

18 (19.6)

11 (12.0)

Bronchitis

7 (16.3)

1 (2.3)

1 (5.6)

25 (20.8)

3 (2.5)

10 (10.9)

1 (1.1)

Headache

7 (16.3)

1 (5.6)

19 (15.8)

19 (20.7)

Nausea

7 (16.3)

3 (16.7)

19 (15.8)

16 (17.4)

Asthenia

7 (16.3)

3 (16.7)

2 (11.1)

19 (15.8)

3 (2.5)

15 (16.3)

2 (2.2)

Nasopharyngitis

7 (16.3)

16 (13.3)

7 (16.3)

16 (13.3)

Edema peripheral

7 (16.3)

2 (11.1)

14 (11.7)

1 (0.8)

17 (18.5)

Fall

7 (16.3)

2 (4.7)

12 (10.0)

1 (0.8)

10 (10.9)

Insomnia

5 (11.6)

1 (2.3)

3 (16.7)

1 (5.6)

32 (26.7)

6 (5.0)

23 (25.0)

2 (2.2)

Vomiting

5 (11.6)

3 (16.7)

18 (15.0)

2 (1.7)

8 (8.7)

1 (1.1)

Both new hypertension and worsening hypertension were included in the preferred term ‘hypertension’; hypertension in medical history was 41.0% in Isa-Kd vs. 18% in Kd in patients with RI and 24.7% in Isa-Kd vs. 12.1% in Kd in patients without RI. d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; K: carfilzomib; RI: renal impairment; TEAE: treatment-emergent adverse event.

Table 5. Hematologic abnormalities determined by laboratory analysis in patients with and without renal impairment in the isatuximab (Isa) carfilzomib (K) dexamethasone (d) (Isa-Kd) and Kd arms – safety population. Hematologic laboratory abnormalitiesa, n (%) Anemia

eGFR <60 mL/min/1.73 m2 (n = 61) Isa-Kd (n = 43) Any grade

Grade 3

43 (100) 11 (25.6)

Kd (n = 18)

Grade 4 0

Any grade

Grade 3

18 (100) 5 (27.8)

Thrombocytopenia 40 (93.0) 9 (20.9)

8 (18.6) 15 (83.3) 6 (33.3)

Neutropenia

1 (2.3)

24 (55.8) 6 (14.0)

eGFR ≥60 mL/min/1.73 m2 (n = 212)

7 (38.9)

Isa-Kd (n = 120) Grade 4 0

Any grade

Grade 3

119 (99.2) 23 (19.2)

Kd (n = 92)

Grade 4 0

Any grade

Grade 3

92 (100) 14 (15.2)

2 (11.1) 113 (94.2) 19 (15.8) 10 (8.3) 82 (89.1) 12 (13.0) 0

67 (55.8) 23 (19.2)

2 (1.7)

43 (46.7)

8 (8.7)

Grade 4 0 6 (6.5) 0

Hematologic abnormalities were derived from laboratory analysis, including complete blood count, neutrophil count, platelet count and hemoglobin values. d: dexamethasone; eGFR: estimated glomerular filtration rate; Isa: isatuximab; K: carfilzomib; RI: renal impairment. Haematologica | 107 - June 2022

1405

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

Discussion In MM patients, RI is considered a poor prognostic factor, being associated with earlier mortality and worse OS.5,22,36 As such, there is a critical need for anti-MM therapies that also improve renal function. This prespecified subgroup analysis of the phase III IKEMA study demonstrated that IsaKd is efficacious and has a manageable safety profile in patients with RI. The risk of disease progression or death in patients with RI was 73% lower in the Isa-Kd arm, indicated by the very low HR (HR: 0.27; 95% CI: 0.11–0.66). This result is consistent with the PFS benefit of Isa-Kd observed in the overall IKEMA study population (HR: 0.53; 99% CI: 0.32–0.89)28 and in patients without RI (HR: 0.63; 95% CI: 0.39–1.00). In IKEMA, median PFS observed among patients with RI (eGFR <60 mL/min/1.73 m²) receiving Kd (13.4 months) was similar to results of the Kd arm in the ENDEAVOR subgroup analysis, with a median PFS of 14.9 months in patients with creatinine clearance ≥15 to <50 mL/min (severe/moderate RI).9 Consistent with the PFS results, duration of study treatment exposure was similar in Isa-Kd patients with or without RI (81 and 78.6 months, respectively), whereas it was shorter in Kd patients with RI versus those without RI (35.7 vs. 68.5 months, respectively). The ORR was greater with Isa-Kd than Kd in patients with RI (93.1% vs. 61.1%), whereas patients without RI showed similar ORR in the two study arms (83.6% vs. 89.2%), consistent with the overall IKEMA study population (86.6% IsaKd vs. 82.9% Kd).28 Of note, depth of response was superior with Isa-Kd versus Kd independently of RI status, with respect to CR rate (41.9% Isa-Kd vs. 22.2% Kd in patients with RI; 40.2% Isa-Kd versus 30.1% Kd in patients without RI), ≥VGPR rate (79.1% Isa-Kd vs. 44.4% Kd in patients with RI; 71.3% Isa-Kd vs. 59.1% Kd in patients without RI), and MRD negativity rate (30.2% Isa-Kd vs. 11.1% Kd in patients with RI; 29.5% Isa-Kd vs. 14.0% Kd in patients without RI). Remarkably, compared with Kd, Isa-Kd increased the proportion of patients with RI who achieved both complete (52.0% Isa-Kd vs. 30.8% Kd) and durable (≥60 days; 32.0% Isa-Kd vs. 7.7% Kd) renal responses, and decreased time to first (1.5 months Isa-Kd vs. 6.5 months Kd) and to complete (7.8 months Isa-Kd vs. NC Kd) renal response, suggesting that Isa-Kd allows the achievement of sustainable reversal of RI. Similarly, compared with Kd, fewer patients in the IsaKd arm experienced worsening of renal function or progression to end-stage RI. The addition of Isa to Kd was associated with a manageable safety profile in MM patients with and without RI. Among patients with RI, there was a similar incidence of patients with grade ≥3 TEAE between the two arms (79.1% Isa-Kd vs. 77.8% Kd), whereas this incidence was higher in patients without RI (77.5% Isa-Kd vs. 65.2% Kd). Furthermore, in patients with RI there was a similar incidence of patients with

serious TEAE or TEAE leading to death during study treatment or treatment discontinuation. The higher treatment exposure observed with Isa-Kd versus Kd might have contributed to the higher incidence of grade ≥3 TEAE in patients without RI. Cardiac failure in the overall population was similar between study arms (7.3% all grades and 4.0% grade ≥3 in Isa-Kd versus 6.6% all grades and 4.1% grade ≥3 in Kd), but incidence of any-grade cardiac failure was higher in patients with RI in Isa-Kd (11.6% vs. 5.6%). This can be related to a longer treatment exposure in Isa-Kd (median number of cycles was 20 in Isa-Kd vs. 9 in Kd) and higher carfilzomib relative dose intensity in Isa-Kd (93.1% vs. 84.6%). This difference in incidence disappeared for grade ≥3 events. The most common TEAE in patients with RI treated with Isa-Kd versus Kd were diarrhea, upper respiratory tract infection, hypertension, and fatigue with similar frequency observed in the overall IKEMA population.28 There was no increased incidence of infusion reactions in the RI (37.2% Isa-Kd vs. 5.6% Kd) compared with non-RI (45.8% Isa-Kd vs. 3.3% Kd) subgroups. There are few reports in the literature analyzing the efficacy and toxicity of anti-CD38 monoclonal antibodies in patients with RI. The results of this IKEMA subgroup analysis reinforce the findings of the ICARIA-MM RI subgroup analysis, which showed that addition of Isa to Pd also improved clinical outcomes in patients with RI.27 Median PFS was 9.5 months with Isa-Pd versus 3.7 months with Pd (HR: 0.50; 95% CI: 0.30–0.85) for patients with RI. Isa-Pd also showed greater depth of response in patients with RI, with a 56% ORR with Isa-Pd versus 25% with Pd. Complete renal response rates were achieved in 23 of 32 (71.9%) patients treated with Isa-Pd and eight of 21 (38.1%) treated with Pd; these were durable in ten of 32 (31.3%) and four of 21 (19.0%) of patients treated with Isa-Pd versus Pd, respectively.27 Data about efficacy and safety of daratumumab, a different CD38 monoclonal antibody, in patients with RI are limited. A few, isolated case reports with single dialysis-dependent patients have been published.37-40 Results from a retrospective, multicenter, open-label study designed to evaluate safety and efficacy of daratumumab in RRMM patients with end-stage RI requiring hemodialysis (n=15) reported a median PFS of 8.7 months, OS of 12.2 months, and ORR of 40%.41 The most common grade 3–4 hematologic AE included thrombocytopenia (n=5), anemia (n=4), and neutropenia (n=4). Infusion reactions (n=6) were the most frequent non-hematologic AE.41 Results of an interim analysis of the phase II DARE study, a multicenter, single-arm, open-label study in RRMM patients with severe RI (eGFR <30 mL/min/1.73 m2) or in need of hemodialysis were reported recently.42 Eligible patients had received ≥2 prior treatment lines (including bortezomib- and lenalidomidebased regimens) and presented with ECOG PS score ≤2. At the cut-off date, 35 patients treated with daratumumab and dexamethasone showed a 12-month PFS probability of 50%,

Haematologica | 107 - June 2022

1406

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

an ORR of 45.7%, and a renal response rate of 17.1%. The most common grade 3–4 AE were anemia (17.1%) and hyperglycemia (8.6%). A total of 48.6% of patients had ≥1 grade 3–4 AE and 25.7% of patients experienced ≥1 serious AE.42 Limitations of this IKEMA subgroup analysis include: (i) less than 3% of patients in each arm had severe RI, so the results presented here are mainly applicable to patients with moderate RI, and (ii) RI status at baseline was not a stratification factor, likely resulting in small differences in subgroup size between treatment arms. In summary, addition of Isa to Kd improved PFS and depth of response in patients with relapsed MM and RI, with a manageable safety profile, consistent with the benefit observed in the overall IKEMA study population. More patients treated with Isa-Kd showed reversal of RI and durable renal responses compared with those treated with Kd. Based on these findings, Isa-Kd represents a valuable addition to the therapies used to treat patients with MM-related renal dysfunction. Disclosures MC is part of the speaker’s bureau of Amgen, Janssen, and Sanofi. TM has received research funding from Amgen, Janssen, and Sanofi; consults for GSK. PM has received honoraria from Amgen, Celgene, Janssen, Novartis, and Takeda; has a consulting or advisory role at Amgen, Celgene, Janssen, Novartis, and Takeda. RB has received research funding from AbbVie, Acerta Pharma, Alexion, Amgen, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Celgene, CSL Behring, Daiichi Sankyo, Janssen-Cilag, MorphoSys AG, Pfizer, Rigel Pharmaceuticals, Roche, Sanofi, and Takeda; has received honoraria from Bayer; has a consulting or advisory role at Janssen-Cilag, Roche; is part of the speaker’s bureau of Bayer. LP, C-KM, MRS, and MT have no conflicts of interest to disclose. XL has received honoraria from AbbVie, Amgen, Bristol Myers Squibb, Carsgen Therapeutics Ltd, Celgene, Gilead Sciences, Janssen-Cilag, Karyopharm Therapeutics, Merck, Mundipharma, Novartis, Oncopeptides, Pierre Fabre, Roche, Sanofi, and Takeda; has received non-financial support from Takeda. MM has received research funding from Adaptive, Amgen, Bristol Myers Squibb, Celgene, GlaxoSmithKline, Janssen, Jazz,

Novartis, Sanofi, Stemline Therapeutics, and Takeda; has received honoraria from Adaptive, Amgen, Bristol Myers Squibb, Celgene, GlaxoSmithKline, Janssen, Jazz, Novartis, Sanofi, Stemline Therapeutics, and Takeda; has received non-financial support from Takeda; has a consulting or advisory role at Adaptive, Amgen, Bristol Myers Squibb, Celgene, GlaxoSmithKline, Janssen, Jazz, Novartis, Sanofi, Stemline Therapeutics, and Takeda. RLB has a consulting or advisory role at Celgene/Bristol Myers Squibb, Janssen, Amgen, Sanofi, and Takeda; has received research funding from Celgene/Bristol Myers Squibb. M-LR and SS are employed by Sanofi; may hold stock and/or stock options in the company. LM has a consulting or advisory role at Aixial (contracted by Sanofi). MD has a consulting or advisory role at Amgen, Bristol Myers Squibb, Celgene, Janssen, and Takeda. Contributions IKEMA Study Steering Committee members (TM, PM, MD) and employees of Sanofi (M-LR, SS) contributed to the conception/design of this study. All authors contributed to the provision of study material, data collection and analysis, as well as development/final approval of the manuscript. Acknowledgments The IKEMA study was sponsored by Sanofi. We thank the participating patients and their caregivers, and the study centers and investigators for their contributions to the study. Medical writing support was provided by C. Semighini Grubor, PhD, and S. Mariani, MD, PhD, of Elevate Medical Affairs, contracted by Sanofi Genzyme, for publication support services. Data sharing statement Qualified researchers can request access to patient-level data and related study documents including the clinical study report, study protocol with any amendments, blank case report forms, statistical analysis plan, and dataset specifications. Patient-level data will be anonymized, and study documents will be redacted to protect the privacy of trial participants. Further details on Sanofi’s data-sharing criteria, eligible studies, and process for requesting access are at: https://www.clinicalstudydatarequest.com.

References 1. Clark AD, Shetty A, Soutar R. Renal failure and multiple myeloma: pathogenesis and treatment of renal failure and management of underlying myeloma. Blood Rev. 1999;13(2):7990. 2. Eleutherakis-Papaiakovou V, Bamias A, Gika D, et al. Renal failure in multiple myeloma: incidence, correlations, and prognostic significance. Leuk Lymphoma. 2007;48(2):337-341. 3. San-Miguel JF, Richardson PG, Sonneveld P, et al. Efficacy and safety of bortezomib in patients with renal impairment: results

from the APEX phase 3 study. Leukemia. 2008;22(4):842-849. 4. Yadav P, Cook M, Cockwell P. Current trends of renal impairment in multiple myeloma. Kidney Dis (Basel). 2016;1(4):241-257. 5. Fotiou D, Dimopoulos MA, Kastritis E. Managing renal complications in multiple myeloma. Expert Rev Hematol. 2016;9(9):839-850. 6. Dimopoulos MA, Terpos E, Chanan-Khan A, et al. Renal impairment in patients with multiple myeloma: a consensus statement on behalf of the International Myeloma Working

Haematologica | 107 - June 2022

1407

ARTICLE - IKEMA renal impairment subgroup

M. Capra et al.

Group. J Clin Oncol. 2010;28(33):4976-4984. 7. Dimopoulos MA, Sonneveld P, Leung N, et al. International Myeloma Working Group recommendations for the diagnosis and management of myeloma-related renal impairment. J Clin Oncol. 2016;34(13):1544-1557. 8. Dimopoulos MA, Moreau P, Palumbo A, et al. Carfilzomib and dexamethasone versus bortezomib and dexamethasone for patients with relapsed or refractory multiple myeloma (ENDEAVOR): a randomised, phase 3, open-label, multicentre study. Lancet Oncol. 2016;17(1):27-38. 9. Dimopoulos M, Siegel D, White DJ, et al. Carfilzomib vs bortezomib in patients with multiple myeloma and renal failure: a subgroup analysis of ENDEAVOR. Blood. 2019;133(2):147-155. 10. Dimopoulos MA, Richardson PG, Schlag R, et al. VMP (bortezomib, melphalan, and prednisone) is active and well tolerated in newly diagnosed patients with multiple myeloma with moderately impaired renal function, and results in reversal of renal impairment: cohort analysis of the phase III VISTA study. J Clin Oncol. 2009;27(36):6086-6093. 11. Dimopoulos MA, Roussou M, Gavriatopoulou M, et al. Reversibility of renal impairment in patients with multiple myeloma treated with bortezomib-based regimens: identification of predictive factors. Clin Lymphoma Myeloma. 2009;9(4):302-306. 12. Dimopoulos MA, Roussou M, Gavriatopoulou M, et al. Bortezomib-based triplets are associated with a high probability of dialysis independence and rapid renal recovery in newly diagnosed myeloma patients with severe renal failure or those requiring dialysis. Am J Hematol. 2016;91(5):499-502. 13. Weisel KC, Dimopoulos MA, Moreau P, et al. Analysis of renal impairment in MM-003, a phase III study of pomalidomide + low-dose dexamethasone versus high-dose dexamethasone in refractory or relapsed and refractory multiple myeloma. Haematologica. 2016;101(7):872-878. 14. Gavriatopoulou M, Terpos E, Dimopoulos MA. IMiDs for myeloma induced renal impairment. Oncotarget. 2018;9(84):35476-35477. 15. Dimopoulos M, Weisel K, van de Donk N, et al. Pomalidomide plus low-dose dexamethasone in patients with relapsed/refractory multiple myeloma and renal impairment: results from a Phase II trial. J Clin Oncol. 2018;36(20):20352043. 16. Siegel DS, Weisel KC, Dimopoulos MA, et al. Pomalidomide plus low-dose dexamethasone in patients with relapsed/refractory multiple myeloma and moderate renal impairment: a pooled analysis of three clinical trials. Leuk Lymphoma. 2016;57(12):2833-2838. 17. Dimopoulos M, Alegre A, Stadtmauer EA, et al. The efficacy and safety of lenalidomide plus dexamethasone in relapsed and/or refractory multiple myeloma patients with impaired renal function. Cancer. 2010;116(16):3807-3814. 18. Dimopoulos MA, Cheung MC, Roussel M, et al. Impact of renal impairment on outcomes with lenalidomide and dexamethasone treatment in the FIRST trial, a randomized, open-label phase 3 trial in transplant-ineligible patients with multiple myeloma. Haematologica. 2016;101(3):363-370. 19. Dimopoulos MA, Christoulas D, Roussou M, et al. Lenalidomide and dexamethasone for the treatment of refractory/relapsed multiple myeloma: dosing of lenalidomide according to renal function and effect on renal impairment. Eur J Haematol. 2010;85(1):1-5. 20. Dimopoulos MA, Terpos E, Goldschmidt H, Alegre A, Mark T, Niesvizky R. Treatment with lenalidomide and dexamethasone in patients with multiple myeloma and renal impairment. Cancer Treat Rev. 2012;38(8):1012-1019.

21. Dimopoulos MA, Roussou M, Gkotzamanidou M, et al. The role of novel agents on the reversibility of renal impairment in newly diagnosed symptomatic patients with multiple myeloma. Leukemia. 2013;27(2):423-429. 22. Park S, Han B, Kim K, et al. Renal insufficiency in newlydiagnosed multiple myeloma: analysis according to International Myeloma Working Group consensus statement. Anticancer Res. 2014;34(8):4299-4306. 23. Kyprolis® (carfilzomib) for injection, for intravenous use [prescribing information], Amgen, Thousand Oaks, CA (March 2021). 24. Sarclisa® (isatuximab-irfc) injection, for intravenous use [prescribing information], Sanofi, Bridgewater, NJ (March 2021). 25. European Medicines Agency. Sarclisa, INN-Ixatuximab. Summary of product characteristics. 2021. https://www.ema.europa.eu/en/documents/productinformation/sarclisa-epar-product-information_en.pdf. Accessed May 13, 2021. 26. Attal M, Richardson PG, Rajkumar SV, San-Miguel J, Beksac M, Spicka I; ICARIA-MM study group. Isatuximab plus pomalidomide and low-dose dexamethasone versus pomalidomide and low-dose dexamethasone in patients with relapsed and refractory multiple myeloma (ICARIA-MM): a randomised, multicentre, open-label, phase 3 study. Lancet. 2019;394(10214):2096-2107. 27. Dimopoulos MA, Leleu X, Moreau P, et al. Isatuximab plus pomalidomide and dexamethasone in relapsed/refractory multiple myeloma patients with renal impairment: ICARIA-MM subgroup analysis. Leukemia. 2021;35(2):562-572. 28. Moreau P, Dimopoulos M-A, Mikhael J, et al. Isatuximab, carfilzomib, and dexamethasone in relapsed multiple myeloma (IKEMA): a multicentre, open-label, randomised phase 3 trial. Lancet. 2021;397(10292):2361-2371. 29. Moreau P, Dimopoulos MA, Yong K, et al. Isatuximab plus carfilzomib/dexamethasone versus carfilzomib/dexamethasone in patients with relapsed/refractory multiple myeloma: IKEMA Phase III study design. Future Oncol. 2020;16(2):4347-4358. 30. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130(6):461-470. 31. Rajkumar SV, Harousseau JL, Durie B, et al. Consensus recommendations for the uniform reporting of clinical trials: report of the International Myeloma Workshop Consensus Panel 1. Blood. 2011;117(18):4691-4695. 32. Kumar S, Paiva B, Anderson KC, et al. International Myeloma Working Group consensus criteria for response and minimal residual disease assessment in multiple myeloma. Lancet Oncol. 2016;17(8):e328-346. 33. Lahuerta JJ, Paiva B, Vidriales MB, et al. Depth of response in multiple myeloma: a pooled analysis of three PETHEMA/GEM clinical trials. J Clin Oncol. 2017;35(25):2900-2910. 34. Paiva B, Puig N, Cedena MT, et al. Measurable residual disease by next-generation flow cytometry in multiple myeloma. J Clin Oncol. 2020;38(8):784-792. 35. Dimopoulos MA, Roussou M, Gavriatopoulou M, et al. Cardiac and renal complications of carfilzomib in patients with multiple myeloma. Blood Adv. 2017;1(7):449-454. 36. Augustson BM, Begum G, Dunn JA, et al. Early mortality after diagnosis of multiple myeloma: analysis of patients entered onto the United kingdom Medical Research Council trials between 1980 and 2002--Medical Research Council Adult Leukaemia Working Party. J Clin Oncol. 2005;23(36):9219-9226.

Haematologica | 107 - June 2022

1408

ARTICLE - IKEMA renal impairment subgroup

L. Li et al.

37. Rocchi S, Tacchetti P, Pantani L, et al. Safety and efficacy of daratumumab in dialysis-dependent renal failure secondary to multiple myeloma. Haematologica. 2018;103(6):e277-278. 38. Smyth E, Glavey S, Melotti D, et al. Dialysis independence following single-agent daratumumab in refractory myeloma with renal failure. Ir J Med Sci. 2019;188(3):1079-1080. 39. Moore DC, Arnall JR, Janes A, Pineda-Roman M. Dialysis independence following combination daratumumab, thalidomide, bortezomib, cyclophosphamide, and dexamethasone in multiple myeloma with severe renal failure. Clin Lymphoma Myeloma Leuk. 2020;20(7):e395-398.

40. Jeyaraman P, Bhasin A, Dayal N, Pathak S, Naithani R. Daratumumab in dialysis-dependent multiple myeloma. Blood Res. 2020;55(1):65-67. 41. Cejalvo MJ, Legarda M, Abella E, et al. Single-agent daratumumab in patients with relapsed and refractory multiple myeloma requiring dialysis: results of a Spanish retrospective, multicentre study. Br J Haematol. 2020;190(5):e289-292. 42. Kastritis E, Terpos E, Symeonidis A, et al. Daratumumab with dexamethasone in patients with relapsed/refractory multiple myeloma and severe renal impairment: results on efficacy and safety of the Phase 2 Dare study. Blood. 2020;136(Suppl 1):S48-49.

Haematologica | 107 - June 2022

1409

ARTICLE - Plasma Cell Disorders

Apoptosis reprogramming triggered by splicing inhibitors sensitizes multiple myeloma cells to Venetoclax treatment Debora Soncini,1 Claudia Martinuzzi,1,2 Pamela Becherini,1,2 Elisa Gelli,1 Samantha Ruberti,1 Katia Todoerti,3 Luca Mastracci,2,4 Paola Contini,5 Antonia Cagnetta,2 Antonella Laudisi,1,2 Fabio Guolo,1,2 Paola Minetto,2 Maurizio Miglino,1,2 Sara Aquino,6 Riccardo Varaldo,6 Daniele Reverberi,7 Matteo Formica,2,8 Mario Passalacqua,9 Alessio Nencioni,2,5 Antonino Neri,3,10 Mehmet K. Samur,11 Nikhil C. Munshi,11 Mariateresa Fulciniti,11 Roberto M. Lemoli,1,2# and Michele Cea.1,2# Clinic of Hematology, Department of Internal Medicine (DiMI), University of Genoa, Genoa, Italy; 2IRCCS Ospedale Policlinico San Martino, Genoa, Italy; 3Hematology, Fondazione Cà Granda IRCCS Policlinico, Milan, Italy; 4Department of Integrated Surgical and Diagnostic Sciences, University of Genoa, Genoa, Italy; 5Department of Internal Medicine (DiMI), University of Genoa, Genoa, Italy; 6Hematology and Hematopoietic Stem Cell Transplantation Unit, IRCCS Ospedale Policlinico San Martino, Genoa, Italy; 7U.O. Molecular Pathology, IRCCS Ospedale Policlinico San Martino, Genoa, Italy; 8Department of Surgical Sciences and Integrated Diagnostic (DISC), University of Genoa, Genoa, Italy; 9Department of Experimental Medicine, University of Genoa, Genoa, Italy; 10Department of Oncology and Hemato-oncology, University of Milan, Milan, Italy and 11Jerome Lipper Multiple Myeloma Center, Department of Medical Oncology, Dana Farber Cancer Institute, Harvard Medical School, Boston, MA, USA. 1

Correspondence: Michele Cea michele.cea@unige.it Received: May 21, 2021. Accepted: October 8, 2021. Prepublished: October 21, 2021. https://doi.org/10.3324/haematol.2021.279276 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

RML and MC contributed equally as co-senior authors.

Abstract Identification of novel vulnerabilities in the context of therapeutic resistance is emerging as a key challenge for cancer treatment. Recent studies have detected pervasive aberrant splicing in cancer cells, supporting its targeting for novel therapeutic strategies. Here, we evaluated the expression of several spliceosome machinery components in multiple myeloma (MM) cells and the impact of splicing modulation on tumor cell growth and viability. A comprehensive gene expression analysis confirmed the reported deregulation of spliceosome machinery components in MM cells, compared to normal plasma cells from healthy donors, with its pharmacological and genetic modulation resulting in impaired growth and survival of MM cell lines and patient-derived malignant plasma cells. Consistent with this, transcriptomic analysis revealed deregulation of BCL2 family members, including decrease of anti-apoptotic long form of myeloid cell leukemia-1 (MCL1) expression, as crucial for “priming” MM cells for Venetoclax activity in vitro and in vivo, irrespective of t(11;14) status. Overall, our data provide a rationale for supporting the clinical use of splicing modulators as a strategy to reprogram apoptotic dependencies and make all MM patients more vulnerable to BCL2 inhibitors.

Introduction Multiple myeloma (MM) is a clonal B-cell malignancy characterized by excessive bone marrow plasma cells in association with monoclonal protein.1,2 MM is a heterogeneous disease driven by a large repertoire of molecular abnormalities, which contribute to its diverse clinical behavior. Currently available anti-MM therapies have remarkably improved patient outcome, but resistance is emerging as one of the foremost challenges in the clinical

management of this tumor. Therefore, there is an unmet medical need to define biologic mechanisms of drug resistance, both to enhance efficacy of existing treatments and to facilitate the design of novel strategies.3 A fundamental feature of MM is its striking genomic instability leading to cancer development and clonal evolution.4,5 Consequently, the majority of studies have focused on changes in DNA: unfortunately, the greatest shortcoming of DNA-based approaches is their failure to capture the panoply of RNA editing events. Indeed, the genetic code

Haematologica | 107 - June 2022

1410

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax is translated in the cell through the production of a messenger, mRNA, which usually has the same code of the DNA. A percentage of mRNA are “edited”, so that although they are made as a faithful copy of the DNA, they are modified afterward to change the products of the genetic code. As a result, pre-mRNA processing by alternative splicing (AS) and/or RNA-specific deaminases, markedly increases the complexity of the human transcriptome.6 This process, which is executed in the nucleus by spliceosomes, occurs overwhelmingly in both normal and transformed cells, thus it is the rule rather than the exception. If this process is compromised, the resulting changes in splicing can lead to neoplastic transformation. Indeed, recent studies provide evidence that an abnormally expressed splicing factor machinery as well as its mutations can have oncogenic properties by impacting AS of genes associated with susceptibility and/or progression of cancer.7 Despite these intriguing findings, the global pattern of RNA processing in human cancer genomes has not been systematically characterized, and its functional importance and clinical relevance in cancer remains largely unknown. Thus, RNA splicing deregulation represents an innovative and exciting area of research, in that it might be possible to modify or regulate RNA as novel anticancer strategy. AS is controlled by spliceosome, which is a dynamic molecular machine consisting of small nuclear RNA (snRNA) and various protein complexes that cycle on and off from pre-mRNA during intronic splicing. This nuclear complex is composed of at least 170 proteins and five snRNA associated with proteins forming the U1, U2, U4, U5, and U6 small nuclear ribonucleoproteins (snRNP). It removes an intron from the primary transcript and subsequently joins the exons by a trans-esterification reaction; the intron then undergoes debranching and is subsequently degraded.8 The accuracy of this process is essential for normal cellular function, whereas alternative splicing deregulation often occurs in tumors. As a result, modulation of this tumor hallmark is now emerging as a promising strategy for anti-cancer therapies.9,10 Changes in alternative splicing are frequently caused by point mutations in the splicing factors such as SF3B1, which occurs in a wide range of tumors, including chronic lymphocytic leukemia (CLL), myelodysplastic syndromes (MDS), melanoma and breast cancers, resulting in aberrant splicing.11– 13 Also, MM cells harbor somatic alterations in this driver gene with missense mutations (K700 and K666) observed in about 1.7% of patients.14 Importantly, despite low mutational burden, recent studies suggest RNA splicing deregulation as a driver mechanism for disease progression and drug-resistance occurrence as well.15,16 Here, we explored the impact of small molecule modulators of the spliceosome in perturbing tumor cells survival pathways and provide evidence for a novel combination strategy to treat MM.

D. Soncini et al.

Methods For a more detailed description of the methods used, see the Online Supplementary Appendix. Reagents Meayamycin B and Sudemycin D6 (SD6) were kind gifts from Prof. Kazunori Koide (University of Pittsburgh) and Prof. Thomas Webb (SRI Biosciences, Dept. of Chemistry and Biochemistry UCSC), respectively. E7107 was kindly provided by H3 Biomedicine, Inc. Cell lines and primary tumor specimens Cell lines were obtained from the ATCC or sources indicated in the Online Supplementary Appendix. Intracellular BH3 profiling Intracellular BH3 (iBH3) profiling was performed as described in 17. Briefly, MM cells were pelleted and suspended in MEB buffer with addition of each BH3 peptide treatment with 0.002% w/v digitonin (MS1, mBAD and HRKy peptides were used). Mitochondria in the permeabilized cells were exposed to peptides for 45 minutes (min) at 26°C before fixation with 4% formaldehyde at room temperature for 10 min. After addition of N2 neutralizing buffer for 15 min, cells were stained with Alexa Fluor® 488 Mouse anti-Cytochrome c (Clone 6H2.B4, BD Biosciences) 1:40 in 2% Tween20, 10% bovine serum albumin phosphate-bufferd saline for 2 hours at room temperature and then over night at 4°C. The quantification of Cytochrome c loss induced by each peptide was evaluated by flow cytometry (Navios 10/3, Beckman Coulter). Values indicate the percentage of Cytochrome c-negative cells calculated as in 17. Transcriptome profiling, alternative splicing and pathway analysis RNA samples from H929 cells treated or not with Meayamycin B (3 nM, 8 hours) were isolated in duplicate for each condition with the miRNeasy mini kit (Qiagen, #1038703) and processed using WT PLUS Reagent Kit, according to the manifacturer’s protocol (Thermo Fisher). Wide mRNA-transcriptome profiling was assessed using ClariomD Human array (Thermo Fisher). Robust Multi Array (RMA) normalization on raw data, transcript annotation (Clariom_D_Human.r1.na36.hg38.a1.transcript) and AS analysis were performed using Transcriptome Analysis Console (TAC 4.0) software (Thermo Fisher). AS events were identified in Meayamycin B-treated versus untreated H929 cells after filtering for transcripts expressed in both conditions, exon expressed in at least one condition, Exon Splicing Index of at least 2 (<-2 or >2), exon P-value <0.05. Pathway enrichment analysis of significant alternative spliced genes was performed with the fgsea package in R to-

Haematologica | 107 - June 2022

1411

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

gether with BIOCARTA and REACTOME datasets. Significant pathways were selected based on false discovery rate (FDR) q-value <25%. Global transcript expression levels were submitted on gene expression omnibus (GEO) data repository, under GEO accession number GSE167042.

association with tumor progression route (Online Supplementary Figures S1 and S2). We confirmed higher expression in MM cells by gene expression analysis in primary cells from seven additional MM patients and peripheral blood mononuclear cells (PBMC) from two healthy donors (Figure 1A). We next extended the analysis to the MMRF (MM Research Foundation) CoMMpass study that includes 774 MM patients5,22 profiled by RNA sequencing by standardizing KEGG spliceosome signature with a zscore method,5 which stratified MM cases into three groups: low (n=303), intermediate (n=126) and high (n=345) (Figure 1B). Then we sought to explore any potential impact of aberrant spliceosome signature on clinical outcome and found that, in line with reported data, 16 patients in the high z-score group showed shorter progression-free survival (PFS) and overall survival (OS) when compared with low-z score patients (median PFS 1,300 days vs. 700 days, P=0.0017; median survival not-reached vs. 2,000 days, P=0.00027) (Figure 1C). Importantly, a univariate Cox-model analysis confirmed a spliceosome signature role in predicting OS (P=0.0012) and PFS (P=0.0021), similarly to ISS stage III (P=0.0001), 1q gain/amp (P<0.0001), del13q/RB1 (P=0.0002), and TP53 mutation occurrence (P<0.0001) (Figure 1D; Online Supplementary Figure S3). Overall, our transcriptome analyses revealed a group of MM patients with higher upregulation in the RNA splicing machinery genes which correlated with poor prognosis. The spliceosome is a dynamic molecular machine consisting of several nuclear protein complexes, with the splicing factor SF3B1 responsible for on/off cycle of pre-mRNA during intronic splicing, essential for the correct functioning of this machinery. As result, western blot, immunohistochemistry and immunofluorescence analyses revealed high levels of SF3B1 on MM cell lines and CD138+ from patient-derived bone marrow biopsies (Figure 1E and F; Online Supplementary Figure S4) which was further increased upon interaction with bone marrow stromal cells (BMSC) (Online Supplementary Figure S5). Notably, the greater SF3B1 protein level observed in tumor cells could be explained, at least in part, by the higher proliferation rate (in term of S and G2/M phases residing cells) of MM compared with healthy donor counterpart cells. (Online Supplementary Figure S6). Altogether these data suggest a role for spliceosome deregulation in MM growth.

In vivo mouse models All in vivo experiments were performed in accordance with the laws and institutional guidelines for animal care, approved by the Institutional Animal Care and Use Committee of University Hospital San Martino (protocol #473). Five-week-old female NOD/SCID J mice were acquired from Charles Rivers Laboratories (France) and were acclimatized for 3 weeks. 4,5x106 MM1S cells were injected subcutaneously in both flanks of each mouse. Treatment was initiated when the tumors reached a volume of about 50 mm3. In each experiment, mice were randomly divided to one of the following groups: control (vehicle); Venetoclax 100 mg/kg/day, dissolved in vehicle (60% phosal 50 PG, 30% polyethylene glycol 400, 10% ethanol) and administered by oral gavage; SD6 12 mg/kg/day, formulated in vehicle (10% hydroxypropyl-beta-cyclodextrin [HPCD] in phosphate buffer pH 7.4) and intra-tumorally injected; E7107 2,5 mg/kg/day prepared in vehicle (10% ethanol, 5% Tween-80, Quorum sensing with saline) and intravenously injected (i.v.). Schedules are described in the individual figure legends. Tumor volume was calculated using the formula: tumor volume= (w2 × W) x π/6, where “w” and “W” are “minor side” and “major side” (in mm), respectively. Mice were sacrificed when the tumor reached a volume of about 1,5 cm3. Tumor masses were always isolated at the end of the experiment, weighted and fixed in 10% neutral buffered formalin (v/v) for histology.

Results Splicing machinery is markedly deregulated in multiple myeloma and represents a disease aggressiveness biomarker Genome-wide studies have recently demonstrated that deregulated expression of genes involved in splicing acts as driver event for numerous tumors including MM.7,18 These data have prompted a growing interest for drugging the spliceosome machinery as a novel strategy to improve anti-cancer therapies.14–16 We explored transcription levels of core snRNP assembly genes in primary MM cells from publicly-available collections of gene expression datasets by applying the KEGG spliceosome signature.14,19–21 A progressive transcriptional increase was observed in more advanced disease phases from monoclonal gammopathy of undetermined significance (MGUS) to plasma cell leukemia (PCL) and in human multiple myeloma cell line (HMCL) compared to normal plasma cells, indicating an

Targeting the spliceosome core-element SF3B1 results in broad anti-multiple myeloma activity Comprehensive RNA sequencing-based analyses have revealed profound and significant transcriptome changes, including alternative pre-mRNA splicing between MM and normal samples, with significant impact on overall clinical outcome.14 Despite that, SF3B1 somatic mutations, resulting in loss of function, occur in only 1.7% of MM patients, thus

Haematologica | 107 - June 2022

1412

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

suggesting additional events accounting for the alterations observed in MM cells.15 Therefore, we evaluated the spectrum of alternatively spliced events (ASE) in MMRF database, irrespective of SF3B1 mutation status. We classified the ASE into five categories and analyzed the frequency of pattern changes according to SF3B1 expression level (top vs. bottom quartile) (Figure 2A). Wide variability was observed, with alternative transcription start site (ATSS), intron retentions (IR) and alternative transcription termination site (ATTS) as the most abundant ASE among the SF3B1 highly expressed group, indicating widespread spliceosome deregulation in MM, regardless of specific mutational profiling. Based on these data, we next investigated SF3B1 role in these cells

by genetic perturbation in a panel of MM cell lines. According to the reported activity of this factor in binding and splicing pre-mRNA,9 immunofluorescence analysis for the SR (serine/arginine-rich) protein SC-35 revealed a significant modulation of splicing machinery after SF3B1 knockdown (KD), with KD cells exhibiting reduced number of speckles which resulted larger and darker than control (Figure 2B). Importantly, SF3B1 KD not only affected splicing but also impaired MM cell viability by increasing apoptotic cell death in different MM cell lines and inducing caspase-3 and PARP1 cleavage as well (Figure 2C; Online Supplementary Figure S7). We subsequently tested the activity of the splicing modulator Meayamycin B in a panel of 14 MM cell lines,

Continued on following page. Haematologica | 107 - June 2022

1413

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

Figure 1. Splicing machinery is markedly deregulated in multiple myeloma. A) Heat map showing expression levels (row z-score) of indicated genes among plasma cells derived from patients with multiple myeloma (MM) (n=7) and peripheral blood mononuclear cells (PBMC) from healthy individuals (n=2). The color scale spans the relative gene expression changes standardized on the variance. B) Heatmap of 774 MM patients included in CoMMpass study ordered in 3 groups according to z-score calculated on 124 genes of KEGG spliceosome gene set, as annotated in MSigDB. C) Kaplan–Meier survival curves of high z-score (red) and low z-score (blue) MM patients of CoMMpass cohort (642 patients analyzed) on progression-free survival (PFS) and overall survival (OS) data. Log-rank test P-value and number of samples at risk in each group across time are reported. D) Forest plot based on Cox univariate analysis for OS. Squares represent hazard ratios; bars represent 95% confidence intervals. E) Protein lysates from a panel of MM cell lines (human multiple myeloma cell line [HMCL]) and primary MM patients or healthy donors (HD) were analyzed for SF3B1 expression by western blotting. GAPDH was used as loading control. One representative experiment is shown. F) Representative images of SF3B1 and CD138 immunocytochemistry stain in bone marrow (BM) from MM patients (n=5). Conventional Giemsa staining is also shown. Original magnification × 200, scale bar, 100 mm.

where we observed a significant decrease of cell viability with half maximal inhibitory concentration (IC50) values ranging from to 0.49 to 3.5 nM (Figure 2D and E). Similar results were observed with additional splicing modulators such as Sudemycin D6 and E7107 (Online Supplementary Table S1).23– 25 The effect of Meayamycin B was next tested in primary MM cells, cultured in the absence or presence of the bone marrow microenvironment. As seen in Figure 2F and G and the Online Supplementary Figure S8, primary MM cells were significantly depleted after treatment with Meayamycin B, while a minimal toxicity was observed on normal components of the bone marrow milieu. These data support SF3B1 as a promising therapeutic target in MM.

Spliceosome deregulation affects genome stability of multiple myeloma cells, irrespective of Myc status Genomic instability is a tumor hallmark resulting from deregulated DNA damage response, DNA repair defects, and failure of cell-cycle checkpoints. Recent studies have revealed that alternative splicing reprogramming after DNA damage responses relies on regulation of RNA-binding proteins (RBP), which directly bind specific pre-mRNA and mRNA sequences and act as gatekeepers of genomic integrity. In this context, post-transcriptional RNA processing may add another layer of complexity to the maintenance of genomic stability in MM cells and targeting spliceosome during DNA damage response may rep-

Haematologica | 107 - June 2022

1414

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

resent an innovative approach to sensitize tumor cells to genotoxic agents.26 As shown in Figure 3A and B, we observed that in SF3B1 KD MM cells increased DNA doublestrand breaks and resulted in defective repair mechanisms as shown by accumulation of gH2AX and reduced RAD51 levels, respectively. These changes were similarly observed after drug-treatment with Meayamycin B in a dose- and time-dependent manner (Figure 3C and D) and persist also after pre-incubation with pan-caspase inhibitor zVAD-fmk, thus suggesting a negligible role of apoptosis activation in the genomic instability triggered by spliceosome modulators (data not shown). Based on these data, we tested sensitivity of KD-SF3B1 MM cells to melphalan-induced DNA damage and, in line with previous studies,26 we found that drug exposure led to increased gH2AX induction, also in SF3B1-depleted cells, (data not shown) suggesting that huge spliceosome de-

regulation, triggered by SF3B1 targeting, accounts for the enhanced anti-MM activity of genotoxic stress. Recent genome-wide MYC-synthetic lethal screens have identified spliceosome components as candidate genes for synthetic lethality strategies, suggesting these as exploitable vulnerabilities for MYC-driven cancers.27 In order to determine whether MYC presence correlates with Meayamycin B efficacy, we compared its protein level with the specific IC50 value measured for each tested cell line. As shown in Figure 3E no significant correlation (R2=0.1261, P=0.348) was observed thus suggesting that Meayamycin B acts regardless of MYC protein level. Similarly, the ectopic expression of MYC in U266 MM cell lines increased g-H2AX levels (Figure 3F) but slightly enhanced the anti-MM activity of chemical or genetic SF3B1 targeting (Figure 3G and H), thus confirming the marginal role played by this oncogenic program on anti-MM activity of splicing modulators.

Figure 2. Continued on following page.

Haematologica | 107 - June 2022

1415

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

Figure 2. Spliceosome core-element SF3B1 targeting results in anti-multiple myeloma activity. A) Percentage of indicated splicing categories across RNA sequencing data derived from CoMMpass study according to SF3B1 expression levels (top vs. bottom quartile). B) Immunofluorescence staining for SC-35 in scramble and short hairpin (sh) SF3B1-sh RNA H929 multiple myeloma (MM) cells. SF3B1 silencing was validated by western blot as shown in the panel below. SC-35 staining of nuclear speckles is shown: mean fluorescence intensity (MFI) per cell/nucleus of the specific signal was quantified by counting at least 50 nuclei per condition as reported in the histogram below (**P=0.0042, two-sided Student t-test). Scale bar, 10 µm. C) MTS assay of MM1S (top) and H929 (bottom) lentivirally transduced with shSF3B1 (#1,#3 and #4) or sh scramble. Cell viability was measured at indicated time point after transduction. Western blot analyses were performed at day 3, confirming decreased SF3B1 protein levels and apoptotic cell death features (PARP1 and caspase 3 cleavage). Data are representative of at least 3 independent experiments. D) Cell viability curves compare a panel of 14 MM cell lines’ sensitivity to Meaymicin B (nM) for 48 h (n=3 technical replicates; mean +/− standard deviation [SD]) E) Evaluation of PARP, caspase 3, MCL1 and GAPDH by western blot on indicated MM cell lines treated with increasing doses of MeaymicinB for 24 hours. F) Treatment of primary bone marrow aspirate samples from MM patients (n=5) at various doses of MeaymicinB (0.1-30 nM) for 48 hours shows significant cytotoxicity of CD138+ tumor cells (n=3 technical replicates; mean +/− SD). G) Ex vivo evaluation of Meaymicin B in total bone marrow cells from one representative MM patient. After red cell lysis, cells were stained with Annexin V, DAPI and CD38 monoclonal antibody to identify viable as well as apoptotic myeloma (CD38 positive/CD45 negative) and normal (CD38 negative/CD45 negative) cells.

Splicing inhibition remodels mitochondrial apoptotic dependencies in multiple myeloma cells In order to shed light on the molecular mechanisms mediating the effects of splicing modulators on MM cell growth, we evaluated the transcriptomic profiles in MM cells treated with Meayamycin B. As shown in Figure 4A, we observed increased AS events compared to untreated cells, with intron retention (IR) and cassette exon (CE) as primary events across multiple targets (41.97% and

28.26%, respectively). Splicing outliers including alternative 5’ and 3’ donor site as well as complex events were observed less frequently. A comprehensive enrichment analysis of BioCarta annotated pathways was employed to identify cellular pathways affected by this treatment. Significant enrichment was observed among biological processes involved in regulation of cell-cycle, disease progression, cell division and DNA replication, with apoptosis-related pathways identified as top-ranked events

Haematologica | 107 - June 2022

1416

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

(Figure 4B). A gene level investigation, based on exon splicing index, revealed eleven genes of apoptosis pathway significantly mis-spliced: CHK1, BAX and BIRC3 were found to be interested by intron retention events, CASP10, CASP8 and CASP3 as affected by alternative cassette exons inclusion, PARP1 and MCL1 among targets with complex events and alternative 5’ donor site, respectively (Figure 4C). Moreover, we confirmed a significant change in MCL1 transcripts with depletion of the anti-apoptotic long isoform (MCL-1L), and accumulation of pro-apoptotic short isoform (MCL-1s) following exposure to both Meaymicin B or Sudemycin D6 (Online Supplementary Figure S9). Other BCL2 family genes, including

BCL2 and BCLxL, were not affected (Figure 4D). A similar MCL-1 exon 2-skipped transcript occurred in SF3B1 silenced cells, with no modifications on BCL2 and BCLxL as well (Figure 4E); these changes were also confirmed at protein level (Figure 4F). Based on these data, we performed the intracellular BH3 (iBH3) profiling to investigate the mitochondrial apoptotic dependencies triggered by Meayamycin B or Sudemycin D6.28 As shown in Figure 4G both inhibitors resulted in significant increase of Cytochrome c release following exposure to mBAD and HRKy BH3 peptides, indicating a functional shift in mitochondrial dependencies from MCL1 towards BLC2/BCLxL anti apoptotic members.

Figure 3. Continued on following page.

Haematologica | 107 - June 2022

1417

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

Figure 3. Spliceosome deregulation affects genome stability of multiple myeloma cells irrespective of Myc status. A) Detection of gH2AX and Q-nuclear was measured by confocal microscopy in H929 cells expressing short hairpin RNA (shRNA) (clone #3 and #4) targeting SF3B1 or control. Each panel includes representative foci-containing cells graph, over 3 experiments. (**P=0.004, ***P<0.001; two-sided Student t-test). B) H929 cells were engineered to express an anti-SF3B1 shRNA (3 clones). Next, SF3B1, RAD51 and gH2AX protein levels were detected by immunoblotting. C) Detection of gH2AX and Q-nuclear by confocal microscopy of H929 cells ex-cultured with or without increased dosed of Meayamycin B (1-3 mM) for 24 hours. Each panel includes representative foci-containing cells graph, over 3 experiments (*P=0.01, ****P<0.0001; two-sided Student t-test). (A and C) scale bar, 50 mm. D) Western blot analysis of DNA damage response markers (RAD51 and gH2AX) after Meaymicin B treatment over a range of doses (upper panel) and timing (3nM) (lower panel) in H929 cells. E) Relative expression of Myc protein plotted vs. Meaymicin B cytotoxicity half maximal inhibitory concentration (IC50) values. The Pearson correlation coefficient (r) and the P-value, calculated using GraphPad Prism Version 5 analysis software, are indicated; F) immunoblot for cMyc, SF3B1 and gH2AX protein levels in isogenic U266 cells (pLV empty) or cMyc overexpressing (pLV cMyc) cells; G) these cells were treated with growing doses of Meayamycin B for 48 hours. Cell viability was measured with MTS assay and presented as a percentage of control. H) Cell viability analysis of pLV empty (pLV cMyc) U266 cells transduced with short hairpin RNA (shRNA) clones containing the target sequence of SF3B1 (clone#1) or scrambled control. (**P=0.004; two-sided Student t-test).

Splicing modulations increases sensitivity of multiple myeloma cells to the BCL2 inhibitor Venetoclax Cancer cells are dependent on multiple anti-apoptotic proteins to promote their survival, therefore targeting more than one protein results in synergistic effects in several preclinical models of solid and hematological tumors.29–32 In such a scenario, the dysregulated expression of BCL2 family members observed after splicing modulator treatment, prompted us to harness apoptosis as a strategy for enhancing anti-MM activity of these agents. We therefore evaluated the ability of splicing modulators to enhance effects of the BCL2 specific inhibitor Venetoclax against MM cell growth and survival. Low dose of this

BH3-mimetic and Meayamycin B or Sudemycin D6 combo was synergistic in several MM cells, including those carrying t(11;14), which has been reported to affect anti-MM activity of Venetoclax. (Figure 5A and B; Online Supplementary Figure S10-11).33 As result, PARP1 and caspase 3 cleavage occurred after these stimuli at much extent compared with single agent exposure (Figure 5C). These results were also confirmed in SF3B1 KD MM cells (Figure 5D). We next examined the effects of the combination therapy on primary plasma cells (PC) collected from MM patients: as shown in Figure 5E we confirmed the increased sensitivity to Venetoclax of MM PC exposed to SD6 compared with single agent treatment, regardless of

Haematologica | 107 - June 2022

1418

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

specific stage-disease, cytogenetic abnormalities, or previous drug exposure. Importantly, no significant effects were observed on PBMC derived from the same MM patients, suggesting tumor specificity of drugs combination. The impact of this strategy was finally tested on primary MM cells in the presence of their bone marrow microenvironment with slight toxicity observed on normal bone marrow components (Figure 5F; Online Supplementary Figure S12). In order to strengthen clinical relevance of our findings, we next screened different agents currently used for MM treatment in combination with splicing modulators, by performing head-to-head comparison analysis. As shown in the Online Supplementary Figure S13, Venetoclax was readily identified as the best sensitizer of MM cells to Meaymicin B activity with a Combination Index resulting to be the lowest among tested drugs. Collectively, these data support splicing modulation as a strategy to increase Venetoclax sensitivity of MM cells. Based on these results, we next focused on drug resistance mechanisms by investigating whether splicing pathways affect Venetoclax resistance of MM cells. To this aim, we assessed a panel of 10 MM cell lines and measured absolute IC50 to capture the efficacy and potency of the drug (Figure 5G). We then

listed all cell lines according to Venetoclax sensitivity and used CCLE RNA sequencing data to compare gene expression profiling of resistant versus sensitive cell lines; significant de-regulated genes were submitted to the GSEA software to find most affected pathways. Importantly, this analysis revealed that gene sets associated with RNA biology, RNA processing, and splicing pathways, were strongly enriched in sensitive cell lines, based on negative normalized enriched score (NES) and low qvalue, thus supporting their link with venetoclax-sensitivity (Figure 5H). Splicing inhibitors synergize with Venetoclax in vivo in multiple myeloma xenograft murine models We finally evaluated the efficacy of the combination regimen in vivo in MM xenograft murine models. In the first model, mice were treated with Venetoclax (100 mg/kg, oral administration, once a day for 15 days) and SD6 (12 mg/kg, intratumorally, once a day for 10 days) alone or in combination. SD6 dosing was chosen based on previously reported efficacy and safety profiles in xenograft lymphoma models, while Venetoclax was used at 100 mg/kg to reduce its potency as single agent. Treatment was well

Figure 4. Continued on following page. Haematologica | 107 - June 2022

1419

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

Figure 4. BCL2 family member deregulation outlines splicing modulators activity on multiple myeloma cells. A) Pie chart showing the proportion of significant splice changes derived from ClariomD data of multiple myeloma (MM) cell lines treated with Meayamycin B compared with dimethyl sulfoxide (DMSO)-treated controls. Yellow slice indicates significant intron cassette exon events; purple slice indicates significant intron retention events; gray indicate all other complex categories of splice events. B) The top 10 pathways included in BioCarta gene sets enriched by fgsea R package among the 1,000 most significant mis-spliced genes in Meayamycin B-treated cells. C) Bubble plot shows the enrichment scores, P-values and the types of aberrant splicing event in the top mis-spliced genes of apoptosis pathway after Meayamycin B treatment compared with control cells. D and E) Reverse transcription polymerase chain reaction (RT-PCR) analyses of H929 cells after 6-hour treatment with growing doses of Meayamycin B (3-6 nM) and Sudemycin D6 (1-3µM) (D) or different time points after lentiviral transduction with scramble control or SF3B1 specific shRNAs clone#1 (E), to assess levels of MCL1 (L, long and S, short isoforms), BCLxL and BCL2. GAPDH is used as internal control. The length of the main amplified isoforms is indicated as base pairs (bp). F) Western blot analysis of MCL1, BCL2, BCLxL, PARP and cleaved caspase3 protein expression upon treatment with Meayamycin B and SD6 in H929 cells. GAPDH was used as loading control. G) Heatmap of percentage Cytochrome c loss, as quantified by flow cytometry on indicated MM cell lines after 6-hour treatment with different chemicals as compared with DMSO-treated controls (purple, lowest value; yellow, highest value). After each treatment, MM cells were exposed to BH3 mimetic peptides (MS1 10 uM, mBAD and HRKy 100 uM for all cell lines except for MM1S where 50 µM mBAD and HRKy were used) for 45 minutes at room temperature and subsequently stained for fluorescence activated cell sorting analysis. Haematologica | 107 - June 2022

1420

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

tolerated during the 2-weeks treatment (Online Supplementary Figure S14). The superior effect of the combination over single agents in reducing tumor volume and weight, (Figure 6A and B) was observed starting after 1 week of therapy. The decrease in tumor growth was associated with improved OS of animals treated with the combination therapy, as compared with vehicle and single agents (Figure 6C, P<0.05). Remarkably, animals treated with Venetoclax monotherapy had a similar survival to that of vehicle-treated animals (P=0.15), confirming the lack of sensitivity of this model to Venetoclax treatment alone. In line with this data, immunohistochemical analyses revealed a dramatic increase of cleaved caspase-3 positive MM cells, with marked reduction of mitotic percentage (Figure 6D and E). Additionally, increased apoptotic bodies were observed in tumors derived from mice treated with the Venetoclax combination (Figure 6E). We also evaluated in vivo efficacy of Venetoclax combination with the clinical grade SF3b-targeting splicing modulator E7107.34 Mice injected with human MM1S cells were treated with vehicle, E7107 (2.5 mg/kg intravenous, once a day for 5 days), Venetoclax (100 mg/kg oral administration, once a day for 5 days) and their combination. As observed in our previous model, E7107-treated tumors resulted in significant tumor regression compared to control mice and mice treated with Venetoclax alone and, the

effect of E7107 was further enhanced by the combination with Venetoclax, as shown in Figure 6F (P<0.05). The mean tumor volume of mice treated with combination was lower than single agent treated mice bearing tumors with higher activity at day 54. Improved outcome was also observed for the combination group of mice, with a median survival significantly longer compared to those treated with vehicle or monotherapies (combo vs. vehicle P<0.05; combo vs. Venetoclax P<0.001; combo vs. E7107 P<0.05) (Figure 6G). Overall, these data support the potential benefits of splicing inhibition to enhance anti-MM effects of Venetoclax.

Discussion The process of RNA splicing regulates gene function in normal cells, but only recently its deregulation has been implicated in driving the development and/or maintenance of cancer.9,35 Indeed, comprehensive genomic analyses have revealed accumulation of transcriptome changes, including aberrantly spliced products during tumor progression and drug resistance occurrence, thus identifying spliceosome machinery deregulation as crucial for tumor cells growth and making its targeting an innovative strategy for successful anti-cancer approaches.9 In

Figure 5. Continued on following page.

Haematologica | 107 - June 2022

1421

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

Figure 5. Splicing modulators sensitize multiple myeloma cells to Venetoclax by deregulating BCL2 family-members. Cell viability curves of multiple myeloma (MM) cells treated with combination therapies using Meayamycin B (A) or Sudemycin D6 (B) and Venetoclax. CI synergy score (calculated with CalcuSyn software) for each set of drugs combination is indicated. Data are presented as mean ± standard deviation (SD) (n=3). (***P≤0.001, ****P≤ 0.0001; unpaired t-test). C) Immunoblots for phosphoSF3B1, SF3B1, PARP, caspase 3, and GAPDH on MM cell lines following each stimulus (indicated in figure) at 24 hours. D) Apoptotic cell death assessed with flow cytometry analysis after Annexin V/propidium iodide (PI) staining of AMO-1 cells SF3B1-silenced (nucleofected with specific small interfering RNA (siRNA) or control cells (siRNA scramble) treated with Venetoclax (7.5 µM) for 48 hours. Displayed are data represented as mean +/− SD in all (n=3). E) CD138+ cells (left) and peripheral blood mononuclear cells (PBMC) (right) collected from MM patients were treated with indicated doses of SD6, Venetoclax (0.5 µM) and their combination for 48 hours. Cell viability was measured by CTG assay. Cells deriving from the same patient are represented with same color in each graph. F) Flow plots of 1 representative MM patient sample. Corresponding sensitivity of MM-gated cells (CD38+/CD45-) and bone marrow stromal cells (BMSC) (CD38-/CD45-) to Venetoclax, SD6and cotreatment are shown. G) Indicated human multiple myeloma cell line (HMCL) (n=10) were treated with different doses of Venetoclax (Ven) for 24 hours and cell survival was assessed by CTG. Half maximal inhibitory concentration (IC50) analysis was performed with GraphPad software. n=3 independent experiments. H) Gene set enrichment analysis (GSEA) normalized enriched scores (NES) and falsediscovery rate (FDR) q-values for top enriched pathways, according to Reactome gene set, among de-regulated genes of Venetoclax resistant vs. sensitive cell lines; red square indicates the most significant region with splicing-related pathways accumulation (left); examples of GSEA-derived enrichment plots for genes involved in the splicing machinery (right). Haematologica | 107 - June 2022

1422

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

such a context, small molecules including spliceostatin, sudemycins, pladienolide and its derivatives are emerging as groundbreaking tools to fight cancers, especially those bearing splicing gene mutations. Since the RNA splicing apparatus is frequently affected by a wide range of driving events, including oncogenic MYC-activation and MCL1 addiction, a synthetic-lethal relationship between different tumor “Achille’s heel” and core-spliceosome inhibition has been proposed, resulting in the increased use of splicing modulators for cancer treatment.9,19,27,36 Here, we first confirmed the already reported upregulation of alternative splicing processes in MM cells by analyzing clinically annotated high throughput RNA sequencing data from the CoMMPass study. A pervasive splicing aberration was found in these cells to be associated with aberrant expression of RNA splicing machinery components and to significantly impact on overall clinical outcome. In line with these data, proliferating MM cells exhibited greater levels of spliceosome component SF3B1 than normal cells suggesting its role for tumor cell growth. Alternative splicing affects a wide range of genes with prevalent impairment of various cellular processes, including cell- cycle. Indeed, splicing isoforms have been reported for key cellcycle factors including cell division cycle 25 (CDC25), aurora kinase B (AURKB), CDC-Like kinase 1 (CLK1) and CDK2. 37,38 Thus, tumor cell proliferation rate might be respon-

sible for high SF3B1 levels observed in these cells, but more investigations are needed. Despite the low mutational burden reported for splicing factors in MM, a striking anti-tumor effects of splicing modulators was observed to be associated with a significant change in MCL1 spliced isoforms in treated cells than specific control. These results are in accordance with other reports showing differential sensitivity to splicing modulation in BCL2 family genes, providing novel insights into mechanisms for spliceosome-targeted based approaches in tumors.29 The effect on MCL1 caused by SF3B1 targeting resulted in a cellular addiction shifting from MCL1 to BCL2 thus providing a rationale for combination of splicing modulators with the BCL2 inhibitor Venetoclax to achieving a synergistic activity against MM cell growth and viability. Indeed, a significant impact of this combination was observed on MM cells both in vitro and in vivo, while sparing non-cancerous cells thus providing evidence for a favorable therapeutic index. Apoptosis is a highly regulated intracellular process, essential for normal cells survival and development; defects in regulation of this process result in a number of abnormalities, including cancer.39 Several B-cell malignancies such as chronic lymphocytic leukemia (CLL), lymphomas and plasma cell neoplasms, present different interaction between subgroups of the BCL2-family members includ-

Figure 6. Continued on following page.

Haematologica | 107 - June 2022

1423

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

Figure 6. Splicing inhibitors synergize with Venetoclax in vivo in multiple myeloma xenograft murine models A) MM1S cells (4,5×106 cells/mouse) were implanted in both flanks of female NOD/SCID J mice (8 weeks of age). Tumor-bearing mice were randomized and treated with vehicle (n=10); Venetoclax (100 mg/kg; n=10) administered by oral gavage once a day for 15 days; SD6 (12 mg/kg; n=10) intratumorally once a day for 10 days and their combination (n=10). A significant delay in tumor growth in combination-treated mice was noted compared to vehicle-treated control mice (***P≤0.001). Data are mean tumor volume ± standard error of the mean (SEM). B) At the end of the experiment, mice were sacrificed, and tumor masses were imaged and weighed. (n.s.: not significant, *P≤0.05, **P≤0.01). C) Kaplan-Meier analysis showing significant survival benefit for mice treated with the combo, compared with single-agent treatment SD6 (P=0.0169, Log-rank Mantel-Cox test). D) Histogram pots show caspase 3-positive cells and mitosis observed in 10 observations in tumors harvested from mice treated with indicated stimuli. (*P≤0.05, **P≤0.01). E) Immunohistochemical analysis for hematoxylin and eosin (H&E) and cleaved caspase 3 in xenografts tumors harvested from mice treated with control, SD6, Venetoclax or co-treatment. Tumor sections from treated and untreated mice stained with H&E were also analyzed for apoptotic bodies formation by using higher magnification. Red arrows indicate apoptotic bodies in each panel. Scale bar, 50 µm. F) MM1S xenograft-bearing, 8-week-old female NOD/SCID J mice were treated with vehicle (n=6); E7107 (2.5 mg/kg; n=9) intravenous once a day for 5 days; Venetoclax (100 mg/kg; n=8) oral administration once a day for 5 days and their combination (n=11). A significant delay in tumor growth in co-treated mice was noted compared to E7107-treated control mice (*P=0.0455). Bars indicate mean ± SEM. G) Kaplan-Meier survival plot showing significant increase in survival of mice receiving the combination of E7107 plus Venetoclax compared to E7107 single agent-treated mice (combo vs. E7107 P=0.0295). n indicates the number of tumors per treatment group. Data were analyzed by two-tailed Student’s t-test (A, B, D, F) or by Log-rank Mantel-Cox test (C,G).

ing the pro-survival BCL2 (BCL2, MCL1, BCL2A1 and BCLxL) and the BH3-only proteins causing apoptosis. In these tumors, drug exposure results in perturbed intrinsic apoptotic pathway with mitochondrial apoptotic dependencies reprogramming and cell-death evasion. A proof-of-concept for this strategy is represented by the breakthrough or small molecules mimicking BH3-only proteins, which have totally changed the treatment landscape of CLL and acute myeloid leukemia.40 Also MM exhibit high dependence on apoptosis as suggested by efficacy of phase I clinical trial with Venetoclax in monotherapy (clinicaltrials gov. Identifier: NCT02755597) in RRMM patients, mainly in those harboring the t(11;14) translocation.33 Unfortunately, changes in apoptosis regulators result in clinical resis-

tance, thus supporting alternative strategies for these patients, including multiple apoptotic proteins inhibition.41,42 40% of MM patients harbor gain or amplification status of 1q21 which makes these cells addicted to MCL1 and extremely vulnerable to its targeting, as suggested by preclinical studies.30,43 However, MCL1 has proven to be more challenging to target than BCL2, despite the long- standing interest in designing potent and selective MCL1 inhibitors for therapeutic use that have resulted in poor clinical benefits with few trials still ongoing (clinicaltrials gov. Identifier: NCT02992483, NCT02675452 and NCT03465540). We demonstrate here that modulation of splicing by selectively targeting MCL1, results in enhanced BCL2 activity on MM cells, and supports the synergistic

Haematologica | 107 - June 2022

1424

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

activity observed in combination with Venetoclax. Remarkably, no significant effects were observed on nontumor cells derived from the same MM patient and treatment was well tolerated in mice as well, suggesting a safe therapeutic window of tested combinations. Emerging studies suggest that tumor cells, including MM, hijack spliceosome activity to evade anti-tumor therapies: mis-spliced forms of binding protein CRBN and broadscale intron retention have been observed in MM patients with immunomodulatory drugs or protease inhibitor resistance occurrence, respectively.16,44 Thus, splicing interference represents an exploitable vulnerability to overcome drug resistance but still deserves further investigations. Here, we confirm that alternative splicing is largely perturbed in MM cells and constitutes an attractive source for novel therapeutic strategies. Low dose of splicing modulators, by shifting apoptotic dependencies, results in enhanced BH3 mimetic Venetoclax susceptibility, irrespective of specific genomic landscapes including splicing factor-mutations, c-MYC-expression or t(11;14) status, providing rationale for the use of this combination for the treatment of all MM patients.

Contributions DS and MC designed the research, performed experiments, analyzed the data and wrote the manuscript; CM, PB, EG, SR, PC and DR performed experiments and analyzed the data; PB performed mice experiments; LM performed IHC analyses; AC, AL, FG, MM, PM, SA, RV, MF and AlN provided patient samples; MP performed immunofluorescence analysis; AnN, KT, MS, MF and NM performed genomic analyses; MC and RML revised the final version of manuscript. Acknowledgments The authors thank Dr. Jeremy Ryan and Prof. Anthony Letai for BH3 profiling experiments setup. They acknowledge also the MMRF for the access to CoMMpass study data. Funding This work was supported in part by the Associazione Italiana per la Ricerca sul Cancro (AIRC, MYFG #18491, IG #2328 to MC and MYFG # 21552 to AC), Italian Ministry of Health (GR-2016-02361523 to AC, and 5x1000 funds 2016 to MC) NIH grant P01-155258-07 (to MF, MKS, and NCM), Associazione Italiana Leucemie Linfomi e Mieloma (AIL sezione di Genova) and University of Genoa, Genoa, Italy.

Disclosures No conflicts of interest to disclose.

References 1. Palumbo A, Anderson K. Multiple myeloma. N Engl J Med. 2011;364(11):1046-1060. 2. Kumar SK, Rajkumar V, Kyle RA, et al. Multiple myeloma. Nat Rev Dis Prim. 2017;3(1):17046. 3. Leung-Hagesteijn C, Erdmann N, Cheung G, et al. Xbp1sNegative tumor B cells and pre-plasmablasts mediate therapeutic proteasome inhibitor resistance in multiple myeloma. Cancer Cell. 2013;24(3):289-304. 4. Cea M, Cagnetta A, Adamia S, et al. Evidence for a role of the histone deacetylase SIRT6 in DNA damage response of multiple myeloma cells. Blood. 2016;127(9):1138-1150. 5. Soncini D, Minetto P, Martinuzzi C, et al. Amino acid depletion triggered by ʟ-asparaginase sensitizes MM cells to carfilzomib by inducing mitochondria ROS-mediated cell death. Blood Adv. 2020;4(18):4312-4326. 6. Ast G. How did alternative splicing evolve? Nat Rev Genet. 2004;5(10):773-782. 7. Graubert TA, Shen D, Ding L, et al. Recurrent mutations in the U2AF1 splicing factor in myelodysplastic syndromes. Nat Genet. 2012;44(1):53-57. 8. Krämer A. The structure and function of proteins involved in mammalian pre-mRNA splicing. Annu Rev Biochem. 1996;65:367-409. 9. Lee SCW, Abdel-Wahab O. Therapeutic targeting of splicing in cancer. Nat Med. 2016;22(9):976-986. 10. Webb TR, Joyner AS, Potter PM. The development and application of small molecule modulators of SF3b as therapeutic agents for cancer. Drug Discov Today. 2013;18(12):43-49.

11. Yoshida K, Sanada M, Shiraishi Y, et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature. 2011;478(7367):64-69. 12. Quesada V, Conde L, Villamor N, et al. Exome sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in chronic lymphocytic leukemia. Nat Genet. 2012;44(1):47-52. 13. Maguire SL, Leonidou A, Wai P, et al. SF3B1 mutations constitute a novel therapeutic target in breast cancer. J Pathol. 2015;235(4):571-580. 14. Bauer MA, Ashby C, Wardell C, et al. Differential RNA splicing as a potentially important driver mechanism in multiple myeloma. Haematologica. 2021;106(3):736-745. 15. Walker BA, Mavrommatis K, Wardell CP, et al. Identification of novel mutational drivers reveals oncogene dependencies in multiple myeloma. Blood. 2018;132(6):587-597. 16. Huang HH, Ferguson ID, Thornton AM, et al. Proteasome inhibitor-induced modulation reveals the spliceosome as a specific therapeutic vulnerability in multiple myeloma. Nat Commun. 2020;11(1):1931. 17. Ryan J. A Guide to BH3 Profiling METHOD 2: iBH3 PROFILING. 2017. 18. Adamia S, Bar-Natan M, Haibe-Kains B, et al. NOTCH2 and FLT3 gene mis-splicings are common events in patients with acute myeloid leukemia (AML): new potential targets in AML. Blood. 2014;123(18):2816-2825. 19. Koh CM, Bezzi M, Low DHPP, et al. MYC regulates the core premRNA splicing machinery as an essential step in lymphomagenesis. Nature. 2015;523(7558):96-100. 20. Adamia S, Abiatari I, Amin SB, et al. The effects of MicroRNA

Haematologica | 107 - June 2022

1425

ARTICLE - Splicing inhibitors sensitize MM to Venetoclax

D. Soncini et al.

deregulation on pre-RNA processing network in multiple myeloma. Leukemia. 2020;34(1):167-179. 21. Zamani-Ahmadmahmudi M, Dabiri S, Nadimi N. Identification of pathway-based prognostic gene signatures in patients with multiple myeloma. Transl Res. 2017;185:47-57. 22. Laganà A, Perumal D, Melnekoff D, et al. Integrative network analysis identifies novel drivers of pathogenesis and progression in newly diagnosed multiple myeloma. Leukemia. 2018;32(1):120-130. 23. Kotake Y, Sagane K, Owa T, et al. Splicing factor SF3b as a target of the antitumor natural product pladienolide. Nat Chem Biol. 2007;3(9):570-575. 24. Kaida D, Motoyoshi H, Tashiro E, et al. Spliceostatin A targets SF3b and inhibits both splicing and nuclear retention of premRNA. Nat Chem Biol. 2007;3(9):576-583. 25. Folco EG, Coil KE, Reed R. The anti-tumor drug E7107 reveals an essential role for SF3b in remodeling U2 snRNP to expose the branch point-binding region. Genes Dev. 2011;25(5):440-444. 26. Marchesini M, Ogoti Y, Fiorini E, et al. ILF2 Is a Regulator of RNA splicing and DNA damage response in 1q21-amplified multiple myeloma. Cancer Cell. 2017;32(1):88-100. 27. Hsu TYT, Simon LM, Neill NJ, et al. The spliceosome is a therapeutic vulnerability in MYC-driven cancer. Nature. 2015;525(7569):384-388. 28. Certo M, Moore VDG, Nishino M, et al. Mitochondria primed by death signals determine cellular addiction to antiapoptotic BCL-2 family members. Cancer Cell. 2006;9(5):351-365. 29. Ruefli-Brasse A, Reed JC. Therapeutics targeting Bcl-2 in hematological malignancies. Biochem J. 2017;474(21):3643-3657. 30. Kotschy A, Szlavik Z, Murray J, et al. The MCL1 inhibitor S63845 is tolerable and effective in diverse cancer models. Nature. 2016;538(7626):477-482. 31. Moujalled DM, Pomilio G, Ghiurau C, et al. Combining BH3mimetics to target both BCL-2 and MCL1 has potent activity in pre-clinical models of acute myeloid leukemia. Leukemia. 2019;33(4):905-917. 32. Prukova D, Andera L, Nahacka Z, et al. Cotargeting of BCL2 with venetoclax and MCL1 with S63845 is synthetically lethal in vivo in relapsed mantle cell lymphoma. Clin Cancer Res. 2019;25(14):4455-4465.

33. Kumar S, Kaufman JL, Gasparetto C, et al. Efficacy of venetoclax as targeted therapy for relapsed/refractory t(11;14) multiple myeloma. Blood. 2017;130(22):2401-2409. 34. Hong DS, Kurzrock R, Naing A, et al. A phase I, open-label, single-arm, dose-escalation study of E7107, a precursor messenger ribonucleic acid (pre-mRNA) splicesome inhibitor administered intravenously on days 1 and 8 every 21 days to patients with solid tumors. Invest New Drugs. 2014;32(3):436444. 35. Chen S, Benbarche S, Abdel-Wahab O. Splicing factor mutations in hematologic malignancies. Blood. 2021;138(8):599-612. 36. Aird D, Teng T, Huang C-L, et al. Sensitivity to splicing modulation of BCL2 family genes defines cancer therapeutic strategies for splicing modulators. Nat Commun. 2019;10(1):137. 37. Petasny M, Bentata M, Pawellek A, et al. Splicing to keep cycling: the Importance of pre-mRNA splicing during the cell cycle. Trends Genet. 2021;37(3):266-278. 38. Jorge J, Petronilho S, Alves R, et al. Apoptosis induction and cell cycle arrest of pladienolide B in erythroleukemia cell lines. Invest New Drugs. 2020;38(2):369-377. 39. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100(1):57-70. 40. DiNardo CD, Jonas BA, Pullarkat V, et al. Azacitidine and Venetoclax in previously untreated acute myeloid leukemia. N Engl J Med. 2020;383(7):617-629. 41. van Delft MF, Wei AH, Mason KD, et al. The BH3 mimetic ABT737 targets selective Bcl-2 proteins and efficiently induces apoptosis via Bak/Bax if Mcl-1 is neutralized. Cancer Cell. 2006;10(5):389-399. 42. Gupta VA, Barwick BG, Matulis SM, et al. Venetoclax sensitivity in multiple myeloma is associated with B cell gene expression. Blood. 2021;137(26):3604-3615. 43. Slomp A, Moesbergen LM, Gong JN, et al. Multiple myeloma with 1q21 amplification is highly sensitive to MCL-1 targeting. Blood Adv. 2019;3(24):4202-4214. 44. Gooding S, Ansari-Pour N, Towfic F, et al. Multiple cereblon genetic changes are associated with acquired resistance to lenalidomide or pomalidomide in multiple myeloma. Blood. 2021;137(2):232-237.

Haematologica | 107 - June 2022

1426

ARTICLE - Red Cell Biology & its Disorders

Correction of RNA splicing defect in b654-thalassemia mice using CRISPR/Cas9 gene-editing technology Dan Lu,1 Xiuli Gong,1 Yudan Fang,1 Xinbing Guo,1 Yanwen Chen,1 Fan Yang,1 Guijun Zhao,1 Qingwen Ma,1 Yitao Zeng1 and Fanyi Zeng1,2,3 Shanghai Institute of Medical Genetics, Shanghai Children’s Hospital, Shanghai Jiao Tong University; 2Department of Histo-Embryology, Genetics and Developmental Biology, Shanghai Jiao Tong University School of Medicine and 3Key Laboratory of Embryo Molecular Biology, Ministry of Health & Shanghai Key Laboratory of Embryo and Reproduction Engineering, Shanghai, China 1

Correspondence: Fanyi Zeng fzeng@vip.163.com Received: December 22, 2020. Accepted: September 23, 2021. Prepublished: October 28, 2021. https://doi.org/10.3324/haematol.2020.278238 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

Abstract b654-thalassemia is a prominent Chinese subtype of b-thalassemia, representing 17% of all cases of b-thalassemia in China. The molecular mechanism underlying this subtype involves the IVS-2-654 C→T mutation leading to aberrant bglobin RNA splicing. This results in an additional 73-nucleotide exon between exons 2 and 3 and leads to a severe thalassemia syndrome. Herein, we explored a CRISPR/Cas9 genome editing approach to eliminate the additional 73nucleotide by targeting both the IVS-2-654 C→T and a cryptic acceptor splice site at IVS-2-579 in order to correct aberrant b-globin RNA splicing and ameliorate the clinical b-thalassemia syndrome in b654 mice. Gene-edited mice were generated by microinjection of sgRNA and Cas9 mRNA into one-cell embryos of b654 or control mice: 83.3% of live-born mice were gene-edited, 70% of which produced correctly spliced RNA. No off-target events were observed. The clinical symptoms, including hematologic parameters and tissue pathology of all of the edited b654 founders and their offspring were significantly improved compared to those of the non-edited b654 mice, consistent with the restoration of wild-type b-globin RNA expression. Notably, the survival rate of gene-edited heterozygous b654 mice increased significantly, and liveborn homozygous b654 mice were observed. Our study demonstrated a new and effective gene-editing approach that may provide groundwork for the exploration of b654-thalassemia therapy in the future.

Introduction b-thalassemia is one of the most common inherited hematopoietic disorders resulting from the absence or deficient synthesis of the b-globin chain of hemoglobin. It poses an important public health problem, and there have been over 200 different mutations in the b-globin locus reported worldwide.1 The absence or reduction of b-globin chains results in an imbalanced ratio of synthesis of a/b-globin chains. The excess free a chains precipitate in red cells and reduce erythrocyte membrane elasticity, thus causing hemolytic anemia. Patients with severe b-thalassemia depend on lifelong blood transfusion to sustain their life and suffer the risk of tissue damage in multiple organs stemming from iron accumulation.2-4 The b654 mutation is one of the most common mutations in the Chinese population, accounting for 17% of all cases

of b-thalassemia in China.5 It is a C→T substitution at IVS2, position 654, which produces a new 5’ donor splice site and activates a cryptic 3’ acceptor splice site at position 579, leading to insertion in the RNA message of an extra 73-nucleotide (nt) exon between exons 2 and 3 which contains a premature termination codon.6 This produces a severe reduction in the amount of normal processed bglobin mRNA and b-globin chains, ultimately resulting in severe b-thalassemia symptoms in b654 patients.5 Our interests have been to pursue various gene therapy treatment options in an effort to cure b654-thalassemia by increasing the synthesis of normal b-globin chains. We have previously explored an antisense RNA strategy to reduce abnormal b654-globin mRNA, combination therapy of RNA interference and antisense RNA to balance a/b-globin gene expression, as well as adding a functional b-globin gene using a lentiviral vector.7-9 These approaches yielded gratifying results, ameliorating symptoms of b-thalasse-

Haematologica | 107 - June 2022

1427

ARTICLE - Gene-editing therapy for b654-thalassemia mia in b654 mice. Other recent studies aimed to restore normal b-globin expression by targeting the IVS-2-645 mutation site and also yielded promising results.10,11 The highly clustered, regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein 9 (Cas9), derived from an element of the bacterial immune system,12,13 has proven to be a more efficient gene-editing tool compared with the other newly developed editing technologies such as transcription activator-like effector nucleases (TALEN). CRISPR/Cas9 has the advantages of high gene-targeting efficiency, low cost, and easy operability. This makes it a very attractive technology for gene therapeutic applications to correct genetic disorders.14-16 It would be highly desirable to improve the targeting efficiency by using CRISPR/Cas9 to disrupt the aberrant splicing site directly. Unfortunately, for the b654 point-mutation site, no simple and appropriate single guide (sg)RNA is available for such a purpose. We hypothesized that targeting both sites (IVS-2654 C→T and IVS-2-579) by CRISPR/Cas9 would effectively correct the aberrant splicing. Here, we report the efficacy of CRISPR/Cas9 genome editing to create targeted elimination of the extra 73-nt exon in b654 mice. The resulting gene-edited mice showed improvement of normal b-globin gene expression, and their clinical thalassemia symptoms were significantly alleviated, illustrating the potential of this strategy for translation into a clinical therapy for b654 thalassemia in the future.

Methods b654-thalassemia mice The b654-thalassemia mice (b654 mice, also known as Hbbth4 /Hbb+, JAX #003250) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). These mice are heterozygous for a human gene with a bIVS-2-654 splicing mutation and the normal mouse b-globin allele. Clinically, b654 mice mimic the symptoms of moderate to severe forms of b-thalassemia, for example, low red blood cell counts, inefficient erythropoiesis, increased red blood cell destruction, and splenomegaly.17 Mouse lines developed in this study are designated as follows: b654-Ctrl (b654 mice with the same genotype as those purchased from the Jackson Laboratory); b654-E (b654 Edited: mice exposed to CRISPR/Cas9 editing at the onecell stage); b654-ER (b654 Edited Responsive: mice edited which subsequently showed corrected RNA splicing); b654ENR (b654 Edited Non-Responsive: mice edited but in which the resulting alterations were ineffective on RNA splicing); b654-EF (b654 Edited Failed: mice exposed to CRISPR/Cas9 but with failure of alteration of any target sequences). All the animal procedures were performed in strict accordance with the Guide for the Care and Use of Laboratory Animals. All experimental procedures were approved

D. Lu et al. by the Animal Care and Use Committee of Shanghai Children’s Hospital. Generation and screening of gene-edited b654-E mice Male b654-Ctrl mice were mated with wild-type (WT) female mice, and the resulting fertilized eggs were collected at the late one-cell stage and subjected to microinjection with the sgRNA (50 ng/mL) and Cas9 mRNA (50 ng/mL) (Invitrogen) and transferred into the Fallopian tubes of pseudopregnant mice. WT and b654-E mice were subject to the following screening at about 3 weeks of age. Quantitative polymerase chain reaction (PCR) was used to measure the copy numbers of normal mouse b-globin gene and to differentiate WT (with two copies) and b654 (that carry only one normal mouse b-globin gene) offspring mice. Additionally, human sequence PCR primers were designed to amplify the IVS-2 region to identify mice carrying the human allele after CRISPR/Cas9 treatment. The resulting PCR products were sequenced to determine the effects of gene editing. Expression of the human b654-globin gene Expression of the human b654-globin gene in mice was examined at the transcription level by reverse transcriptase PCR and at the translational level by western blot. For reverse transcriptase PCR, cDNA was synthesized from total RNA by M-MLV reverse transcriptase (Promega, USA) according to the manufacturer’s instructions and subjected to PCR analysis with IVS-2-654 specific primers HBG-L and HBG-R, which produce an amplicon of 399 bp for the normal cDNA, or 472 bp for the b654-globin cDNA. The mouse glyceraldehyde-3-phosphate dehydrogenase (Gapdh) gene with a 530 bp amplified fragment was used as a control. Western blot analysis was performed as described previously,18 with primary human b-globin monoclonal antibody (Abcam, Cambridge, USA; 1:2000 diluted) and horseradish peroxidase-conjugated goat anti-mouse IgG as a secondary antibody (Dako, Denmark; 1:2000). Polyclonal rabbit anti-human a-globin antibody (Abcam, Cambridge, USA; 1:2000) was used to detect human and mouse a-globin as an internal control. Details of plasmid constructions, off-target analysis, in vitro transcription of sgRNA, cell culture and transfection, hematologic analysis, histopathology analysis, genomic sequencing, and statistical analysis are available in the Online Supplementary Methods.

Results Designing and screening of sgRNA targeting the b654-globin splicing mutation Based on predicted high on-target efficiency and low off-

Haematologica | 107 - June 2022

1428

ARTICLE - Gene-editing therapy for b654-thalassemia

D. Lu et al.

Legend on following page. Haematologica | 107 - June 2022

1429

ARTICLE - Gene-editing therapy for b654-thalassemia

D. Lu et al.

Figure 1. Single-guide RNA screening and gene-editing profiles in 293T cells. (A) Schematic of the single-guide (sg)RNA design within the human b-globin chain. The red arrow indicates the position of IVS-2-654 C→T, and the green arrow is IVS-2-579. G1 is the sgRNA designed to target the downstream IVS-2-654. G2 and G3 are sgRNA designed to target the upstream IVS-2-579. bL/b-R are polymerase chain reaction (PCR) primer pairs. (B) Gel electrophoresis of PCR amplicons from 293T cells transfected with different sgRNA pairs. Lane M, 100 bp DNA ladder; lane 1, 293T cell control; lanes 2-5, 293T cells transfected with water (2), Cas9 (3), G1+G2 (4), and G1+G3 (5). Deletion rates, determined by the ratio of 480 bp band intensity vs. 616 bp band intensity, are shown below each lane. (C) Representative G1+G2 gene-editing DNA profiles. The underlined sequences are G1 and G2. A total of 223 clones were sequenced. The percentage of each alteration observed is shown on the right side. The yellow highlighted sequences represent the protospacer adjacent motifs. The red sequences represent inserted bases. The bordered sequences represent substitution nucleotides.

target effects determined by the CRISPOR program,19 three sgRNA (G1, G2 and G3) that target the aberrant donor or acceptor site (Figure 1A and Online Supplementary Table S1) were selected to be tested in pairs. After cloning sgRNA into the pX459 plasmid vector, two pools containing CRISPR plasmid pairs (G1+G2 or G1+G3) were used to create two double-strand breaks targeted deletion in 293T cells. The dual sgRNA-mediated deletions were verified by PCR using the primer pairs b-L/b-R depicted in Figure 1A. The G1+G2 pool produced a higher deletion efficiency, up to 55% (Figure 1B), and was thus selected for the experiments that follow. PCR products were subcloned and used for Sanger sequencing (Figure 1C). A total of 223 clones were examined. Large deletions (>100 bp) occurred in 96 clones (43.0%), and two inversion clones (0.9%), five substitution clones (2.2%), 68 WT sequence clones (30.5%) and 52 short insertion or deletion clones (23.3%) were identified. The 140 bp deletion was most frequent among the large deletion clones and accounted for one quarter of all the deletions noted. In addition, single cuts repaired by non-homologous end joining (NHEJ) were observed at just the 5’ or 3’ target sites, with frequencies of 17.9% and 0.4%, respectively. Off-target sites were predicted using the CRISPOR program. The top ten sites ranked by the cut-frequency determination off-target score for each sgRNA were subjected to targeted deep sequencing (Online Supplementary Table S2 and Online Supplementary Figure S1). There was no evidence of off-target events at these sites. Generation and characterization of b654-E mice The b654-Ctrl mice are heterozygous for a human gene with a bIVS-2-654 splice mutation and the normal mouse bglobin allele. They typically display moderate to severe clinical symptoms of b-thalassemia17 and a reduced survival rate (26.8%).18 Normally, no homozygous b654-Ctrl mice survive postnatally.17 By mating b654-Ctrl male mice with WT female mice, a total of 142 fertilized eggs were obtained. After microinjecting the G1+G2 sgRNA and Cas9 mRNA into the zygotes, 123 embryos were transplanted into pseudopregnant mice. A total of 37 offspring were born, and tail vein blood samples were collected 3 weeks after birth (Online Supplementary Table S3). Quantitative PCR with mouse-spe-

cific b-globin primers was used to measure mouse allele copy number. Twelve of the 37 mice had only one copy of the mouse b-major globin gene and thus were b654-E mice. The remaining mice had two copies of the b-major globin gene and were therefore WT mice (Online Supplementary Figure S2). PCR analysis of the 12 mice with b654-specific primers was consistent with the copy number measurements (Figure 2A and Online Supplementary Figure S2). The PCR products were subjected to Sanger sequencing. Ten of the 12 mice (83%) had alterations at the targeted loci (Online Supplementary Table S3). Multiple PCR subclones from each mouse were sequenced, and six out of the ten mice showed a single deletion while the remaining four mice were mosaic with multiple alterations (Figure 2B). RNA expression of edited genes To investigate whether genetic deletion of the extra exon is sufficient to restore normal b-globin splicing and expression, we analyzed RNA from peripheral blood cells of the 12 b654-E offspring mice. Reverese transcriptase PCR showed that the correct b-globin transcript (399 bp) was present in seven gene-edited mice, among which six showed correctly spliced mRNA exclusively, and one (mouse ID 59) showed DNA mosaicism that produced about 60% expression of correctly spliced mRNA (i.e., b654ER mice). The remaining three mice also displayed genetic mosaicism, but all mRNA examined showed the incorrect splicing message with the 73-nt additional exon, and these mice were, therefore, b654-ENR (Figure 2C). Of the six mice with effective, non-mosaic gene-editing, four have successful deletion of both the IVS-2-654 aberrant splicing acceptor site and the IVS-2-579 splicing donor site, while the other two had deletion of either the IVS-2-654 or IVS-2-579 site but not both (Figure 2B). Western blot analysis using a b-globin monoclonal antibody that detects only WT human b-globin showed the presence of human b-globin in the seven b654-ER mice. No human bglobin expression was observed in the three b654-ENR mice. These data are consistent with the reverse transcriptase PCR results. The corrected human b-globin protein levels in the seven b654-ER mice were comparable (less than 20% difference) (Figure 2D). Reverse-phase high-performance liquid chromatography was then used to quantify the amounts of mouse and normal human b-globin chains (for

Haematologica | 107 - June 2022

1430

ARTICLE - Gene-editing therapy for b654-thalassemia

D. Lu et al.

Legend on following page. Haematologica | 107 - June 2022

1431

ARTICLE - Gene-editing therapy for b654-thalassemia

D. Lu et al.

Figure 2. Identification of human b-globin expression in b654-E mice. (A) Polymerase chain reaction (PCR) of genomic DNA identified b654-E mice. The unedited amplicon is 616 bp. Lane M: 100 bp DNA ladder; lane 1: no template control; lane 2: b654-Ctrl mouse; lane 3: wild-type mouse (WT); lanes 4-15: b654-E mice obtained following gene editing. The lane numbers corresponding to the Mouse ID are listed in panel B and are consistent across all panels of Figure2C, D. (B) Gene-editing sequence profiles. The underlined sequences are G1 and G2, and the yellow highlighted sequences are protospacer adjacent motifs. There is one founder embryo for each mouse ID. The red sequences represent inserted bases. (C) Reverse transcriptase PCR analysis of human b-globin RNA expression. The 399 bp band indicates the correctly spliced human b-globin (hub-globin) transcript, and the 472 bp band indicates the aberrant b-globin transcript. The mouse glyceraldehyde-3-phosphate dehydrogenase (Gapdh) gene with a 530 bp amplicon was used as an internal control. Lane M: 100 bp DNA ladder; lane 1: human; lane 2: WT mouse; lane 3: b654-Ctrl mouse; lanes 4-15 as described above for 2A. The ratios of correctly spliced amplicons vs. total amplicons (correct + aberrant) were measured by gel densitometry. (D) Western blot analysis of human b-globin protein expression. Lane 1: human; lane 2: b654-Ctrl mouse; lane 3: WT mouse; lanes 4-15: as described for 2A above. Mouse a-globin (ma-globin) served as an internal control. The intensity ratios of hub-globin vs. ma-globin are shown below each lane.

Table 1. Hematologic analyses. Group

RBC (106/mL)

HGB (g/L)

HCT (%)

MCV (fL)

MCH (pg)

MCHC (g/L)

RET (%)

9.2±1.1*

136.6±18.2*

44.6±5.7*

48.3±1.0*

14.8±0.4*

305.9±8.7*

3.3±0.6*

b654-ER

9.6±1.2*

145.0±17.5*

47.5±6.2*

49.5±1.6*

15.1±0.6*

304.4±6.5

3.8±1.2*

b654-ENR

6.0±0.6**

76.3±0.1**

22.8±2.9**

37.7±1.1**

12.8±0.9**

337.7±32.7**

15.9±3.1**

b654-Ctrl

6.5±0.7

82.3±7.6

25.7±2.3

39.4±2.2

12.6±0.8

320.8±14.3

19.2±3.6

Values represent mean ± standard deviation. Statistically significant differences are indicated for WT, b654-ER, or b654-ENR compared to b654-Ctrl. *P<0.01, **P>0.05. N: number of mice tested; RBC: red blood cell count; HGB: hemoglobin; HCT: hematocrit; MCV: mean corpuscular volume; MCH: mean corpuscular hemoglobin; MCHC: mean corpuscular hemoglobin concentration; RET: reticulocyte count.

b654-ER mice), and human peptides ranged from 50.1% to Histopathology of the spleens showed that, compared to 52.5% of total b-globin (Online Supplementary Figure S3). b654-Ctrl mice and b654-ENR mice, b654-ER mice showed significant reductions of red pulp and the marginal zone of 654-ER Improvement of hematologic indices in b mice white pulp was distinct (Figure 3C). The hemosiderin in Hematologic parameters were measured once every 2 the spleens of the b654-ER mice was dramatically reduced weeks for 3 months. Elevated levels of red blood cells, he- compared to that in the b654-Ctrl mice and b654-ENR mice (Figmoglobin, hematocrit, mean corpuscular volume, mean ure 3D). Iron accumulation was rarely observed in the corpuscular hemoglobin, and mean corpuscular hemoglo- livers of the b654-ER mice (Figure 3E). bin concentration were noted in the b654-ER mice when The extent of splenomegaly was investigated after discompared to b654-Ctrl mice (P<0.01), while these parameters secting and weighing spleens from the b654-E mice. Both were not significantly different compared to those in WT the weight and sizes of spleens of b654-ER mice were remice (P>0.05) (Table 1). In addition, there was a dramatic markably reduced compared to values for the b654-Ctrl and reduction in reticulocyte numbers, from 19.17% to 3.78%. b654-ENR mice. There was no significant difference between Altogether, this indicated a significant improvement of the values for b654-ER mice and WT mice (Figure 4). anemia symptoms in the seven b654-ER mice. Peripheral blood smears indicated that anisocytosis and Elevated survival rate and stable therapeutic effects in poikilocytosis were reduced, and almost no target cells offspring of b654-ER mice were observed in the b654-ER mice (Figure 3A). The abnormal To investigate whether the therapeutic effect can be inbone marrow proliferation and erythropoiesis with an inherited and stable in the progenies of b654-ER mice, two creasing proportion of nucleated cells usually seen in b654- founder mice (ID 52 and ID 53) were mated with WT mice Ctrl samples were also decreased in b654-ER mice (Figure 3B). to obtain F1 offspring. F2 animals were generated by fullsibling matings. The survival rates of b654-Ctrl F1 and F2 Improvement of morphological changes in spleen and generations were only 29.0% and 43.4%, respectively, due 654-ER liver in b mice to the severity of the thalassemia syndrome (Table 2).17 To determine the extent of the improvement of extrame- Interestingly, elevated survival rates of 45.6% and 72.0% dullary hematopoiesis in b654-ER mice, a morphological were observed in the F1 and F2 generations of b654-ER examination of spleens and livers from b654-ER 7-month- mice, respectively (Table 2), indicating a clinically signifiold mice and age-matched b654-Ctrl mice was conducted. cant improvement of survival rate of b654-ER mice followHaematologica | 107 - June 2022

1432

ARTICLE - Gene-editing therapy for b654-thalassemia ing gene editing. Furthermore, our gene-editing therapy resulted in the survival of homozygous b654-ER mice, while the unedited homozygous b654-ctrl mice are not usually born live. Hematologic parameters were also measured once every 2 weeks for 3 months to monitor anemia in offspring. The red cell indices of heterozygous b654-ER mice showed significant improvement compared with the levels found in

D. Lu et al. the b654-Ctrl mice, and there were no significant differences in the levels of red blood cells, hemoglobin and reticulocytes between heterozygous b654-ER mice and WT mice (Online Supplementary Table S4). Off-target analysis of b654-ER mice To identify off-target events after gene editing, whole-genome sequencing was performed on the seven b654-ER mice

Figure 3. Red blood cell morphology and tissue pathology in representative b654-E mice. (A) Peripheral blood smears stained with Wright-Giemsa. WT, wild-type mouse; b654-ER mouse ID 64 (142 bp deletion); b654-ENR mouse ID 54 (2 bp or 4 bp deletion); b654-Ctrl mouse. The black arrow points to poikilocytosis, and the red arrow points to target cells in b654-Ctrl mice. Scale bars are 50 mm in all images. (B) Bone marrow smears stained with Wright-Giemsa. (C) Spleen histological sections stained with hematoxylineosin. RP: red pulp; WP: white pulp. (D) Spleen histological sections stained with hematoxylin-eosin. Black arrows point to hemosiderin granules found in b654-Ctrl mice. (E) Ferrocyanide iron staining of liver samples. Blue staining in the right two panels indicates iron accumulation. Haematologica | 107 - June 2022

1433

ARTICLE - Gene-editing therapy for b654-thalassemia

D. Lu et al.

Figure 4. Improvement of splenomegaly in b654-ER mice. (A) Spleen mass in b654-ER mice. The spleen relative mass is represented as % body mass. **Statistically significant difference P<0.01; ns: not significant. Wild-type (WT) n=18; b654-ER n=4 (mouse ID 52, 53, 64 and 59); b654-ENR n=3 (mouse ID 54, 74, and 76); b654 n=10. (B) Representative spleen images. 1: spleen from a WT mouse; 2-3: spleens from b654-ER mice 52 and 53; 4-5: spleens from b654-Ctrl mice.

and three b654-Ctrl mice. Structural variants (including large deletions and insertions, duplications, translocations, and inversions), short insertions or deletions (indels), as well as single nucleotide variants were determined after alignment and comparison to the mm10 mouse reference genome. There were no significant differences in the average number of structural variants and indels between b654-ER mice and b654-Ctrl (Figure 5A, B and Online Supplementary Table S5). Total single nucleotide variant counts were significantly different (Figure 5C) with fewer variants in b654-ER mice. Thirtyfive off-target sites predicted by the CRISPOR program (19 for G1 and 16 for G2) (Online Supplementary Table S6) were examined in detail in the whole genome sequencing data. No sequence variations were observed at the predicted offtarget sites in any of the seven b654-ER mice.

Discussion

Previous gene editing approaches to target the IVS2-654 C→T mutation site have been reported.10,11 However, the CRISPR/Cas9 tactic can be limited by the requirement for a protospacer adjacent motif sequence, so proper sgRNA directly targeting the mutation sites may not always be available. To find alternative target candidates, the splicing donor site (at IVS-2-579) and the splicing acceptor site (at IVS-2-654) were both considered. Normal human b-globin splicing in b654-ER mice was observed after the extra 73-nt exon sequence was eliminated by this approach. The clinical b-thalassemia symptoms were remarkedly relieved, and the survival rate of b654-ER mice improved significantly. Various sequence profiles in gene-edited human hematopoietic stem cells were observed when multiple sgRNA were used.23,24 In this study, the percentage of dual cut deletions of both aberrant splicing sites in founder b654-ER mice and 293T cells was consistent (more than 40%). Table 2. The elevated survival rate of b654 mice following gene

b654-thalassemia is one of the most common b-thalassemias found in Chinese populations and is caused by abnormal RNA processing. According to the Human Gene Mutation Database (http://www.hgmd.cf.ac.uk/ac/index.php), 10%-15% of point mutations in genetic diseases can cause abnormal mRNA splicing,20 and similar RNA processing defects caused by a point mutation were also found in other b-thalassemia diseases (e.g., position at IVS-2-705 and IVS-2-745).21,22 Therefore, the strategy employed to treat b654-thalassemia could also be expanded to other genetic diseases caused by abnormal RNA splicing for a range of clinical disorders. Here, we report the restoration of normal splicing in b654 mice using CRISPR/Cas9-mediated NHEJ with two sgRNA and described therapeutic efficacy in treating b654-thalassemia.

editing.

Mouse

b654 Heterozygous Homozygous Total Survival b654 b654 rate (%)

45.61*

F1 controls

155

29.03

72.00**

F1 controls

122

43.44

F1 mice were produced by crossing b654-ER with wild-type mice. Controls were produced by mating b654-Ctrl with wild-type mice. The F2 generation was generated by full-sibling matings of b654-ER mice. Controls were produced by full-sibling matings of b654-Ctrl mice. *P<0.05, **P<0.01.

Haematologica | 107 - June 2022

1434

ARTICLE - Gene-editing therapy for b654-thalassemia

D. Lu et al.

Figure 5. Whole-genome sequence analysis of CRISPR/Cas9 off-target alterations. The total counts of sequence variants relative to the mouse mm10 reference genome are plotted for three b654-Ctrl mice (Control) and seven experimental mice (b654-ER). Variant categories include: (A) large structural variants; (B) short indels; and (C) single nucleotide variants. The horizontal line is the group average, the vertical line is ±1 standard deviation, *Statistically significant difference P<0.05, and data symbols are consistent for individual mice across the three plots (detailed in the table at the right of Figure 5A). SV: structural variants; indels: insertions or deletions; SNV: single nucleotide variants.

Deletions involving single cuts repaired by NHEJ at only the 5’ or 3’ sites were observed in some b654-E mice, and the single deletion sizes ranged from 4 bp to 79 bp. The therapeutic efficacy appeared to be related to the location of the deletion. Gene-edited mice with deletions including one or two aberrant splicing sites showed correction of the abnormal splicing pathway. It is interesting to note that the deletion of the IVS-2-579 splicing donor site alone can restore the normal expression of b-globin. Thus, the cryptic splicing donor site could be an alternative target for human genetic diseases caused by splice site mutations. One mosaic b654-E mouse with 55% deletion of the extra exon DNA produced 60% correctly spliced mRNA, and a therapeutic effect was observed. This suggests that partially effective gene editing may be sufficient to relieve the disease symptoms. Similar dose effects were found in previous work in which 30% chimerism by induced pluripotent stem cells in b654-thalassemia mice reversed the pathology of anemia.25 Therefore, mosaicism provides a good model

to explore the dosage effect of gene editing in b654-thalassemia gene therapy. A study of human embryo gene editing reported mosaicism was prone to include off-target genetic modifications, which occurred after the first cell cycle.21 More effort is needed to restrict the activity of Cas9 at the single-cell stage and reduce on-target and off-target mosaicism; such work may include developing anti-CRISPR proteins.26 Safety issues concerning CRISPR-mediated gene therapy must be considered. Off-target events may result in the production of unwanted mutations.27,28 Several recent studies have shown that double-strand breaks induced by Cas9 can result in more extensive genetic changes, including loss of partial or whole chromosomes,28 frequent loss of heterozygosity in human embryos,29 chromothripsis in human hematopoietic stem cells,30 and more complex genomic rearrangements in mouse embryos or embryonic stem cells.31,32 Our study did not detect an obvious increase in sequence variants in the b654-ER mice, and these results are consistent with those of studies in other mouse em-

Haematologica | 107 - June 2022

1435

ARTICLE - Gene-editing therapy for b654-thalassemia bryos33 and hematopoietic stem cells.11,34 On the other hand, off-target events may become apparent with expanded sample size, and a lack of overall increases in variation does not preclude the occurrence of individual pathogenic changes. We did not observe alterations at any of the in silico predicted off-site targets after editing cultured cells and embryos. Recent efforts have been made to increase the specificity of the CRISPR system by several orders of magnitude through adding an engineered hairpin secondary structure onto the spacer region of sgRNA.35 Alternative tools that do not require double-strand breaks, such as base or primer editing,36,37 may relieve the worries of adverse outcomes associated with double-strand breaks. Protospacer adjacent motif limitation and base changes at sites flanking the targeted base are also aspects amenable to continuing improvement of CRISPR/Cas9 technology for future clinical therapy. In this study, the sgRNA and Cas9 mRNA were injected into single-cell embryos to examine the therapeutic efficacy for b654-thalassemia. As human germline therapy is prohibited due to the technical challenges of editing efficiency and safety, as well as ethical and legal factors, somatic cell therapy, such as autologous hematopoietic stem cell transplantation with gene modification, is considered to be a practical clinical alternative. Autologous hematopoietic stem cells modified with a lentiviral vector carrying the human b-globin gene, or gene editing targeting the enhancer of BCL11A, have proven effective in clinical trials, and a large number of patients have received successful somatic gene therapy.38-42 The consequences of somatic cell gene editing seem more predictable and acceptable since fewer chromosomal changes in hematopoietic stem cells via NHEJ were found in human embryo studies.30,43 Thus, future careful examination employing the strategy presented in this study for somatic gene therapy in hematopoietic stem cells could lead us one step closer to clinical therapy for b654-thalassemia.

D. Lu et al. In conclusion, our study showed that a CRISPR/Cas9-mediated editing approach to eliminate the extra 73-nt exon could correct the aberrant splicing pathway in b654 mice, significantly relieve the symptoms of the b654 phenotype, and elevate survival rates of heterozygote b654-ER mice. Notably, the live birth of homozygous b654-ER mice (no live homozygous b654-ctrl mice observed previously) also demonstrates the value of our strategy. The gene-editing study presented here appears to be both effective and efficient and could be applied to somatic cell engineering as a novel b654-thalassemia therapeutic approach. Disclosures No conflicts of interest to disclose. Contributions DL designed and performed experiments, analyzed the data, and drafted the manuscript; XG, YF, XG, and YC conducted experiments; FY and GZ performed the genomic and targeted deep sequencing data analysis; QM and YT reviewed the design and edited the manuscript; FZ designed the project and drafted the manuscript. Funding This study was supported by grants from the National Key Research and Development Program of China (2019YFA0801402, 2016YFC1000503 and 2016YFC0905102), China National Basic Research Program (2014CB964703), National Natural Science Foundation of China (82000186), Shanghai Science and Technology Support Program (13431901300), Shanghai Sailing Program (18YF1420300), Clinical Medical Center and Key Disciplines Construction Program of Shanghai (2017ZZ02019), the Experimental Animals Project of Shanghai Municipality (18140901600 and 18140901601), and KingMed Diagnostics Project of Academician Workstation (2017B090904030).

References 1. Giardine B, Borg J, Viennas E, et al. Updates of the HbVar database of human hemoglobin variants and thalassemia mutations. Nucleic Acids Res. 2014;42(Database issue):D1063D1069. 2. Modell B, Darlison M. Global epidemiology of haemoglobin disorders and derived service indicators. Bull World Health Organ. 2008;86(6):480-487. 3. Origa R. Beta-thalassemia. Genet Med. 2017;19(6):609-619. 4. Taher AT, Weatherall DJ, Cappellini MD. Thalassaemia. Lancet. 2018;391(10116):155-167. 5. Huang SZ, Zhou XD, Zhu H, Ren ZR, Zeng YT. Detection of betathalassemia mutations in the Chinese using amplified DNA from dried blood specimens. Hum Genet. 1990;84(2):129-131. 6. Zhang JZ, Cai SP, He X, et al. Molecular basis of betathalassemia in south China. Strategy for DNA analysis. Hum

Genet. 1988;78(1):37-40. 7. Xie SY, Ren ZR, Zhang JZ, et al. Restoration of the balanced alpha/beta-globin gene expression in beta654-thalassemia mice using combined RNAi and antisense RNA approach. Hum Mol Genet. 2007;16(21):2616-2625. 8. Xie SY, Li W, Ren ZR, Huang SZ, Zeng F, Zeng YT. Correction of b654-thalassaemia mice using direct intravenous injection of siRNA and antisense RNA vectors. Int J Hematol. 2011;93(3):301310. 9. Xie S, Li W, Ren Z, et al. Amelioration of beta654-thalassemia in mouse model with the knockdown of aberrantly spliced betaglobin mRNA. J Genet Genomics. 2008;35(10):595-601. 10. Fang Y, Cheng Y, Lu D, et al. Treatment of beta(654)thalassaemia by TALENs in a mouse model. Cell Prolif. 2018;51(6):e12491.

Haematologica | 107 - June 2022

1436

ARTICLE - Gene-editing therapy for b654-thalassemia 11. Xu S, Luk K, Yao Q, et al. Editing aberrant splice sites efficiently restores b-globin expression in b-thalassemia. Blood. 2019;133(21):2255-2262. 12. Cong L, Ran FA, Cox D, et al. Multiplex genome engineering using CRISPR/Cas systems. Science. 2013;339(6121):819-823. 13. Mali P, Yang L, Esvelt KM, et al. RNA-guided human genome engineering via Cas9. Science. 2013;339(6121):823-826. 14. Cornu TI, Mussolino C, Cathomen T. Refining strategies to translate genome editing to the clinic. Nat Med. 2017;23(4):415-423. 15. Dunbar CE, High KA, Joung JK, Kohn DB, Ozawa K, Sadelain M. Gene therapy comes of age. Science. 2018;359(6372):eaan4672. 16. Wang L, Li L, Ma Y, et al. Reactivation of gamma-globin expression through Cas9 or base editor to treat betahemoglobinopathies. Cell Res. 2020;30(3):276-278. 17. Lewis J, Yang B, Kim R, et al. A common human beta globin splicing mutation modeled in mice. Blood. 1998;91(6):2152-2156. 18. Li W, Xie S, Guo X, et al. A novel transgenic mouse model produced from lentiviral germline integration for the study of bthalassemia gene therapy. Haematologica. 2008;93(3):356-362. 19. Concordet JP, Haeussler M. CRISPOR: intuitive guide selection for CRISPR/Cas9 genome editing experiments and screens. Nucleic Acids Res. 2018;46(W1):W242-W245. 20. Faustino NA, Cooper TA. Pre-mRNA splicing and human disease. Genes Dev. 2003;17(4):419-437. 21. Orkin SH, Kazazian HH Jr, Antonarakis SE, et al. Linkage of betathalassaemia mutations and beta-globin gene polymorphisms with DNA polymorphisms in human beta-globin gene cluster. Nature. 1982;296(5858):627-631. 22. Dobkin C, Pergolizzi RG, Bahre P, Bank A. Abnormal splice in a mutant human beta-globin gene not at the site of a mutation. Proc Natl Acad Sci U S A. 1983;80(5):1184-1188. 23. Ye L, Wang J, Tan Y, et al. Genome editing using CRISPR-Cas9 to create the HPFH genotype in HSPCs: an approach for treating sickle cell disease and b-thalassemia. Proc Natl Acad Sci U S A. 2016;113(38):10661-10665. 24. Mettananda S, Fisher CA, Hay D, et al. Editing an a-globin enhancer in primary human hematopoietic stem cells as a treatment for b-thalassemia. Nat Commun. 2017;8(1):424. 25. Yang G, Shi W, Hu X, et al. Therapeutic effects of induced pluripotent stem cells in chimeric mice with b-thalassemia. Haematologica. 2014;99(8):1304-1311. 26. Pawluk A, Davidson AR, Maxwell KL. Anti-CRISPR: discovery, mechanism and function. Nat Rev Microbiol. 2018;16(1):12-17. 27. Ledford H. CRISPR gene editing in human embryos wreaks chromosomal mayhem. Nature. 2020;583(7814):17-18. 28. Zuccaro MV, Xu J, Mitchell C, et al. Allele-specific chromosome removal after Cas9 cleavage in human embryos. Cell. 2020;183(6):1650-1664.

D. Lu et al. 29. Alanis-Lobato G, Zohren J, McCarthy A, et al. Frequent loss-ofheterozygosity in CRISPR-Cas9–edited early human embryos. Proc Natl Acad Sci U S A. 2021;118(22):e2004832117. 30. Leibowitz ML, Papathanasiou S, Doerfler PA, et al. Chromothripsis as an on-target consequence of CRISPR-Cas9 genome editing. Nat Genet. 2021;53(6):895-905. 31. Kosicki M, Tomberg K, Bradley A. Repair of double-strand breaks induced by CRISPR-Cas9 leads to large deletions and complex rearrangements. Nat Biotechnol. 2018;36(8):765-771. 32. Adikusuma F, Piltz S, Corbett MA, et al. Large deletions induced by Cas9 cleavage. Nature. 2018;560(7717):E8-E9. 33. Zuo E, Sun Y, Wei W, et al. Cytosine base editor generates substantial off-target single-nucleotide variants in mouse embryos. Science. 2019;364(6437):289-292. 34. Wu Y, Zeng J, Roscoe BP, et al. Highly efficient therapeutic gene editing of human hematopoietic stem cells. Nat Med. 2019;25(5):776-783. 35. Kocak DD, Josephs EA, Bhandarkar V, Adkar SS, Kwon JB, Gersbach CA. Increasing the specificity of CRISPR systems with engineered RNA secondary structures. Nat Biotechnol. 2019;37(6):657-666. 36. Komor AC, Kim YB, Packer MS, Zuris JA, Liu DR. Programmable editing of a target base in genomic DNA without doublestranded DNA cleavage. Nature. 2016;533(7603):420-424. 37. Anzalone AV, Randolph PB, Davis JR, et al. Search-and-replace genome editing without double-strand breaks or donor DNA. Nature. 2019;576(7785):149-157. 38. Thompson AA, Walters MC, Kwiatkowski J, et al. Gene therapy in patients with transfusion-dependent b-thalassemia. N Engl J Med. 2018;378(16):1479-1493. 39. Kwiatkowski JL, Thompson AA, Rasko JEJ, et al. Long-term clinical outcomes of lentiglobin gene therapy for transfusiondependent b-thalassemia in the Northstar (HGB-204) study. Blood. 2019;134(Suppl_1):4628. 40. Locatelli F, Walters MC, Kwiatkowski JL, et al. Lentiglobin gene therapy for patients with transfusion-dependent b-thalassemia (TDT): results from the phase 3 Northstar-2 and Northstar-3 studies. Blood. 2018;132(Suppl 1):1025. 41. Thompson AA, Walters MC, Kwiatkowski JL, et al. Northstar-2: updated safety and efficacy analysis of lentiglobin gene therapy in patients with transfusion-dependent b-thalassemia and nonb0/b0 genotypes. Blood. 2019;134(Suppl_1):3543. 42. Frangoul H, Altshuler D, Cappellini MD, et al. CRISPR-Cas9 gene editing for sickle cell disease and b-thalassemia. N Engl J Med. 2021;384(3):252-260. 43. Turocy J, Adashi EY, Egli D. Heritable human genome editing: research progress, ethical considerations, and hurdles to clinical practice. Cell. 2021;184(6):1561-1574.

Haematologica | 107 - June 2022

1437

ARTICLE - Red Cell Biology & its Disorders

Ionophore-mediated swelling of erythrocytes as a therapeutic mechanism in sickle cell disease Athena C. Geisness1,* Melissa Azul,1,2* Dillon Williams,1 Hannah Szafraniec,1 Daniel C. De Souza,3,4,5 John M. Higgins,3,4,5 and David K. Wood1 Department of Biomedical Engineering, University of Minnesota Twin Cities, Minneapolis, MN; 2Department of Pediatric and Adolescent Medicine, Division of Pediatric HematologyOncology, Mayo Clinic, Rochester, MN; 3Department of Systems Biology, Harvard Medical School, Boston, MA; 4Department of Pathology, Massachusetts General Hospital, Harvard Medical School, Boston, MA and 5Center for Systems Biology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA 1

*ACG and MA contributed equally as co-first authors.

Correspondence: David K. Wood dkwood@umn.edu Received: March 8, 2021. Accepted: October 18, 2021. Pre published: October 28, 2021. https://doi.org/10.3324 haematol.2021.278666 ©2022 Ferrata Storti Foundation Haematologica material is under a license CC BY-NC

Abstract Sickle cell disease (SCD) is characterized by sickle hemoglobin (HbS) which polymerizes under deoxygenated conditions to form a stiff, sickled erythrocyte. The dehydration of sickle erythrocytes increases intracellular HbS concentration and the propensity of erythrocyte sickling. Prevention of this mechanism may provide a target for potential SCD therapy investigation. Ionophores such as monensin can increase erythrocyte sodium permeability by facilitating its transmembrane transport, leading to osmotic swelling of the erythrocyte and decreased hemoglobin concentration. In this study, we treated 13 blood samples from patients with SCD with 10 nM of monensin ex vivo. We measured changes in cell volume and hemoglobin concentration in response to monensin treatment, and we perfused treated blood samples through a microfluidic device that permits quantification of blood flow under controlled hypoxia. Monensin treatment led to increases in cell volume and reductions in hemoglobin concentration in most blood samples, though the degree of response varied across samples. Monensin-treated samples also demonstrated reduced blood flow impairment under hypoxic conditions relative to untreated controls. Moreover, there was a significant correlation between the improvement in blood flow and the decrease in hemoglobin concentration. Thus, our results demonstrate that a reduction in intracellular HbS concentration by osmotic swelling improves blood flow under hypoxic conditions. Although the toxicity of monensin will likely prevent it from being a viable clinical treatment, these results suggest that osmotic swelling should be investigated further as a potential mechanism for SCD therapy.

Introduction Sickle cell disease (SCD) is an inherited blood disorder that affects approximately 100,000 Americans in the United States and decreases a patient’s life expectancy by 30 years.1 The disease is caused by a genetic mutation in the b-globin gene which produces sickle hemoglobin (HbS). HbS can polymerize under deoxygenated conditions forming stiff, sickled red blood cells (sRBC).2 The presence of sRBC contributes to the key elements of SCD pathology: hemoglobin polymerization, endothelial dysfunction, sterile inflammation, leading to overall disruption of blood flow particularly in the microvasculature.3 These processes ultimately give rise to the vast clinical manifestations seen in SCD including vaso-occlusive episodes (VOE), acute chest syndrome, and stroke. The complex pathophysiology of SCD requires the development of treatments that target one or more of the molecular disease pathology mechanisms.4-5

Given that HbS polymerization is essential in the pathophysiology of SCD, treatments to prevent HbS polymerization continue to be an area of investigation for therapeutic development.6-8 Hydroxyurea, the first drug approved by the Food and Drug Administration (FDA) for the treatment of SCD, induces production of fetal hemoglobin (HbF), an anti-sickling hemoglobin. Though available for decades, patients treated with hydroxyurea experience variable clinical benefit and are subject to ongoing monitoring given its hematologic side effects.9 Recently, another antisickling agent, voxelotor, was approved by the FDA for the treatment of SCD. Voxelotor stabilizes the oxygenated form of HbS by increasing hemoglobin’s oxygen affinity, preventing polymerization when exposed to deoxygenated conditions.10,11 Research is still ongoing to determine voxelotor’s potential side effects and clinical benefit. In its phase III clinical trial, patients randomized to receiving voxelotor did experience an increase in hemoglobin after

Haematologica | 107 - June 2022

1438

ARTICLE - RBC swelling improves sickle blood flow 6 months of use, but they did not have a reduction in VOE.11 While treatments that inhibit polymerization such as voxelotor or hydroxyurea have been successful, therapeutics that target polymerization without affecting hemoglobin oxygen affinity or targeting the hematopoietic niche may provide similar clinical benefit without the side effects seen with these drugs. An alternative mechanism to inhibit HbS polymerization is the reduction of intracellular HbS concentration within a sRBC. Small decreases in HbS concentration can lead to slower polymerization rates that are longer than sRBC capillary transit time.8,12 Previous studies to reduce HbS concentration include using antidiuretic hormone and a low sodium (Na) diet to reduce plasma osmolality and Na concentration. This caused hypotonic swelling of sRBC and a reduction in mean cell hemoglobin concentration (MCHC), ultimately leading to decreased erythrocyte sickling observed in three patients.13 However, maintaining the necessary level of hyponatremia was impractical and results could not be reproduced in later studies.14,15 Rather than decreasing plasma Na and osmolality as a method to reduce MCHC and HbS polymerization, increasing intracellular Na and osmolality may be more feasible. This produces similar osmotic swelling effects and decreases MCHC without the difficulties of sustaining low plasma Na concentrations. In order to study this mechanism and its potential benefit in SCD, ionophores that increase the erythrocyte permeability to Na, such as monensin, can be used to facilitate intracellular Na transport. Monensin selectively binds to Na+ ions and facilitates its electrogenic transport across the erythrocyte membrane, creating an osmotic gradient and causing an influx of fluid intracellularly.16-18 Previous work in sRBC treated with monensin have demonstrated that monensin is effective at increasing mean corpuscular volume (MCV), decreasing MCHC, and increasing deformability of sRBC.19-21 These studies provided a basis for understanding the molecular effects of monensin on RBC, however, they did not examine how these molecular changes impact the mechanics of RBC flow under physiologic conditions. In this study, we use monensin as a model compound to investigate osmotic swelling to reduce MCHC as a potential mechanism for SCD therapy development. We aim to characterize the effects of sRBC osmotic swelling and reduced MCHC on sRBC rheologic oxygen dependence using a microfluidic device designed to recapitulate the physiological environment of the microvasculature. We compare the rheological response to hypoxia in our microfluidic device between blood samples treated with monensin and untreated controls. In order to further quantify its effect, we correlate MCV, MCHC, and the rheological response to hypoxia. By studying the effect monensin may have on rheology, we build upon previous monensin studies and are now able to better capture the complex pathophysio-

A.C. Geisness et al. logic changes in blood flow that occur with deoxygenated conditions in a physiologically relevant system, gaining a more comprehensive understanding of the potential therapeutic mechanism and its in vivo effects.

Methods Monensin treatment All study protocols were approved by the Institutional Review Board (IRB). In preparation for monensin treatment (Figure 1A), RBC were washed three times by centrifugation in Buffer A solution (104 mM NaCl, 32 mM Na2HPO4, 8 mM KH2PO4, 5.5 mM dextrose, 1g/L bovine serum albumin [BSA]; pH 7.4, 305 mOsm) with techniques previously published.18 Dextrose and BSA components of the Buffer A solution were added on the day of experiments. Samples were resuspended in Buffer A with 0.01% EtOH and 10 nM monensin (420701, BioLegend) to achieve 25% hematocrit (hct) and incubated at 37°C for 12 hours (hrs). A concentration of 10 nM monensin was chosen based on previous studies demonstrating optimal cellular effect without increased hemolysis.19 Preliminary research shows that incubation in Buffer A between 6 and 24 hrs limits RBC swelling in controls to less than 5% MCV. While the effect of 0.01% EtOH on RBC has been previously studied to be insignificant,22-24 a control resuspended in Buffer A and 0.01% EtOH was used for each sample in this study. After incubation, the sample was washed with phosphate-buffered saline (PBS) to remove extracellular monensin and resuspended in PBS to achieve 25% hct prior to rheology measurements. Details of blood sample collection, storage, and methods of obtaining laboratory values are provided in the Online Supplementary Appendix. Data collection and analysis Device design, fabrication, and experimental set up have been previously published24-26 and is detailed in the Online Supplementary Appendix. Continuous rheological data were captured using a high-speed camera (GS3-U3-23S6M-C, FLIR) at a frame rate of 500-600 FPS (frames per second) at 40x magnification. Blood flow velocity measurements were collected using a contrast detection algorithm developed in MATLAB based on the Kanade-Lucas-Tomasi algorithm.27-29 The velocity of thousands of contrasting points per frame were identified and averaged to obtain an average velocity per frame. Representative data in Figure 1B demonstrates blood flow velocity under normoxic and hypoxic conditions for a control and monensin treated sample. Each sample was exposed to 1 minute of normoxia (160 mmHg) and then 1 minute of hypoxia (0 mmHg). This oxygenation-deoxygenation cycle was then repeated for a total of 3 cycles. Average steady state (SS) velocity at nor-

Haematologica | 107 - June 2022

1439

ARTICLE - RBC swelling improves sickle blood flow moxia or hypoxia was determined by averaging the velocities of the three cycles at each oxygen tension for each sample. The average SS velocity value was used to determine two velocity metrics used for analysis: velocity response and recovery. Velocity response is defined by the difference between the average SS velocity at 160 mmHg and 0 mmHg oxygen. The response is normalized by the sample’s average velocity at 160 mmHg oxygen tension (Figure 1C) and indicates the magnitude of velocity reduc-

A.C. Geisness et al. tion during deoxygenation. Velocity recovery is defined by the difference in velocity response between treatment and the untreated control and indicates the change in velocity reduction during deoxygenation due to treatment. Statistical analysis A Wilcoxon signed-rank test is used to establish significant difference between control and treatment groups (n=13). A Pearson’s product moment correlation coefficient is

Figure 1. Data collection and analysis. An overview of methods of sample preparation, data collection, and defining rheological variables used throughout the study. (A) Schematic of the monensin treatment workflow. (B) Representative image of raw velocity data (below) as it relates to oxygen tension (above) from a single sickle cell disease (SCD) patient sample. In the bottom panel of (B), blood flow velocity is compared between monensin treatment (red) and the untreated control (blue). The average oxygenated shear rate during experiments was 355 s-1, within physiologic range for channel dimensions.46 In this sample, it appears that the velocity at normoxia of the monensin-treated condition is lower than that of the untreated condition. In order to address the differences in normoxic velocities between treatment conditions, conductance of all 13 samples was calculated to determine if additional variables were present contributing to normoxic velocities (Online Supplementary Figure S3). There were no significant differences in conductance at normoxia between treatment conditions in all samples, indicating velocity differences at normoxia were related to driving pressure. (C) The oxygenated (160 mmHg) and deoxygenated (0 mmHg) sections of the collected velocity data in (B), normalized by the average oxygenated steady state velocity for the representative sample. The representative single patient data in (C) demonstrates a 13% velocity response for a monensin-treated sample and a 33% response for the untreated control. This corresponds to a velocity recovery of 20% after monensin treatment. Velocity response is calculated using the difference between oxygenated (160 mmHg) and deoxygenated (0 mmHg) velocities and velocity recovery is calculated by the difference in the control and monensin treated response. SN: supernatant; PBS: phosphate-buffered saline, SS: steady-state. Haematologica | 107 - June 2022

1440

ARTICLE - RBC swelling improves sickle blood flow used to describe the linear correlation (n=13). Significance was defined by a P≤0.05.

Results Mean cell hemoglobin concentration strongly correlates with rheologic response to hypoxia A total of 13 samples from patients with SCD were obtained and used in experimentation. A summary of corresponding patient demographic, clinical, and baseline laboratory data are shown in Table 1. Several different sickle cell genotypes were included in the cohort. When oxygenation was decreased from 160 mmHg to 0 mmHg, all untreated sickle samples responded with velocity reduction to a specific steady state velocity. When oxygen tension was restored back to 160 mmHg, blood flow velocity then increased and returned to its SS velocity prior to deoxygenation. Similar velocity response was replicated with repeated cycles of deoxygenation. The conductance of each sample at normoxia and hypoxia were calculated in each treatment condition to ensure non-significant differences in sample preparation and device variability between experiments (Online Supplementary Figure S2). In

Table 1. Patient demographics and laboratory values. Variable Genotype SS SC SB0 SB+ Age (years) Median (range, 23-42) Sex Male Female Hydroxyurea Yes No Clinical History Transfusion* Hospitalizations* Laboratory Values Hematocrit (%) MCV (fL) MCHC (%) HbA (%) HbS (%) HbF (%)

N (13) 7 2 3 1 27 6 7 12 1 Median

Range

2 1

0-4 0-2

21.9 101.4 33.3 15.8 67.8 7.5

13.1-24.1 73.0-119.3 29.7-36.2 7.2-46 38.2-86.5 0.5-21.6

* # of transfusions or hospitalizations within the last year. N: number; MCV: mean corpuscular volume; MCHC: mean cell hemoglobin concentration; Hb: hemoglobin.

A.C. Geisness et al. contrast, oxygen-dependent velocity was not observed in healthy, AA, blood controls (Online Supplementary Figure S3A). In order to identify a parameter which may dictate a sample’s velocity response to hypoxia, we first determined if MCHC and MCV were independent variables within the 13 untreated sickle samples. There was no correlation between MCV and MCHC (Figure 2A, r=-0.008, P=0.982), as MCHC is maintained between 30-36 g/dL within RBC of varying sizes. These MCHC values are consistent with low to normal adult MCHC values typically seen in SCD30 and support previous work demonstrating that native MCV values have no correlation with native MCHC values in SCD.31 This supports that MCV and MCHC are likely independent variables and may individually influence the rheological response. In order to determine the influence of these variables on samples’ velocity response, we compared MCV or MCHC with each sample’s velocity response. There is a slight negative relationship between velocity response and MCV amongst the 13 untreated samples, though this correlation was not significant (Figure 2B, r=0.13, P=0.660). There was, however, a significant positive relationship when correlating velocity response and MCHC (Figure 2C, r=0.83, P=0.001), as untreated samples with lower MCHC had smaller velocity responses when exposed to hypoxia. Collectively, this data corroborates previous work by others demonstrating that cell volume does not strongly correlate with the rheological response and rather it is hemoglobin concentration that is strongly correlated with sample blood flow response.32 Monensin increases sickled red blood cells mean corpuscular volume, decreases mean cell hemoglobin concentration, and reduces hypoxia-induced polymerization In order to determine the effect of monensin on sRBC, MCV and MCHC values pre- and post-treatment were collected of all 13 SCD samples and shown in Figure 3A and B. Overall, the monensin-treated samples had significantly increased MCV (Figure 3A) and decreased MCHC (Figure 3B) when compared to the controls (P<0.01). The significant effects in MCV and MCHC were also observed when treating three healthy, AA blood controls with monensin as well (Online Supplementary Figure S3B). Throughout the sample cohort, the effect of monensin on MCV and MCHC widely varied between samples. Some samples had large differences in MCV and MCHC after monensin treatment (sample ID: 11 and 12). Sample 11 experienced the largest change in both MCV and MCHC after treatment and was from a patient with HbSB+ thalassemia. Sample 12 also experienced large changes in MCV and MCHC and came from a patient who had been recently transfused with a severe clinical phenotype. Others demonstrated only minor changes (sample ID: 3 and 7). Sample 3 came from a patient with SC disease and sample 7 was from a patient

Haematologica | 107 - June 2022

1441

ARTICLE - RBC swelling improves sickle blood flow with HbSB0. Given that monensin drives cell swelling, measured by MCV, and decreases hemoglobin concentration, measured by MCHC, these results reflect the degree in which the sample was affected by monensin. Given the observed changes in MCV and MCHC with monensin treatment, to further demonstrate the mechanism of cell swelling to reduce sickle pathophysiology, we analyzed the morphology of cells from three additional sickle blood samples under shear flow and controlled oxygen tension using a previously published microfluidic chip.33 A full description of the device and the methodology as well as the samples’ baseline hematological laboratory data can be found in the Online Supplementary Appendix (Online Supplementary Table S1). Monensin reduced the fraction of cells containing polymer when exposed to hypoxic oxygen tensions in all samples (Online Supplementary Figure S4). However, similar to the effect observed on MCV and MCHC, the amount monensin reduced polymerized cells in hypoxia varied between samples. Monensin treatment improves rheological response to hypoxia In order to quantify the effect of monensin on sRBC blood flow velocity in a hypoxic environment, we examined the velocity response to 0 mmHg oxygen tension between treated and untreated samples for the 13 SCD blood

A.C. Geisness et al. samples. In monensin-treated samples, there was a significant decrease in velocity response with deoxygenation compared to that of untreated controls (P<0.01, Figure 3C), indicating the efficacy of monensin in decreasing sRBC sensitivity to hypoxia. However, there was variability in the degree of response to monensin treatment across all samples. For example, in sample ID 11, monensin treatment eliminated almost all blood flow velocity oxygen dependence demonstrated by no velocity response to hypoxia compared to a 40% response in the control. This contrasts with sample ID 3 and 5, where there was no monensin effect on blood flow velocity response when compared to the untreated control. In AA samples, there was no change in velocity response between monensintreated and untreated controls, despite the significant changes in MCV and MCHC after monensin treatment (Online Supplementary Figure S3). Reduction in mean cell hemoglobin correlates with improved rheologic response to hypoxia In order o objectively determine whether the magnitude of the monensin-induced changes to MCHC or MCV affects the magnitude of change in velocity response to hypoxia, we compared the linear correlation between MCV or MCHC change induced by monensin and the velocity recovery of each sample. By using the absolute change in MCV and

Figure 2. Dependent parameter analysis. Correlative data from 13 untreated sickle cell disease (SCD) samples to determine the relationship, if any, between mean corpuscular volume (MCV) and mean cell hemoglobin concentration (MCHC) and each variable’s relation to sample velocity response. A Pearson correlation coefficient analysis was used to determine the strength of the linear relationship and a two-tailed analysis of the Pearson coefficient was used to determine significance of the correlation. (A) No correlation was identified between MCV and MCHC (r=-0.008, P=0.982), establishing MCV and MCHC as independent variables. (B) MCV as it relates to velocity response. No correlation was identified between MCV values and velocity response (r=-0.13, P=0.660). A slope of -0.001 and -0.1 are found in (A) and (B) respectively. The slope for these figures is provided for clarification of the scale for (A and B). (C) MCHC and velocity response had a significant correlation (r= 0.83, P<0.001). Haematologica | 107 - June 2022

1442

ARTICLE - RBC swelling improves sickle blood flow

A.C. Geisness et al.

Figure 3. Monensin treatment efficacy and variability. Summary data of monensin treatment. (A and B) The effect of monensin on mean cell hemoglobin concentration (MCHC) and mean corpuscular volume (MCV). The monensin-treated group had significantly higher MCV and lower MCHC compared to the control group. (C) There was a significantly lower velocity response in the monensin-treated group compared to that in the control group. Significance between control and monensin-treated groups was determined using a Wilcoxon signed-rank test and indicated by the asterisks (*) denoting P<0.01. Sample ID represents the de-identified patient ID corresponding to the sample. Error bars indicate the standard deviation in velocity response over 3 oxygenation/deoxygenations cycles. SS: steady state.

MCHC, the analysis removes the variability of each sample’s initial MCV and MCHC and controls for patients’ baseline heterogenous clinical severity. First, we ensured correlation between MCV increases and MCHC decreases in the monensin-treated samples. In Figure 4A, there was a significant positive correlation between MCV change and MCHC change (r=0.91, P<0.001), in that large MCV increases due to monensin corresponded with large MCHC reductions. This demonstrates cell swelling is an effective method to decrease MCHC. When comparing the degree in which MCV was increased by monensin and velocity recovery we found a significant positive correlation (Figure 4B, r=0.87, P<0.001). A more significant positive relationship is seen when comparing the degree to which monensin decreased MCHC and sample recovery (Figure 4C, r=0.96, P<0.001), in that the largest improvements in

sample velocity response to hypoxia correlated with larger reductions in MCHC. These relationships reveal that the degree to which monensin affects sample sensitivity to hypoxia is strongly dependent on the degree to which the MCHC is reduced.

Discussion In this study, we examined osmotic cell swelling to decrease intracellular HbS concentration as a potential mechanism to be targeted for future therapeutic development in SCD. We used a model Na ionophore compound, monensin, to treat SCD blood samples ex vivo. The samples were exposed to hypoxic conditions in a microfluidic device while blood flow was quantified. Though

Haematologica | 107 - June 2022

1443

ARTICLE - RBC swelling improves sickle blood flow

A.C. Geisness et al.

Figure 4. Linear correlation analysis. In order to determine which effect of monensin was driving sample velocity recovery, a correlation analysis to define the relationship between velocity recovery and monensin-induced mean corpuscular volume (MCV) or mean cell hemoglobin concentration (MCHC) change was completed. A Pearson correlation coefficient analysis was conducted to determine the strength of the linear relationship. (A) Change in MCV and change in MCHC was significantly positively correlated (r=0.91, P<0.001). (B) Velocity recovery and MCV change also were positively significantly correlated (r=0.87, P<0.001). (C) The strongest correlation was found between velocity recovery and MCHC change (r=0.96, P<0.01).

previous studies conducted in the early 1980’s demonstrated monensin’s ability to decrease sRBC MCHC,19 monensin’s impact on dynamic blood flow, particularly in hypoxic environments, was not explored. Furthermore, deformability measurements were made through ektacytometry19 which may allude to improved rheology, however, monensin’s global effect on blood flow was not directly measured. Additionally, more recent research reports ektacytometry deformability measurements are unreliable in predicting the ability of RBC perfusion of a microvascular network.34 Given that MCHC reduction was initially hypothesized to reduce polymerization and decrease vasoocclusion, investigating monensin’s effect on dynamic flow rheology is critical to understanding whether the mechanism has a role in future therapeutic development. Using our microfluidic platform, we were able to observe and quantify monensin’s effect of reduced MCHC on blood flow by measuring rheologic variables in a dynamic, physiologically relevant system. In our study, we not only found that monensin decreased MCHC, but we also report that it significantly decreased the sensitivity of SCD blood flow to hypoxia compared to controls. The velocity recovery with monensin treatment varied between samples which correlated to the variation in monensin-induced changes in MCHC. Through the rheological measurements obtained in this study, we provide both novel insight into the capability of this mechanism in the prevention of vaso-occlusion but also provide findings to suggest patient response variability.

While osmotic swelling and increases in MCV are the primary effects of the ionophore treatment, we found that the magnitude of MCHC reduction is the primary parameter modifying blood flow response to hypoxia. This is demonstrated by the significant correlation found between reduction in MCHC and reduction in velocity response (Figure 4C). When comparing response with MCV, the relationship is not as strongly correlated (Figure 4B). This corroborates existing studies which demonstrate that polymerization rates are extremely dependent on HbS concentration.6-7 Additionally, we found that in the four samples that demonstrated insignificant change in velocity response when treated with monensin compared to controls (sample ID: 3, 5, 7, and 10), there was less than a 5% MCHC decrease. It is unclear what caused the observed patient variability in MCV/MCHC response to monensin, however our data suggests that decreasing the MCHC by 5% has a significant impact on the sample’s blood flow in hypoxia. Future studies that examine reduction of MCHC as a mechanism for SCD treatment should use a minimal threshold in MCHC reduction to guide drug efficacy. This study investigated the mechanism of osmotic swelling to decrease intracellular MCHC thereby decreasing HbS polymerization. Similarly, previous SCD drug trials have focused on decreasing intracellular MCHC by inhibiting ion channels that are involved in the pathologic dehydration of sRBC. Clotrimazole, an inhibitor of the Gardos channel

Haematologica | 107 - June 2022

1444

ARTICLE - RBC swelling improves sickle blood flow which contributes to sRBC water loss, demonstrated reduced sRBC dehydration, decreased MCHC, and mild improvements in hemoglobin when taken by five SCD patients.35,36 Patients on Senicapoc, a similar Gardos channel inhibitor, also experienced significant increases in hemoglobin and hematocrit in a phase II clinical trial. Despite these promising results, studies involving Senicapoc were terminated early due to limited efficacy when no significant improvement in the rate of VOE were seen in those taking Senicapoc compared to those on placebo.37 While reductions in MCHC may correlate with reductions in polymerization-induced hemolysis and increased hemoglobin, it does not appear to be as correlated to frequency of VOE and implies that hemolysis and vaso-occlusion are distinct, yet intertwined, pathologic mechanisms. Further, although reductions in MCHC have been shown to decrease rigidity and stiffness of sRBC,38 over-swelling of sRBC by exposure to hypotonic solutions has also led to increased vaso-occlusions in in vitro models.39 Therefore, while the mechanism of osmotic swelling has been beneficial to distinct aspects of SCD pathology, it may not be effective or potentially problematic. Given that previous compounds have attempted to exploit a similar mechanism tested in this study but have ultimately proven unsuccessful due to a lack of reduction in VOE, the ability to determine how a compound affects blood flow, particularly in deoxygenated conditions like that of a VOE, is important in predicting its potential clinical success. Therapeutics that demonstrate improvement in overall blood rheology rather than on a single SCD pathophysiologic mechanism are likely to provide more benefit in reducing VOE frequency. By using microfluidic technology in this study, we can characterize the effect of monensin on velocity response and demonstrate that the mechanism improves rheological behavior. Due to its recognized narrow therapeutic window, use of monensin in veterinary medicine has led to several toxicities and accidental death, and therefore would not be an ideal agent for human study.40 However, this study motivates development of other compounds that may have similar osmotic swelling effects on erythrocytes for SCD treatment.41 A potential concerning side effect of this treatment mechanism includes increased blood viscosity due to increased cell volume and hct. With increased viscosity, the potential for VOE is amplified.41 Viscosity, however, is dependent on several factors such as hct, RBC deformability and aggregation.41,42 Previous studies demonstrate with decreased MCHC, sRBC deformability increases.43 Additionally, it has been reported that cell stiffness has a greater influence on blood viscosity than hct41 and as previously mentioned, monensin treatment has demonstrated increased sRBC deformability when compared to untreated controls.8 Patients treated with hydroxyurea also experience an increase in MCV without association of worsening outcomes.44 Additionally, although our study

A.C. Geisness et al. does not control for post-monensin increase in hct, we still demonstrate improved rheological behavior. Therefore, while osmotic sRBC swelling may increase viscosity, RBC deformability appears to be stronger determining factor in overall blood viscosity. Our study is limited by primarily demonstrating correlative relationships with little experimentation on causation. However, previous work demonstrating increases in RBC deformability and decreases in sRBC fraction with monensin treatment, support the reduced rheological response to hypoxia observed in this study. While we demonstrate that decreasing MCHC by erythrocyte osmotic swelling successfully reduces sRBC oxygen dependent flow, this study does not capture the complex biological interactions between the many other cellular components involved in the pathophysiology of SCD. For example, in our experiments, we used washed red cells rather than whole blood. Previous studies found that monensin had a reduced effect on sRBC in the presence of plasma and required significantly increased concentrations to replicate improved deformability, indicating that the drug likely binds across plasma constituents.19 While using patient plasma may be helpful in determining concentration of monensin to observe rheological improvement, this study was to demonstrate the efficacy of the mechanism of MCHC reduction to improve sRBC rheological behavior. Should further studies exploring this mechanism be conducted, or for future therapeutic development, agents that specifically target RBC Na permeability by RBC-specific cation channels or transport would be of priority to reduce effect on other potential cellular components. Additionally, the endothelium is particularly of interest given that increased RBC deformability has been shown to increase endothelial adhesion and may contribute to the occurrence of VOE and clinical severity.45,46 Further studies examining the effect of MCHC reduction on adhesion and inflammation are needed to assess the potential benefit of this therapeutic mechanism. As discovery of SCD pathophysiology reveals more intricate biological pathways and multifaceted systems simultaneously at play, the approach to treatment may require an equally multifaceted, multi-drug approach. By combining a therapy that reduces MCHC and polymerization with an anti-adhesion therapy, perhaps further benefit may be achieved. Studies are needed to determine the advantages of a multi-agent approach, however similar strategies have already been successful in HIV therapy, cardiology, and oncology. Conclusion The reduction of MCHC through osmotic swelling of a sickle erythrocyte can effectively decrease the rheological dependence on oxygenation. Blood flow velocity measurements within microfluidic channels, of physiologic dimen-

Haematologica | 107 - June 2022

1445

ARTICLE - RBC swelling improves sickle blood flow sions, indicate a strong correlation between MCHC reduction and reduction of blood flow sensitivity to hypoxia. These relationships may indicate the potential efficacy of regulating MCHC as a targeted mechanism for SCD therapeutic development.

A.C. Geisness et al. Acknowledgments We would like to thank Chhaya Patel, Hasmukh Patel, and Yvonne Datta for assistance with blood sample collection according to IRB protocol 2006P000066/PHS and STUDY0000003.

Disclosures No conflicts of interest to disclose.

Funding Portions of this work were conducted in the Minnesota Nano Center, which is supported by the National Science FoundaContributions tion through the National Nano Coordinated Infrastructure ACG, MA, DW, HS, and DCD performed research and data Network (NNCI) under Award Number ECCS-1542202. We analysis; ACG, DW, HS, DCD, DKW designed the studies; ACG, would like to thank the National Heart Lung and Blood InMA, DW, DCD, JMH and DKW wrote the manuscript. stitute for their support under grant no. HL132906.

References 1. Piel FB, Steinberg MH, Rees DC. Sickle cell disease. N Engl J Med. 2017;376(16):1561-1573. 2. Barabino GA, Platt MO, Kaul DK. Sickle cell biomechanics. Annu Rev Biomed Eng. 2010;12:345-367. 3. Sundd P, Gladwin MT, Novelli EM. Pathophysiology of sickle cell disease. Annu Rev Pathol. 2019;14:263-292. 4. Riley TR, Boss A, McClain D, Riley TT. Review of medication therapy for the prevention of sickle cell crisis. P T. 2018;43(7):417-437. 5. Matte A, Zorzi F, Mazzi F, Federti F, Olivieri O, Francheschi LD. New therapeutic options for the treatment of sickle cell disease. Mediterr J Hematol Infect Dis. 2019;11(1):e2019002. 6. Ferrone FA. Polymerization and sickle cell disease: a molecular view. Microcirculation. 2004;11(2):115-128. 7. Hofrichter J, Ross PD, Eation WA. Kinetics and mechanism of deoxyhemoglobin S gelation: a new approach to understanding sickle cell disease. Proc Natl Acad Sci USA. 1974;71(12):4864-4868. 8. Eaton WA, Bunn HF. Treating sickle cell disease by targeting HbS polymerization. Blood. 2017;129(20):2719-2726. 9. McGann PT, Ware RE. Hydroxyurea for sickle cell anemia: what have we learned and what questions still remain? Curr Opin Hematol. 2011;18(3):158-165. 10. Oksenberg D, Dufu K, Patel MP, et al. GBT440 increases haemoglobin oxygen affinity, reduces sickling and prolongs RBC half-life in a murine model of sickle cell disease. Br J Haematol. 2016;175(1):141-153. 11. Vichinsky E, Hoppe CC, Ataga AI, et al. A phase 3 randomized trial of Voxelotor in sickle cell disease. N Engl J Med. 2019;381(6):509-519. 12. Eaton WA, Hofrichter J. Hemoglobin S gelation and sickle cell disease. Blood. 1987;70(5):1245-1266. 13. Rosa RM, Bierer BE, Thomas R, et al. A study of induced hyponatremia in the prevention and treatment of sickle-cell crisis. N Engl J Med. 1980;303(20):1138-1143. 14. Leary M, Abramson N. Induced hyponatremia for sickle-cell crisis. N Engl J Med. 1981;304(14):844-845. 15. Charache S, Walker WG. Failure of Desmopressin to lower serum sodium or prevent crisis in patients with sickle cell anemia. Blood. 1981;58(5):892-896. 16. Stillwell W. Chapter 19: Membrane Transport. An Introduction to Biological Membranes 2nd Edition. 2016;440-443. 17. Freedman JC. Chapter 4 - Ionophores in Planar Lipid Bilayers. Cell Physiology 4th Edition. 2012;61-66. 18. Aowicki D, Huczyński A. Structure and antimicrobial properties

of monensin A and its derivatives: summary of the achievements. Biomed Res Int. 2013;2013:742149. 19. Clark MR, Mohandas N, Shohet SB. Hydration of sickle cells using the sodium ionophore monensin. A model for therapy. J Clin Invest. 1982;70(5):1074-1080. 20. Li Q, Henry ER, Hofrichter J, et al. Kinetic assay shows that increasing red cell volume could be a treatment for sickle cell disease. Proc Natl Acad Sci USA. 2017;114(5):E689-E696. 21. Lu L, Li Z, Li H, Li X, Vekilov PG, Karniadakis GE. Quantitative prediction of erythrocyte sickling for the development of advanced sickle cell therapies. Sci Adv. 2019;5(8):eaax3905. 22. Sonmez M, Ince HY, Yalcin O, et al. The effect of alcohols on red blood cell mechanical properties and membrane fluidity depends on their molecular size. PLoS One. 2013;8(9):e76579. 23. Lee S, Park H. Best-Popescu C, Jang S, Park Y. The effects of ethanol on the morphological and biochemical properties of individual human red blood cells. PLoS One. 2015;10(12):e0145327. 24. Wood DK, Soriano A, Mahadevan L, Higgins JM, Bhatia SN. A biophysical indicator of vaso-occlusive risk in sickle cell disease. Sci Transl Med. 2012;4(123):123ra26. 25. Lu X, Wood DK, Higgins JM. Deoxygenation reduces sickle cell blood flow at arterial oxygen tension. Biophys J. 2016;110(12):2751-2758. 26. Valdez JM, Datta YH, Higgins JM, Wood DK. A microfluidic platform for simultaneous quantification of oxygen-dependent viscosity and shear thinning in sickle cell blood. APL Bioeng. 2019;3(4):046102. 27. Lucas BD, Kanade T. An iterative image registration technique with an application to stereo vision. IJCAI. 1981;674-679. 28. Tomasi C, Kanade T. Detection and tracking of point features. Carnegie Mellon University Technical Report CMU-CS-91-132. 1999 April. 29. Shi J, Tomasi C. Good Features to Track. IEEE Conference on Computer Vision and Pattern Recognition. 1994;593-600. 30. MCHC Blood Test – Low, High, What does it Mean. actforlibraries.org. 2017. 31. Serjeant G, Serjeant B, Stephens A, et al. Determinants of haemoglobin level in steady-state homozygous sickle cell disease. Br J Haematol. 1996:92(1):143-149. 32. Kaul DK, Liu XD. Rate of deoxygenation modulates rheologic behavior of sickle red blood cells at a given mean corpuscular hemoglobin concentration. Clin Hemorheol Microcirc. 1999;21(2):125-135.

Haematologica | 107 - June 2022

1446

ARTICLE - RBC swelling improves sickle blood flow 33. Di Caprio G, Schonbrun E, Goncalves BP, Valdez JM, Wood DK, Higgins JM. High-throughput assessment of hemoglobin polymer in single red blood cells from sickle cell patients under controlled oxygen tension. Proc Natl Acad Sci USA. 2019;116(50):25236-25242. 34. Sosa JM, Nielsen ND, Vignes SM, Tanya GC, Shevkoplyas SS. The relationship between red blood cell deformability metrics and perfusion of an artificial microvascular network. Clin Hemorheol Microcirc. 2014;57(3):275-228. 35. Brugnara C, Gee B, Armsby CC, et al. Therapy with oral clotrimazole induces inhibition of the Gardos channel and reduction of erythrocyte dehydration in patients with sickle cell disease. J Clin Invest. 1996;97(5):1227-1234. 36. Brugnara C. Sickle cell dehydration: pathophysiology and therapeutic applications. Clin Hemorheol Microcirc. 2018;68(23):187-204. 37. Ataga KI, Reid M, Ballas SK, et al. Improvements in haemolysis and indicators of erythrocyte survival do not correlate with acute vaso-occlusive crises in patients with sickle cell disease: a phase III randomized, placebo-controlled, double-blind study of the Gardos channel blocker Senicapoc (ICA-17043). Br J Haematol. 2011;153(1):92-104. 38. Evans E, Mohandas N, Leung A. Static and dynamic rigidities of normal and sickle erythrocytes. Major influence of cell hemoglobin concentration. J Clin Invest. 1984;73(2):477-488.

A.C. Geisness et al. 39. Carden MA, Fay ME, Lu X, et al. Extracellular fluid tonicity impacts sickle red blood cell deformability and adhesion. Blood. 2017;130(24):2654-2663. 40. Connes P, Alexy T, Detterich, J, Romana M, Hardy-Dessources MD, Ballas SK. The role of blood rheology in sickle cell disease. Blood Rev. 2016;30(2):111-118. 41. Novilla MN. The veterinary importance of the toxic syndrome induced by ionophores. Vet Hum Toxicol. 1992;34(1):66-70. 42. Nader E, Skinner S, Romana M, et al. Blood rheology: key parameters, impact on blood flow, role in sickle cell disease and effects of exercise. Front Physiol. 2019;10:1329. 43. Chien S. Determinants of blood viscosity and red cell deformability. Scand J Clin Lab Invest. 1981;156:7-12. 44. Ferster A, Vermylen C, Cornu G, et al. Hydroxyurea for treatment of severe sickle cell anemia: a pediatric clinical trial. Blood. 1996;88(6):1960-1964. 45. Hebbel RP, Moldow CF, Steinberg MH. Modulation of erythrocyte-endothelial interactions and the vaso-occlusive severity of sickling disorders. Blood. 1981;58(5):947-952. 46. Kaul DK, Chen D, Zhan J. Adhesion of sickle cells to vascular endothelium is critically dependent on changes in density and shape of the cells. Blood. 1994;83(10):3006-3007. 47. Sakariassen KS, Orning L, Turitto VT. The impact of blood shear rate on arterial thrombus formation. Future Sci OA. 2015;1(4):FSO30.

Haematologica | 107 - June 2022

1447

LETTER TO THE EDITOR

Direct and indirect anti-leukemic properties of activity-on-target interferons for the treatment of T-cell acute lymphoblastic leukemia T-cell acute lymphoblastic leukemias (T-ALL) are rare aggressive hematologic tumors resulting from the malignant transformation of T-cell progenitors in the thymus.1 T-ALL treatment currently consists of high-dose multi-agent chemotherapy, potentially followed by hematopoietic stem cell transplantation for high-risk patients.2 Unfortunately, even with such a harsh and long treatment regime, the outcome of T-ALL patients with primary refractory or early relapsed leukemia remains poor. Novel therapeutic strategies to treat those high-risk patients remain an important unmet medical need.3 Type-I interferons (IFN-I) have a long history in the treatment of cancer, including hematologic malignancies.4 The pleiotropic anticancer effects induced by IFN-I result from a combination of: (i) direct cancer cell growth inhibition by inducing cell cycle arrest, apoptosis, and differentiation; and (ii) indirect effects by the activation of the immune system involving antigen presentation by dendritic cells and priming of cytotoxic CD8+ T cells.5 As such, this cytokine, with both direct and indirect immunostimulatory anti-cancer effects, is unique in its kind. Nevertheless, IFN-I therapy experienced variable and unpredictable success in the clinic. Its clinical application is severely impeded by a complex pattern of adverse side-effects, due to the multifaceted pattern of activity of IFN-I.6 Therefore, safe exploitation of the anti-cancer potential of IFN-I requires strategies to direct its activity to selected target cells only. To document the direct anti-cancer effect of IFN-I on T-ALL, we stimulated nine human T-ALL cell lines with increasing concentrations of recombinant human interferon alpha-2 (hIFNa2) for 4 days and measured the effect on their in vitro growth. A variable anti-proliferative effect was observed, with no correlation between sensitivity to IFN-I and immunophenotype or molecular subtype (Online Supplementary Figure S1). Consistent with previous studies,7,8 the anti-proliferative effect was associated with activation of the JAK/STAT1 pathway as documented by the increased levels of phosphorylated STAT1 (pSTAT1) determined by flow cytometry. We hypothesized that elevated pSTAT1 levels upon IFN-I stimulation could be used as a biomarker to stratify between IFN-I-sensitive and non-sensitive T-ALL patients. To validate our findings in primary T-ALL patients’ material, we stimulated three patient-derived xenografts (PDX) samples with 100 ng/mL hIFNa2 for 30 min and quantified pSTAT1 levels (Figure 1A). As for the cell lines, a variable response to IFN-I was observed. Subsequently, we transplanted the IFN-I-responsive (PDX#1) and non-responsive (PDX#3) PDX samples

into immunodeficient NOD-scid IL2Rgnull (NSG) mice. As soon as evidence of leukemia progression was observed (>5% human CD45+ [hCD45+] cells in peripheral blood), mice (5 mice/group) were treated for 7 consecutive days with intraperitoneal injections of 30 mg hIFNa2/mouse or vehicle (100 mL phosphate-buffered saline). A significant anti-leukemic effect was observed only for the IFN-I-responsive PDX#1, with a decrease in the percentage of hCD45+ cells in the peripheral blood at the end of the 7-day treatment regime and a prolonged leukemia-free survival of the mice (Figure 1B). Of note, IFNa2 has strict species specificity and human IFNa2 is unable to activate the mouse receptor complex,7,9 indicating that the observed anti-leukemic effects can only be due to cell-intrinsic direct effects on the human leukemia cells. To further document the indirect immune-mediated antileukemic properties of IFNa2, we generated an experimental T-ALL model that could be exploited to monitor leukemia progression both in an immunocompetent background as well as an immunodeficient background. To do this, we intercrossed the Lck-cretg/+;Ptenfl/fl spontaneous murine T-ALL model10 with a ROSA26-eGFP/luciferase reporter line11 on a pure C57BL/6 background and derived eight primary murine T-ALL cell lines from diseased mice (Online Supplementary Figure S1C). As for the human TALL cell lines, a variable anti-proliferative effect with increased percentages of apoptotic cells was observed when murine T-ALL were treated for 3 days in vitro with increasing concentrations of recombinant mouse interferon alpha-2 (mIFNa2) (Online Supplementary Figure S1D, E). Subsequently, we transplanted the murine T-ALL cell lines in both immunodeficient NSG as well as immunocompetent syngeneic mice and treated them with intraperitoneal injections of 30 mg mIFNa2/mice or vehicle for 7 consecutive days (Figure 1C). As the injected T-ALL lymphoblasts express the eGFP-Firefly luciferase reporter from the Rosa26 promoter, leukemic burden could be efficiently monitored using in vivo bioluminescence imaging.11 mIFNa2 treatment resulted in a significant reduction of the leukemic burden, and prolonged survival of both immunodeficient and immunocompetent mice transplanted with murine T-ALL. Notably, the therapeutic efficacy of mIFNa2 was consistently higher in the presence of an intact immune system (Figure 1C), suggesting an additive immune-mediated antileukemic effect. To uncouple these direct and putative indirect anti-leukemic properties of mIFNa2, we used our recently developed cell-type specific Activity-on-Target

Haematologica | 107 - June 2022

1448

LETTER TO THE EDITOR

Figure 1. Anti-leukemic effects of interferon alpha-2 on the progression of human and murine T-cell acute lymphoblastic leukemia in vivo. (A) Flow cytometric analysis of phospho-STAT1 (pSTAT1) levels in three established T-cell acute lymphoblastic leukemia (T-ALL) patient-derived xenografts (PDX) after 30 min. ex vivo stimulation with 100 ng recombinant human interferon alpha-2 (hIFNa2). (B) Direct anti-leukemic effects of recombinant hIFNa2 treatment (7 consecutive days with 30 mg/mouse) versus vehicle on type I interferon-responsive (PDX#1) and non-responsive (PDX#3) T-ALL patient-derived xenograft model transplanted in NSG mice (5 mice/group). Effect on leukemia progression was measured by flow cytometric quantification of the percentage of hCD45+ cells in peripheral blood (left) and Kaplan-Meier survival curves (right). (C) Anti-leukemic effects of recombinant murine interferon alpha-2 (mIFNa2) treatment (7 consecutive days; 30 mg/mouse) versus vehicle control (phosphate-buffered saline, PBS) on in vivo progression of murine T-ALL cell line transplanted in immunodeficient NSG animals or immunocompetent syngeneic C57BL/6 animals. Leukemia burden was quantified via in vivo bioluminescence imaging during treatment (right) and by Kaplan-Meier survival curves (left). Haematologica | 107 - June 2022

1449

LETTER TO THE EDITOR interferons (AcTaferons; AFN).7,12-15 To obtain these AFN, we fused the hIFNa2Q124R, a human IFNa2 mutation that enables binding to the murine interferon receptor complex with a 100-fold reduced affinity compared to wild-type (WT) mIFNa2, to VHH single domain antibodies targeting the murine CD8 (mCD8-AFN). We specifically chose the

mCD8-AFN, as this CD8a epitope is not only present on 40% of T-ALL cases, but also on mouse classical dendritic cell type 1 (cDC1) which are particularly important for eliciting a CD8+ cytotoxic T lymphocyte response to kill cancerous cells upon IFNa2 stimulation.12 Evaluating the therapeutic effects of the mCD8-AFN on both CD8+ versus

Figure 2. mCD8-AFN has direct but no indirect immune-mediated anti-leukemic effects on murine T-cell acute lymphoblastic leukemia. (A) Anti-leukemic effect of mCD8-AFN treatment (7 consecutive days; 30 mg/mouse) versus vehicle (phosphatebuffered saline, PBS) on in vivo progression of CD8– and CD8+ murine T-cell acute lymphoblastic leukemia (T-ALL) cell lines transplanted in immunodeficient NSG mice or (B) in syngeneic C57BL/6 mice. Leukemic burden was quantified via in vivo bioluminescence imaging during treatment regime (right) and Kaplan-Meier survival curves (left). Haematologica | 107 - June 2022

1450

LETTER TO THE EDITOR

Figure 3. mClec9A-AFN has indirect but no direct anti-leukemic effects on murine T-cell acute lymphoblastic leukemia. (A) Anti-leukemic effect of mClec9A-AFN treatment (7 consecutive days; 30 mg/mouse) versus vehicle (phosphate-buffered saline, PBS) on in vivo progression of CD8– and CD8+ murine T-cell acute lymphoblastic leukemia (T-ALL) cell lines transplanted in immunodeficient NSG mice or (B) in syngeneic C57BL/6 mice. Leukemic burden was quantified via in vivo bioluminescence imaging during treatment regime (right) and Kaplan-Meier survival curves (left).

CD8– T-ALL in an immunocompetent versus immunodeficient background, allowed us to dissect the roles of the direct versus indirect anti-leukemic actions of IFNa2. To compare our results to those of a WT IFNa2 with a similar molecular weight as the AFN, we coupled this cytokine to a single-domain antibody targeting BcII10, an epitope that is absent in mice (mIFNa2-WT). As expected, the mCD8AFN had only an anti-proliferative effect on the CD8+ TALL and not on the CD8– T-ALL cell lines in vitro (Online Supplementary Figure S2A-C), confirming its cell-specificity. To evaluate the cell-specific direct anti-leukemic properties of mCD8-AFN on the growth of murine CD8+ T-ALL in an in vivo context, we transplanted both CD8+ and CD8– T-ALL cell lines into immunodeficient NSG mice. As expected, the mCD8-AFN had a significant direct anti-leukemic effect only on the CD8+ T-ALL cell line, with a significant decrease of bioluminescence during the treatment and prolonged survival (Figure 2A). Of note, the direct antileukemic effect of the mCD8-AFN was more efficient than that of WT mIFNa2 (Online Supplementary Figure S3). As

AFN do not efficiently bind the ubiquitously expressed interferon receptor complex, they will not be cleared from the circulation before reaching their desired target population. This so-called ‘sink-effect’ can explain the improved direct anti-leukemic effect seen with the AFN. As the CD8a epitope of the mCD8-AFN is also present on cDC1 cells, which have been shown to be superior in antigen cross-presentation, and on effector CD8+ T cells necessary to kill tumor cells, we hypothesized that the mCD8-AFN would also be able to induce an indirect antileukemic response. To test this, we repeated the experiment in an immunocompetent background. As an indirect immune-mediated effect would be independent of the immunophenotype of the T-ALL cells, we hypothesized that, in this setting, we would observe an anti-leukemic effect of the mCD8-AFN on both CD8– and CD8+ T-ALL. However, no effect could be seen on the CD8– T-ALL, only on the CD8+ T-ALL (Figure 2B), indicating that mCD8-AFN has only direct anti-leukemic T-ALL properties, and is not able to induce an indirect immune-mediated anti-leukemic response.

Haematologica | 107 - June 2022

1451

LETTER TO THE EDITOR The inability of mCD8-AFN to elicit an immune-mediated anti-leukemic response could be either due to mature T-cell aplasia and leukemia progression in these mice, or due to the intrinsic nature of the AFN. We therefore tested the anti-leukemic properties of a second AFN, mClec9a-AFN, which we previously used to efficiently induce a cDC1-mediated anti-tumor response in the context of melanoma, breast carcinoma and lymphoma.12 As Clec9A is not expressed in normal or malignant T cells, no direct anti-leukemic properties could be observed in vitro (Online Supplementary Figure S2A-C) or in vivo in immunodeficient NSG mice (Figure 3A). To evaluate the indirect immune-mediated anti-leukemic effects of mClec9A-AFN, we repeated the experiment in syngeneic C57BL/6 mice (7 mice/group). This time, a significant indirect effect on leukemic progression could be seen in the transplanted mice, irrespective of whether they were transplanted with CD8+ or CD8– T-ALL (Figure 3B). These data demonstrate that mClec9A-AFN is able to induce a strong immune-mediated anti-tumor response. However, we currently cannot explain why we saw a difference between the ability of mCD8-AFN and mClec9A-AFN to induce an indirect immune-mediated anti-leukemic effect. In the mouse, the Clec9A-epitope is also present on CD8– immune cell subtypes, which may contribute to the immune response. Alternatively, binding of the CD8AFN on cytotoxic CD8+ T cells may (partially) hamper their cytotoxic functions, although the anti-CD8 singledomain antibody that was used for the design of the mCD8-AFN was demonstrated to be a non-neutralizing antibody.15 Finally, the adverse side-effects (body weight loss and hematologic deficiencies) observed in mice treated with mIFNa2-WT were significantly reduced in mice treated with the AcTaferons (Online Supplementary Figure S2D, E), as shown before.7,12,13 In summary, we used preclinical mouse models of T-ALL to show that immunocytokines with cell-specific activity can preserve the anti-leukemic properties of IFNa2 with a concomitant reduction of systemic toxic side-effects. As such, these AcTaferons can be used as a direct anti-leukemic agent in combination with classical chemotherapy, or as off-the-shelf targeted immunotherapy for T-cell malignancies.

Smedt,1,3 Sara T’Sas,1,3 André Almeida,1,3 Willem Daneels,1,6 Pieter Van Vlierberghe1,3,# and Jan Tavernier1,3,4,5,# 1

Cancer Research Institute Ghent (CRIG), Ghent University;

Department of Diagnostic Sciences, Ghent University; 3Department

of Biomolecular Medicine, Ghent University; 4VIB-UGent Center for Medical Biotechnology; 5Orionis Biosciences BV and 6Department of Hematology, Ghent University Hospital, Ghent, Belgium *SG and AC contributed equally as co-first authors. #

PVV and JT contributed equally as co-last authors.

Correspondence: PIETER VAN VLIERBERGHE - pieter.vanvlierberghe@ugent.be https://doi.org/10.3324/haematol.2021.278913 Received: April 5, 2021. Accepted: July 2, 2021. Prepublished: October 14, 2021. Disclosures JT and AC are affiliated with Orionis Biosciences. JT holds equity interests and receives financial research support from Orionis Biosciences. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Contributions SG, AC, PVV and JT conceived and designed the experiments and wrote the paper. SG, AJ, TP, ST and RDS performed experiments. SG and ST analyzed data. SG, TP and AA generated the mouse T-cell acute lymphoblastic leukemia cell lines. AC and JT produced the cytokines. SG, AC, WD, PVV and JT discussed results and critically reviewed the manuscript. SG and PVV directed and supervised the research. All authors approved the final manuscript. Acknowledgments We thank Reza Hassanzadeh Ghassabeh (VIB Nanobody Core) for the selection of the VHH. JT is a recipient of UGent Methusalem and ERC Advanced (CYRE, n. 340941) grants. Funding This work was supported by the Stichting ME TO YOU, the Ghent University Research Fund (BOF-UGent) and the Research Foundation Flanders (FWO).

Authors

Data-sharing statement Original data can be made available in response to a reasonable, 1,2,3,*

Steven Goossens,

1,3,4,5,*

Anje Cauwels,

1,3

Tim Pieters,

Renate De

written request to the corresponding author.

References 1. Meijerink JP. Genetic rearrangements in relation to immunophenotype and outcome in T-cell acute lymphoblastic leukaemia. Best Pract Res Clin Haematol. 2019;23(3):307-318.

2. Hofmans M, Suciu S, Ferster A, et al. Results of successive EORTCCLG 58 881 and 58 951 trials in paediatric T-cell acute lymphoblastic leukaemia (ALL). Br J Haematol. 2019;186(5):741-753.

Haematologica | 107 - June 2022

1452

LETTER TO THE EDITOR 3. Hunger SP, Mullighan CG. Acute lymphoblastic leukemia in children. N Engl J Med. 2015;373(16):1541-1552. 4. Parker BS, Rautela J, Hertzog PJ. Antitumour actions of interferons: implications for cancer therapy. Nat Rev Cancer. 2016;16(3):131-144. 5. Demerdash Y, Kain B, Essers MAG, King KY. Yin and yang: the dual effects of interferons on hematopoieisis. Exp Hematol. 2021;96:1-12. 6. Jonasch E, Haluska FG. Interferon in oncological practice: review of interferon biology, clinical applications, and toxicities. Oncologist. 2001;6(1):34-55. 7. Garcin G, Paul F, Staufenbiel M, et al. High efficiency cellspecific targeting of cytokine activity. Nat Commun. 2014;5:3016. 8. Lesinski GB, Anghelina M, Zimmerer J, et al. The antitumor effects of IFN-alpha are abrogated in a STAT1-deficient mouse. J Clin Invest. 2003;112(2):170-180. 9. Weber H, Valenzuela D, Lujber G, Gubler M, Weissmann C. Single amino acid changes that render human IFN-a2 biologically active on mouse cells. EMBO J. 1987;6(3):591-598.

10. Suzuki A, Yamaguchi MT, Ohteki T, et al. T cell-specific loss of Pten leads to defects in central and peripheral tolerance. Immunity. 2001;14(5):523-534. 11. Pieters T, T’Sas S, Demoen L, et al. Novel strategy for rapid functional in vivo validation of oncogenic drivers in haematological malignancies. Sci Rep. 2019;9(1):10577. 12. Cauwels A, Van Lint S, Paul F, et al. Delivering type I interferon to dendritic cells empowers tumor eradication and immune combination treatments. Cancer Res. 2018;78(2):463-474. 13. Cauwels A, Van Lint S, Garcin G, et al. A safe and highly efficient tumor-targeted type I interferon immunotherapy depends on the tumor microenvironment. Oncoimmunology. 2017;7(3):e1398876. 14. Cauwels A, Van Lint S, Catteeuw D, et al. Targeting interferon activity to dendritic cells enables in vivo tolerization and protection against EAE in mice. J Autoimmun. 2019;97:70-76 15. Huyghe L, Van Parys A, Cauwels A, et al. Safe eradication of large established tumors using neovasculature-targeted tumor necrosis factor-based therapies. EMBO Mol Med. 2020;12(2):e11223.

Haematologica | 107 - June 2022

1453

LETTER TO THE EDITOR

A third anti-SARS-CoV-2 mRNA dose does not overcome the pejorative impact of anti-CD20 therapy and/or low immunoglobulin levels in patients with lymphoma or chronic lymphocytic leukemia Patients with non-Hodgkin lymphoma (NHL) or chronic lymphocytic leukemia (CLL) share immune deficiencies due to the biological features of these diseases and their treatment. They are at risk of developing severe and/or prolonged forms of coronavirus disease 2019 (COVID-19).1,2 Moreover, their seroconversion rate after infection by the etiological agent, severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) or after vaccination is low, in particular, after recently receiving anti-CD20 monoclonal antibodies.3,4 Addition of a third vaccine dose is, therefore, recommended for non-responding patients or those with a low serological response. Here, we show that the addition of a third dose does not modify the proportion of seropositivity after vaccination in a population of patients with NHL/CLL when treated with anti-CD20 therapies. The antibody response to mRNA SARS-CoV-2 vaccine in NHL/CLL patients has been shown to be negatively affected by their treatments, in particular, when treatment is ongoing during the vaccination. In a recent cohort of nearly 400 CLL patients, the seropositivity rate was 61% in treatment-naïve patients versus 23% in patients treated with Bruton tyrosine kinase (BTK) inhibitors, 24% in those treated with Bcl2 inhibitors, and, remarkably, only 5% in those who received anti-CD20 monoclonal antibodies in the year before vaccination.4 The results were similar among patients with NHL.5,6 In a cohort of 162 lymphoma patients, the proportion of seropositivity was 51% overall and decreased to 12% within the first 12 months following treatment with anti-CD20 antibodies.5 Boosting solid organ transplanted patients, who are at risk of severe COVID-19 after two doses of vaccine,7 with a third dose of mRNA vaccine was shown to significantly improve their immune response (seropositivity of 55% after third dose versus 18% in a placebo group).8.9 Although, the addition of a third dose of vaccine for immunocompromised patients has been recommended in France since April 2021, there has been no evaluation of this strategy in patients with hematologic malignancies. We therefore conducted a single-center study in our institution to investigate determinants of the antibody response to anti-SARS-CoV-2 Pfizer-BioNTech® BNT162b2 or Moderna® mRNA-1273 vaccines among adults with a current or past diagnosis of lymphoma/CLL. The addition of a third dose was at the discretion of each physician. The two-dose vaccination courses were Pfizer (n=58) and

Moderna (n=6) and the three-dose vaccination courses were Pfizer (n=35) and Moderna/Moderna/Pfizer (n=1). The median interval between the first and second vaccine doses was 28 days while that between the second and third vaccine doses was 61 days. Serology was performed at least 2 weeks after the last vaccination. Data were extracted from the patients’ medical charts and included their demographics, lymphoma/CLL history and treatment, vaccination dates and biological variables (immunoglobulin [Ig] G level, total lymphocytes, B-cell, and T-cell counts [determined by a Beckman Coulter® Aquios Tetra system], and SARS-CoV-2 anti-spike serology (determined by ECLIA Elecsys anti-SARS-CoV2 S, Roche®]). Negative serological tests were defined as antibody titers <0.8 U/mL. The patients’ baseline characteristics are presented using medians and interquartile ranges (IQR) or ranges for quantitative variables and frequencies and percentages for categorical variables. The association between patients' serological status and other variables was assessed using a Wilcoxon or Fisher test, as appropriate, and univariate logistic regression. Multivariate logistic regression was performed for variables for which the univariate two-sided significance was <5% and for which there was a minimum of two patients per category for categorical variables (for reasons of algorithmic convergence). Statistical analysis was performed using R and the environment for statistical computing version 4·03 (R Core Team, 2020). The study was conducted in accordance with the Declaration of Helsinki, and approved by the ethics committee of our institution. In total, 100 patients with B-cell NHL (n=51), T-cell NHL (n=4), CLL (n=33) or Hodgkin lymphoma (HL) (n=12) were analyzed (Table 1). The median age of the patients was 66.6 (range, 19.6-92.2) years and the female/male ratio was 0.43. The median interval between serology and the last vaccine injection (second or third) was 47 days. Fiftyone patients (51%) showed negative serology, with the proportion of seronegativity being 55%, 50%, 67% and 39% in patients with B-cell NHL, T-cell NHL, HL and CLL, respectively (P=0.35). Ten patients had a history of COVID-19, and of these, three needed hospitalization. One of these patients developed COVID-19 after failing to seroconvert following his second vaccine injection. Patients with negative serology had significantly lower lymphocyte

Haematologica | 107 - June 2022

1454

LETTER TO THE EDITOR Table 1. Characteristics of patents with lymphoma/chronic lymphocytic leukemia according to their serological response to anti-SARS-CoV-2 vaccination.

Positive serology, n (%) 49 (49.0)

Negative serology, n (%) 51 (51.0)

66.9 (54.9-76.9)

66.6 (54.9-76.9)

67.0 (56.3-77.0)

Female

15 (50.0)

Male

34 (48.6)

36 (51.4)

40 (45.1)

50 (54.9)

Yes

9 (90.0)

1 (10.0)

34 (53.1)

30 (46.9)

15 (41.7)

21 (58.3)

47.5 (33.0-78.25)

53.0 (37.0- 84.0)

45.0 (28.5-72.0)

B-NHL

23 (45.1)

28 (54.9)

T-NHL

2 (50.0)

4 (33.3)

8 (66.7)

CLL

20 (60.6)

13 (39.4)

≥1.5x109/L

3.7 (2.2-4.9) 81

3.8 (2.83-5.4) 41 (50.6)

3.4 (1.9-4.7) 40 (49.4)

<1.5x109/L

2 (20.0)

8 (80.0)

≥1.5x109/L

1.0 (0.6-2.0) 35

1.4 (0.7-2.3) 23 (65.7)

0.8 (0.5-1.4) 12 (34.3)

<1.5x109/L

20 (36.4)

35 (63.6)

0.6 (0.3-1.2)

0.7 (0.4-1.4)

0.5 (0.3-1.1)

All patients, n. 100

OR (95% CI) P value Univariable

Multivariable model 1

model 2

0.09 (0.00-0.50) P=0.02

1.59 (0.70-3.67) P=0.27

1.41 (0.47-4.32) P=0.54

1.32 (0.46-3.77) P=0.61

Demographic characteristics Age in years, median (IQR)

1.00 (0.97-1.02) P=0.76

Gender 1.06 (0.45-2.50) P=0.90

Previous COVID-19

N. of vaccine injections

N. of days between first vaccination and serology, median (IQR)

1.00 (0.99-1.00) P=0.30

Lymphoma characteristics reference 0.82 (0.09-7.28) P=0.85 1.64 (0.46-6.80) P=0.46 0.53 (0.22-1.29) P=0.17

Biological characteristics Neutrophils* Median (IQR)

reference 4.10 (0.96-8.29) P=0.09

Lymphocytes** Median, IQR

T cells*** Median, IQR

reference 3.35 (1.40-8.37) P<0.01

Table 1. Continued on following page Haematologica | 107 - June 2022

1455

LETTER TO THE EDITOR <1.1x109/L

13 (54.2)

11 (45.8)

reference 1.42 (0.54-3.77) P=0.48

<1.1x109/L

25 (45.5)

30 (54.5)

B cells^ Median, IQR

0.0

(0.0-0.2)

(0.0-0.9)

(0.0-0.0)

≥0.2x109/L

16 (88.9)

2 (11.1)

reference

<0.2x109/L

23 (37.1)

39 (62.9)

13.57 (3.45-90.84) P=0.001

10.73 (2.48-76.28) P=0.005

7.0 (5.1-9.3)

8.2 (6.7-10.8)

5.5 (4.6-7.6)

≥6 g/L

35 (64.8)

19 (35.2)

reference

<6 g/L

8 (22.2)

28 (77.8)

6.45 (2.55-17.80) P<0.001

5.51 (1.88-18.04) P=0.003

7.33 (2.64-22.64) P<0.001

40 (60.6)

26 (39.4)

Yes

9 (26.5)

25 (73.5)

35 (47.3)

39 (52.7)

reference

IgG^^ Median, IQR

Treatment history within preceding 1 year B-cell depleting agent° reference 4.27 (1.77-11.05) P=0.002

reference 6.25 (2.15-20.36) P<0.001

Targeted therapy°° No

14 (53.8)

12 (46.2)

0.77 (0.31-1.88) P=0.566

29 (53.7)

25 (46.3)

reference

Yes

20 (43.5)

26 (56.5)

1.51 (0.69-3.35) P=0.309

20 80

17 (85.0) 32 (40.0)

3 (15.0) 48 (60.0)

Yes Chemotherapy°°°

Any therapy (among shown) No Yes

reference 8.50 (2.60-8.56) P=0.001

Missing values: *9, **10, ***21, ^20, ^^26. °Rituximab (n=30), obinutuzumab (n=4); °°ibrutinib and/or venetoclax (n=20), masitinib (n=1), pembrolizumab (n=3), lenalidomide (n=2); °°°R-CHOP and R-CHOP-like (n=39), R-DHAX (n=7); R-bendamustine (n=7), ABVD (n=6), BEACOPP (n=4), rituximab/fludarabin/cyclophosphamide (n=4), the total exceeds the number of patients since some patients received more than one line of chemotherapy. OR: odds ratio; 95% CI: 95% confidence interval; IQR: interquartile range; COVID-19: coronavirus disease 2019; NHL: non-Hodgkin lymphoma; HL. Hodgkin lymphoma; CLL: chronic lymphocytic leukemia; NA: not available because of the absence of algorithmic convergence.

counts (median 0.8x109/L versus 1.4x109/L; P=0.006), Bcell counts (median 0.00x109/L versus 0.04x109/L; P<0.001) and IgG levels (median 5.5x109/L versus 8.2x109/L; P<0.001) than seropositive patients. They were also less likely to have a history of COVID-19: odds ratio (OR)=0.09, 95% confidence interval (95% CI): 0.00-0.50; P=0.02. Patients who had received any treatment within the year before their first vaccine injection (anti-CD20 [n=34], chemotherapy [n=46], or targeted therapy [n=26]) had a

higher risk of seronegativity than other patients (n=20) (OR=8.50 [95% CI: 2.60-38.56]; P=0.001). Among treatments, anti-CD20 was strongly associated with negative serology (OR=4.27, [95% CI: 1.77-11.05]; P=0.002). Time between first vaccine injection and anti-CD20 administration was significantly associated with patients’ serological status (P<10-3). Most patients who had their last antiCD20 administration within 1 year prior to first vaccine injection did not seroconvert (25 out of 34, 74%). Out of

Haematologica | 107 - June 2022

1456

LETTER TO THE EDITOR

Figure 1. Time interval between the first vaccine injection and the last (if prior to vaccination) or first (if after vaccination) antiCD20 administration. Time is represented by vertical bars and the color code represents patients’ serological status. Patients who had their last anti-CD20 administration before their first vaccine injection are represented at the left hand side of the vertical dotted line. Patients who had their last anti-CD20 administration after their first vaccine injection are represented at the right hand side of the vertical dotted line. The color code of the vertical bars represents patients’ serological status. The total number of vaccine doses received prior to serology is represented by colored points (2 doses or 3 doses).

these 25 negative patients, 12 (48%) had three vaccine injections (Figure 1). Among the nine patients who seroconverted, four had received three injections (44%). Among the 14 patients who had their first anti-CD20 administration after their first vaccine injection, seven did not seroconvert. Of note these seven patients had their first anti-CD20 administration within 45 days after their first vaccine injection. Overall, there was no association between the number of vaccine injections and seroconversion: 58.3% of patients who had three doses remained seronegative versus 46.9% of patients who had two doses (OR=1.59 [95% CI: 0.70-3.67]; P=0.27). The titers of anti-spike serology according to anti-CD20 status and the number of vaccine administrations are presented in Figure 2. Regardless of the number of vaccine administrations, patients who had their last anti-CD20 administration within the year before the first vaccine injection had low anti-spike titers (median titer: 0.4 U/mL), patients who had their first antiCD20 administration after their first vaccine injection had intermediate titers of anti-spike (median titer: 1.0 u/mL), while other patients had higher levels of anti-spike (median titer: 47.6 U/mL). As anti-CD20 administration and B-cell counts correlated strongly, we performed

multivariable analyses using two models, model 1 based on B-cell counts, and model 2, based on anti-CD20. Low IgG levels (OR=7.33, [95% CI: 2.64-22.64]; P<0.001) and recent anti-CD20 treatment (OR=6.25, [95% CI: 2.15-20.36]; P<0.001) were both associated with seronegativity. The results were similar with low B-cell counts (Table 1). Overall, recent anti-CD20 therapy (low B-cell counts) and low IgG levels are the main independent factors associated with a poor serological response after anti-SARSCoV-2 mRNA vaccination in CLL/lymphoma patients. Administration of a third vaccine dose did not overcome the poor serological response observed in patients who had anti-CD20 treatment within 1 year prior to their first vaccine injection or low IgG levels. However, due to its single-center character, the limited number of patients and the heterogeneity of the lymphoid malignancies described, results of our study should be confirmed by other studies. Moreover, clinical studies examining the impact of the lack of serological conversion on the incidence and severity of COVID-19 are needed. Indeed, several studies have reported the presence of T-cell responses after vaccination in patients treated with rituximab for autoimmune diseases,10-12 despite an altered humoral response and low seroconversion levels. Never-

Haematologica | 107 - June 2022

1457

LETTERS TO THE EDITOR

Figure 2. Anti-spike titers with respect to anti-CD20 treatment and the number of vaccine injections. The natural logarithm of the anti-spike titer is represented with respect to anti-CD20 treatment status. Red: patients without a history of anti-CD20 treatment or patients who had their last administration more than 1 year prior to the first vaccine injection; green: patients who had their last anti-CD20 administration within 1 year prior to the first vaccine injection; and (blue) patients who had their first vaccine injection prior to their first anti-CD20 administration.

theless, the clinical benefit of the cellular immune response in COVID-19 is yet to be elucidated. The minimum time between the last dose of anti-CD20 and vaccination3,5 to observe patients’ serological response remains controversial: in line with our results previously published studies have suggested a minimum of 12 months whereas others have reported a time interval of 3 months.13 In the meantime, strategies other than increasing the number of vaccine injections are needed to protect patients with B-cell depletion against COVID-19. First, SARS-CoV-2 vaccination should be proposed before the onset of lymphoma/CLL therapy in all non-critical clinical situations. Strategies of maintenance treatment in follicular and mantle-cell lymphoma with anti-CD20 therapy should be discussed for each patient, depending on comorbidities, disease status and response to vaccination. Individuals with CLL/lymphoma should continue social distancing and barrier measures. Systematic vaccination of their entourage and hospital workers would also directly benefit patients. If COVID-19 infection occurs after vaccination, convalescent plasma therapy14 or neutralizing monoclonal antibodies against SARS-CoV-215 could be discussed as future strategies for CLL/lymphoma patients with a low humoral immune response. Finally, prophylactic administration of neutralizing monoclonal antibodies is currently recommended in France for vaccinated seronegative patients and should also be evaluated.

Authors Milena Kohn,1* Marc Delord,2* Maureen Chbat,1* Amina Guemriche,1 Fatiha Merabet,1 Anne-Laure Roupie,1 Naelle Lombion,3 Hassan Farhat,1 Thomas Longval,1 Aurélie Cabannes-Hamy,1 Juliette Lambert,1 Stéphanie Marque-Juillet,4 Victoria Raggueneau,4 Jennifer Osman,4 Marc Spentchian,4 Sophie Rigaudeau,1 Philippe Rousselot1 and Caroline Besson1.5 1

Service d’Hématologie Oncologie, Centre Hospitalier de Versailles,

Le Chesnay; 2DRCI, Centre Hospitalier de Versailles, Le Chesnay; 3

Service d’hématologie, Hôpital de Poissy, Poissy; 4Service de

Biologie, Centre Hospitalier de Versailles, Le Chesnay and 5UVSQ, Inserm, CESP, Villejuif, France *MK, MD and MC contributed equally as co-first authors. Correspondence: Caroline Besson - cbesson@ch-versailles.fr https://doi.org/10.3324/haematol.2021.280026 Received: September 17, 2021. Accepted: November 4, 2021. Prepublished: November 18, 2021. Disclosures No conflicts of interest to disclose.

Haematologica | 107 - June 2022

1458

LETTERS TO THE EDITOR Contributions MK, MC, AG, FM, ALR, NL, HF, TL ACH, JL, SR and PR included the patients, collected the data, and approved the article. MD performed the statistical analysis. CB and MD designed the study. MK, MD and CB wrote the article. SMJ, JO, VR and MS performed the biological analyses.

Acknowledgments We thank the Centre Hospitalier de Versailles for their contribution to editing.

References 1. Duléry R, Lamure S, Delord M, et al. Prolonged in-hospital stay and higher mortality after Covid-19 among patients with nonHodgkin lymphoma treated with B-cell depleting immunotherapy. Am J Hematol. 2021;96(8):934-944. 2. Mato AR, Roeker LE, Lamanna N, et al. Outcomes of COVID-19 in patients with CLL: a multicenter international experience. Blood. 2020;136(10):1134-1143. 3. Herishanu Y, Avivi I, Aharon A, et al. Efficacy of the BNT162b2 mRNA COVID-19 vaccine in patients with chronic lymphocytic leukemia. Blood. 2021;137(23):3165-3173. 4. Benjamini O, Rokach L, Itchaki G, et al. Safety and efficacy of BNT162b mRNA COVID-19 vaccine in patients with chronic lymphocytic leukemia. Haematologica. 2022;107(3):625-634. 5. Gurion R, Rozovski U, Itchaki G, et al. Humoral serologic response to the BNT162b2 vaccine is abrogated in lymphoma patients within the first 12 months following treatment with anti-CD20 antibodies. Haematologica. 2022;107(3):715-720. 6. Perry C, Luttwak E, Balaban R, et al. Efficacy of the BNT162b2 mRNA COVID-19 vaccine in patients with B-cell non-Hodgkin lymphoma. Blood Adv. 2021;5(16):3053-3061. 7. Caillard S, Chavarot N, Bertrand D et al. Occurrence of severe COVID-19 in vaccinated transplant patients. Kidney Int. 2021;100(2):477-479. 8. Kamar N, Abravanel F, Marion O, Couat C, Izopet J, Del Bello A. Three doses of an mRNA Covid-19 vaccine in solid-organ tran-

splant recipients. N Engl J Med. 2021;385(7):661-662. 9. Hall VG, Ferreira VH, Ku T, et al. Randomized trial of a third dose of mRNA-1273 vaccine in transplant recipients. N Engl J Med. 2021;385(13):1244-1246. 10. Prendecki M, Clarke C, Edwards H, et al. Humoral and T-cell responses to SARS-CoV-2 vaccination in patients receiving immunosuppression. Ann Rheum Dis. 2021;80(10):1322-1329. 11. Benucci M, Damiani A, Infantino M, et al. Presence of specific T cell response after SARS-CoV-2 vaccination in rheumatoid arthritis patients receiving rituximab. Immunol Res. 2021;69(4):309-311. 12. Mrak D, Tobudic S, Koblischke M, et al. SARS-CoV-2 vaccination in rituximab-treated patients: B cells promote humoral immune responses in the presence of T-cell-mediated immunity. Ann Rheum Dis. 2021;80(10):1345-1350. 13. Marchesi F, Pimpinelli F, Giannarelli D, et al. Impact of antiCD20 monoclonal antibodies on serologic response to BNT162b2 vaccine in B-cell non-Hodgkin's lymphomas. Leukemia. 2022;36(2):588-590. 14. Hueso T, Pouderoux C, Péré H, et al. Convalescent plasma therapy for B-cell-depleted patients with protracted COVID-19. Blood. 2020;136(20):2290-2295. 15. Weinreich DM, Sivapalasingam S, Norton T, et al. REGN-COV2, a neutralizing antibody cocktail, in outpatients with Covid-19. N Engl J Med. 2021;384(3):238-251.

Haematologica | 107 - June 2022

1459

LETTER TO THE EDITOR

The highly selective Bruton tyrosine kinase inhibitor acalabrutinib leaves macrophage phagocytosis intact Treatment of chronic lymphocytic leukemia (CLL), has been transformed by the Bruton tyrosine kinase inhibitor (BTKi) ibrutinib (IBR). Although BTKi treatment mobilizes CLL cells from lymphoid organs into the circulation,1 where they are highly susceptible to clearance by antiCD20 monoclonal antibodies (mAb),2 randomized clinical trials showed no advantage of adding the anti-CD20 mAb rituximab to IBR for treatment of CLL.2,3 IBR inhibition of antibody-dependent cellular phagocytosis (ADCP) by splenic and liver macrophages, the principal mechanism of action of anti-CD20 mAb,4 might explain this result. Macrophage ADCP requires immunoglobulin Fcg receptors (FcR) binding to mAb-opsonized target cells. Because FcR signaling may involve BTK,5 BTKi could prevent the therapeutic effects of anti-CD20 mAb. Indeed, studies have found that IBR decreases ADCP in vitro.6,7 In contrast, the highly selective BTKi acalabrutinib (ACALA) with fewer offtargets than IBR ( ~5 vs. >20)8, 9 does not significantly decrease in vitro ADCP.6,7 We hypothesized that suppression of ADCP by IBR is mediated by off-target inhibition unique to IBR. (Figure 1A). ACALA also causes less lymphocytosis than IBR,10 leading to the hypothesis that IBR, but not ACALA, inhibits phagocytosis of apoptotic cells (efferocytosis)11 by a more generalizable off-target effect on phagocytosis. In order to test these hypotheses, we measured human monocyte-derived macrophage (hMDM) phagocytic engulfment and processing of CLL cell targets in vitro.7, 12 Human specimen collection and usage was conducted with written informed consent after approval of the University of Rochester Research Subjects Review Board according to the ethical guidelines of the Declaration of Helsinki. We examined the effect of IBR or ACALA on hMDM ADCP of anti-CD20 mAb opsonized CLL cells in real time using live cell time-lapse video imaging that visualizes phagocytic engulfments as “voids” in Cell Tracker Deep Red (CTDR, Thermo Fisher Scientific)-labeled hMDM (Online Supplementary Figure S1).12 Images of 10 mg/mL rituximab-mediated ADCP of CLL cells by CTDR-labeled hMDM (20:1 CLL:hMDM ratio) with or without IBR or ACALA were captured every 4 minutes (min) over 2.8 hours (h) in 18 replicate experiments. In order to enable detection of off-target effects, the range of BTKi concentrations (serial dilutions from 0.41-100 mM) was chosen to encompass and span above the mean clinical peak free drug concentration (Cmax is ~0.5 mM and ~1.2 mM for IBR and ACALA, respectively1,10). Imaging showed a visually apparent reduction in ADCP after 60 min with IBR but not ACALA (Figure 1B). Further examples of ADCP time-lapse images

and videos of IBR or ACALA treatment with combined CTDR and Phase channels or CTDR channel alone are available in data sharing statement (Supplementary Figure DS1; Supplementary Videos DS5 to 8). ADCP engulfment events were quantified by void index and graphically displayed as previously described.12,13 Rituximab alone caused the ADCP void index to rapidly increase and approach maximum by 1 h, which is the initial engorgement phase observed during ADCP kinetics (0 mM, Figure 1C; Online Supplementary Figure S2A).13 IBR exhibits a concentrationdependent inhibition of ADCP from 0.41-100 mM during this phase, while ACALA only inhibited at the highest 100 mM concentration (Figure 1C; Online Supplementary Figure S2A). In order to analyze this kinetic inhibition, ADCP was summarized as area under the curve (AUC) in the first hour of treatment. Serial dilutions of each drug compared to untreated showed that IBR significantly inhibited ADCP at all measured concentrations (0.41-100 mM, P<0.05; Figure 1D). In contrast, ACALA did not significantly inhibit ADCP at concentrations <100 mM (P>0.05; Figure 1D). Moreover, comparison of relative ACALA versus IBR inhibition as a ratio showed significantly higher inhibition by IBR at all measured drug concentrations (0.41-100 mM, P<0.01; Figure 1E), which may be slightly overestimated due to an unexpected increase in ADCP at low ACALA concentrations (Figure 1C). These results confirm and extend previous studies using indirect or semi-quantitative single time point observations that suggest that IBR but not ACALA inhibits ADCP.6,7,14 Because ACALA is a more selective inhibitor of BTK, these data imply that BTK inhibition is not responsible for the decreased ADCP measured in hMDM treated with IBR. Inhibition of ADCP by IBR is likely the result of IBR-specific off-target effects (Figure 1A). In order to determine if IBR specific off-target effects broadly alter phagocytosis, we studied the effect of IBR and ACALA on hMDM efferocytosis, an antibody-independent form of phagocytosis that does not involve BTK (Figure 1A).11 Initial measurements of efferocytosis were done by flow cytometry at a single time point. Treatment of hMDM with IBR or ACALA at concentrations ranging from 1.25-10 mM showed no significant effect of either drug on the percentage of hMDM efferocytosis (P>0.05; Figure 2A). In order to determine if either drug had any effects on the kinetics of efferocytosis, we used the live cell time-lapse video imaging approach using pHrodo Red-labeled apoptotic CLL cells as targets for phagocytosis by CTDR-labeled hMDM (20:1 CLL:hMDM ratio) either untreated or with IBR or ACALA (2-fold serial dilutions from 10-1.25 mM). Images of duplicate or triplicate wells for each drug con-

Haematologica | 107 - June 2022

1460

LETTER TO THE EDITOR

Figure 1. Bruton tyrosine kinase inhibitor effects on phagocytosis and antibody-dependent cellular phagocytosis measurements. (A) Schematic of ibrutinib (IBR) or acalabrutinib (ACALA) inhibition of phagocytosis (antibody-dependent cellular phagocytosis [ADCP] or efferocytosis) shows ADCP (left) proceeding via target cell antigen (Ag) binding to antibody (Ab) with the Fc portion binding to the Fc receptor (FcR) on the macrophage, resulting in FcR signaling via BTK and/or other molecules that may be offtargets inhibited by IBR and/or ACALA. Efferocytosis (right) proceeds via phosphatidylserine (PS) exposure on apoptotic cell surfaces that are bound by macrophage PS receptors (PS R), resulting in signaling via potential off-target molecules. Both pathways lead to common phagocytic machinery and subsequent phagolysosomal processing via potential off-target molecules. IBR offtarget inhibition affects FcR-mediated signaling and downstream phagolysosomal processing (gold rectangles). (B) Representative time-lapse video images (~8% subsection of full field of view of the CTDR channel) of anti-CD20 mAb-mediated ADCP of chronic lymphocytic leukemia (CLL) cells by human monocyte-derived macrophage (hMDM) visually show phagocytic engulfment voids at 60 minutes (min) with 0.0, 1.2 and 33 mM of IBR (top) or ACALA (bottom), but with less voids at 33 mM IBR. Duplicate 0.0 mM image (+ Voids) highlights voids with yellow spots. Scale bar =25 mM. Images of all drug concentrations are available in the data sharing statement (Supplementary Figure DS1). (C) Phagocytic engulfments were measured by phagocytic void index (means (n=18) + standard error) every 4 min and plotted for IBR (left) and ACALA (right) at indicated drug concentrations for the first 1.0 h of ADCP. (D) ADCP measured by AUC of each void index curve (1.0 h) was calculated via the trapezoidal rule, log10-transformed, and modeled as a function of drug concentration, dose and their interaction. All hypothesis tests were performed at the 2-sided 0.05 level using SAS 9.4 (SAS Institute, Inc. Cary, NC, USA). Means and associated 95% confidence intervals (CI) of mixed modelbased contrasts of each concentration with untreated (0 mM) shows relative ADCP (*P<0.05, **P<0.01, ***P<0.001). Dotted line marks no change relative to 0 mM with values below indicating inhibition. (E) The difference between ACALA and IBR was assessed by mixed model-based contrasts with means and associated 95% CI shown for each concentration (**P<0.01, ***P<0.001). Dotted line marks drug equivalence with values above indicating more inhibition of ADCP by IBR than ACALA.

centration were collected every 4 min over 2.8 h in seven replicate experiments. Efferocytosis was visible with no apparent differences after drug treatment (2.5 mM or 10 mM, 60 min; Figure 2B). Further examples of efferocytosis time-lapse images and videos with IBR or ACALA treatment are available in data sharing information (Supplementary Figure DS3; Supplementary Videos DS9, DS10, DS13 and DS14). Quantitation of engulfments by void index indicate that IBR and ACALA did not alter engulfment kinetics across all drug concentrations (1.25–10 mM; Figure 2B; Online Supplementary Figure S2B). Quantitation by AUC of the void index plot over the first 1 h showed no significant difference between untreated and treatment with any concentration of either drug (P>0.05; Figure 2D). Because IBR or ACALA did not inhibit efferocytosis, these

data suggest that BTK and off-target molecules inhibited by IBR are not involved in initial signaling mediated by the receptors for apoptotic cells in efferocytosis or common downstream phagocytosis pathway signaling (Figure 1A). This lack of effect of IBR and ACALA on apoptotic cell phagocytosis in vitro suggests that inhibition of efferocytosis does not cause the greater lymphocytosis seen clinically with IBR treatment,1,10 and provided an opportunity to study the effects of IBR or ACALA on downstream phagolysosomal processing. Phagosomes with internalized cargo transition to acidified phagolysosomes for processing.12 The kinetics of this process can be measured by target cells labeled with pH-sensitive dyes, such as pHrodo Red, which increase in intensity with decreasing pH.12 Efferocytosis by hMDM

Haematologica | 107 - June 2022

1461

LETTER TO THE EDITOR

Figure 2. BTK inhibitors have little effect on efferocytosis. (A) Percent human monocyte-derived macrophage (hMDM) that engulfed eFluor670+ apoptotic chronic lymphocytic leukemia (CLL) cells after treatment with 0–10 mM drug is shown after 1 hour (h) by flow cytometry with no significant change for ibrutinib (IBR) or acalabrutinib (ACALA) (mean [n=4] + standard error). Cytocholasin D (CytoD) prevents phagocytosis and serves as a control. (B) Representative time-lapse video images (~8% subsection of full field of view of the CTDR channel) of efferocytosis show no visual decrease in phagocytic engulfment voids following treatment (0.0, 2.5, 10 mM) with IBR (top) or ACALA (bottom). Duplicate 0.0 mM image (+ Voids) highlights voids with yellow spots. Scale bar =25 mM. Images of all drug concentrations are available in data sharing statement (Supplementary Figure DS3). (C) Efferocytosis phagocytic engulfments were measured by phagocytic void index (means [n=7] + standard error) every 4 minutes (min) and plotted for IBR (left) and ACALA (right) at indicated drug concentrations for the first 1.0 h of efferocytosis. (D) Efferocytosis, as measured by the area under the curve (AUC) of the 1st h of each void index curve, was calculated and modeled as in Figure 1D. Means and associated 95% confidence intervals (CI) of mixed model-based contrasts of each concentration with untreated are shown. Dotted line indicates no change relative to 0 mM. There was insufficient evidence of any difference from untreated (P>0.05).

of pHrodo Red-labeled apoptotic CLL cells produces a readily visualized increase in dye intensity after 2 h that colocalizes with phagocytic voids (Figure 3A). Effects of IBR or ACALA drug treatment on this change in dye intensity were not easily visualized on inspection as illustrated in representative 2.5 mM and 10 mM images (Figure 3A). Further examples of IBR or ACALA treated phagolysosomal processing time-lapse images and videos are available in the data sharing statement (Supplementary Figure DS4; Supplementary Videos DS11, DS12, DS15 and DS16). In order to quantitate pHrodo Red dye intensity levels normalized to macrophage number, the dye intensity index was calculated and plotted over time (Figure 3B; Online Supplementary Figure S2C).12,13 The kinetics of the dye intensity index demonstrated the expected delay relative to the void index (Figure 2C; Online Supplemen-

tary Figure S2B), because phagolysosomal processing occurs after phagocytic engulfments.12 Higher concentrations of IBR but not ACALA exhibited a decrease in the dye index relative to untreated (Figure 3B; Online Supplementary Figure S2C). For phagolysosomal processing, the AUC was calculated for 2 h and the drug concentration curves were analyzed. Compared to untreated, there was a progressive decrease in phagolysosomal processing with increasing IBR but not ACALA concentrations, which was significant at 10 mM (P=0.008; Figure 3C). When comparing the effects of ACALA versus IBR, there was a progressive increase in this ratio with increased drug concentration that was significant at 10 mM (P=0.0002; Figure 3D). Since ACALA had no effect on phagolysosomal processing, these results suggest that BTK signaling is not essential for this process and that

Haematologica | 107 - June 2022

1462

LETTER TO THE EDITOR

Figure 3. Phagolysosomal processing is inhibited by ibrutinib, but not by acalabrutinib. Live cell time-lapse video imaging of efferocytosis was collected as in Figure 2. (A) In order to illustrate the appearance of pHrodo Red dye intensity, time-lapse video images of the combined Cell Tracker Deep Red (CTDR) and pHrodo Red channels (CTDR + pHrodo) or pHrodo Red channel alone (pHrodo) from the representative well and drug concentrations in Figure 2 are shown for untreated (0 and 120 minutes [min]) and ibrutinib (IBR) (left) or acalabrutinib (ACALA) (right) treatment (120 min). Scale bar =25 mM. Images of all drug concentrations are available in data sharing statement (Supplementary Figure DS4). (B) Live cell time-lapse high-content microscopy imaging collected from experiments shown in Figure 2 was measured by dye intensity index to assess phagolysosomal processing. Efferocytosis dye intensity index time course over 2.0 hours is shown for IBR (left) and ACALA (right) with drug concentrations indicated. The mean of 6 experiments for each time-point measured in duplicate or triplicate is shown with positive standard error bars. (C) Phagolysosomal processing during efferocytosis as measured by area under the curve (AUC) of the first 2 hours of each dye intensity curve, was modeled as a function of drug concentration, dose and their interaction. Means and associated 95% confidence intervals (CI) of mixed model-based contrasts of each concentration with untreated are shown. Dotted line indicates no change relative to 0 mM. Significant difference from untreated was only observed at 10 mM IBR concentration (**P<0.01). (D) Mixed model-based contrasts were built to assess the difference between drugs at every concentration. Means and associated 95% CI are shown. Dotted line indicates drug equivalence. Values above dotted line indicate more inhibition of efferocytosis by ibrutinib than acalabrutinib. Significant difference between drugs was only seen at 10 mM drug concentration (***P<0.001).

IBR effects are mediated by off-target inhibition (Figure 1A). IBR obstruction of phagolysosomal processing could result in delay of upstream phagocytosis. Future studies will be needed to study IBR-inhibited off-target(s) in phagocytosis. These data show that short-term highly selective BTK inhibition in vitro by ACALA does not alter macrophage functions of mAb mediated ADCP, antibody-independent efferocytosis, or phagolysosomal processing. In contrast, IBR significantly inhibited ADCP over a wide range of drug concentrations (0.41–100 mM). Thus, inhibition of ADCP, the principal mechanism of therapeutic anti-CD20 mAb efficacy,4 could explain the lack of clinical benefit for the addition of anti-CD20 mAb rituximab to IBR for treatment of CLL.2,3 These data suggest that addition of antiCD20 mAb to a more selective BTKi would be a preferable choice in the treatment of patients with Bcell malignancies such as CLL. Furthermore, newer highly selective reversible BTKi, would be of interest to assess in combination with anti-CD20 mAb therapy.15

Authors Jonathan J. Pinney,1,2 Sara K. Blick-Nitko,3 Andrea M. Baran,4 Derick R. Peterson,4 Hannah E. Whitehead,1,2 Raquel Izumi,5 Veerendra Munugalavadla,6 Karl R. VanDerMeid,7,8 Paul M. Barr,7,8 Clive S. Zent,7,8 Michael R. Elliott1,2,9 and Charles C. Chu7,8 Center for Vaccine Biology and Immunology, and 2Department of

Microbiology and Immunology, University of Rochester, Rochester, NY; 3Department of Pathology and Laboratory Medicine, University of Rochester, Rochester, NY; 4Department of Biostatistics and Computational Biology, University of Rochester Medical Center, Rochester, NY; 5Acerta Pharma, a member of the AstraZeneca Group, South San Francisco, CA; 6AstraZeneca, South San Francisco, CA; 7Department of Medicine, University of Rochester Medical Center, Rochester, NY, 8Wilmot Cancer Institute, University of Rochester Medical Center, Rochester, NY and 9Center for Cell Clearance and the Department of Microbiology, Immunology, and Cancer Biology, University of Virginia, Charlottesville, VA, USA.

Haematologica | 107 - June 2022

1463

LETTER TO THE EDITOR Correspondence:

Margaret Lindorfer (UVA) for critical feedback throughout the

Charles C. Chu - charles_chu@urmc.rochester.edu

project; Deb Gilbertson and Cecile Krejsa for critical discussions on

Michael R. Elliott - mre4n@virginia.edu

this project; Rachel Pinney for advice on figure design; CLL patients

https://doi.org/10.3324/haematol.2021.279560

for participation in our research and donation of blood specimens;

Received: July 7, 2021.

Sharon Lewinski RN and Tania Orzol NP for specimen collection,

Accepted: February 9, 2022.

and the New York-Penn branch of the American Red Cross for

Prepublished: February 17, 2022.

supplying healthy donor leukocytes.

Disclosures

Funding

AMB received research funding from Acerta Pharma / AstraZeneca;

This study was supported by research funding from Acerta Pharma

RI has equity ownership in Acerta Pharma and AstraZeneca; was

LLC to MRE, CCC, PMB, AMB, and CSZ, by grants to MRE from NIH

employed by Acerta Pharma (during time of study); and holds

(grant number: AI114554, DK119285), and the University of Rochester

patents from intellectual property for Acerta Pharma including

Research Award, Wilmot Cancer Institute and to CSZ (University of

patents surrounding acalabrutinib; VM is employed by Acerta

Rochester Research Award, Wilmot Cancer Institute), and by

Pharma (a member of the AstraZeneca group) and has equity

philanthropic donation from Ms. Elizabeth J. Aaron to CSZ. JJP was

ownership in AstraZeneca and Gilead Sciences; KRV received

supported in part by the University of Rochester Immunology

research funding from Acerta Pharma / AstraZeneca; PMB consults

Training T32 Grant from NIH (grant number: AI007285).

for Pharmacyclics LLC / AbbVie, AbbVie, Genentech, Gilead, Merck, Seattle Genetics, Verastem, AstraZeneca, Celgene, Morphosys, TG

Data sharing statement

Therapeutics, and Janssen; and receives research funding from

Additional data sharing available upon request. In order to further

Pharmacyclics LLC / AbbVie, TG Therapeutics, and AstraZeneca; CSZ

illustrate the phagocytic quantitation method, Supplementary

received research funding from Acerta Pharma / AstraZeneca,

Videos DS1 to 4 are available in support of Online Supplementary

Mentrik Biotech, and TG Therapeutics; MRE received research

Figure S1. In order to complement Figure 1B, representative time-

funding from Acerta Pharma / AstraZeneca; CCC has equity

lapse video images of ADCP phagocytic engulfment voids for all

ownership in Pfizer and received research funding from Acerta

drug concentrations of IBR or ACALA are available in the

Pharma / AstraZeneca and TG Therapeutics. All other authors report

Supplementary Figure DS1 and examples of full-length time-lapse

no conflicts of interest.

videos using 3.7

mM concentration of IBR or ACALA with combined

CTDR and Phase channels or CTDR channel alone are available in Contributions

Supplementary Videos DS5 to 8. Preparation of apoptotic CLL cells

MRE, CSZ, JJP, and CCC planned the study strategy; JJP, SKB-N,

by venetoclax treatment for efferocytosis assays is available in

and CCC performed experiments; AMB and DRP performed

Supplementary Figure DS2. In order to complement Figure 2B,

statistical analyses; PMB and CSZ provided clinical samples; HEW

representative time-lapse video images of efferocytosis phagocytic

and KRV provided cells and reagents for experiments; MRE, CSZ,

engulfment voids for all drug concentrations of IBR or ACALA are

JJP, AMB, DRP and CCC interpreted data; MRE, CSZ, RI, VM, CCC

available in Supplementary Figure DS3 and examples of full-length

acquired funding, managed collaboration, and reviewed manuscript;

time-lapse videos using 10

MRE, CSZ, JJP, AMB, DRP and CCC wrote the manuscript.

combined CTDR and Phase channels or CTDR channel alone are

mM concentration of IBR or ACALA with

available in Supplementary Videos DS9, DS10, DS13 and DS14. In Acknowledgments

order to complement Figure 3A, representative time-lapse video

The authors are grateful to members of the Center for Vaccine

images of efferocytosis changes in phagolysosomal processing dye

Biology and Immunology, University of Rochester Medical Center

intensity for all drug concentrations of IBR or ACALA are available in

(URMC), Wilmot Cancer Institute (URMC), Department of Pathology

Supplementary Figure DS4 and examples of full-length time-lapse

and Laboratory Medicine (URMC), and the Center for Cell Clearance,

videos using 10

University of Virginia (UVA) for critical feedback on this project and

CTDR and Phase channels or CTDR channel alone are available in

manuscript; Genentech for anti-CD20 mAb; Drs. Ron Taylor and

Supplementary Videos DS11, DS12, DS15 and DS16.

mM concentration of IBR or ACALA with combined

References 1. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-42. 2. Burger JA, Keating MJ, Wierda WG, et al. Safety and activity of ibrutinib plus rituximab for patients with high-risk chronic lymphocytic leukaemia: a single-arm, phase 2 study. Lancet Oncol. 2014;15(10):1090-1099.

3. Rogers A, Woyach JA. BTK inhibitors and anti-CD20 monoclonal antibodies for treatment-naive elderly patients with CLL. Ther Adv Hematol. 2020;11:2040620720912990. 4. Grandjean CL, Montalvao F, Celli S, et al. Intravital imaging reveals improved Kupffer cell-mediated phagocytosis as a mode of action of glycoengineered anti-CD20 antibodies. Sci Rep. 2016;6:34382.

Haematologica | 107 - June 2022

1464

LETTER TO THE EDITOR 5. Nimmerjahn F, Ravetch JV. Fcgamma receptors as regulators of immune responses. Nat Rev Immunol. 2008;8(1):34-47. 6. Da Roit F, Engelberts PJ, Taylor RP, et al. Ibrutinib interferes with the cell-mediated anti-tumor activities of therapeutic CD20 antibodies: implications for combination therapy. Haematologica. 2015;100(1):77-86. 7. VanDerMeid KR, Elliott MR, Baran AM, Barr PM, Chu CC, Zent CS. Cellular cytotoxicity of next-generation CD20 monoclonal antibodies. Cancer Immunol Res. 2018;6(10):1150-1160. 8. Barf T, Covey T, Izumi R, et al. Acalabrutinib (ACP-196): a covalent Bruton tyrosine kinase inhibitor with a differentiated selectivity and in vivo potency profile. J Pharmacol Exp Ther. 2017;363(2):240-252. 9. Honigberg LA, Smith AM, Sirisawad M, et al. The Bruton tyrosine kinase inhibitor PCI-32765 blocks B-cell activation and is efficacious in models of autoimmune disease and B-cell malignancy. Proc Natl Acad Sci U S A. 2010;107(29):13075-13080. 10. Byrd JC, Harrington B, O'Brien S, et al. Acalabrutinib (ACP-196)

in relapsed chronic lymphocytic leukemia. N Engl J Med. 2016;374(4):323-332. 11. Arandjelovic S, Ravichandran KS. Phagocytosis of apoptotic cells in homeostasis. Nat Immunol. 2015;16(9):907-917. 12. Chu CC, Pinney JJ, Whitehead HE, et al. High-resolution quantification of discrete phagocytic events by live cell timelapse high-content microscopy imaging. J Cell Sci. 2020;133(5):jcs237883. 13. Pinney JJ, Rivera-Escalera F, Chu CC, et al. Macrophage hypophagia as a mechanism of innate immune exhaustion in mAb-induced cell clearance. Blood. 2020;136(18):2065-2079. 14. Golay J, Ubiali G, Introna M. The specific Bruton tyrosine kinase inhibitor acalabrutinib (ACP-196) shows favorable in vitro activity against chronic lymphocytic leukemia B cells with CD20 antibodies. Haematologica. 2017;102(10):e400-e403. 15. Wen T, Wang J, Shi Y, Qian H, Liu P. Inhibitors targeting Bruton's tyrosine kinase in cancers: drug development advances. Leukemia. 2021;35(2):312-332.

Haematologica | 107 - June 2022

1465

LETTER TO THE EDITOR

Diagnosis of acute promyelocytic leukemia based on routine biological parameters using machine learning Acute promyelocytic leukemia (APL) has become the most curable subtype of acute myeloid leukemia (AML) since the introduction of all-trans retinoic acid and arsenic trioxide.1 However, this disease is still characterized by a high rate of early death (10-17%), mainly due to severe coagulopathy.2,3 In order to avoid these early deaths, immediate treatment initiation is recommended, either with all-trans retinoic acid or with chemotherapy in case of hyperleucocytosis.4 Hence, a fast and accurate diagnosis is mandatory to allow early recognition and treatment of APL. Cytology is the fastest technique for the diagnosis of APL, while the definitive confirmation requires the observation of the t(15;17) translocation or the PML-RARA fusion mRNA amplification, which induces further delay. Cytogenetic and molecular confirmation can be more difficult when other partners of RARA are implicated,5 and even more challenging in exceptional cases of viral insertion in the RARA gene, as recently described.6 The cytological diagnosis of APL is usually straightforward, when multiple bundles of Auer rods are observed in the blasts cells. However, the microgranular variant might be more difficult to diagnose, even for experienced hematopathologists. In some cases, myeloperoxidase deficiency in the blast cells further complicates the recognition of APL.7,8 Moreover, cytology requires a long training to recognize rare diseases such as APL, and this expertise is not always available. We hypothesized that routine biological parameters might fuel an artificial intelligence to identify APL without a high level of cytological expertise. We collected 34 basic biological parameters in all the APL patients diagnosed in Lyon University Hospital during the period from 2013 to 2020 (n=76), and in patients with non-promyelocytic AML matched according to the year of diagnosis (n=146). Altogether, these patients constituted the cohort 1 (n=222). All the APL cases were confirmed by cytogenetic and/or quantitative reverse transcription polymerase chain reaction (RT-qPCR) amplification of the PML-RARA fusion transcript. The biological parameters were measured during the first 2 days of hospital referral, before any treatment initiation. Missing data were imputed by the variable’s median value.9 The basic hematology and hemostasis parameters were available for most of the patients, but there were more missing data concerning the biochemical parameters (Online Supplementary Table S1). The cohort was randomly split into a training (80%, n=177) and a test cohort (20%, n=45). Different classification algorithms were then compared (XGBoost, random forest, gradient boosting classifier, adaboost classifier, decision tree, logistic regression and support vector machine), considering APL

diagnosis as a binary outcome and using 5X cross-validation to select the more stable models. No normalization of the data was used, because both strategies tested (StandardScaler, MinMaxScaler) had a negative impact on the performances of the algorithm. Hyperparameter tuning was performed using GridSearchCV. All analyses were performed using Python v3.7. Among the different artificial intelligence strategies tested, XGBoost’s gradient boosting algorithm achieved the highest performances in the test cohort, with an area under the receiver operator curve (ROC) of 0.95 (Figure 1A, see the Online Supplementary Appendix for methodological details). Of note, learning curves reached a plateau with 80-100 patients, meaning that no major refinements is expected with an increase in the size of the cohort (data not shown). Using this model, we established artificial intelligence for promyelocytic leukemia (AIPL), an open-source tool with a graphical user interface to evaluate the probability of APL diagnosis (https://github.com/Nico-Facto/Leukemia-AplClassification) and propose a ready-to-use web application (https://share.streamlit.io/nico-facto/leukemia-aplclassification/main/Leucemie_app.py). The eight parameters required to run AIPL are the following: age, white blood cells (absolute value), lymphocytes (% of total leucocytes), neutrophil polynuclear count (absolute value), mean corpuscular volume (MCV), mean corpuscular hemoglobin concentration (MCHC), prothrombin time ratio, and fibrinogen concentration. In order to validate the AIPL tool, its performances were assessed in three independent retrospective validation cohorts from three other hospitals (cohorts 2, 3, and 4) which comprised 44 (including 15 APL), 258 (including 46 APL), and 63 (including 32 APL) patients, respectively. A prospective cohort (cohort 5) was also collected in the Lyon University Hospital with 50 (including 10 APL) new AML diagnoses referred during a 6-month period. AIPL showed a very high discrimination ability both in the merged (n=415 patients, AUC =0.96, Figure 1B) and in individual cohorts (Figure 1C). Importantly, AIPL output is not only a classification (APL vs. non-APL), but also a confidence score reflecting how much the conclusion can be trusted. As expected, the confidence score was significantly higher in cases where the prediction was correct compared to cases where the AIPL prediction failed (mean 95% vs. 85%, Mann-Withney test P<0.0001, Figure 2A). Hence, the AIPL confidence score could be used to determine for which patient the prediction of AIPL is reliable in routine use. For 244 (59%) patients with a high confidence score (above 99%), the accuracy of AIPL was

Haematologica | 107 - June 2022

1466

LETTER TO THE EDITOR

Figure 1. Development of AIPL for the diagnosis of acute promyelocytic leukemia. (A) Receiver-operator characteristics (ROC) curve of the XGBoost method in the test cohort. (B) ROC curve in all the patients from the validation cohorts. (C) Area under the curve of the ROC curves in each validation cohort, separately.

99.5% (only one false negative case, i.e., an APL patient wrongly classified as non-APL AML). For 114 (27%) patients with an intermediate confidence score (between 85 and 99%), the accuracy was 85% (7 false negative and 10 false positive cases, i.e., a non-APL AML patient wrongly classified as APL), and for 57 (14%) patients with a low confidence score (below 85%), the accuracy dropped to 68% (8 false negative and 10 false positive cases) (Figure 2B). As the data from the retrospective cohorts were acquired on different analyzers, it was possible to assess the impact of the variability in analytical techniques on AIPL performances. When comparing different time periods determined by changes in analyzers, there was no significant variation in the level of confidence scores of AIPL, suggesting that this approach is robust to variations due to analytical processes (Online Supplementary Figure S1). Of note, 16 cases of the microgranular variant of APL were identified in the validation cohorts. Their AIPL confidence scores tended to be lower than the confidence scores of classical APL (86% vs. 92%, ns, Online Supplementary Figure S2), suggesting that further algorithm training using microgranular variants might be interesting. Importantly, AIPL correctly identified the six microgranular cases with a confidence score above 99%. We also assessed the performances of AIPL in patients with other differential diagnosis of APL: aplastic anemia (n=10), acute lymphoblastic leukemia (n=28) and AML with t(8;16), a rare subtype of AML with clinical presentation often resembling APL (n=9). AIPL classified only one case aplastic anemia and one case of acute lymphoblastic leukemia as APL, but with

confidence score below 87%. Hence, with the proposed threshold of confidence score of 99%, there was no false positive diagnostic of APL in these challenging differential diagnoses. In order to further interpret the predictions from AIPL, Shapley additive explanation (SHAP) was used to illustrate the impact of the different parameters according to their value obtained from the individual cases of all the validation cohorts.10 In Figure 2C, the parameters are ordered from top to bottom according to their importance in the classification, and each individual measure is colored according to its impact on the final classification. The high performances of AIPL rely on some expected parameters such as fibrinogen, prothrombin time ratio, or polynuclear neutrophil count (Figure 2C). Unexpectedly, two parameters of red blood cells (MCV and MCHC) were highly discriminant between APL and non-APL AML, even if the mean values of these parameters remained in the normal ranges (mean value of MCV 89 fL vs. 96 fL, mean value of MCHC 349 g/L vs. 334 g/L in APL and non-APL cases, respectively). This observation, together with a report of PML-RARA expression in burst forming unit-erythroid (BFU-E) derived from APL patients,11 raises the hypothesis that PML-RARA cells contribute to erythropoiesis. To conclude, this work demonstrates that machine learning based on routine biological parameters provides a fast and accurate help for the diagnosis of APL in the majority of cases. Given the unmet need to improve the reliability of APL diagnosis, other strategies based on incorporation of cell population data generated during complete blood

Haematologica | 107 - June 2022

1467

LETTER TO THE EDITOR

Figure 2. Validation of artificial intelligence for promyelocytic leukemia. (A) Mean confidence score of artificial intelligence for promyelocytic leukemia (AIPL) in cases correctly or wrongly classified in all the validation cohorts. (B) AIPL performances according to the confidence score. (C) Shapley additive explanation (SHAP) representation of the parameters used by AIPL. The parameters are ordered from top to bottom according to their importance in the classification, and each individual measure is colored according to its impact on the final classification (the more a point is red, the more it favors the classification as APL)

cell count with cytometry based analyzers,12 or deep learning analysis of blood smears, have also been developed.13 As these approaches rely on parameters not used in AIPL, combining them with AIPL could help further increase diagnostic accuracy. Another interesting possibility could be the addition of other biological parameters such as the fibrinolysis marker D-Dimers, which were not included in this retrospective study due to excessive missing data. AIPL might represent a very important complement to cytological expertise, and could allow early diagnosis of APL in settings where this expertise is not available on a 24/7 basis, or not available at all, such as in developing countries. The consequences of misclassification could be excessive treatment with ATRA in case of false positive result, or delay in ATRA initiation in case of false negative result. Using AIPL with the proposed threshold of 99% of confidence score, this risk is very low but should not be forgotten. All-trans retinoic acid could hence be initiated in patients with a high probability of APL according to AIPL prediction without waiting for diagnostic confirmation in specialized laboratories, thus preventing early death from coagulopathy. However, an important limitation of AIPL is that its ability to distinguish APL from other differential diagnoses such as acute lymphoblastic leukemia or aplastic anemia has not been formally assessed in this study. In order to make this tool available, a web user interface has been created (available at https://share.streamlit.io/nico-facto/leukemia-apl-clas-

sification/main/Leucemie_app.py) to use AIPL in the case of patients with myeloid blast on the blood smear. It allows to instantaneously classify patients as APL or nonAPL and provides a confidence score. Of course, this result does not supplant the need to evaluate the bone marrow and to formally demonstrate the presence of the t(15;17) translocation or the PML-RARA fusion transcript. Given the importance of early treatment initiation in these patients, we hope that AIPL will contribute to decrease early mortality in APL patients.

Authors Estelle Cheli,1 Simon Chevalier,2 Olivier Kosmider,3 Marion Eveillard,4 Nicolas Chapuis,3 Adriana Plesa,1 Maël Heiblig,5 Lydie Andre,2 Jenny Pouget,4 Pascal Mossuz,2 Olivier Theisen,4 Vincent Alcazer,5 Dan Gugenheim,6 Nicolas Autexier6 and Pierre Sujobert1,7 Hospices Civils de Lyon, Hôpital Lyon Sud, Service d'Hématologie

Biologique, Lyon; 2CHU Grenoble Alpes, Service d’Hématologie Biologique, Grenoble; 3APHP, Hôpital Cochin, Service d’Hématologie Biologique, Cochin; 4CHU Nantes, Service d’Hématologie Biologique, Nantes; 5Hospices Civils de Lyon, Hôpital Lyon Sud, Service d'Hématologie Clinique, Lyon; 6Groupe onepoint, Bordeaux and Université Claude Bernard Lyon 1, Faculté de Médecine et de

Maïeutique Lyon Sud, Charles Mérieux, Lymphoma Immunobiology Team, Lyon, France.

Haematologica | 107 - June 2022

1468

LETTER TO THE EDITOR Correspondence:

LA, JP, PM, OT and MH collected the data; NA, VA and DG developed

PIERRE SUJOBERT - pierre.sujobert@chu-lyon.fr

the AI and the web interface; PS, EC, NA and DG analyzed the data;

https://doi.org/10.3324/haematol.2021.280406

PS wrote the paper and all authors revised and approved the final

Received: November 23, 2021.

manuscript.

Accepted: February 11, 2022. Prepublished: February 24, 2022.

Acknowledgements The authors thanks Mrs Verena Landel from Hospices Civils de Lyon

Disclosures

for English editing, as well as Dr Nicolas Duployez, Dr Damien

NA and DG were employed by Groupe onepoint when the work was

Bodet, Dr Thomas Boyer and Dr Pascale Flandrin for their help in

conducted. All other authors have no conflicts of interest to

assessing AIPL performances on patients with t(8;16) AML.

declare. Data sharing statement Contributions

The data that support the findings of this study are available on

PS, AP, EC and DG designed the study; EC, SC, OK, ME, NC, AP, PM,

request from the corresponding author.

References 1. Lo-Coco F, Avvisati G, Vignetti M, et al. Retinoic acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med. 2013;369(2):111-121. 2. Park JH, Qiao B, Panageas KS, et al. Early death rate in acute promyelocytic leukemia remains high despite all-trans retinoic acid. Blood. 2011;118(5):1248-1254. 3. Rahmé R, Thomas X, Recher C, et al. Early death in acute promyelocytic leukemia (APL) in French centers: a multicenter study in 399 patients. Leukemia. 2014;28(12):2422-2424. 4. Sanz MA, Fenaux P, Tallman MS, et al. Management of acute promyelocytic leukemia: updated recommendations from an expert panel of the European LeukemiaNet. Blood. 2019;133(15):1630-1643. 5. Grimwade D, Biondi A, Mozziconacci MJ, et al. Characterization of acute promyelocytic leukemia cases lacking the classic t(15;17): results of the European Working Party. Groupe Français de Cytogénétique Hématologique, Groupe de Français d’Hematologie Cellulaire, UK Cancer Cytogenetics Group and BIOMED 1 European Community-Concerted Action “Molecular Cytogenetic Diagnosis in Haematological Malignancies.” Blood. 2000;96(4):1297-1308. 6. Astolfi A, Masetti R, Indio V, et al. Torque teno mini virus as a cause of childhood acute promyelocytic leukemia lacking PML/RARA fusion. Blood. 2021;138(18):1773-1777. 7. Cui W, Qing S, Xu Y, Hao Y, Xue Y, He G. Negative stain for myeloid

peroxidase and Sudan black B in acute promyelocytic leukemia (APL) cells: report of two patients with APL variant. Haematologica. 2002;87(5):ECR16. 8. Heiblig M, Paubelle E, Plesa A, et al. Comprehensive analysis of a myeloperoxidase-negative acute promyelocytic leukemia. Blood. 2017;129(1):128-131. 9. Acuña E, Rodriguez C. The Treatment of Missing Values and its Effect on Classifier Accuracy. In: Banks D, McMorris FR, Arabie P, Gaul W, editors. Classification, Clustering, and Data Mining Applications. Berlin, Heidelberg: Springer. 2004. p639-647. 10. Lundberg SM, Erion GG, Lee S-I. Consistent Individualized Feature Attribution for Tree Ensembles. arXiv.https://doi.org/10.48550/arXiv.1802.03888 preprint, [not peer-reviewed] 11. Takatsuki H, Sadamura S, Umemura T, et al. PML/RAR alpha fusion gene is expressed in both granuloid/macrophage and erythroid colonies in acute promyelocytic leukaemia. Br J Haematol. 1993;85(3):477-482. 12. Haider RZ, Ujjan IU, Shamsi TS. Cell population data–driven acute promyelocytic leukemia flagging through artificial neural network predictive modeling. Trans Oncol. 2020;13(1):11-16. 13. Sidhom J-W, Siddarthan IJ, Lai B-S, et al. Deep Learning for distinguishing morphological features of acute promyelocytic leukemia. Blood. 2020;136(Suppl 1):S10-12.

Haematologica | 107 - June 2022

1469

LETTER TO THE EDITOR

Risk of hepatitis B virus reactivation in chronic lymphocytic leukemia patients receiving ibrutinib with or without antiviral prophylaxis. A retrospective multicentric GIMEMA study Several reports have highlighted the risk of hepatitis B virus (HBV) reactivation in patients with lymphoproliferative disorders undergoing cytotoxic treatment. This risk is particularly relevant in chronic lymphocytic leukemia (CLL) patients with occult HBV infection (OBI), especially during treatment with anti-CD20 monoclonal antibodies. Moreover, CLL is one of the B-cell lymphoproliferative diseases with the highest risk of HBV reactivation. Since chemo-immunotherapy treatments (CIT) of CLL patients with OBI are associated with an intermediate to high risk of HBV reactivation,1,2 antiviral therapy as prophylaxis is recommended. Currently, lamivudine is universally used as prophylaxis due to its low cost and toxicity profile. As HBV reactivation can occur up to 12-18 months after the end of the chemotherapy,3 the antiviral prophylaxis is indicated from the beginning of the specific CLL chemotherapy up to 18 months after the end of treatment. HBV reactivation has also been anecdotally reported in CLL patients with OBI treated with a B-cell receptor inhibitor (BCRi), such as ibrutinib, although the evidences remain limited to individual reports and the incidence of HBV reactivation in this setting is still unknown.4-7 Ibrutinib seems to modestly increase the risk of infections in general and to be associated with a moderate risk for HBV reactivation (1-10%);8 however, it is unclear whether HBV reactivation is due to ibrutinib treatment per se, considering that this drug is often used in patients who have previously been subjected to CIT. Another unanswered question is whether the cumulative risk of a HBV reactivation is high enough to ask for a routine HBV-DNA monitoring or a HBV prophylaxis during or after ibrutinib therapy. We performed a retrospective analysis of 109 CLL patients with OBI from 22 Italian GIMEMA centers treated with single agent ibrutinib prior to 31 January 2019, with at least 1 year of follow-up from the first administration. These patients were identified among a cohort of 789 CLL patients who had been analyzed for HBV serum markers before starting ibrutinib, resulting in an overall prevalence of OBI seropositivity of 14%. Inclusion criteria included a CLL diagnosis according to the IwCLL guidelines9 and HBV serum markers suggestive of a seropositive OBI (HBsAg negative, presence of antibodies towards the HBV core antigen [anti-HBc] with or without antibodies towards the HBsAg [anti-HBs] in the

serum)10 at the time of ibrutinib initiation. All enrolled patients were HBV-DNA negative at baseline. Patients who had a concomitant HCV infection, HIV infection and/or any other liver disease were excluded. The primary endpoint of the study was the rate of HBV reactivation, defined as a HBsAg seroconversion and/or an increase of serum HBV-DNA by at least one log above the lower limit of detection of the assay, with or without liver injury, assessed by serum alanine aminotransferase levels.12 For all patients, serological markers for HBV infection (including HBsAg, anti-HBs antibody, anti-HBc antibody, HBeAg, and anti-HBe antibody) and serum HBVDNA were assayed prior to the start of treatment with ibrutinib and every 3 months thereafter. The study was conducted in agreement with the Declaration of Helsinki and was approved by the Local Ethical Committee of each participating institution. Documented informed consent was obtained for all patients included in the study before they were registered or randomized at the GIMEMA Data Center. Data were collected from the medical files and entered into case record forms by treating physicians. Study data were collected and managed using REDCap electronic data capture tools hosted GIMEMA Foundation.11,12 Among the 109 enrolled patients, one was excluded because of missing information regarding the management of OBI during ibrutinib. For the 108 analyzed patients, baseline demographic and disease characteristics for the two cohorts segregated by therapy (i.e., prophylactic antiviral therapy with lamivudine administered at the standard dose of 100 mg daily [n=73] vs. monitoring of HBV serum markers [n=35]) are shown in Table 1. At the start of ibrutinib treatment, nine, 51 and 43 patients were in Binet stage A progressive, B and C, respectively; five patients had missing data. Twenty-five (23%) patients were treatment-naïve at the start of ibrutinib, whereas 83 (77%) had been previously treated; among the latter, 42 (39%), 18 (17%) and 23 (21%) patients had received one, two or more than two lines of CIT, respectively. In the group of previously treated patients, 52 started ibrutinib more than 12 months after the last chemotherapy, while 31 received it prior than 12 months from the end of chemotherapy. The median duration of ibrutinib treatment was 12 months (range, 1-64 months). Only two of the 108 patients (1.9%) witnessed a HBV reac-

Haematologica | 107 - June 2022

1470

LETTER TO THE EDITOR Table 1. Patients’ characteristics and results.

Prophylactic antiviral therapy with lamivudine and HBV-DNA monitoring

22 Italian GIMEMA centers Overall, 108 patients

Characteristics

Time to ibrutinib, n (%) 108/108

Overall reactivation by therapy, n (%) Details for patients with reactivation, n

No = 35

Yes = 73

P-value

Sex: M/F, n

73/35

23/12

50/23

0.83

Median age (range)

64 (39-83)

63 (48-81)

65 (39-83)

0.41

Binet stage, n (%) 103/108

A B C

9 (9) 51 (49) 43 (42)

1 (3) 21 (60) 13 (37)

8 (12) 30 (44) 30 (44)

0.20

IGHV, n (%) 81/108

Unmutated Mutated

57 (70) 24 (30)

22 (81) 5 (19)

35 (65) 19 (35)

0.20

FISH, n (%) 103/108

Normal karyotype Del 13q Tris 12 Del 11q Del 17p

31 (30) 19 (18) 12 (12) 13 (13) 28 (27)

11 (31) 7 (20) 3 (9) 4 (11) 10 (29)

20 (29) 12 (18) 9 (13) 9 (13) 18 (26)

0.97

Ibrutinib 1st line

25 (23)

8 (23)

17 (23)

After less than 12 months from the last treatment

31 (29)

13 (37)

18 (25)

After more than 12 months from the last treatment

52 (48)

14 (40)

38 (52)

Overall reactivation

2 (1.9)

1 (2.9)

1 (1.4)

Ibrutinib 1st line

Ibrutinib 2nd or subsequent lines*

0.50

0.55

*After more than 12 months from the last treatment. HBV: hepatitis B virus; M: male; F: female; IGHV: immunoglobulin heavy chain variable region; FISH: fluorescence in situ hybridization.

tivation, one occurring in the HBV prophylaxis group (1/73, 1.4%) and another in the HBV monitoring group (1/35, 2.9%) (P=0.55); the two patients had been previously treated with CIT (2/83, 2.4%). Both reactivations were detected during the first 6 months of ibrutinib treatment. The patient who experienced a reactivation in the prophylactic lamivudine group was a 69-year-old male who started treatment with ibrutinib and lamivudine 15 months after receiving front-line treatment with rituximab and second-line treatment with fludarabine-rituximab. The serological status at the start of ibrutinib was as follows: HBcAb, HBsAb and HBeAb positive, HBsAg and HBeAg negative, and HBV-DNA undetectable. After 1 month of ibrutinib treatment, the patient showed a detectable HBV-DNA (76 UI/mL), whereas the other serum HBV markers were unchanged and serum transaminase levels remained within the normal range. During the following months, the patient was carefully monitored and

all liver function parameters remained normal. Antiviral therapy with entecavir at the dose of 0.5 mg/daily was started after 7 months of ibrutinib treatment when the HBV-DNA raised to 350 UI/mL, and ibrutinib was reduced to 280 mg/daily because of severe diarrhea. HBV-DNA became undetectable after 3 months from the beginning of entecavir. Ibrutinib treatment was stopped after 1 year because of atrial fibrillation. Two months later, the patient started treatment with venetoclax that is still ongoing together with entecavir administration. The patient who experienced a reactivation in the HBVDNA monitoring group was a 59-year-old male who started treatment with ibrutinib 12 months after receiving the last CIT with fludarabine, cyclophosphamide and rituximab. At baseline, he was HBcAb positive, HBsAb, HBeAb and HBsAg negative, and HBV-DNA was undetectable. After 3 months of ibrutinib treatment, he developed a detectable HBV-DNA at 741 UI/mL and HBeAb became

Haematologica | 107 - June 2022

1471

LETTER TO THE EDITOR positive; the other serological HBV markers remained unchanged and all liver function tests, including serum transaminases, remained normal. Entecavir therapy was administered at the dose of 0.5 mg daily without ibrutinib modifications. During the follow-up, HBV-DNA became undetectable after one month of therapy, whereas HBeAb became negative 1 month later. The patient continued ibrutinib and entecavir treatment until now. The data presented in this study demonstrate a high prevalence of seropositive OBI (14%) in Italian CLL patients treated with ibrutinib, further emphasizing that the management of these patients represents a relevant clinical problem. The higher prevalence of OBI in CLL patients treated with Ibrutinib that we found, when compared with previuos data, which reported it between 8% and 10%,13,14 is possibly explained with closer monitoring of the HBV status before initiating Ibrutinib. The current guidelines of the European Conference on Infection in Leukemia (ECIL-5) and the American Gastroenterology Association (AGA) on the prevention of HBV reactivation provide no recommendations with regard to the management and the need for antiviral prophylaxis of patients with seropositive occult HBV infection treated with a BCRi.15,2 A recent recommendation has been issued by acknowledging the intermediate risk of HBV reactivation and advising prophylaxis with lamivudine in HBsAgnegative, anti-HBc-positive patients starting ibrutinib treatment. However, given that ibrutinib may be administered continuously for years, the toxicity associated with prolonged antiviral prophylaxis and drug-resistance may be significant.8 Previous studies on much smaller series of patients reported a variable incidence of reactivation ranging from 0% to 13%.4,6,7 In our cohort, with a median duration of ibrutinib treatment of 12 months, only two of the 108 patients developed a HBV reactivation with a cumulative incidence of 1.9%. Both patients had been previously treated with CIT, indicating that the rare cases of HBV reactivation are not necessarily related to the immunomodulatory effect of ibrutinib, but more likely to the persistent immunosuppressive effects of the previous CIT. One of the two patients was in the prophylactic lamivudine group, suggesting that lamivudine prophylaxis may reduce but does not eliminate the risk of reactivation. Considering that the risk of reactivation in HBV monitored patients is very low, the option of monitoring at 3 monthintervals the trend of HBV serum markers with the possibility to start treatment with entecavir in case of a HBV reactivation seems the most reasonable and cost-effective option, also in terms of decreased risk of adverse events from long-term treatment with entecavir. In conclusion, we confirm that HBV reactivation may rarely occur during ibrutinib treatment in OBI/CLL patients, mainly if not only in patients previously treated with CIT.

Based on the easy management with entecavir in case of HBV reactivation, we recommend, for CLL patients with OBI during ibrutinib treatment, 3-months interval monitoring of HBV serum markers rather than HBV prophylaxis.

Authors Idanna Innocenti,1* Gianluigi Reda,2* Andrea Visentin,3 Marta Coscia,4 Marina Motta,5 Roberta Murru,6 Riccardo Moia,7 Massimo Gentile,8 Elsa Pennese,9 Francesca Maria Quaglia,10 Francesco Albano,11 Ramona Cassin,2 Marina Deodato,12 Claudia Ielo,13 Anna Maria Frustaci,12 Alfonso Piciocchi,14 Arianna Rughini,14 Valentina Arena,14 Daniela Di Sevo,14 Annamaria Tomasso,15 Francesco Autore,1 Giovanni Del Poeta,16 Lydia Scarfò,17 Francesca Romana Mauro,13 Alessandra Tedeschi,12 Livio Trentin,3 Maurizio Pompili,18 Robin Foà,13 Paolo Ghia,17 Antonio Cuneo19 and Luca Laurenti1,15 Sezione di Ematologia, Dipartimento di Diagnostica per Immagini,

Radioterapia Oncologica ed Ematologia, Fondazione Policlinico Universitario A. Gemelli IRCCS, Roma; 2U.O.C. Ematologia, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milano; Department of Medicine, Hematology and Clinical Immunology

Branch, University of Padova, Padova; 4Division of Hematology, A.O.U. Città della Salute e della Scienza di Torino, Torino; 5SC Ematologia, ASST Spedali Civili, Brescia; 6Hematology and Stem Cell Transplantation Unit, Ospedale A. Businco, ARNAS “G.Brotzu”, Cagliari; 7Division of Hematology, Department of Translational Medicine, University of Eastern Piedmont, Novara; 8UOC Ematologia AO di Cosenza, Presidio Ospedaliero Annunziata, Cosenza; 9U.O.S.D. Centro Diagnosi e Terapia dei Linfomi, Dipartimento OncologicoEmatologico, Presidio Ospedaliero "Spirito Santo", Pescara; Department of Medicine, Section of Hematology, University of

Verona, Verona; 11Department of Emergency and Organ Transplantation (D.E.T.O.), Hematology Section, University of Bari "Aldo Moro", Bari; 12ASST Grande Ospedale Metropolitano Niguarda Hospital, Milano; 13Hematology, Department of Translational and Precision Medicine, Sapienza University, Policlinico Umberto I, Roma; 14GIMEMA Foundation, Roma; 15Sezione di Ematologia, Dipartimento di Scienze Radiologiche ed Ematologiche, Università Cattolica del Sacro Cuore, Roma; 16UOC Unit of Hematology and Stem Cell Transplantation, AOU Policlinico Tor Vergata Roma, Roma; Strategic Research Program on CLL, Università Vita Salute and

IRCSS Ospedale San Raffaele, Milano; 18Dipartimento di Medicina Interna e Gastroenterologia, Fondazione Policlinico Universitario A. Gemelli IRCCS, Università Cattolica del Sacro Cuore, Roma and Hematology Section, Department of Medical Sciences, University

of Ferrara, Ferrara, Italy. *II and GR contributed equally as co-first authors Correspondence: Luca Laurenti - luca.laurenti@unicatt.it https://doi.org/10.3324/haematol.2021.280325

Haematologica | 107 - June 2022

1472

LETTER TO THE EDITOR Received: November 11, 2021.

VA, DDS, AT, FA, GDP, LS, FRM, AT, LT, MP, RF, PG, AC and LL

Accepted: February 11, 2022.

collected the data; AP performed the statistical analysis. All the

Prepublished: February 24, 2022.

authors reviewed the manuscript for important intellectual contents, approved the final version of the manuscript and

Disclosures

supervised the project.

AV has participated in scientific board meetings held by Janssen; LT has participated in scientific board meetings and has received

Data sharing statement

funding from Janssen. All other authors have no conflicts of

Study data were collected and managed using REDCAP (Research

interest to disclose.

Electronic Data Capture), a web-based software platform designed to support data capture for research studies. The data presented in

Contributions

this study are not available. Study protocol can be required to the

II, GR, AV, MC, MM, RM, RM, MG, EP, FMQ, FA, RC, MD, CI, AMF, AP, AR,

corresponding author (luca.laurenti@unicatt.it ).

References 1. Perrillo RP, Gisg R, Falck-Ytter YT. American Gastroenterological Association Institute technical review on prevention and treatment of hepatitis B virus reactivation during immunosuppressive drug therapy. Gastroenterology. 2015;148(1):221-244. 2. Reddy KR, Beavers KL, Hammond SP, Lim JK, Falck-Ytter YT. American Gastroenterological Association Institute guideline on the prevention and treatment of hepatitis B virus reactivation during immunosuppressive drug therapy. Gastroenterology. 2015;148(1):215-219. 3. Di Bisceglie AM, Lok AS, Martin P, Terrault N, Perrillo RP, Hoofnagle JH. Recent US Food and Drug Administration warnings on hepatitis B reactivation with immune-suppressing and anticancer drugs: just the tip of the iceberg? Hepatology. 2015;61(2):703-711. 4. Hammond SP, Chen K, Pandit A, Davids MS, Issa NC, Marty FM. Risk of hepatitis B virus reactivation in patients treated with ibrutinib. Blood. 2018;131(17):1987-1989. 5. İskender G, İskender D, Ertek M. Hepatitis B Virus Reactivation under ibrutinib treatment in a patient with chronic lymphocytic leukemia. Turk J Haematol. 2020;37(3):208-209. 6. Tedeschi A, Frustaci AM, Mazzucchelli M, Cairoli R, Montillo M. Is HBV prophylaxis required during CLL treatment with ibrutinib? Leuk Lymphoma. 2017;58(12):2966-2968. 7. Innocenti I, Morelli F, Autore F, et al. HBV reactivation in CLL patients with occult HBV infection treated with ibrutinib without viral prophylaxis. Leuk Lymphoma. 2019;60(5):13401342. 8. Wang B, Mufti G, Agarwal K. Reactivation of hepatitis B virus infection in patients with hematologic disorders. Haematologica. 2019;104(3):435-443.

9. 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 [published correction appears in Blood. 2008 Dec 15;112(13):5259]. Blood. 2008;111(12):5446-5456. 10. Raimondo G, Locarnini S, Pollicino T, et al. Update of the statements on biology and clinical impact of occult hepatitis B virus infection. J Hepatol. 2019;71(2):397-408. 11. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)-a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377381. 12. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. 2019;95:103208. 13. Innocenti I, Morelli F, Autore F, et al. HBV reactivation in CLL patients with occult HBV infection treated with ibrutinib without viral prophylaxis. Leuk Lymphoma. 2019;60(5):13401342. 14. Rossi D, Sala L, Minisini R, et al. Occult hepatitis B virus infection of peripheral blood mononuclear cells among treatment-naive patients with chronic lymphocytic leukemia. Leuk Lymphoma. 2009;50(4):604-611. 15. Mallet V, Van Bömmel F, Doerig C, et al. Management of viral hepatitis in patients with haematological malignancy and in patients undergoing haemopoietic stem cell transplantation: recommendations of the 5th European Conference on Infections in Leukaemia (ECIL-5). Lancet Infect Dis. 2016;16(5):606-617.

Haematologica | 107 - June 2022

1473

LETTER TO THE EDITOR

Do liberal thresholds for red cell transfusion result in improved quality of life for patients undergoing intensive chemotherapy for acute myeloid leukemia? A randomized crossover feasibility study Patients with acute myeloid leukemia (AML) on intensive treatment are dependent on red blood cell (RBC) transfusion, but the evidence base defining optimal transfusion threshold support in AML is very weak. We conducted a multicenter, randomized controlled feasibility trial in patients transfused below 70 (restrictive, R) or 90g/L (liberal, L) hemoglobin. The aim of the study was to assess the feasibility of randomizing to two hemoglobin thresholds in the setting of AML and assessing quality of life (QoL) among participants. We incorporated a novel crossover design between cycles of intensive chemotherapy, thus

allowing patients to serve as their own control, and assessed QoL at multiple intervals. Patients were eligible if ≥18 years, and treated with curative intent for newly diagnosed or relapsed AML, or myelodysplasia with excess of blasts (MDS-EB). Patients with prior MDS were excluded. Participants were randomized (1:1, web-based service, stratified by center) by day 5 of chemotherapy in cycle 1, to restrictive transfusion (threshold 70 g/L; target 71-80 g/L) or liberal transfusion (threshold 90 g/L, target 91-100 g/L); participants then crossed over to the alternative policy for the second cycle

Table 1. Baseline characteristics.

Baseline characteristics - N (%) or median (q1-q3) Restrictive then liberal (n=21)

Liberal then restrictive (n=22)

Total (n=43)

62 (46-68)

61 (52-68)

61 (48-68)

Male

8 (38)

14 (64)

22 (51)

Ethnic origin Caucasian Black Hispanic Asian

19 (90) 0 (0) 1 (5) 1 (5)

19 (86) 1 (5) 0 (0) 2 (9)

38 (88) 1 (2) 1 (2) 3 (7)

ECOG status 0 1 2 3 4

13 (62) 8 (38) 0 (0) 0 (0) 0 (0)

13 (59) 8 (36) 0 (0) 0 (0) 1 (5)

26 (60) 16 (37) 0 (0) 0 (0) 1 (2)

Cytogenetic category Not reported Favorable risk Standard risk Poor risk

1 (5) 2 (10) 11 (52) 7 (33)

0 (0) 2 (9) 14 (64) 6 (27)

1 (2) 4 (9) 25 (58) 13 (30)

21 (100) 0 (0)

20 (91) 2 (9)

41 (95) 2 (5)

Hemoglobin (g/L)

82 (77-92)

80 (74-94)

81 (75-94)

Platelets (x109/L)

55 (25-112)

75 (26-154)

55 (25-139)

White cell count (x109/L)

5 (2-23)

2 (1-14)

2 (1-20)

Total RBC units in 8 weeks prior to randomisation

2 (1-2)

1 (0-2)

Received RBC transfusion prior to randomisation, n (%)

9 (43)

7 (32)

16 (37)

Age (years)

AML subtype First presentation Relapsed

ECOG: Eastern Cooperative Oncology Group Performance Status; RBC: red blood cells; q1-q3= quartile 1 to quartile 3; AML: acute myeloid leukemia. Haematologica | 107 - June 2022

1474

LETTER TO THE EDITOR

Figure 1. Pretransfusion hemoglobin.

(groups R/L and L/R, respectively). Follow-up was until the end of the second cycle, or 42 days from start of cycle two, whichever was earlier. Primary outcomes were: i) percentage pretransfusion hemoglobins below threshold of the assigned strategy (predefined compliance ≥70%); and ii) difference of at least 15g/L between mean pretransfusion hemoglobin in the two strategies. Secondary outcomes included adherence to protocol and QoL; survival at 3 months; transfusion related adverse events. QoL questionnaires EORTC QLQ-C30 and EQ-5D5L were administered before, during and following each cycle. Participants also completed daily QoL scores (adapted from ECOG score1 and from EQ 5D-5L2). Participants were blinded. The sample size was 36, based on 10% attrition and the need for 31 patients to detect 15g/L difference between the two strategies (estimated group standard deviation 20 g/L, 90% power, intraclass correlation coefficient 0.25 and significance level 0.05; t-test). The analysis was intention to treat. A mixed linear regression model, with a random participant and period effect, was used to test for a difference between the groups and for evidence of a period effect. Secondary outcomes were analyzed using summary statistics. All participants completing at least one QoL questionnaire at baseline and at least one other were included in QoL analysis. 43 patients were randomized from 84 eligible (51.2%), receiving 75 chemotherapy cycles at eight UK hospitals between May 2017 to August 2018 (Online Supplementary

Figure S1). Twenty-one participants were allocated to group R/L and 22 to group L/R. Overall, 37 participants followed the restrictive strategy and 38 the liberal strategy. Baseline characteristics were similar between the two groups (Table 1). Pretransfusion hemoglobin was below threshold in 91% of all transfusion episodes; 77.2% for transfusions in restrictive (95% confidence interval [CI]: 70.7-82.8) and 99.3% in liberal (95% CI: 97.6-99.9) cycles. As both groups had compliance ≥70%, this demonstrates predefined feasibility. The unadjusted mean pretransfusion hemoglobin was 68.7 g/L and 83.4 g/L for the restrictive and liberal strategies respectively. After adjusting for participant as a random effect and the period effect (which was found to be significant (P=0.01), with hemoglobin values lower in cycle 1 than 2) the adjusted mean difference was 15.1 g/L (95% CI: 13.9-16.2; P<0.001) (Figure 1). No statistically significant carry-over effect was found (P=0.2). Greater numbers of transfusions were required in the liberal arm to maintain the higher hemoglobin compared to restrictive arm. Median RBC units transfused per participant over all cycles was 13 (inter quartile range [IQR], 1017), with six units (IQR, 4-9) for the restrictive strategy and nine (IQR, 7-11) liberal. Fourteen transfusions were given for symptoms above the designated hemoglobin threshold during the restrictive strategy and none during the liberal strategy. Numbers of transfusions given according to protocol, based on hemoglobin, were 447 of 537 (83.2%); 152 (70.7%) for the restrictive strategy and 295 (91.6%) liberal.

Haematologica | 107 - June 2022

1475

LETTER TO THE EDITOR Table 2. Secondary outcomes.

Restrictive

Liberal

Overall

155 91

258 108

413 199

6 (4-9)

9 (7-11)

13 (10-17)

21 (100.0) 13 (81.3)

22 (100.0) 16 (100.0)

43 (100.0) 29 (90.6)

RBC given as an inpatient, %

98.5%

97.0%

97.6%

RBC given as single units, % of inpatient transfusions

84.9%

87.2%

86.3%

52 (40-65) 55 (42.5-65)

55 (40-75) 55 (37-78)

65 (45-80) 65 (45-80)

60 (42-80) 60 (40-80)

152 (70.7%)

295 (91.6%)

447 (83.2%)

18% (9-34)

Thrombotic events, n (%)

0 (0.0)

2 (5.3)

2 (4.7)

Grade 3 or 4 bleeding, n (%)

1 (2.7)

2 (5.3)

3 (4.0)

Syncopal events, n (%)

0 (0.0)

16 (43.2%)

22 (57.9%)

27 (62.8)

298

369

667

7 (6-10)

8 (4-14)

8 (5-12)

Red cell usage Transfusion episodes Cycle 1 Cycle 2 Median (IQR) RBC transfused per participant Participants with ≥1 transfusion, n (% of all participants) Cycle 1 Cycle 2

Number of transfusions given for symptoms Median (IQR) self-evaluated health score for symptomatic transfusions Prior to transfusiona Post transfusionb Median (IQR) self-evaluated health score for non-symptomatic transfusions Prior to transfusiona Post transfusionb

Protocol deviations Transfusions given according to hemoglobin trigger, n (%)

Safety and other outcomesc All cause mortality at 3 months, % (95% CI)

≥1 blood culture verified bacterial infection, n (%) Platelet units transfused, n Median (IQR) platelet units transfused per participant

within 24 hours; bat least 24 hours, cSerious adverse events defined as: death, life threatening adverse event, events requiring admission to hospital or prolongation of hospitalization, or resulting in significant disability, including severe sepsis, admission to intensive care unit, major organ dysfunction (single or multi-organ), transient ischemic attack, thromboembolic and ischemic events and acute transfusion reactions. RBC: red blood cells; IQR: interquartile range; CI: confidence interval. a

When the hemoglobin was below threshold, transfusions were given on 152 of 187 (81.3%) and 295 of 447 (66.0%) occasions for the restrictive and liberal strategies respectively. Completion rate for QoL questionnaires was high but reduced over time; 93.0% participants completed both questionnaires at the start and 55.2% at the end. Compliance was similar for both questionnaires. Daily visual analogue scores were completed on 70.3% occasions. No clear overall correlations between daily hemoglobin and daily QoL scores were found, although in an exploratory analysis, QoL scores appeared to favor the liberal threshold in the second treatment cycle (Online Supplementary Figure S2). There was little change in the self-evaluated visual analogue score prior to and post symptom-triggered transfusion; median score 52 (IQR, 40.0-65.0) pre

and 55 (IQR, 42.5-65.0) post. For non-symptomatic transfusions the score was 60 (IQR, 42-80) pre and 60 (IQR, 40-80) post (see Table 1). Safety and other secondary outcomes are reported in Table 2. Numerically higher numbers of culture-verified infections, thrombotic events and grade 3 or 4 bleeding were seen during liberal cycles, although numbers are small. We have previously identified the need for further trials to identify optimal transfusion support for patients with AML.3 In our study, we have successfully recruited patients with AML to follow liberal or restrictive transfusion strategies and demonstrated feasibility of a randomized crossover trial. Only two earlier small randomized trials to date have explored the impact of transfusion thresholds on outcomes in AML. A feasibility trial of patients with

Haematologica | 107 - June 2022

1476

LETTER TO THE EDITOR acute leukemia compared thresholds of 80 g/L versus 120 g/L, but no QoL measures were compared, and the liberal thresholds would be considered high by current standards.4 Another small study compared 70 g/L versus 80 g/L, but the hemoglobin differences between arms appeared too small to generate separation between arms.5 The TRIST study examined hemoglobin thresholds of 70 g/L versus 90 g/L in 300 patients undergoing stem cell transplantation, but the findings cannot be extrapolated to patients with AML.6 Recommendations in guidelines generally advocate for restrictive transfusion policies7,8,9 but QoL is a critical outcome in transfusion-dependent patients with bone marrow failure. Risks including bleeding, thrombosis and immunomodulatory effects need to be weighed against potential QoL benefits for higher hemoglobin thresholds. The numerically higher rate of culture-verified infections in patients on the liberal transfusion strategy needs to be assessed in further studies applying consistent definitions, and of note, in one previously published metaanalysis of health-care associated infection, rates were also reported to be higher in liberal transfusion threshold arms.10 A number of additional learning points were identified that may help inform the design of further studies based on our results. We anticipated some hesitancy to consent and hence allowed 5 days from the start of treatment until randomization; 60.9% declined consent at the outset, the reasons reflecting the burden of decisions at diagnosis (Online Supplementary Table S1). Four participants declined to continue with the study after recruitment, most commonly following at least one cycle of chemotherapy, at one recruiting site, and seemingly because of a perception that the 70 g/L threshold was too difficult to tolerate. Our study re-iterates the importance of support for clinical teams and close monitoring to support protocol adherence. Our study also supports a need for economic evaluations in larger studies. Numerically the excess transfusions in the liberal arm exceeded the two units that may have been anticipated to initially raise the hemoglobin from 70 g/L to 90 g/L, suggesting there may be other on-going contributing factors to RBC requirements. This observation is a recurrent finding,11 and requires investigation regarding RBC loss and/or suppression of erythropoiesis when the hemoglobin is maintained at a higher level. In summary, we have demonstrated the feasibility of randomizing to two hemoglobin thresholds in patients with AML treated with intensive chemotherapy. Given the subjectivity of QoL measurement, the ability for each participant to serve as their own control was an advantage of our study. The findings from our crossover study may be used to inform a larger definitive threshold study addressing QoL as a primary outcome.

Authors Suzy Morton,1,2 Mallika Sekhar,3,4 Heather Smethurst,5 Ana Mora,5 Renate L. Hodge,5 Cara L. Hudson,6 Joseph Parsons,6 Valerie Hopkins5 and Simon J. Stanworth7,8,9 Department of Clinical Hematology, University Hospitals

Birmingham NHS Foundation Trust, Birmingham; 2Medical Department, NHS Blood and Transplant, Birmingham; 3Department of Hematology, University College London Hospitals NHS Foundation Trust, London; 4Royal Free London NHS Foundation Trust, Department of Hematology, London; 5Clinical Trials Unit, NHS Blood and Transplant, Cambridge; 6Clinical Trials Unit, NHS Blood and Transplant, Bristol; 7Department of Clinical Hematology, Oxford University Hospitals NHS Foundation Trust, Oxford; 8Clinical Department, NHS Blood and Transplant, Oxford and 9Radcliffe Department of Medicine, University of Oxford, and NIHR Oxford Biomedical Research Center, Oxford, UK Correspondence: SUZY MORTON - suzy.morton@nhsbt.nhs.uk https://doi.org/10.3324/haematol.2021.279867 Received: October 12, 2021. Accepted: February 16, 2022. Prepublished: February 24, 2022. Disclosures No conflicts of interest to disclose. Contributions SS, SM and MS designed and analyzed the study; RH and VH provided database support; CH and JP performed statistical analysis; HS and AM managed the trial. All authors approved the manuscript. Acknowledgments The authors would like to thank the hospital staff and research teams who conducted this trial and all recruited participants. Participating sites and investigators: Royal Stoke University Hospital: J. Graham; Norfolk and Norwich University Hospitals: H. Lyall; Queen Elizabeth Hospital Birmingham: S. Morton; Russells Hall Hospital: C. Taylor; Sandwell General Hospital: S. Pancham; Royal Oldham Hospital: A. Allameddine; Torbay Hospital: P. Roberts; University College Hospital London: M. Sekhar. We acknowledge A. Newland, P. White, G. Murphy (Data and Safety Monitoring Committee) and S. Knapper, J. Birchall, H. Kaur (Trial Steering Committee). Funding The study was supported by the NHS Blood and Transplant Clinical Trials Unit (H. Thomas, A. Deary, L. Pankhurst). The study was registered (clinicaltrials gov. Identifier: ISRCTN 96390716), received ethics approval (REC number 16/WM/0406) and adopted onto the UK’s National Institute for Health Research (NIHR) Clinical Research Network Portfolio. This manuscript is independent research funded by

Haematologica | 107 - June 2022

1477

LETTER TO THE EDITOR NHS Blood and Transplant (REF15/86). The views expressed in this publication are those of the author(s) and not necessarily those of NHS Blood and Transplant. Data-sharing statement The datasets generated during and/or analysed during the current

study will be available upon request from the NHSBT Clinical Trials Unit after de-identification (text, tables, figures and appendices) 9 months after publication and ending 5 years following article publication. Data will be shared with investigators whose use of the data has been assessed and approved by an NHSBT review committee as a methodologically sound proposal.

References 1. Buccheri G, Ferrigno D, Tamburini M. Karnofsky and ECOG performance status scoring in lung cancer: a prospective, longitudinal study of 536 patients from a single institution. Eur J Cancer. 1996;32A(7):1135-1141. 2. Devlin N, Shah K, Feng Y, Mulhern B, van Hout B. Valuing healthrelated quality of life: an EQ-5D-5L value set for England. Health Econ. 2018;27(1):7-22. 3. Loke J, Lowe DM, Miller LJ, et al. Supportive care in the management of patients with acute myeloid leukaemia: where are the research needs? Br J Haematol. 2020;190(3):311-313. 4. Webert KE, Cook RJ, Couban S, et al. A multicenter pilotrandomized controlled trial of the feasibility of an augmented red blood cell transfusion strategy for patients treated with induction chemotherapy for acute leukemia or stem cell transplantation. Transfusion. 2008;48(1):81-91. 5. DeZern AE, Williams K, Zahurak M, et al. Red blood cell transfusion triggers in acute leukemia: a randomized pilot study. Transfusion. 2016;56(7):1750-1757. 6. Tay J, Allan DS, Chatelain E, et al. Liberal versus restrictive red blood cell transfusion thresholds in hematopoietic cell transplantation: a randomized, open label, phase III,

noninferiority trial. J Clin Oncol. 2020;38(13):1463-1473. 7. National Institute for Health and Care Excellence. Blood Transfusion (NICE Guideline NG24) [Internet]. 2015 [Cited 10/05/2021]. Available from https://www.nice.org.uk/guidance/ng24. 8. Carson JL, Guyatt G, Heddle NM, et al. Clinical practice guidelines from the AABB: red blood cell transfusion thresholds and storage. JAMA. 2016;316(19):2025-2035. 9. National Blood Authority Australia. The Patient Blood Management Guidelines Companions. [Internet] 2014. [Cited 10/05/2021]. Available from https://www.blood.gov.au/system/files/documents/patient_bloo d__management_guidelines_companions.pdf 10. Rohde JM, Dimcheff DE, Blumberg N, et al. Health careassociated infection after red blood cell transfusion: a systematic review and meta-analysis. JAMA. 2014;312(19):13171326. 11. Stanworth SJ, Killick S, McQuilten ZK, et al. Red cell transfusion in outpatients with myelodysplastic syndromes: a feasibility and exploratory randomised trial. Br J Haematol. 2020;189(2):279-290.

Haematologica | 107 - June 2022

1478

LETTER TO THE EDITOR

Reduced immunogenicity of a third COVID-19 vaccination among recipients of allogeneic hematopoietic stem cell transplantation Previous allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a risk factor for severe COVID-19 with mortality rates that may exceed 20%.1,2 The efficiency of two doses of mRNA-based COVID-19 vaccines is reportedly lower in allo-transplanted patients than in healthy controls with rates of seronegativity or failure to seroconvert in 15-31%.3-6 In a study of allo-transplanted patients with insufficient responses to two doses of the BNT162b2 (Pfizer-BioBTech) mRNA vaccine, only 48% of patients reached a putative threshold (4,160 arbitrary units [AU]/mL, corresponding to 590 World Health Organization [WHO] standard binding antibody units [BAU]/mL) of immunoglobulin G (IgG) against the receptor-binding domain of the spike 1 (S1) protein (anti-RBD) following a third vaccine dose.7 Forty recipients of allo-HSCT for hematological malignancies were identified in local transplant registries of the Region Western Götaland (population of approximately 1.7 million) and accepted participation in this sub-study within the DurIRVac study (clinicaltrials gov. Identifier: Eu-

draCT no. 2021-000349-42) at the Sahlgrenska University Hospital. All participants gave written informed consent before enrolment. The DurIRVac study was approved by the Swedish Ethical Review Authority (permit no. 202003276, 2021-00374 and 2021-00539) and by the Swedish Medical Products Agency (permit no. 5.1-2021-11118). All patients fulfilled national criteria from the Public Health Agency of Sweden (www.folkhalsomyndigheten.se) for receiving a third dose namely: (i) having undergone transplantation within 3 years or (ii) having ongoing immunosuppressive treatment for graft-versus-host-disease (GvHD). The European Society for Blood and Marrow transplantation (EBMT) guidelines for COVID-19 vaccination were also followed (www.ebmt.org, version 6.0, May 31, 2021). Three patients were excluded based on previously confirmed COVID-19. All enrolled patients (n=37) had received two doses of COVID-19 mRNA vaccine ≥8 weeks prior to screening. The median time from transplantation to the third vaccination was 23 months (min-max 6-191). Twenty-one (57%)

Table 1. Patient and treatment characteristics by responders and non-responders to the third dose of mRNA vaccine.

Anti-RBD IgG in serum

T-cell reactivity (S1-γ)

Positive Negative Positive Negative P-value1 P-value2 (≥14 BAU/mL) (<14 BAU/mL) (≥10 pg/mL IFN-γ) (<10 pg/mL IFN-γ) All patients5 (n=37)

31 (84 %)

6 (16%)

19 (51 %)

18 (48 %)

21/10

3/3

13/6

7/11

60 (19-78)

63 (32-72)

64 (19-70)

60 (26-78)

123 (56–157)

139 (127–174)

123 (74-149)

127 (56-174)

26 (6–188)

19 (13–191)

22 (7-188)

30 (6-191)

Ongoing IST6 (yes/no)

20/11

5/1

9/10

15/3

0.0464

Ongoing prednisone (yes/no)

16/15

2/4

12/7

13/5

0.034

Vaccine (Pfizer/Moderna) Age in years at vaccination median (range) Median days between dose 2 and 3 (range) Median months from allo-HSCT, (range)

0.013

All comparisons refer to patients responding or not responding to third dose vaccination by anti-RBD IgG or SARS-CoV-2-specific T-cell reactivity. 3MannWhitney U-test. 4Chi-square test. 5Patients had received allogenic HSCT for acute myeloid leukemia (n=15 patients), acute lymphoblastic leukemia (n=4), myelodysplastic syndrome (n=5), myelofibrosis (n=4), chronic myeloid leukemia (n=4), atypical chronic myeloid leukemia (n=1), myeloma (n=1), chronic lymphocytic leukemia (n=1), Hodgkin disease (n=1), and STAT-1 immune deficiency (n=1). 6Immunosuppressive therapy comprising prednisone (n=20 patients), photopheresis (n=5), ibrutinib (n=4), ruxolitinib (n=5), photopheresis (n=5), dasatinib (n=2), cyclosporine (n=2), daratumumab (n=1), imatinib (n=1), carfilzomib (n=1), and/or ponatinib (n=1). allo-HSCT: allogeneic hematopoietic stem cell transplantation; BAU: binding antibody untits; IgG: immunoglobulin G; S1: spike 1 protein; IST: immunosuppressive therapy; IFN-g: interferon-g. 1,2

Haematologica | 107 - June 2022

1479

LETTER TO THE EDITOR of participants had chronic GvHD and 25 (68%) received immunosuppressive therapy (IST) at the time of vaccination (Table 1). Patients were given the same vaccine as in their initial two doses, i.e., either BNT162b2 (Pfizer-BioBTech Comirnaty®; n=24) or mRNA-1273 (Moderna Spikevax®; n=13) at a median 127 days (min-max 56-174) after the second dose. Peripheral blood was collected immediately before and 4 weeks (median 24 days, range, 19-30) after the third vaccination. Patients completed a questionnaire 2 weeks after the third dose to assess side-effects, categorized according to the CTCAE (common terminology criteria for adverse events) standards. Severity of GvHD was additionally assessed from medical records. Chemiluminescent microparticle immunoassays were performed on serum using the automated Alinity system for analysis of IgG antibodies against RBD (SARS-CoV-2 IgG II Quant, Abbott, Illinois, USA) with levels reported in the WHO international standard BAU/mL (quantitative detection range, 14-5,680 BAU/mL), which correlate well with neutralizing antibody levels.8 In order to assess T-cell responses 1 mL of peripheral blood, collected in heparinized tubes, was stimulated with peptides spanning the N-terminal spike 1 (S1) domain of the SARS-CoV-2 surface glycoprotein. After 2 days of incubation at 37°C plasma was recovered for analysis of interferon-g (IFN-g) by enzymelinked immunosorbant assay (ELISA). This assay captures SARS-CoV-2-specific reactivity of CD4+ and CD8+ T cells with high specificity and sensitivity.9 S1-induced IFN-g production is presented with levels in unstimulated samples subtracted using a limit of detection of 10 pg/mL. Statistical analyses were performed using SPSS statistical software package (version 24) or GraphPad Prism software (version 9). The majority (31/37, 84%) of allo-HSCT patients responded to the third dose vaccination by increased anti-RBD IgG levels (Figure 1A). A subgroup (12/37, 32%) achieved very high antibody levels (>5,680 BAU/mL). However, among the

14 patients seronegative prior to the third dose vaccination, six (42%) remained seronegative 4 weeks after the third vaccine dose (Figure 1A). All patients who were seropositive before the third dose (23/37, 62%) achieved antibody responses exceeding 100 BAU/mL, a level above which has been proposed to provide protection against COVID-19.10 The characteristics of responders and non-responders to the third vaccine dose are detailed in Table 1. No significant differences in serological responses were noted among patients with or without chronic GvHD or ongoing IST. Regarding T-cell immune response, 18 of 37 (49%) were devoid of measurable responses 4 weeks after the third vaccination (Figure 1B). T-cell responses tended to be lower in patients with chronic GvHD and were significantly diminished in patients receiving IST, in particular among those receiving prednisone (Table 1). Seronegativity prior to the third dose predicted poor humoral and cellular responses after vaccination. Treatment with IST was associated with insufficient T-cell responses, more so than time from transplantation. Furthermore, four of five (80%) of patients on ruxolitinib showed no T-cell reactivity. Of note, among the 14 patients who were seronegative for antiRBD IgG prior to the third dose, 11 (79%) also lacked a Tcell response after three vaccine doses, compared with seven of 23 (30%) among those seropositive prior to the third dose (P<0.01, chi-square test). Seronegativity prior to the third vaccination was non-significantly associated with ongoing GvHD (9/14 vs. 12/23 in seropositive patients) and IST (10/14 vs. 15/23). Additionally, a lower fraction of patients mounted SARS-CoV-2 specific T-cell responses than developed anti-RBD IgG after three vaccinations (P<0.01, chi-square test). Of six patients who remained seronegative after three vaccine doses, five (83%) were also devoid of specific T-cells. Vaccine-reported adverse events were observed in 15 (41%) patients after the third dose, with the majority of these categorized as mild local injection-related reactions. No exacerbations of GvHD were noted.

B Figure 1. Serological and virus-specific T-cell responses to the spike 1 protein receptor region of SARS-CoV-2 before and after the third dose of of COVID-19 vaccines in allo-transplanted patients. (A) Shows serum levels of immunoglobulin G (IgG) against the receptor-binding domain (RBD). (B) Shows interferon-g (IFNg) production in supernatant plasma following stimulation of whole blood with spike 1 peptides, reflecting reactivity of SARSCoV-2-specific T cells. The upper dotted line represents the cutoff value of 590 binding antibody units [BAU]/mL (i.e., corresponding to 4,160 Abbott Arbitrary Units [AU]/mL) while the middle dotted line corresponds to 100 BAU/ml and the lower dotted line represents the limit of detection (LOD) for respective assay. Statistical comparison by Wilcoxon matched pairs test (n=37). P-values are two-sided and are designated as follows: **P<0.01, ****P<0.0001.

Haematologica | 107 - June 2022

1480

LETTER TO THE EDITOR The main findings in this study were that a significant fraction of allo-transplanted patients failed to produce anti-RBD IgG (16%) and that 48% of patients did not mount measurable SARS-CoV-specific T cells despite three vaccinations. Our results confirm and extend a previous report of insufficient anti-RBD responses among allo-transplanted patients7 to imply that the inherent and treatment-induced T-cell deficiency associated with allotransplantation may translate into lack of COVID-19 mRNA vaccine efficacy. The interval between dose 2 and 3 was longer among patients remaining seronegative following the third dose, implying that a shorter interval between vaccinations may improve responses. Our results additionally suggest that the SARS-CoV-2specific T-cell response to vaccination is more affected than the humoral response among allo-transplanted patients, based on the finding that a significantly higher fraction of patients showed complete deficiency of T-cell responsiveness to SARS-CoV-2-derived peptides compared with those remaining seronegative. Using the same T-cell assay, we have previously shown that 13 of 13 (100%) of healthy donors developed detectable T-cell responses 4 weeks after the second SARS-CoV-2 vaccine dose.11 Notably, 35% of allo-transplanted patients lacked T-cell reactivity against S1 peptides despite mounting anti-RBD IgG. The clinical relevance of the observed T-cell deficiency remains to be established. In conclusion, the third dose of COVID-19 mRNA vaccine resulted in elevated antibody titres and measurable SARSCoV-2-S1 T-cell responses in many allo-transplanted patients. However, a substantial proportion of patients did not respond by antibody formation and/or SARS-CoV-2-specific T cells, highlighting the need for additional preventive measures and continued vigilance in this cohort.

Research, Department of Infectious Diseases, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg; 3Region Västra Götaland, Sahlgrenska University Hospital, Department of Clinical Microbiology, Gothenburg; 4

Department of Infectious Diseases, Institute of Biomedicine,

Sahlgrenska Academy, University of Gothenburg, Gothenburg and 5

Department of Cellular Therapy and Allogeneic Stem Cell

Transplantation, Karolinska Comprehensive Cancer Center, Karolinska University Hospital, Huddinge and Division of Hematology, Department of Medicine, Huddinge, Karolinska Institutet, Stockholm, Sweden. Correspondence: MARTIN LAGGING - martin.lagging@medfak.gu.se https://doi.org/10.3324/haematol.2021.280494 Received: December 9, 2021. Accepted: February 22, 2022. Prepublished: March 3, 2022. Disclosures No conflicts of interest to disclose. Contributions SE and ML were responsible for designing and writing the protocol, conducting the study, extracting and analysing data, interpreting results, writing the letter, updating reference lists and creating the table and figure; AM and KH were responsible for designing and writing the protocol, extracting and analyzing data, interpreting results, writing the letter, updating reference lists and creating the table and figure; MN participated in interpreting results, writing the letter, and creating the figure; HGW, MA, and AT performed the Tcell assays as well as participated in extracting and analyzing data and interpreting results; KV, JW, and JR participated in extracting and analysing data and interpreting results; TB, MB, and PL participated in designing and writing the protocol, interpreting results, writing the letter, and updating the reference list.

Authors

Funding This work was supported by the Swedish Medical Research Council

Sigrun Einarsdottir, Anna Martner, Malin Nicklasson, Hanna Grauers

(Vetenskapsrådet; grant number 2021-04779) and ALF Funds at

Wiktorin,2 Mohammad Arabpour,2,3 Andreas Törnell,2 Krista Vaht,1

Sahlgrenska University Hospital (grant number ALFGBG-438371).

Jesper Waldenström,3,4 Johan Ringlander,3,4 Tomas Bergström,3,4 Mats Brune,1 Kristoffer Hellstrand,2,3 Per Ljungman5 and Martin

Data sharing statement

Lagging

The original data and protocols may be made available to other

3,4

investigators after contact with the corresponding author. 1

Department of Hematology and Coagulation, Institute of

Medicine, Sahlgrenska Academy, University of Gothenburg,

Clinical trial registration

Gothenburg; TIMM Laboratory, Sahlgrenska Center for Cancer

EudraCT 2021-000349-42

References 1. Sharma A, Bhatt NS, St Martin A, et al. Clinical characteristics and outcomes of COVID-19 in haematopoietic stem-cell

transplantation recipients: an observational cohort study. Lancet Haematol. 2021;8(3):e185-e193.

Haematologica | 107 - June 2022

1481

LETTER TO THE EDITOR 2. Ljungman P, de la Camara R, Mikulska M, et al. COVID-19 and stem cell transplantation; results from an EBMT and GETH multicenter prospective survey. Leukemia. 2021;35(10):28852894. 3. Redjoul R, Le Bouter A, Beckerich F, Fourati S, Maury S. Antibody response after second BNT162b2 dose in allogeneic HSCT recipients. Lancet. 2021;398(10297):298-299. 4. Bergman P, Blennow O, Hansson L, et al. Safety and efficacy of the mRNA BNT162b2 vaccine against SARS-CoV-2 in five groups of immunocompromised patients and healthy controls in a prospective open-label clinical trial. EBioMedicine. 2021;74:103705. 5. Dhakal B, Abedin S, Fenske T, et al. Response to SARS-CoV-2 vaccination in patients after hematopoietic cell transplantation and CAR T-cell therapy. Blood. 2021;138(14):1278-1281. 6. Ram R, Hagin D, Kikozashvilli N, et al. Safety and immunogenicity of the BNT162b2 mRNA COVID-19 vaccine in patients after allogeneic HCT or CD19-based CART therapy - a

single-center prospective cohort study. Transplant Cell Ther. 2021;27(9):788-794. 7. Redjoul R, Le Bouter A, Parinet V, Fourati S, Maury S. Antibody response after third BNT162b2 dose in recipients of allogeneic HSCT. Lancet Haematol. 2021;8(10):e681-e683. 8. Kristiansen PA, Page M, Bernasconi V, et al. WHO international standard for anti-SARS-CoV-2 immunoglobulin. Lancet. 2021;397(10282):1347-1348. 9. Brand I, Gilberg L, Bruger J, et al. Broad T cell targeting of structural proteins after SARS-CoV-2 infection: high throughput assessment of T cell reactivity using an automated interferon gamma release assay. Front Immunol. 2021;12:688436. 10. Hall VG, Ferreira VH, Ku T, et al. Randomized trial of a third dose of mRNA-1273 vaccine in transplant recipients. N Engl J Med. 2021;385(13):1244-1246. 11. Tornell A, Grauers Wiktorin H, Ringlander J, et al. Rapid cytokine release assays for analysis of SARS-CoV-2-specific T cells in whole blood. J Infect Dis. 2022 Jan 12 [Epub ahead of print].

Haematologica | 107 - June 2022

1482

LETTER TO THE EDITOR

Tesidolumab (LFG316) for treatment of C5-variant patients with paroxysmal nocturnal hemoglobinuria Paroxysmal nocturnal hemoglobinuria (PNH) is caused by clonal expansion of hematopoietic stem cells that carry a somatic mutation in the X-linked gene PIG-A (phosphatidylinositol glycan anchor biosynthesis class A). The mutation leads to a deficiency of glycosylphosphatidylinositol-anchored membrane proteins.1-3 The loss of membrane-associated complement regulatory proteins CD55 and CD59 increases susceptibility of red blood cells and platelets to complement-mediated lysis, leading to hemolytic anemia, thrombophilia and reduced life expectancy in untreated patients.2,4 The currently approved monoclonal antibodies targeting C5 eculizumab5 and ravulizumab6 significantly reduce intravascular hemolysis and transfusion dependency and improve the life expectancy of PNH patients.7,8,9 However, there are several, mainly Japanese patients who do not respond to eculizumab,10 due to a variant C5 protein sequence with an arginine-to-histidine change at position 885 (Arg885His) that prevents eculizumab binding.1,10 Ravulizumab binds to the same epitope as eculizumab,6 so it cannot bind to this C5 variant, either. Tesidolumab (LFG316) is a fully-human IgG1/λ anti-C5 monoclonal antibody of 143 kDa (without glycosylation), with a half-life in humans of approximately 9 days. Tesidolumab blocks cleavage of C5 and prevents subsequent formation of the membrane attack complex. Crystal structure analysis of tesidolumab Fab complexed to C5 demonstrated that tesidolumab bound to a distinct epitope to that of eculizumab and ravulizumab, distant to Arg 885 (Figure 1A). Consistent with this binding mode, tesidolumab inhibited both variant and non-variant C5 activation in a functional assay (Figure 1B). An open-label, single-arm, multicenter, proof-of-concept phase II trial was conducted at seven centers in three countries to test the efficacy of tesidolumab in patients with variant and non-variant C5. The study comprised three treatment periods, namely a 4-week treatment period 1 (days 1-29), followed by an optional 48-week treatment period 2 (days 30-365), after which an interim analysis was performed, followed by an optional treatment period 3 that allowed a maximal treatment extension up to week 312. The primary endpoint was serum lactate dehydrogenase (LDH) reduction on day 29, and secondary endpoints involved monitoring of safety, tolerability, and tesidolumab pharmacokinetics. Exploratory endpoints included the assessment of hemoglobin levels, blood transfusion requirements, free hemoglobin, reticulocyte counts, bilirubin and FACIT fatigue score and pharmacodynamics measurements including sC5b-9 and the CH50 assay (a measure of ex vivo serum hemolytic activity). Adult PNH patients with a PNH clone size of ≥10% and

serum LDH levels ≥1.5-fold above the upper limit of normal (ULN) were included in the study. Additional requirements were vaccination against Neisseria meningitidis types A, C, Y and W-135 and, if available and acceptable by local regulations, vaccination against Neisseria meningitidis type B at least 2 weeks prior to first dosing. Treatment with corticosteroids and/or other immunosuppressive regimens could continue if indicated for treatment of autoimmune disease (e.g., aplastic anemia). Key exclusion criteria were history of recurrent meningitis or meningococcal meningitis despite vaccination, active infection, history of hematopoietic stem cell transplantation, positive HIV test, known or suspected hereditary complement deficiency, and severe concurrent co-morbidities (e.g., advanced cardiac disease, severe pulmonary arterial hypertension). Cytopenic patients with neutrophils <0.5x109/L or platelets <30x109/L were excluded to avoid confounding significant bone marrow failure. All centers received approval from independent ethics committees and regulatory bodies. The study was conducted in accordance with the principles of the Declaration of Helsinki. Signed informed consent was obtained from each patient before any study-related procedures were undertaken. Trial information was published on https://ClinicalTrials.gov before first patient first visit, clinicaltrails gov. Identifier: NCT02534909, with the investigational product tesidolumab designated as LFG316. Nine patients (5 C5-variant, 4 C5 non-variant; 4 females, 5 males) were enrolled between Sep 2015 and March 2017 (Table 1). Five of the nine patients were transfusion-dependent (3 of these were C5-variant patients), having received an average of 16.2 (range 8-30; standard deviation [SD] 8.9) units of erythrocytes in the year prior to screening. The patients had an average PNH clone size (type III erythrocytes) at baseline of 29.5 % (SD 10.5%), mean LDH of 1,270 U/L (SD 520) and mean hemoglobin levels of 93.6 g/L (SD 25.5). Six of the patients had previously used eculizumab, but had stopped eculizumab at least 2 years before starting tesidolumab. Patients received intravenous tesidolumab every second week at a dose of 20 mg per kg of body weight, infused over approximately 2 hours in period 1 and over 40 minutes to 2 hours thereafter. This dose was selected based on modeling of PK and PD data from a prior phase I clinical study in healthy subjects (data on file). At the time of cutoff, the nine PNH patients were successfully treated with tesidolumab for an average of 405 days (range 210-505 days), and eight of nine patients had completed treatment periods 1 and 2 (day 365). Tesidolumab therapy rapidly decreased LDH levels and sustained the decrease over the first year of therapy in all pa-

Haematologica | 107 - June 2022

1483

LETTER TO THE EDITOR tients (Figure 2A). The mean relative LDH reduction from baseline was 79.2% (SD 8.9%) at week 4 (n=8) and 78.8% (SD 11.4%) at week 52 (n=8) (Figure 2B). LDH decline coincided with a meaningful decrease in transfusion requirement and increase in hemoglobin levels. Up until the cutoff, only three patients (one with variant C5) required red blood

cell transfusions after initiation of tesidolumab, with an average of 1.5 units/year each (Figure 2C). Concomitantly, mean hemoglobin levels increased from 93.6 g/L (SD 25.5) at baseline to 112.8 g/L (SD 24.7) at week 52, with two patients achieving hemoglobin levels of >120 g/L (Figure 2D). With the exception of reticulocytes, other markers of

Table 1. Patient characteristics and adverse events.

Patient Age (years) A B C D E F G H I Mean

36 42 45 66 35 27 36 45 39 41.2

Adverse events

Sex

Race

C5 status

Male Female Male Female Female Female Male Male Male

Caucasian Caucasian Asian Asian Asian Asian Asian Asian Caucasian

Non-variant Non-variant Variant Variant Variant Non-variant Variant Variant Non-variant

Tesidolumab 20 mg/kg N=9 nE, nS (%)

Weight (kg)

BMI (kg/m2)

91.0 68.0 75.8 46.8 63.9 56.5 67.1 99.9 108.0 75.2

29.7 25.6 24.8 22.3 21.9 21 24.1 31.9 31.6 25.9

Preferred term

Tesidolumab 20 mg/kg. N=9, nS (%)

Total AE AE of mild intensity

73, 7 (77.8) 59, 7 (77.8)

Any AE Headache Nasopharyngitis

7 (77.8) 4 (44.4) 4 (44.4)

AE of moderate intensity

12, 6 (66.7)

Nausea

2, 1 (11.1)

Abdominal pain upper Back pain

AE of severe intensity

Preferred term

87.0 120.0 145.0 80.0 87.0 91.0 69.0 63.0 100.5 93.6

Tesidolumab 20 mg/kg. N=9, nS (%)

Serious AE

3 (33.3)

Infection

1 (11.1)

2 (22.2) 2 (22.2)

Study drug-related AE Headache

3 (33.3)

Migraine

2 (22.2)

17, 4 (44.4)

Serious AE

2, 2 (22.2)

Cystitis

2 (22.2)

Migraine

2 (22.2)

Neutropenia Oedema

2 (22.2) 2 (22.2)

Study drug-related AE leading to study discontinuation

2084 714 1690 938 887 1368 1937 776 1018 1268

1 (11.1)

Study drug-related AE

11 0 30 20 0 8 12 0 0 9

Enterocolitis viral

Blood creatine phosphokinase increased

AE leading to discontinuation of study treatment

Units Lactate Hemoglobin transfused dehydrogenase (g/L) in last year (U/L)

2 (22.2)

Neutropenia

2 (22.2)

Atrioventricular block first degree

1 (11.1)

Enterocolitis viral

1 (11.1)

Leukopenia

1 (11.1)

Migraine with aura

1 (11.1)

Nausea Sinus bradycardia

1 (11.1) 1 (11.1)

nE: number of events; nS: number of subjects; AE: adverse events; BMI: body mass index. Haematologica | 107 - June 2022

1484

LETTER TO THE EDITOR

Figure 1. Crystallographic structure of the complex between complement protein 5 (C5) and tesidolumab Fab and in vitro activity on non-variant C5 and variant C5. (A) Human C5 (CompTech) was mixed with a 2.5-fold molar excess of tesidolumab Fab. The complex was purified by size exclusion chromatography and crystallized by vapor diffusion. Crystals diffracting to 3.5Å resolution were grown at 292 K from 0.07 M sodium acetate pH 4.6, 5.6% polyethylene glycol 4,000, 30% glycerol. A full diffraction data set was measured at the Swiss Light Source (Villigen), with a Pilatus pixel detector. The crystals were in space group P43 with one C5 complex per asymmetric unit. The structure was determined by molecular replacement, using a published C5 structure (PDB entry 3CU7) and a high-resolution structure of the tesidolumab Fab determined in house, and refined using standard crystallographic methods. Highlighted to the right is the variant location. (B) Inhibition of non-variant C5 (solid curves) and variant C5 (hatched curves) by tesidolumab (blue) or eculizumab (red). Tesidolumab blocks both C5 types, whereas variant C5 is resistant to eculizumab. The alternative pathway Wieslab Assay (EuroDiagnostica, Malmö, Sweden) was performed according to the manufacturer’s instructions using human C5-depleted serum (CompTech, Tyler, TX) spiked with 7 mg/mL nonvariant or variant C5 (CompTech, Tyler, TX, or Novartis in-house production). Tesidolumab or eculizumab were tested at concentrations ranging from 0.02-15 mg/mL. Percent complement activity was calculated using the formula: (sample – negative control)/(positive control – negative control) x100.

hemolysis showed similar improvements (data not shown). Free hemoglobin, which averaged 411 mg/L (SD 224 mg/L) at baseline (upper limit of detection was 450 mg/L), was reduced to 153 mg/L (SD 66 mg/L). Total bilirubin decreased from 19.4 mmol/L (SD 10.1 mmol/L) at baseline to 16.4 mmol/L (SD 9.1 mmol/L) at week 52. Type III erythrocytes increased from 29.5% (SD 10.5%) at baseline to 42.5% (SD 18.5%) at the end of year 1. Reticulocytes were at 129x109/L (SD 71x109/L) at baseline and remained elevated during tesidolumab therapy (139x109/L [SD 71x109/L] at week 52). Trough concentrations of tesidolumab increased slightly over time from 134 mg/mL (SD 31 mg/mL) after the first dose to 184 mg/mL (SD 32 mg/mL) at week 4 and 264 mg/mL (± 72 mg/mL) at week 52 (Figure 2E). Tesidolumab efficiently suppressed terminal complement activity. Soluble C5b-9 was reduced from 138 ng/mL at baseline (SD 62 ng/mL) to 40 ng/mL (SD 17 ng/mL) at 4 weeks and 64 ng/mL ± 9 ng/mL after 1 year of therapy (data not shown). Ex vivo CH50 assays showed a reduction of complement activity from 115 mg eq/mL (SD 35 mg eq/mL) at baseline to 1 mg eq/mL (SD 1 mg eq/mL) at week 4 and stabilized at about 2 mg eq/mL (range 1-3 mg eq/mL) at week 52 (data not shown). Importantly, tesidolumab treatment improved quality of life

(QoL) in all patients, as measured by the Functional Assessment of Chronic Illness Therapy (FACIT) fatigue score,11 the mean of which increased from 38.6 (SD 10.0) at baseline to 44.0 (SD 9.1) at week 4 and to 46.9 (SD 6.5) at week 52 (Figure 2F). Adverse events are presented in Table 1. Overall, tesidolumab was well tolerated. Seven of nine patients reported adverse events, the majority of which were mild. Two serious adverse events were reported, one acute infection of moderate severity on day 41 deemed unrelated to the study drug, the other a viral enterocolitis classified as severe and suspected to be related to the study drug. Both individuals responded well to therapy and findings resolved without complications. The most frequent study drug-related adverse events (preferred terms) were headache (n=3), migraine (n=2), and neutropenia (n=2). No thromboembolic events were observed throughout the study (data not shown), and D-dimer levels were reduced from 0.37 FEU/L (SD 0.29 FEU/L) at baseline to 0.20 FEU/L (SD 0.05 FEU/L) at week 52. In addition, tesidolumab had no apparent effect on platelet counts and renal function (eGFR) was stable in all patients. In summary, tesidolumab had a favorable safety profile and

Haematologica | 107 - June 2022

1485

LETTER TO THE EDITOR

Figure 2. Key clinical endpoints. (A) Tesidolumab treatment reduced LDH in all patients. The spike in lactate dehydrogenase(LDH) after week 30 in subject F coincided with an episode of viral enterocolitis (reported as serious adverse event [SAE]) causing a 1week delay of tesidolumab infusion. (B) Tesidolumab treatment reduced LDH levels by an average of about 75% compared to baseline. (C) Tesidolumab treatment reduced erythrocyte transfusion need compared to the year before therapy. Blood transfusions (units) were counted starting 12 months prior to the screening visit until the cutoff date and calculated as annual rate (units/year) before () and after () the start of tesidolumab treatment. Two patients given transfusions after the first dose of tesidolumab received erythrocyte units 1-2 months after initiation of therapy. The third individual was transfused 7 months after start of therapy. No further blood transfusions were given. (D) Tesidolumab treatment increased hemoglobin levels in both C5-variant and non-variant PNH patients, but only 2 patients reached hemoglobin levels observed for healthy individuals (>120 g/L). (E) Tesidolumab concentration in blood. Shown are peak and pre-dose levels for the first three treatments, followed by pre-dose levels only until the end of the study, Functional Assessment of Chronic Illness Therapy (FACIT) score (F) improved during the course of the study. LLOQ: lower level of quantification; ULN: upper limit of normal; LLN: lower limit of normal. Haematologica | 107 - June 2022

1486

LETTER TO THE EDITOR was efficacious for PNH patients with either variant or nonvariant C5. Significant decreases of transfusion dependency and reductions of LDH concentrations to near normal levels were observed in all patients. As observed for eculizumab,12,13 there was evidence of residual extravascular hemolysis. Reticulocyte counts and bilirubin levels remained elevated throughout treatment. Further, a proportion of type III red blood cells were coated with C3 fragments suggesting residual proximal complement pathway activation. These cells are expected to be susceptible to extravascular hemolysis as demonstrated by the lack of hemoglobin normalization in patients undergoing anti-C5 therapy. Thus, additional therapy may be necessary to achieve an optimal response in these patients.

Novartis has produced tesidolumab for commercial purposes for patients with complement-driven disorders. The academic authors have the following to disclose: YK received research funding by Novartis, Chugai and Alexion, is a member of advisory committees for Chugai and Alexion, speaker bureau for Alexion and consults for Chugai. JN received research funding from Novartis, Chugai and Alexion, is a member of advisory committees for Alexion and Chugai, speaker bureau for Alexion. YI is on the speaker bureau for Alexion Pharma LLC, Jansen Pharmaceutical KK and Eisai KK. LG receieved research funding by Novartis. MM, JM, HN and KA have non conflicts of interest to disclose. Novartis is committed to sharing with qualified external researchers, access to patient-level data, and supporting the clinical documents from eligible studies. These requests are reviewed and approved by an independent review panel on the basis of scientific merit. All data provided are

Authors

anonymized to respect the privacy of patients who have participated in the trial in line with applicable laws and regulations. The availability of this trial data is according to the criteria and 1*

Jun-ichi Nishimura, Kiyoshi Ando, 4

Masayoshi Masuko, Hideyoshi

process described on www.clinicalstudydatarequest.com.

Noji, Yoshikazu Ito, Jiri Mayer, Laimonas Griskevicius, Christoph Bucher,8 Florian Müllershausen,8 Peter Gergely,8 Izabela Rozenberg,8 8

Anna Schubart, Raghav Chawla, Jean-Michel Rondeau, Michael 8

Roguska, Igor Splawski, Mark T. Keating, Leslie Johnson, 8

Rambabu Danekula, Morten Bagger, Yoko Watanabe, Börje 8

1,9

Haraldsson and Yuzuru Kanakura

Contributions YK and JN co-designed the study together with Novartis and were co-authors of all aspects of this paper; KA, MM, HN, YI, JM and LG were principal investigators that closely monitored their patients; IS, MR and MTK initiated the identification of tesidolumab, and progressed its development through engineering, characterization,

Department of Hematology and Oncology, Osaka University Graduate 2

selection, and preclinical toxicology studies; MR conceived the

School of Medicine, Suita, Japan; Department of Hematology and

idea of testing tesidolumab in the C5-variant PNH patients; LJ was

Oncology, Tokai University, Isehara, Japan

responsible for the generation of variant C5 reagents and in vitro

Department of Hematology, Endocrinology and Metabolism, Niigata 4

analytical data comparing tesidolumab and eculizumab on variant

University Medical and Dental Hospital, Niigata, Japan; Department

C5; BH, CB, FM and PG co-designed the study and BH and RC were

of Cardiology and Hematology, Fukushima Medical University,

the global Novartis Medical Leads of the study; YW was the

Fukushima, Japan; Department of Hematology, Tokyo Medical

Novartis Medical Lead in Japan; IR was the Novartis Clinical Trial

University, Shinjuku-ku, Tokyo, Japan; 6Department of Internal

Lead. MB designed and analyzed the pharmacokinetic parts of the

Medicine, Hematology and Oncology, Masaryk University Hospital and

study; AS was the Novartis Research Lead for tesidolumab; JMR

Faculty of Medicine, Brno, Czech Republic; Hematology, Oncology

did the crystallographic analysis of tesidolumab; RD was the

and Transfusion Medicine Center, Vilnius University Hospital Santaros

statistical expert of the study; AS, MR, RC, BH and IS wrote the

Klinikos and Institute of Clinical Medicine, Vilnius University, Vilnius,

manuscript and all authors reviewed it and contributed to its

Lithuania; Novartis Institutes for BioMedical Research, Basel,

finalization.

Switzerland, Cambridge, USA, and Novartis Pharma KK, Toranomon Minato-ku, Tokyo, Japan and 9Sumitomo Hospital, Osaka, Japan

Funding This study was sponsored by Novartis Institute of Biomedical

*JN and YA contributed equally as co-first authors.

Research.

Correspondence:

Data-sharing statement

Jun-ichi Nishimura - junnishi@bldon.med.osaka-u.ac.jp

Novartis is committed to sharing with qualified external researchers,

https://doi.org/10.3324/haematol.2020.265868

access to patient-level data and supporting clinical documents from

Received: August 6, 2020.

eligible studies. These requests are reviewed and approved by an

Accepted: March 1, 2022.

independent review panel on the basis of scientific merit. All data

Prepublished: March 10, 2021.

provided are anonymized to respect the privacy of patients who have participated in the trial in line with applicable laws and

Disclosures

regulations. The availability of this trial data is according to the

The following authors were employees at Novartis at the time of the

criteria and process described on

study: CB, FM, PG, IR, AS, JMR, MR, IS, MTK, LJ, MB, YW, RD, RC and BH.

www.clinicalstudydatarequest.com.

Haematologica | 107 - June 2022

1487

LETTER TO THE EDITOR

References 1. Kanakura Y, Kinoshita T, Nishimura J-I. Paroxysmal Nocturnal Hemoglobinuria - From Bench to Bedside: First edition: Springer Japan. 2017. 2. Devalet B, Mullier F, Chatelain B, Dogné JM, Chatelain C. Pathophysiology, diagnosis, and treatment of paroxysmal nocturnal hemoglobinuria: a review. Eur J Haematol. 2015;95(3):190-198. 3. Brodsky RA. Paroxysmal nocturnal hemoglobinuria. Blood. 2014;124(18):2804-2811. 4. Hillmen P, Muus P, Dührsen U, et al. Effect of the complement inhibitor eculizumab on thromboembolism in patients with paroxysmal nocturnal hemoglobinuria. Blood. 2007;110(12):41234128. 5. Kaplan M. Eculizumab (Alexion). Curr Opin Investig Drugs. 2002;3(7):1017-1023. 6. Lee JW, Kulasekararaj AG. Ravulizumab for the treatment of paroxysmal nocturnal hemoglobinuria. Expert Opin Biol Ther. 2020;20(3):227-237. 7. Hillmen P, Young NS, Schubert J, et al. The complement inhibitor eculizumab in paroxysmal nocturnal hemoglobinuria. N Engl J Med. 2006;355(12):1233-1243.

8. Martí-Carvajal AJ, Anand V, Cardona AF, Solà I. Eculizumab for treating patients with paroxysmal nocturnal hemoglobinuria. Cochrane Database Syst Rev. 2014(10):CD010340. 9. Lee JW, Sicre de Fontbrune F, Wong Lee Lee L, et al. Ravulizumab (ALXN1210) vs eculizumab in adult patients with PNH naive to complement inhibitors: the 301 study. Blood. 2019;133(6):530-539 10. Nishimura J, Yamamoto M, Hayashi S, et al. Genetic variants in C5 and poor response to eculizumab. N Engl J Med. 2014;370(7):632-639. 11. Ueda Y, Obara N, Yonemura Y, et al. Effects of eculizumab treatment on quality of life in patients with paroxysmal nocturnal hemoglobinuria in Japan. Int J Hematol. 2018;107(6):656-665. 12. Risitano AM, Notaro R, Marando L, et al. Complement fraction 3 binding on erythrocytes as additional mechanism of disease in paroxysmal nocturnal hemoglobinuria patients treated by eculizumab. Blood. 2009;113(17):4094-4100. 13. Subías Hidalgo M, Martin Merinero H, López A, et al. Extravascular hemolysis and complement consumption in paroxysmal nocturnal hemoglobinuria patients undergoing eculizumab treatment. Immunobiology. 2017;222(2):363-371.

Haematologica | 107 - June 2022

1488

COMMENT

Comment on: “Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients” We read with great interest the study by Brown et al.1 analyzing the cardiovascular events during acalabrutinib treatment within four clinical trials. After a median time of acalabrutinib treatment of 24.9 months in the polled analysis, 38 of 762 chronic lymphocytic leukemia (CLL) patients developed atrial fibrillation (AF) and/or atrial flutter (15 with treatment-naïve and 23 with relapsed/refractory CLL).1 Rates of AF decreased over time, but were higher among patients with a prior history of arrhythmia as well as in elderly patients. Bruton’s tyrosine kinase (BTK) inhibitors are widely used for the treatment of B-cell malignancies, including CLL. Continuous treatment with the first in class BTK inhibitor, ibrutinib, has been associated with increased incidence of cardiovascular events like AF, atrial flutter and new onset or worse previous arterial hypertension compared to the control healthy population.2-4 The risk of developing ibrutinib-induced AF can be predicted by some AF risk scores developed.3,5-7 As reported in Table 1, (i) the Framingham score is based on age, male sex, body mass index, systolic pressure, treatment for hypertension, PR interval, significant murmur, heart failure;3 (ii) the Shanafelt score includes age, sex, valvular heart disease and hypertension;6 (iii) the Italian score includes age, sex, non-valvular cardiopathy, valvular heart disease, hypo/hyper-thyroidism, chronic lung diseases, diabetes mellitus and previous grade 3-4 infections.5 In addition, colleagues from the Mayo clinic analyzed the incidence and management of AF among 290 CLL patients treated with ibrutinib.8 After a median time on ibrutinib of 19 months, the authors concluded that all the three scores were able to identify patients with an increased risk to develop AF but “based on lower Akaike information criteria (AIC), the Italian score (AIC=513) was best at predicting the risk of treatment-emergent AF versus the Mayo CLL risk score (AIC=524) and the Framingham risk score (AIC=530)”.8 However, in the paper by Brown et al. only the Shanafelt risk score was applied to acalabrutinib-treated patients.1 Given the non-inferiority of the Italian AF score, or slightly better prediction accuracy, we would like to suggest stratifying and comparing the cumulative incidence of AF of acalabrutinib-treated patients with both scores. Recently updated results from the Elevate RR trial have been published.9 This pivotal phase III clinical trial compared acalabrutinib versus ibrutinib in relapsed-refractory

Table 1. Atrial fibrillation scores. Variables

Framingham Shanafelt score score

Italian score

Age

Male sex

Body mass index

Systolic pressure

Treatment for hypertension

ECG PR interval

Heart murmur

Heart failure

Valvular heart disease

Hypertension

Cardiopathy

Hypo/hyper-thyroidism

Chronic lung disease

Diabetes mellitus

G3-4 infections

ECG: elctrocardiography; PR interval: time from the onset of the P wave to the start of the QRS complex.

patients with high-risk cytogenetics, displaying superimposable efficacy but a better safety profile of acalabrutinib in the management of CLL patients compared to ibrutinib.9 In particular AF, hypertension, diarrhea, arthralgia and muscle spasms were less frequent in the acalabrutinib arm, while headache and cough were more common in the acalabrutinib than the ibrutinib arm.9 Whether second generation BTK inhibitors, like acalabrutinib, decrease AF rate to all risk classes, in particular to score ≥5, compared to ibrutinib is unknown but is a relevant unmet clinical need.

Authors Andrea Visentin and Livio Trentin Hematology and Clinical Immunology Unit, Department of Medicine, University of Padua, Padova, Italy

Haematologica | 107 - June 2022

1489

COMMENT Disclosures No conflicts of interest to disclose.

Correspondence: ANDREA VISENTIN- andrea.visentin@unipd.it https://doi.org/10.3324/haematol.2021.280199

Contributions AV and LT reviewed the literature and co-wrote the comment.

Received: October 14, 2021. Accepted: November 5, 2021. Prepublished: November 18, 2021.

References 1. Brown JR, Byrd JC, Ghia P, et al. Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients. Haematologica. 2022;107(6):1335-1346. 2. Ahn IE, Brown JR. Targeting Bruton's tyrosine kinase in CLL. Front Immunol. 2021;12:687458. 3. Wiczer TE, Levine LB, Brumbaugh J, et al. Cumulative incidence, risk factors, and management of atrial fibrillation in patients receiving ibrutinib. Blood Adv. 2017;1(20):1739-1748. 4. Visentin A, Campello E, Scomazzon E, et al. Dabigatran in ibrutinib-treated patients with atrial fibrillation and lymphoproliferative diseases: experience of 4 cases. Hematol Oncol. 2018;36(5):801-803. 5. Visentin A, Deodato M, Mauro FR, et al. A scoring system to predict the risk of atrial fibrillation in chronic lymphocytic

leukemia. Hematol Oncol. 2019;37(4):508-512. 6. Shanafelt TD, Parikh SA, Noseworthy PA, et al. Atrial fibrillation in patients with chronic lymphocytic leukemia (CLL). Leuk Lymphoma. 2017;58(7):1630-1639. 7. Brown JR, Moslehi J, O'Brien S, et al. Characterization of atrial fibrillation adverse events reported in ibrutinib randomized controlled registration trials. Haematologica. 2017;102(10): 1796-1805. 8. Archibald WJ, Rabe KG, Kabat BF, et al. Atrial fibrillation in patients with chronic lymphocytic leukemia (CLL) treated with ibrutinib: risk prediction, management, and clinical outcomes. Ann Hematol. 2021;100(1):143-155. 9. Byrd JC, Hillmen P, Ghia P, et al. Acalabrutinib versus ibrutinib in previously treated chronic lymphocytic leukemia: results of the first randomized phase III trial. J Clin Oncol. 2021;39(31):3441-3452.

Haematologica | 107 - June 2022

1490

RESPONSE

Response to Comment on: “Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients” I would like to thank Drs. Visentin and Trentin for their insightful response to our recently published article entitled “Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients.”1 In their response, Drs. Visentin and Trentin suggest an analysis of our pooled cardiovascular data for acalabrutinib using the Italian atrial fibrillation (AF) risk score2 in addition to the analysis that we reported using the Shanafelt AF risk score.1 They refer to the report by Archibald and colleagues,3 which compared the risk of AF with ibrutinib using three prediction tools: the Framingham Heart Study AF score,4 the Mayo chronic lymphocytic leukemia (CLL) AF score (also known as the Shanafelt AF risk score),5 and the Italian AF risk score.2 That analysis demonstrated good performance of all three tools based on clear separation of time to AF in each risk group; however, based on lower Akaike information criteria (estimate of prediction error), the Italian AF risk score was best able to predict risk of developing AF.3 We agree that additional data assessing risk of AF are needed in the context of Bruton tyrosine kinase inhibitor therapy, particularly in high-risk patient subgroups. In our pooled analysis of acalabrutinib data,1 we analyzed the incidence of de novo AF/flutter according to Shanafelt AF risk category in order to be consistent with a previous analysis demonstrating increased incidence and risk of de novo AF with increasing Shanafelt AF risk category in ibrutinib-treated patients.6 The Shanafelt AF risk category (0–1, 2–3, 4, and ≥5) is based on factors that were independently associated with AF in their cohort: older age (2 points for age 65–74 years; 3 points for age ≥75 years), male sex (1 point), valvular heart disease (2 points), and hypertension (1 point).1,5 Our analysis of acalabrutinib data showed a notable increase in AF incidence only among patients with the highest Shanafelt risk scores, with an incidence of 13% reported for Shanafelt risk category ≥5 compared with 2% to 5% for the lower Shanafelt risk categories (0–4).1 While these data suggested that the incidence of AF by Shanafelt risk category was lower for acalabrutinib compared with data previously reported for ibrutinib (Shanafelt risk categories ≥5 [15%] and 0–4 [4% to 9%]),6 our findings indicate that the lower risk categories may be less informative for assessing AF risk in the context of acala-

brutinib therapy. Compared with the Shanafelt AF risk score, the Italian AF risk score (categories: 0, 1-2, 3-4, and ≥5) weights older age less heavily (>65 years: 1 point), uses the same weighting for male sex (1 point) and valvular disease (2 points), excludes hypertension, and includes several additional factors that reflect comorbidities relevant to AF in this population (cardiomyopathy: 3 points; hyperthyroidism: 1 point; chronic lung disease: 1 point; diabetes mellitus: 1 point; and severe infections: 1 point).2,3 Given the potential for greater stratification of higherrisk patients using the Italian score, we agree that a comparative analysis of the Shanafelt AF risk score and the Italian AF risk score using data from our pooled cardiovascular safety analysis of acalabrutinib in patients with CLL is of interest. We are currently assessing these data.

Author Jennifer R. Brown Dana-Farber Cancer Institute, Boston, MA, USA Correspondence: JENNIFER R. BROWN- jennifer_brown@dfci.harvard.edu https://doi.org/10.3324/haematol.2021.280310 Received: November 3, 2021. Accepted: November 5, 2021. Prepublished: November 18, 2021. Disclosures None additional beyond those already cited in the main manuscript. For the original manuscript this is in response to, support was provided by the National Cancer Institute R35 CA197734, Four Winds Foundation, the D. Warren Brown Foundation, the Intramural Research Program of the National Heart, Lung, and Blood Institute. Additionally, the original work this is in response to was supported by Acerta Pharma, a member of the AstraZeneca Group and medical writing assistance, funded by AstraZeneca, was provided by Robert J. Schoen, PharmD, and Cindy Gobbel, PhD, of Peloton Advantage, LLC, an OPEN Health company. JRB has served as a consultant for Abbvie, Acerta/Astra-Zeneca, BeiGene, Bristol Myers Squibb/Juno/Celgene, Catapult, Dynamo, Eli Lilly, Genentech/Roche,

Haematologica | 107 - June 2022

1491

RESPONSE Hutchmed, Janssen, Kite, Loxo, MEI Pharma, Morphosys AG,

received research funding from Gilead, Loxo/Lilly, Sun,

Nextcea, Novartis, Octapharma, Pfizer, Pharmacyclics, Rigel, TG

Verastem/SecuraBio, and TG Therapeutics; and served on data

Therapeutics and Verastem; received honoraria from Janssen;

safety monitoring committees for Invectys and Morphosys.

References 1. Brown JR, Byrd JC, Ghia P, et al. Cardiovascular adverse events in patients with chronic lymphocytic leukemia receiving acalabrutinib monotherapy: pooled analysis of 762 patients. Haematologica. 2022;107(6):1335-1346. 2. Visentin A, Deodato M, Mauro FR, et al. A scoring system to predict the risk of atrial fibrillation in chronic lymphocytic leukemia. Hematol Oncol. 2019;37(4):508-512. 3. Archibald WJ, Rabe KG, Kabat BF, et al. Atrial fibrillation in patients with chronic lymphocytic leukemia (CLL) treated with ibrutinib: risk prediction, management, and clinical outcomes. Ann Hematol. 2020;100(1):143-155.

4. Wiczer TE, Levine LB, Brumbaugh J, et al. Cumulative incidence, risk factors, and management of atrial fibrillation in patients receiving ibrutinib. Blood Adv. 2017;1(20):1739-1748. 5. Shanafelt TD, Parikh SA, Noseworthy PA, et al. Atrial fibrillation in patients with chronic lymphocytic leukemia (CLL). Leuk Lymphoma. 2017;58(7):1630-1639. 6. Brown JR, Moslehi J, O’Brien, et al. Characterization of atrial fibrillation adverse events reported in ibrutinib randomized controlled registration trials. Haematologica. 2017;102(10):1796-1805.

Haematologica | 107 - June 2022

1492

haematologica Journal of the Ferrata Storti Foundation

haematologica.org